Evaluating the Feasibility of AI-Predicted mpMRI Image Features for Predicting Prostate Cancer Aggressiveness: a Multicenter Study

Evaluating the Feasibility of AI-Predicted mpMRI Image Features for Predicting Prostate Cancer Aggressiveness: a Multicenter Study | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Evaluating the Feasibility of AI-Predicted mpMRI Image Features for Predicting Prostate Cancer Aggressiveness: a Multicenter Study Kexin Wang, Ning Luo, Zhaonan Sun, Xiangpeng Zhao, Lilan She, and 6 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4252741/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Objective To evaluate the feasibility of utilizing artificial intelligence (AI)-predicted multiparametric MRI (mpMRI) image features for predicting the aggressiveness of prostate cancer (PCa). Materials and methods A total of 878 PCa patients from 4 hospitals were retrospectively collected, all of whom had pathological results after radical prostatectomy(RP).A pre-trained AI algorithm was used to select suspected PCa lesions and extract lesion features for model development. The study evaluated five prediction methods, including 1) A clinical model of clinical features and image features of suspected PCa lesions selected by AI algorithm, 2)the PIRADS category, 3)a conventional radiomics model, 4) a radiomics model based on deep learning, 5)biopsy pathology. Results In the externally validated dataset, the deep learn-based radiomics model showed the highest area under the curve (AUC 0.700 to 0.791).It exceeded clinical model (AUC 0.597 to 0.718), traditional radiomic model (AUC 0.566 to 0.632), PIRADS score (AUC 0.554 to 0.613) and biopsy pathology (AUC 0.537 to 0.578). And the AUC predicted by the model did not show statistically significant difference among the three externally verified hospitals (P > 0.05). Conclusion Deep-radiomics models utilizing AI-extracted image features from mpMRI images can potentially be used to predict PCa aggressiveness, demonstrating a generalized ability for external validation. Health sciences/Diseases/Cancer Health sciences/Diseases/Urogenital diseases Gleason score prostate cancer multiparametric magnetic resonance imaging deep learning radiomics Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Introduction Prostate cancer (PCa) is the second most common cancer in men worldwide [ 1 ] . At present, how to avoid overtreatment of inert prostate cancer was a clinical challenge [ 2 ] . Since the selection of a PCa treatment plan and its prognostic value are closely related to its biological behavior, early prediction and accurate judgment of whether the patient's lesion is invasive are clinically important for obtaining the best treatment plan and effect [ 3 ] . Gleason score (GS) is a commonly used and effective indicator of the invasiveness of PCa. In 2014, the International Society of Urology Pathology (ISUP) proposed the use of the Gleason score group (GG), which was divided into five groups according to the GS, which can better guide clinical work [ 4 ] . Transrectal ultrasound (TRUS)-guided biopsy is often used to determine preoperative GS, but due to the randomness of biopsy sampling, it was not possible to accurately evaluate prostate cancer lesions and evaluate tumor heterogeneity. For the above reasons, GS obtained by preoperative biopsy is different from GS obtained by radical prostatectomy (RP), which may cause patients to be overtreated or undertreated [ 5 , 6 ] . Advances in prostate multiparametric magnetic resonance imaging (mpMRI) and prostate imaging reporting and data system (PI-RADS v2.1) have improved the ability to detect and evaluate prostate cancer. Related research [ 7 ] has shown that the apparent diffusion coefficient (ADC) of the PCa region is a better indicator of prostate cancer aggressiveness than is the TRUS biopsy GS. However, the analysis of prostate mpMRIs relies mainly on the experience and subjective judgment of radiologists, resulting in increased variability and decreased reliability of the results [ 8 ] . Radiomics research involves the extraction of high-throughput and quantitative features from multimodal medical images and the use of machine learning algorithms to transform these features into high-dimensional mining information related to tumor pathophysiology, which may aid in clinical diagnosis and decision-making [ 9 , 10 ] . Radiomic models based on mpMRI had been used for risk stratification of PCa patients [ 11 – 14 ] . Several studies have shown that the features extracted from T2WI and ADC are helpful for classifying the Gleason score [ 15 – 17 ] . At present, manual delineation of the region of interest (ROI) is still the main method of radiometric analysis and inevitably consumes certain human resources and time. Most related research reports [ 13 , 18 , 19 ] at this stage are limited to single-center research. The generalization ability of relevant models and how to reach the level of clinical application still need to be discussed and verified. In this study, we used a pre-trained AI algorithm to automatically segment PCa lesions [ 20 ] and extract mpMRI features. We developed various models and evaluated their feasibility for predicting PCa invasiveness before surgery. In addition, the clinical efficacy of the prediction model was evaluated and validated on a multicenter dataset. Materials and methods Data enrollment This retrospective study received approval from the institutional review board (IRB number: 2021060), all methods were performed in accordance with the relevant guidelines.The Biomedical Research Ethics Committee of Peking University First Hospital granted a waiver for written patient informed consent. The data were collected retrospectively from four distinct hospitals. The study included patients who underwent prostate mpMRI followed by RP between November 2017 and December 2022. This encompassed gathering mpMRI images and pertinent clinical information, such as age, PSA levels, mpMRI results, biopsy pathology, and RP pathology reports. The exclusion criteria were as follows: 1) prior endocrine therapy, 2) benign prostate hyperplasia according to RP pathology, 3) missing PSA data, 4) incomplete biopsy pathology records, 5) incomplete MR images, and 6) evident image artifacts. The process of data enrollment is depicted in Fig. 1. Figure 1 Data enrollment process Flowchart depicting the process of data inclusion. The training dataset comprises data from Hospital_1, while the external validation dataset includes data from Hospital_2, Hospital_3, and Hospital_4. All the data adhered to the same inclusion and exclusion criteria. In total, the training dataset consisted of 345 included patients, and the external validation dataset included 533 patients, 18 from Hospital_2, 231 from Hospital_3, and 284 from Hospital_4. Reference standard All patients underwent both biopsy and RP, and pathology samples were available for all patients. Pathology was assessed by experienced pathologists following the guidelines of the International Society of Urological Pathology (ISUP) group. The reference standard was established using the RP pathology results, categorizing the patients into three groups: ISUP_1, ISUP_2, and ISUP_3 ~ 5. MR scanning protocols The mpMRI images in this study were acquired with 13 MR scanners from five different vendors. Detailed information regarding the MR scanners and image acquisition protocols can be found in Table 1 . Table 1 MR scanning protocols used in different hospitals Overall Training dataset External validation dataset (N = 533) P value (N = 878) Hospital_1 Hospital_2 Hospital_3 Hospital_4 (N = 345) (N = 18) (N = 231) (N = 284) MR scanner Magnetic field < 0.001 1.436 T 5 (0.6%) 0 (0%) 0 (0%) 0 (0%) 5 (1.8%) 1.5 T 229 (26.1%) 19 (5.5%) 0 (0%) 45 (19.5%) 165 (58.1%) 3.0 T 644 (73.3%) 326 (94.5%) 18 (100%) 186 (80.5%) 114 (40.1%) Station name < 0.001 GE-Scanner_1 262 (29.8%) 262 (75.9%) 0 (0%) 0 (0%) 0 (0%) GE-Scanner_2 37 (4.2%) 0 (0%) 0 (0%) 37 (16.0%) 0 (0%) GE-Scanner_3 159 (18.1%) 0 (0%) 1 (5.6%) 44 (19.0%) 114 (40.1%) PHILIPS_Scanner_1 38 (4.3%) 38 (11.0%) 0 (0%) 0 (0%) 0 (0%) PHILIPS_Scanner_2 1 (0.1%) 1 (0.3%) 0 (0%) 0 (0%) 0 (0%) PHILIPS_Scanner_3 1 (0.1%) 1 (0.3%) 0 (0%) 0 (0%) 0 (0%) SIEMENS_Scanner_1 14 (1.6%) 0 (0%) 14 (77.8%) 0 (0%) 0 (0%) SIEMENS_Scanner_2 43 (4.9%) 0 (0%) 0 (0%) 43 (18.6%) 0 (0%) SIEMENS_Scanner_3 64 (7.3%) 0 (0%) 0 (0%) 64 (27.7%) 0 (0%) SIEMENS_Scanner_4 19 (2.2%) 19 (5.5%) 0 (0%) 0 (0%) 0 (0%) SIEMENS_Scanner_5 3 (0.3%) 0 (0%) 3 (16.7%) 0 (0%) 0 (0%) SIEMENS_Scanner_6 208 (23.7%) 0 (0%) 0 (0%) 43 (18.6%) 165 (58.1%) UIH_Scanner_1 29 (3.3%) 24 (7.0%) 0 (0%) 0 (0%) 5 (1.8%) DWI/ADC protocol B value (10^-6 s/mm2) < 0.001 800 8 (0.9%) 6 (1.7%) 1 (5.6%) 0 (0%) 1 (0.4%) 1000 288 (32.8%) 0 (0%) 2 (11.1%) 8 (3.5%) 278 (97.9%) 1200 97 (11.0%) 0 (0%) 0 (0%) 96 (41.6%) 1 (0.4%) 1400 377 (42.9%) 285 (82.6%) 1 (5.6%) 87 (37.7%) 4 (1.4%) 1500 52 (5.9%) 0 (0%) 12 (66.7%) 40 (17.3%) 0 (0%) 2000 56 (6.4%) 54 (15.7%) 2 (11.1%) 0 (0%) 0 (0%) Repetition time (ms) < 0.001 Median [Q1,Q3] 3000 [2630,3800] 2680 [2650,3000] 2500 [2500,3660] 4310 [3000,4850] 3000 [2090,3000] Echo time (ms) < 0.001 Median [Q1,Q3] 61.7 [59.9,70.0] 61.0 [59.8,61.5] 91.0 [82.8,91.0] 70.0 [63.0,74.0] 70.0 [59.7,70.0] Pixel bandwidth (MHz) < 0.001 Median [Q1,Q3] 1950 [1540,1950] 1950 [1950,1950] 1300 [1300,1500] 1630 [1540,1950] 1540 [1540,1950] Slice thickness (mm) < 0.001 Median [Q1,Q3] 3.50 [3.00,4.00] 4.00 [4.00,4.00] 4.00 [4.00,4.00] 3.00 [3.00,3.50] 3.50 [3.00,3.50] Slice spacing (mm) < 0.001 Median [Q1,Q3] 4.00 [3.50,4.00] 4.00 [4.00,4.00] 4.80 [4.80,4.80] 3.60 [3.30,4.00] 3.50 [3.50,4.00] Reconstruction diameter (ms) < 0.001 Median [Q1,Q3] 230 [200,240] 240 [220,240] 250 [250,250] 221 [200,260] 200 [200,340] Pixel spacing (ms) < 0.001 Median [Q1,Q3] 1.25 [0.938,1.63] 0.938 [0.938,0.938] 1.30 [1.30,1.30] 1.33 [0.877,1.63] 1.56 [1.33,1.79] T2WI protocol Repetition time (ms) < 0.001 Median [Q1,Q3] 3560 [3130,4470] 3130 [3020,3500] 4000 [4000,4000] 3710 [3340,4120] 5400 [3410,5540] Echo time (ms) < 0.001 Median [Q1,Q3] 105 [88.9,112] 88.8 [87.3,93.0] 100 [100,100] 108 [96.0,112] 112 [106,112] Pixel bandwidth (MHz) < 0.001 Median [Q1,Q3] 163 [163,200] 163 [163,163] 203 [203,203] 190 [160,200] 200 [163,200] Slice thickness (mm) < 0.001 Median [Q1,Q3] 3.50 [3.00,4.00] 4.00 [4.00,4.00] 3.00 [3.00,3.00] 3.00 [3.00,3.50] 3.50 [3.00,3.50] Slice spacing (mm) < 0.001 Median [Q1,Q3] 4.00 [3.50,4.00] 4.00 [4.00,4.00] 3.60 [3.60,3.60] 3.60 [3.50,4.00] 3.50 [3.50,4.00] Reconstruction diameter (ms) < 0.001 Median [Q1,Q3] 200 [200,240] 240 [240,240] 178 [178,178] 200 [200,220] 200 [200,200] Pixel spacing (ms) < 0.001 Median [Q1,Q3] 0.469 [0.391,0.625] 0.469 [0.469,0.469] 0.625 [0.625,0.625] 0.391 [0.344,0.625] 0.625 [0.391,0.625] Image features extracted by AI algorithms Following anonymization, the DICOM files were converted to NIFTI format using dicom2nii.py (Python 3.5) and then input into pretrained AI software. For detailed technical information about the AI software, please refer to the supplementary material. For each enrolled patient, the largest lesion predicted by the AI algorithms on the mpMRI images was designated the index lesion for analysis. The lesion's location was classified as solely in the peripheral zone (PZ), solely in the transitional zone (TZ), or in both the peripheral and transitional zones (PZ + TZ). The following image features were computed and utilized in the prediction model: 1) lesion volume, 2) diameter of the lesion, signifying the largest diameter of the lesion, 3) mean ADC value of the lesion (ADClesion), 4) mean signal intensity of the lesion on DWI (DWIlesion), and 5) mean signal intensity of the lesion on T2WI (T2WIlesion) (as illustrated in Fig. 2 ). Step 1 involves selecting and inputting anonymized mpMRI images, which encompass T2WI, DWI, and ADC maps. In Step 2, the pre-trained AI models are employed to identify suspicious areas indicative of PCa. Step 3 entails identifying the index lesion by selecting the largest lesion from the predicted labels. The index lesion was annotated as the red areas on the images. Step 4 encompasses extracting image features from the index lesion, encompassing 1) clinical image features such as lesion volume, lesion diameter, ADC value, DWI signal intensity, and T2WI intensity, 2) conventional radiomic features, and 3) deep radiomic features. In Step 5, various prediction methods are developed, including a) biopsy prediction, which forecasts ISUP grouping based on biopsy pathology; b) PIRADS prediction, which anticipates ISUP grouping based on the PIRADS category; c) a clinical model, which incorporates age, PSA, PIRADS category, and clinical image features to predict ISUP grouping; d) radiomics model, which forecasts ISUP grouping using conventional radiomic features obtained from the index lesion; and e) deep-radiomics model, which forecasts ISUP grouping using deep-radiomic features extracted from the index lesion. Prediction methods To predict the ISUP_1, ISUP_2, and ISUP_3 ~ 5 classes, five prediction methods were developed, namely, the PIRADS category, biopsy pathology, a clinical model, a radiomics model, and a deep-radiomics model. The steps involved in developing the prediction methods are illustrated in Fig. 2 . The training dataset utilized was sourced from Hospital_1, while the external validation dataset included data from Hospital_2, Hospital_3, and Hospital_4. For the construction of the clinical model, clinical variables were employed as covariates. These included age, PSA levels, PIRADS score, and image features of the suspected clinically significant prostate cancer (csPCa) lesion, which included position, volume, diameter, ADC value, signal intensity on DWI, and signal intensity on T2WI. To construct the radiomics model, image features were derived from regions of interest (ROIs) on the ADC maps using the PyRadiomics package in Python ( https://pyradiomics.readthedocs.io/en/latest/changes.html ). Z score normalization was applied to standardize the extracted features, and Pearson correlation coefficients (PCCs) were computed to identify highly correlated features. Features with a PCC value exceeding 0.99 were eliminated to mitigate multicollinearity. The Kruskal‒Wallis test was then employed to select the features for the final radiomics model. For the classifier, the least absolute shrinkage and selection operator (LASSO) algorithm was utilized. To construct the deep-radiomics model, image features were extracted using a deep learning algorithm. The construction process involved several key steps: 1) Preprocessing ADC maps: Initial preprocessing included normalizing the intensities of the ADC maps. 2) ROI resampling: ROIs were resampled to ensure a consistent voxel size. 3) Deep Feature Extraction: A pretrained deep learning model was used to extract features from the segmented ROIs. 4) Feature Dimension Reduction: The resulting channel feature maps underwent dimension reduction by filtering with the maximum value, resulting in a set of 512 one-dimensional features. 5) Model Construction: After extracting the deep features, the construction of the deep-radiomics model followed a procedure similar to that of the radiomics model, as previously described. For model tuning, 5-fold cross-validation was employed to select the optimal value of the hyperparameter α, which controls the strength of the L1 penalty. A grid search was performed over a range of α values, and the one yielding the highest cross-validated accuracy was chosen. This process identified a subset of significant predictors associated with ISUP classes. Prediction efficacy evaluation The external dataset was used to assess the prediction effectiveness of the methods via receiver operating characteristic (ROC) analysis. This evaluation was conducted by computing the area under the ROC curve (AUC). Statistical analysis The statistical analysis was conducted using R 4.1.3 software. Quantitative variables with a normal distribution are presented as the mean ± standard deviation, while those with a nonnormal distribution are presented as the median [Q1, Q3]. Categorical variables are presented as frequencies. The normality of the variables was assessed using the Kolmogorov‒Smirnov test, associations between categorical variables were examined using the chi-square test, and differences between multiple groups were compared using the Kruskal‒Wallis test. The DeLong test was used to compare differences in the area under the curve (AUC) among the five types of prediction methods. A P value less than 0.05 was considered to indicate statistical significance. Results Clinical characteristics A total of 878 eligible patients participated in the study, with an average age of 68.7 ± 6.6 years. Among these patients, 345 were recruited from Hospital_1, and 18, 231, and 284 were from Hospital_2, Hospital_3, and Hospital_4, respectively. All clinical variables exhibited statistically significant differences among the four hospitals (all P < 0.001), except for age (P = 0.878). Further details regarding the clinical variables among the different datasets can be found in Table 2 . Table 2 Clinical characteristics of patients in the training and external validation cohorts Overall Training cohort External validation cohort P value (N = 878) Hospital_1 Hospital_2 Hospital_3 Hospital_4 (N = 345) (N = 18) (N = 231) (N = 284) Age (year) 0.878 Mean(SD) 68.7 (6.63) 68.5 (6.58) 69.4 (6.71) 68.8 (6.73) 68.7 (6.64) PSA (ng/dl) < 0.001 Median [Q1,Q3] 11.6 [7.90,19.1] 10.9 [7.06,15.5] 9.93 [6.35,14.0] 12.8 [8.82,23.9] 12.8 [8.20,22.5] PIRADS PIRADS_1 2 (0.2%) 1 (0.3%) 0 (0%) 1 (0.4%) 0 (0%) < 0.001 PIRADS_2 87 (9.9%) 12 (3.5%) 4 (22.2%) 4 (1.7%) 67 (23.6%) PIRADS_3 216 (24.6%) 45 (13.0%) 8 (44.4%) 57 (24.7%) 106 (37.3%) PIRADS_4 286 (32.6%) 173 (50.1%) 0 (0%) 57 (24.7%) 56 (19.7%) PIRADS_5 287 (32.7%) 114 (33.0%) 6 (33.3%) 112 (48.5%) 55 (19.4%) PCa location < 0.001 Not measurable 137 (15.6%) 110 (31.9%) 1 (5.6%) 12 (5.2%) 14 (4.9%) PZ 80 (9.1%) 38 (11.0%) 3 (16.7%) 14 (6.1%) 25 (8.8%) PZ + TZ 491 (55.9%) 121 (35.1%) 10 (55.6%) 172 (74.5%) 188 (66.2%) TZ 170 (19.4%) 76 (22.0%) 4 (22.2%) 33 (14.3%) 57 (20.1%) Lesion volume < 0.001 Median [Q1,Q3] 1.60 [0.717,3.93] 0.737 [0,1.55] 2.73 [0.976,4.68] 3.25 [1.44,7.62] 2.44 [1.16,5.25] Lesion diameter < 0.001 Median [Q1,Q3] 24.9 [17.8,33.5] 18.2 [0,25.4] 27.5 [23.7,37.2] 31.1 [23.1,40.0] 28.1 [21.3,36.2] ADC lesion (10^6 mm2/s) < 0.001 Median [Q1,Q3] 802 [663,907] 738 [0,865] 848 [757,957] 816 [714,909] 840 [739,926] DWI lesion < 0.001 Median [Q1,Q3] 120 [46.3,404] 52.2 [0,85.7] 313 [176,398] 266 [80.5,510] 281 [102,506] TWI lesion < 0.001 Median [Q1,Q3] 324 [119,1150] 127 [0,297] 579 [436,1490] 490 [160,1580] 499 [274,1520] Biopsy result < 0.001 No cancer 8 (0.9%) 0 (0%) 0 (0%) 1 (0.4%) 7 (2.5%) ISUP_1 132 (15.0%) 49 (14.2%) 0 (0%) 14 (6.1%) 69 (24.3%) ISUP_2 185 (21.1%) 128 (37.1%) 0 (0%) 18 (7.8%) 39 (13.7%) ISUP_3 331 (37.7%) 84 (24.3%) 18 (100%) 159 (68.8%) 70 (24.6%) ISUP_4 133 (15.1%) 43 (12.5%) 0 (0%) 26 (11.3%) 64 (22.5%) ISUP_5 89 (10.1%) 41 (11.9%) 0 (0%) 13 (5.6%) 35 (12.3%) PR result ISUP_1 62 (7.1%) 10 (2.9%) 0 (0%) 27 (11.7%) 25 (8.8%) < 0.001 ISUP_2 300 (34.2%) 148 (42.9%) 6 (33.3%) 63 (27.3%) 83 (29.2%) ISUP_3 241 (27.4%) 93 (27.0%) 3 (16.7%) 66 (28.6%) 79 (27.8%) ISUP_4 108 (12.3%) 40 (11.6%) 4 (22.2%) 27 (11.7%) 37 (13.0%) ISUP_5 167 (19.0%) 54 (15.7%) 5 (27.8%) 48 (20.8%) 60 (21.1%) Based on the RP pathology results, 62 patients were categorized as ISUP_1, 300 as ISUP_2, and 516 as ISUP_3 ~ 5. Except for patient age distribution, which did not differ significantly among the three ISUP classes (P = 0.563), all the other clinical variables demonstrated significant differences among the various ISUP classes (all P < 0.05). Moreover, there was a statistically significant difference in the distribution of ISUP classes among the four hospitals (P < 0.001). Detailed information about the variables in different ISUP classes can be found in Table 3 . The distributions of these variables across different ISUP classes are illustrated in Fig. 3 . Table 3 Clinical characteristics of patients in different ISUP classes Overall ISUP_1 ISUP_2 ISUP_3 ~ 5 P value (N = 878) (N = 62) (N = 300) (N = 516) Age (year) 0.563 Mean(SD) 68.7(6.63) 69.6(6.91) 68.5(6.74) 68.6(6.54) PSA (ng/dl) < 0.001 Median [Q1,Q3] 11.6 [7.90,19.1] 11.9 [9.37,18.4] 10.8 [6.99,15.8] 12.3 [8.18,21.4] PIRADS < 0.001 PIRADS_1 2 (0.2%) 0 (0%) 1 (0.3%) 1 (0.2%) PIRADS_2 87 (9.9%) 16 (25.8%) 29 (9.7%) 42 (8.1%) PIRADS_3 216 (24.6%) 20 (32.3%) 81 (27.0%) 115 (22.3%) PIRADS_4 286 (32.6%) 15 (24.2%) 119 (39.7%) 152 (29.5%) PIRADS_5 287 (32.7%) 11 (17.7%) 70 (23.3%) 206 (39.9%) PCa location < 0.001 Not measurable 137 (15.6%) 7 (11.3%) 86 (28.7%) 44 (8.5%) PZ 80 (9.1%) 7 (11.3%) 39 (13.0%) 34 (6.6%) TZ 170 (19.4%) 14 (22.6%) 67 (22.3%) 89 (17.2%) PZ + TZ 491 (55.9%) 34 (54.8%) 108 (36.0%) 349 (67.6%) Lesion volume < 0.001 Median [Q1,Q3] 1.60 [0.717,3.93] 1.66 [0.869,2.84] 0.887 [0,1.90] 2.37 [0.973,5.87] Lesion diameter < 0.001 Median [Q1,Q3] 24.9 [17.8,33.5] 25.4 [19.1,31.1] 19.6 [0,26.1] 28.1 [20.6,38.1] ADC lesion (10^6 mm2/s) < 0.001 Median [Q1,Q3] 802 [663,907] 820 [732,915] 773 [0,892] 813 [697,907] DWI lesion < 0.001 Median [Q1,Q3] 120 [46.3,404] 101 [54.5,282] 68.6 [0,287] 205 [61.4,489] TWI lesion < 0.001 Median [Q1,Q3] 324 [119,1150] 303 [122,476] 162 [0,576] 434 [132,1390] Biopsy result < 0.001 No cancer 8 (0.9%) 2 (3.2%) 1 (0.3%) 5 (1.0%) ISUP_1 132 (15.0%) 33 (53.2%) 62 (20.7%) 37 (7.2%) ISUP_2 185 (21.1%) 3 (4.8%) 111 (37.0%) 71 (13.8%) ISUP_3 331 (37.7%) 22 (35.5%) 99 (33.0%) 210 (40.7%) ISUP_4 133 (15.1%) 2 (3.2%) 21 (7.0%) 110 (21.3%) ISUP_5 89 (10.1%) 0 (0%) 6 (2.0%) 83 (16.1%) Dataset < 0.001 Training dataset 345 (39.3%) 10 (16.1%) 148 (49.3%) 187 (36.2%) Test dataset 533 (60.7%) 52 (83.9%) 152 (50.7%) 329 (63.8%) Institute name < 0.001 Hospital_1 345 (39.3%) 10 (16.1%) 148 (49.3%) 187 (36.2%) Hospital_2 18 (2.1%) 0 (0%) 6 (2.0%) 12 (2.3%) Hospital_3 231 (26.3%) 27 (43.5%) 63 (21.0%) 141 (27.3%) Hospital_4 284 (32.3%) 25 (40.3%) 83 (27.7%) 176 (34.1%) The PSA map displays the distribution of PSA levels. Columns 1 to 4 show the PSA values for lesions classified as "not measurable," lesions in the PZ, lesions in the TZ, and lesions in both the PZ and TZ, respectively. Column 5 displays the PSA values for all lesions. Rows 1 to 5 display the PSA values for lesions categorized as PIRADS_1 to PIRADS_5, and row 6 shows the PSA values for all lesions. The distribution of measurements for lesion volume, lesion diameter, ADC value, DWI signal intensity, and T2WI signal intensity is presented in a similar manner. Columns 1 to 3 display the measurements for lesions in the PZ, lesions in the TZ, and lesions in both the PZ and TZ, respectively. Column 4 represents the measurements for all lesions. When observed by rows, rows 1 to 4 show the measurements for lesions categorized as PIRADS_2 to PIRADS_5, and row 5 displays the measurements for all lesions. Model development metrics Ordinal logistic regression analysis results indicated that age, PSA level, lesion location, and the ADC value were statistically significant predictors (all P 0.05). The model exhibited a good fit to the data, as evidenced by the Brant test (P > 0.05). Additional details of the ordinal logistic regression model can be found in the supplementary material (Table S 14). The radiomics model was constructed using the LASSO method, which identified 13 critical features (3 first-order features and 10 texture features) for inclusion in the final model. The deep-radiomics model incorporated 13 features extracted from the convolutional network. The development of the radiomics and deep-radiomics LASSO classifiers is depicted in the supplementary material (Figure S 8). The coefficients of the selected features for the radiomics model and the deep-radiomics model are presented in the supplementary material (Table S 15). The prediction results are visualized in the form of a confusion matrix, as illustrated in Fig. 4. Model evaluation The performance of the clinical model, PIRADS category, radiomics model, deep-radiomics model, and biopsy pathology was assessed using ROC analysis, and the results are detailed in Figs. 5 ~ 6 and Tables 4 ~ 5. Table 4 Efficacy of the five prediction methods in the training and external validation datasets AUC ACC SEN SPE PPV NPV Training dataset ISUP_1 (N = 10) Clinical model 0.815 (0.695, 0.934) 0.736 (0.735, 0.737) 0.900 (0.714, 1.000) 0.731 (0.684, 0.779) 0.091 (0.034, 0.148) 0.996 (0.988, 1.004) PIRADS category 0.507 (0.293, 0.722) 0.838 (0.837, 0.838) 0.300 (0.016, 0.584) 0.854 (0.816, 0.892) 0.058 (-0.006, 0.121) 0.976 (0.959, 0.994) Radiomics model 0.852 (0.741, 0.963) 0.849 (0.849, 0.850) 0.800 (0.552, 1.000) 0.851 (0.813, 0.889) 0.138 (0.049, 0.227) 0.993 (0.983, 1.003) Deep-radiomics model 0.882 (0.819, 0.946) 0.739 (0.738, 0.740) 1.000 (1.000, 1.000) 0.731 (0.684, 0.779) 0.100 (0.041, 0.159) 1.000 (1.000, 1.000) Biopsy pathology 0.799 (0.653, 0.946) 0.820 (0.819, 0.821) 0.800 (0.552, 1.000) 0.821 (0.780, 0.862) 0.118 (0.041, 0.194) 0.993 (0.983, 1.003) ISUP_2 (N = 148) Clinical model 0.666 (0.609, 0.723) 0.635 (0.633, 0.636) 0.784 (0.717, 0.850) 0.523 (0.453, 0.593) 0.552 (0.485, 0.620) 0.763 (0.691, 0.835) PIRADS category 0.604 (0.556, 0.653) 0.609 (0.607, 0.610) 0.243 (0.174, 0.312) 0.883 (0.838, 0.928) 0.610 (0.486, 0.735) 0.608 (0.552, 0.665) Radiomics model 0.609 (0.549, 0.668) 0.583 (0.581, 0.584) 0.642 (0.565, 0.719) 0.538 (0.468, 0.608) 0.511 (0.439, 0.583) 0.667 (0.593, 0.740) Deep-radiomics model 0.791 (0.745, 0.837) 0.713 (0.712, 0.714) 0.878 (0.826, 0.931) 0.589 (0.520, 0.658) 0.616 (0.550, 0.682) 0.866 (0.808, 0.923) Biopsy pathology 0.565 (0.515, 0.616) 0.597 (0.596, 0.598) 0.385 (0.307, 0.464) 0.756 (0.696, 0.816) 0.543 (0.448, 0.638) 0.621 (0.559, 0.682) ISUP_3 ~ 5 (N = 187) Clinical model 0.699 (0.644, 0.754) 0.661 (0.660, 0.662) 0.561 (0.490, 0.633) 0.778 (0.714, 0.843) 0.750 (0.678, 0.822) 0.600 (0.533, 0.667) PIRADS category 0.602 (0.553, 0.651) 0.597 (0.596, 0.598) 0.893 (0.849, 0.937) 0.247 (0.180, 0.314) 0.584 (0.527, 0.641) 0.661 (0.540, 0.782) Radiomics model 0.649 (0.591, 0.707) 0.635 (0.633, 0.636) 0.786 (0.727, 0.845) 0.456 (0.378, 0.533) 0.631 (0.569, 0.693) 0.643 (0.554, 0.732) Deep-radiomics model 0.808 (0.764, 0.852) 0.716 (0.715, 0.717) 0.578 (0.507, 0.648) 0.880 (0.829, 0.930) 0.850 (0.788, 0.912) 0.638 (0.574, 0.701) Biopsy pathology 0.599 (0.549, 0.648) 0.614 (0.613, 0.616) 0.786 (0.727, 0.845) 0.411 (0.335, 0.488) 0.613 (0.551, 0.674) 0.619 (0.526, 0.712) External validation dataset ISUP_1 (N = 52) Clinical model 0.718 (0.652, 0.785) 0.717 (0.716, 0.717) 0.635 (0.504, 0.765) 0.726 (0.686, 0.765) 0.200 (0.139, 0.261) 0.948 (0.926, 0.971) PIRADS category 0.554 (0.485, 0.623) 0.475 (0.474, 0.476) 0.673 (0.546, 0.801) 0.453 (0.409, 0.498) 0.117 (0.081, 0.154) 0.928 (0.895, 0.961) Radiomics model 0.632 (0.549, 0.714) 0.829 (0.829, 0.830) 0.365 (0.235, 0.496) 0.879 (0.850, 0.909) 0.247 (0.150, 0.343) 0.928 (0.904, 0.951) Deep-radiomics model 0.791 (0.728, 0.854) 0.871 (0.870, 0.871) 0.538 (0.403, 0.674) 0.906 (0.880, 0.932) 0.384 (0.272, 0.495) 0.948 (0.928, 0.968) Biopsy pathology 0.537 (0.492, 0.582) 0.244 (0.243, 0.245) 0.904 (0.824, 0.984) 0.173 (0.139, 0.206) 0.106 (0.077, 0.134) 0.943 (0.895, 0.992) ISUP_2 (N = 152) Clinical model 0.597 (0.547, 0.646) 0.568 (0.568, 0.569) 0.651 (0.576, 0.727) 0.535 (0.485, 0.586) 0.359 (0.302, 0.415) 0.794 (0.744, 0.843) PIRADS category 0.607 (0.561, 0.653) 0.557 (0.556, 0.558) 0.704 (0.631, 0.777) 0.499 (0.448, 0.549) 0.359 (0.305, 0.414) 0.809 (0.758, 0.859) Radiomics model 0.566 (0.511, 0.621) 0.642 (0.641, 0.642) 0.414 (0.336, 0.493) 0.732 (0.688, 0.777) 0.382 (0.308, 0.456) 0.758 (0.714, 0.802) Deep-radiomics model 0.700 (0.652, 0.748) 0.595 (0.594, 0.596) 0.789 (0.725, 0.854) 0.517 (0.467, 0.567) 0.395 (0.340, 0.450) 0.860 (0.815, 0.905) Biopsy pathology 0.578 (0.539, 0.618) 0.707 (0.707, 0.708) 0.276 (0.205, 0.347) 0.879 (0.847, 0.912) 0.477 (0.373, 0.582) 0.753 (0.713, 0.793) ISUP_3 ~ 5 (N = 329) Clinical model 0.668 (0.622, 0.713) 0.598 (0.598, 0.599) 0.498 (0.444, 0.553) 0.760 (0.701, 0.818) 0.770 (0.713, 0.826) 0.484 (0.430, 0.539) PIRADS category 0.613 (0.569, 0.656) 0.591 (0.590, 0.592) 0.526 (0.472, 0.580) 0.696 (0.633, 0.759) 0.736 (0.680, 0.793) 0.477 (0.420, 0.533) Radiomics model 0.602 (0.552, 0.652) 0.649 (0.648, 0.650) 0.845 (0.806, 0.884) 0.333 (0.269, 0.398) 0.671 (0.626, 0.717) 0.571 (0.483, 0.660) Deep-radiomics model 0.745 (0.703, 0.788) 0.700 (0.699, 0.701) 0.778 (0.733, 0.823) 0.574 (0.506, 0.641) 0.746 (0.700, 0.792) 0.616 (0.547, 0.685) Biopsy pathology 0.554 (0.520, 0.588) 0.629 (0.628, 0.629) 0.875 (0.840, 0.911) 0.230 (0.173, 0.288) 0.647 (0.603, 0.692) 0.534 (0.430, 0.638) Table 5 Efficacy of the five prediction methods across the different hospitals in the external validation datasets AUC ACC SEN SPE PPV NPV Clinical model ISUP_2 (N = 152) Hospital_2 0.575 (0.285, 0.865) 0.611 (0.585, 0.637) 0.500 (0.190, 0.810) 0.750 (0.450, 1.000) 0.714 (0.380, 1.049) 0.545 (0.251, 0.840) Hosptial_3 0.638 (0.564, 0.711) 0.623 (0.621, 0.625) 0.692 (0.580, 0.805) 0.596 (0.522, 0.671) 0.402 (0.311, 0.493) 0.832 (0.765, 0.899) Hosptial_4 0.545 (0.476, 0.615) 0.574 (0.572, 0.576) 0.596 (0.506, 0.687) 0.559 (0.484, 0.633) 0.476 (0.394, 0.557) 0.674 (0.596, 0.751) ISUP_3 ~ 5 (N = 329) Hospital_2 0.575 (0.285, 0.865) 0.611 (0.585, 0.637) 0.750 (0.450, 1.000) 0.500 (0.190, 0.810) 0.545 (0.251, 0.840) 0.714 (0.380, 1.049) Hosptial_3 0.706 (0.639, 0.773) 0.671 (0.669, 0.673) 0.648 (0.571, 0.726) 0.709 (0.613, 0.805) 0.790 (0.717, 0.863) 0.545 (0.452, 0.637) Hosptial_4 0.579 (0.512, 0.645) 0.581 (0.579, 0.583) 0.573 (0.492, 0.654) 0.589 (0.507, 0.670) 0.586 (0.504, 0.667) 0.576 (0.496, 0.657) PIRADS category ISUP_2 (N = 152) Hospital_2 0.675 (0.429, 0.921) 0.667 (0.642, 0.691) 0.800 (0.552, 1.000) 0.500 (0.154, 0.846) 0.667 (0.400, 0.933) 0.667 (0.289, 1.044) Hosptial_3 0.616 (0.550, 0.682) 0.597 (0.595, 0.599) 0.566 (0.485, 0.646) 0.651 (0.550, 0.752) 0.732 (0.650, 0.814) 0.471 (0.381, 0.560) Hosptial_4 0.577 (0.515, 0.639) 0.623 (0.622, 0.625) 0.325 (0.239, 0.411) 0.824 (0.766, 0.881) 0.552 (0.433, 0.671) 0.645 (0.581, 0.709) ISUP_3 ~ 5 (N = 329) Hospital_2 0.675 (0.429, 0.921) 0.667 (0.642, 0.691) 0.500 (0.154, 0.846) 0.800 (0.552, 1.000) 0.667 (0.289, 1.044) 0.667 (0.400, 0.933) Hosptial_3 0.616 (0.550, 0.682) 0.597 (0.595, 0.599) 0.566 (0.485, 0.646) 0.651 (0.550, 0.752) 0.732 (0.650, 0.814) 0.471 (0.381, 0.560) Hosptial_4 0.567 (0.508, 0.626) 0.556 (0.555, 0.558) 0.818 (0.755, 0.881) 0.291 (0.216, 0.366) 0.539 (0.473, 0.605) 0.612 (0.495, 0.729) Radiomics model ISUP_2 (N = 152) Hospital_2 0.525 (0.227, 0.823) 0.667 (0.642, 0.691) 0.900 (0.714, 1.000) 0.375 (0.040, 0.710) 0.643 (0.392, 0.894) 0.750 (0.326, 1.174) Hosptial_3 0.550 (0.471, 0.629) 0.442 (0.439, 0.444) 0.892 (0.817, 0.968) 0.265 (0.198, 0.332) 0.322 (0.254, 0.390) 0.863 (0.768, 0.957) Hosptial_4 0.519 (0.449, 0.588) 0.556 (0.555, 0.558) 0.482 (0.391, 0.574) 0.606 (0.532, 0.679) 0.451 (0.363, 0.539) 0.636 (0.562, 0.710) ISUP_3 ~ 5 (N = 329) Hospital_2 0.512 (0.214, 0.811) 0.667 (0.642, 0.691) 0.375 (0.040, 0.710) 0.900 (0.714, 1.000) 0.750 (0.326, 1.174) 0.643 (0.392, 0.894) Hosptial_3 0.617 (0.542, 0.691) 0.558 (0.556, 0.561) 0.455 (0.374, 0.536) 0.733 (0.639, 0.826) 0.742 (0.651, 0.833) 0.444 (0.362, 0.525) Hosptial_4 0.557 (0.490, 0.624) 0.563 (0.562, 0.565) 0.916 (0.871, 0.962) 0.206 (0.139, 0.272) 0.539 (0.476, 0.602) 0.707 (0.568, 0.847) Deep-radiomics model ISUP_2 (N = 152) Hospital_2 0.725 (0.474, 0.976) 0.722 (0.700, 0.744) 0.900 (0.714, 1.000) 0.500 (0.154, 0.846) 0.692 (0.441, 0.943) 0.800 (0.449, 1.151) Hosptial_3 0.720 (0.651, 0.789) 0.632 (0.630, 0.634) 0.769 (0.667, 0.872) 0.578 (0.503, 0.653) 0.417 (0.328, 0.505) 0.865 (0.801, 0.928) Hosptial_4 0.696 (0.635, 0.756) 0.616 (0.615, 0.618) 0.746 (0.666, 0.826) 0.529 (0.454, 0.604) 0.515 (0.439, 0.591) 0.756 (0.679, 0.833) ISUP_3 ~ 5 (N = 329) Hospital_2 0.713 (0.461, 0.964) 0.667 (0.642, 0.691) 0.875 (0.646, 1.000) 0.500 (0.190, 0.810) 0.583 (0.304, 0.862) 0.833 (0.535, 1.132) Hosptial_3 0.770 (0.709, 0.831) 0.654 (0.652, 0.656) 0.545 (0.464, 0.626) 0.837 (0.759, 0.915) 0.849 (0.777, 0.922) 0.522 (0.438, 0.605) Hosptial_4 0.739 (0.683, 0.796) 0.665 (0.664, 0.667) 0.874 (0.820, 0.928) 0.454 (0.372, 0.536) 0.619 (0.552, 0.686) 0.780 (0.691, 0.870) Biopsy pathology ISUP_2 (N = 152) Hospital_2 0.588 (0.394, 0.781) 0.556 (0.529, 0.582) 0.300 (0.016, 0.584) 0.875 (0.646, 1.000) 0.750 (0.326, 1.174) 0.500 (0.238, 0.762) Hosptial_3 0.527 (0.483, 0.571) 0.706 (0.704, 0.707) 0.123 (0.043, 0.203) 0.934 (0.896, 0.972) 0.421 (0.199, 0.643) 0.731 (0.671, 0.791) Hosptial_4 0.615 (0.555, 0.674) 0.630 (0.629, 0.632) 0.535 (0.444, 0.627) 0.694 (0.625, 0.763) 0.540 (0.448, 0.632) 0.690 (0.621, 0.759) ISUP_3 ~ 5 (N = 329) Hospital_2 0.587 (0.394, 0.781) 0.556 (0.529, 0.582) 0.875 (0.646, 1.000) 0.300 (0.016, 0.584) 0.500 (0.238, 0.762) 0.750 (0.326, 1.174) Hosptial_3 0.536 (0.496, 0.576) 0.641 (0.639, 0.643) 0.945 (0.908, 0.982) 0.128 (0.057, 0.198) 0.646 (0.582, 0.711) 0.579 (0.357, 0.801) Hosptial_4 0.598 (0.541, 0.656) 0.592 (0.590, 0.593) 0.692 (0.617, 0.768) 0.489 (0.407, 0.572) 0.579 (0.505, 0.653) 0.611 (0.521, 0.701) Figure 4a to Fig. 4e represent the confusion matrices of the training dataset, while Fig. 4f to Fig. 4j depict the confusion matrices of the external validation dataset. Figure 4a and Fig. 4f display the predictions of the clinical model. Figure 4b and Fig. 4g show the classification results based on the PIRADS category. Figure 4c and Fig. 4h show the predictions of the radiomics model. Figure 4d and Fig. 4i show the predictions of the deep radiomics model. Finally, Fig. 4e and Fig. 4j illustrate the classification outcomes according to the biopsy results. When observed by columns, the first to three columns display the true values for ISUP_1, ISUP_2, and ISUP_3, respectively, while the fourth column represents the total true values. When observed by rows, from bottom to top, the first to three rows represent the predicted values for ISUP_1, ISUP_2, and ISUP_3, respectively, and the fourth row corresponds to the total predicted values. Figure 5a to Fig. 5c represent the ROC curves of the training dataset, while Fig. 5d to Fig. 5e depict the ROC curves of the external validation dataset. Figure 5a display the ROC curves for various prediction methods targeting ISUP_1. Figure 5b and Fig. 5d illustrate the ROC curves for various prediction methods targeting ISUP_2. Figure 5c and Fig. 5e show the ROC curves for various prediction methods targeting ISUP_3 ~ 5. Figure 6a displays the predictions of the clinical model across different hospitals. Figure 6b shows the classification results based on the PIRADS category across different hospitals. Figure 6c illustrates the predictions of the radiomics model across different hospitals. Figure 6d shows the predictions of the deep-radiomics model across different hospitals. Finally, Fig. 6e shows the classification results based on biopsy results across different hospitals. In Table 4 , evaluation metrics, including AUC, accuracy (ACC), sensitivity (SEN), specificity (SPE), positive predictive value (PPV), negative predictive value (NPV), positive likelihood ratio (PLR), and negative likelihood ratio (NLR), are presented. In the external validation dataset, the AUCs for the prediction methods ranged from 0.597 to 0.718 for the clinical model, 0.554 to 0.613 for the PIRADS score, 0.566 to 0.632 for the radiomics model, 0.700 to 0.791 for the deep-radiomics model, and 0.537 to 0.578 for the biopsy. The ROC curves of the training dataset and the external validation dataset are depicted in Fig. 5. For the prediction of ISUP_1, the AUC of the deep-radiomics model was the highest, surpassing those of the radiomics model, PIRADS score, and biopsy pathology (all P < 0.05); however, the deep-radiomics model did not significantly differ from the clinical model (P = 0.120). For the classification of ISUP_2, the AUC of the deep-radiomics model was the highest compared to those of the other prediction methods (all P < 0.05). For the classification of ISUP_3 ~ 5, the AUC of the deep-radiomics model also outperformed those of the other prediction methods (all P < 0.05). The results of the DeLong test for the comparison of AUCs in discriminating ISUP classes in the external validation dataset are shown in the supplementary material (Table S 16). Table 5 displays the AUCs evaluated across the three hospitals in the external dataset. The AUCs were compared using the Delong test, and the results are outlined in the supplementary material (Table S 17). Since there was no ISUP_1 patient in the data from Hospital_2, a comparison of ISUP_1 could not be performed. For the prediction of ISUP_2, only the AUC of biopsy pathology was significantly different between Hospital_3 and Hospital_4 (P = 0.021). For the prediction of ISUP_3 ~ 5, only the AUC of the clinical model showed a statistically significant difference between Hospital_3 and Hospital_4 (P = 0.008). The PIRADS category, radiomics model, and deep-radiomics model did not significantly differ among the three hospitals in terms of the prediction of the ISUP_2 and ISUP_3 ~ 5 classes (all P > 0.05). The ROC curves for the predictions are illustrated in Fig. 6. Discussion In this study, according to the International Society of Urology (ISUP) grading system, we divided the data into the ISUP 1, ISUP 2 and ISUP 3–5 groups, developed five methods to predict the grouping of PCa ISUP, and evaluated and verified the clinical efficacy of various prediction methods on external multicenter datasets. Our results showed that, compared with other methods, the radiomics model based on deep learning showed the best performance in predicting the invasiveness of PCa (the AUC ranged from 0.700 to 0.791), and there was no statistically significant difference among the other three external validation datasets (all P > 0.05). Our study showed that the radiomics model based on deep learning is promising for providing an objective and noninvasive method for evaluating the invasiveness of PCa, which is helpful for further reducing overtreatment and avoiding unnecessary biopsy and has a certain generalizability. MpMRI can provide important information about the invasiveness of PCa [ 21 , 22 ] , and according to previous studies, ADC value had high value in predicting Gleason score [ 23 , 24 ] . ADC value is calculated on the basis of DWI, which uses the diffusion attenuation of water molecules to reflect the microstructure of living tissues and has good reliability and stability [ 25 ] . The combination of radiomics and ADC in the diagnosis and evaluation of prostate cancer has also been a research hotspot. Several researchers [ 26 ] reported that the accuracy of feature prediction PCa of ADC sequence images can reach 97.39%, which means that better accuracy can be achieved only by extracting image features on ADC maps for prediction. Algohary [ 27 ] et al. reported that intra-tumoral radiomics features on dual parameter MRI can predict the risk of prostate cancer, and the radiomics features are mainly derived from ADC maps. In this study, ADC maps were selected for feature extraction, and two models, traditional radiomics and deep learning radiomics, were developed to evaluate their clinical application value in predicting the invasiveness of PCa. The results of this study showed that the AUC of the traditional radiomics model in the training cohort was 0.609 to 0.882, and the AUC of the three external validation datasets was only 0.566 to 0.632. Several previous single center studies reported that [ 28 – 31 ] the traditional radiomics model performed well in predicting the invasiveness of PCa (AUC value ranged from 0.70 to 0.93); however, this multi-centers study showed low accuracy. The reason may be that there is currently no unified machine learning algorithm for traditional radiomics. The systematic and strict machine learning framework includes feature extraction, classification, cross validation, algorithm selection and statistical analysis. The manual selection of these algorithm steps can affect the efficiency of the model to a certain extent [ 32 ] . Various algorithms have both advantages and disadvantages, and there is currently no unified standard for these algorithms. Improving the reproducibility of traditional radiomics models in different datasets still requires a large number of sample sets for research and validation. Moreover, the deep learning radiomics model showed good efficacy in this study. The model included 13 features extracted from the convolutional network (Table 5 ), and its prediction effect ranged from 0.791 to 0.882 in the training cohort and from 0.700 to 0.791 in the external verification cohort. The use of deep learning algorithms is an important branch of machine learning. Various machine learning algorithms are used on neural networks to learn the internal laws of sample data. Convolutional neural networks (CNNs) are the most representative models in deep learning. Many neural networks with different structures, such as AlexNet, GoogLeNet, ResNet, and DeepLab, have been created. These algorithms can more fully extract data features, thus realizing the algorithm set of various tasks [ 33 ] . Our research results were consistent with previous work. Similarly, Khosravi et al. [ 34 ] developed a model based on GoogLeNet; their research showed that the AUC of the lesions that differentiated between the GG = 1 group and the GG ≥ 2 group was 0.78, while that of lesions that differentiated between the GG ≥ 4 group and the GG = 1 group was 0.86. Aldoj et al. [ 35 ] conducted an analysis based on a multi-channel 3D CNN and used the open source dataset ProstateX to establish a prostate cancer classifier. Research has shown that the sensitivity, specificity, and AUC of the lesions in the GG ≥ 2 group and GG = 1 group were 81.9%, 86.1%, and 0.897, respectively. These results confirmed that deep learning technology can mine high-dimensional features of images, has obvious advantages in building prediction models, provides a new idea for predicting the invasiveness of PCa, and is expected to be applied in the clinical environment in the future. In recent years, although many studies have reported the broad prospects of the application of the radiomics model in prostate cancer [ 13 , 36 , 37 ] , few studies based on multi-center data or focused on invasion have been conducted. Multiple centers of data research verification are highly important for training the applicability and stability of the radiomics model and can promote its clinical application. In this study, there was no statistically significant difference in the prediction of ISUP grouping of PCa lesions by the in-depth learning radiomics model among the three external hospitals (P > 0.05), indicating that the model has strong generalizability among different hospitals, different MRI scanners, different field strengths and different scanning parameters. The reasons may be as follows: (1) This study used a pre-trained AI algorithm to process mpMRI images for focus segmentation and feature extraction, which made the selection of ROIs more accurate and stable and reduced human error. (2) The ISUP groups included in the study all had postoperative RP pathology. Compared with the results of biopsy pathology [ 13 ] , the pathological information in this study was more accurate. Previous studies have shown that the consistency of biopsy pathology and total prostatectomy pathology is only 54% [ 38 ] . (3) In this study, the radiomics model selected lesions and extracted features based on ADC maps. According to previous reports [ 39 ] , there is no statistically significant difference (P = 0.51) in the stability of features extracted from different vendors' scanners and ADC maps with different field strengths, which helps to develop a stable radiomics model and promote clinical application across multiple centers. Our study has several limitations. Although this was a multi-center study, our cohort size was relatively small, and the distribution of patients and ISUP grouping data at each center were unbalanced, which may have affected our findings. Second, this study did not group the PCa in the TZ or PZ. Related studies have reported that [ 40 ] the effective radiomics features for PCa detection and evaluation differ between the TZ and peripheral zone. Future studies could group lesions in the TZ and peripheral zone, extract the radiomics features more accurately, and improve the diagnostic efficiency of the model. This multi-center study confirmed that the deep learning radiomics model using artificial intelligence to extract image features from mpMRI images can potentially predict the invasiveness of PCa and has strong external generalizability; additionally, this model is expected to provide an objective and noninvasive new method for evaluating PCa in the clinic. Abbreviations PCa Prostate cancer MpMRI Multiparametric magnetic resonance imaging AI Artificial intelligence RP Artificial intelligence AUC Area under the curve ROC Receiver operating characteristic ROI Region of interest GS Gleason score GG Gleason score group ISUP International Society of Urology Pathology ADC Apparent diffusion coefficient DWI Diffusion-weighted imaging T2WI T2-weighted imaging PZ Peripheral zone TZ Transitional zone CsPCa Clinically significant prostate cancer Declarations Author contributions Kexin Wang, Ning Luo wrote the main manuscript text. Zhaonan Sun, Xiangpeng Zhao, Lilan She, Zhangli Xing, Yuntian Chen, Chunlei He, Pengsheng Wu, Xiangpeng Wang conducted data collection. ZiXuan Kong reviewed the manuscript. Competing interests The authors declare no competing interests. Data Availability The datasets used and analysed during the current study available from the corresponding author on reasonable request. References WHO. Cancer. 2018. Accessed on October 26, 2020. Available from: https://www.who.int/health-topics/cancer#tab=tab_1 Mottet N BJ, Briers E, Bolla M, Cornford P, De Santis M. European Association of Urology Prostate Cancer Guidelines. 2020. https://uroweb.org/wp-content/uploads/EAU-Guidelines-Prostate-Cancer-2015-v2 Matoso A, Epstein JI (2019) Defining clinically significant prostate cancer on the basis of pathological findings.Histopathology 74(1):135–145 Epstein JI, Egevad L, Amin MB, Delahunt B, Srigley JR, Humphrey PA et al (2016) The 2014 International Society of Urological Pathology (ISUP) Consensus Conference on Gleason Grading of Prostatic Carcinoma:Definition of Grading Patterns and Proposal for a New Grading System. Am J Surg Pathol 40(2):244–252. Lizhi Shao, Ye Yan, Zhenyu Liu, Xiongjun Ye, Haizhui Xia, Xuehua Zhu et al (2020) Radiologist-like artificial intelligence for grade group prediction of radical prostatectomy for reducing upgrading and downgrading from biopsy. Theranostics 10(22):10200–10212. Boorjian SA, Karnes RJ, Crispen PL, Rangel LJ, Bergstralh EJ, Sebo TJ et al (2009) The impact of discordance between biopsy and pathological Gleason scores on survival after radical prostatectomy. J Urol 181: 95–104. Chunmei Li, Min Chen, Jianye Wang, Xuan Wang, Wei Zhang, Chen Zhang (2017) Apparent diffusion coefficient values are superior to transrectal ultrasound-guided prostate biopsy for the assessment of prostate cancer aggressiveness. Acta Radiol 58(2):232–239. Midiri F, Vernuccio F, Purpura P, Pierpaolo A, Tommaso VB (2021) Multiparametric MRI and radiomics in prostate cancer: a review of the current literature. Diagnostics 11(10):1829. Stanzione A, Gambardella M, Cuocolo R, Ponsiglione A, Romeo V, Imbriaco M (2020) Prostate MRI Radiomics: A Systematic Review and Radiomic Quality Score Assessment. Eur J Radiol 129:109095. Castaldo R, Cavaliere C, Soricelli A, Salvatore M, Pecchia L, Franzese M (2021) Radiomic and Genomic Machine Learning Method Performance for Prostate Cancer Diagnosis: Systematic Literature Review. J Med Internet Res 23(4):e22394. Hectors SJ, Chen C, Chen J, Wang J, Sharon G, Miko Y et al (2021) Magnetic resonance imaging radiomics-based machine learning prediction of clinically significant prostate cancer in equivocal PI-RADS 3 lesions. J Magn Reson Imaging 54(5):1466–1473. Chen T, Zhang ZY, Tan SX, Zhang Y, Wei CG, Wang S et al (2022) MRI based radiomics compared with the PI-RADS v2.1 in the prediction of clinically significant prostate cancer: biparametric vs multiparametric MRI. Front Oncol 11:792456. Chen T, Li M, Gu Y, Zhang Y, Yang S, Wei CG et al (2019) Prostate Cancer Differentiation and Aggressiveness: Assessment with a Radiomic-Based Model vs. PI-RADS v2. J Magn Reson Imaging 49(3):875–884. Gabriel A Nketiah, Mattijs Elschot, Tom W Scheenen, Marnix C Maas, Tone F Bathen, Kirsten M Selnæs et al (2021) Utility of T2-weighted MRI texture analysis in assessment of peripheral zone prostate cancer aggressiveness: a single-arm, multicenter study. Sci Rep 11(1):2085. Leandro Donisi, Giuseppe Cesarelli, Anna Castaldo, Davide Raffaele De Lucia, Francesca Nessuno, Gaia Spadarella et al (2021) A Combined Radiomics and Machine Learning Approach to Distinguish Clinically Significant Prostate Lesions on a Publicly Available MRI Dataset. J Imaging 7(10):215. Zhen Kang, Anhui Xu, Liang Wang (2021) Predictive role of T2WI and ADC-derived texture parameters in differentiating Gleason score 3 + 4 and 4 + 3 prostate cancer. J Xray Sci Technol 29(2):307–315. Elena Bertelli, Laura Mercatelli, Chiara Marzi, Eva Pachetti, Michela Baccini, Andrea Barucci et al (2022) Machine and Deep Learning Prediction Of Prostate Cancer Aggressiveness Using Multiparametric MRI. Front Oncol 11:802964 David Bonekamp, Simon Kohl, Manuel Wiesenfarth, Patrick Schelb, Jan Philipp Radtke, Michael Götz et al (2018) Radiomic machine learning for characterization of prostate lesions with MRI: comparison to ADC values. Radiology 289:128–137. Ahmad Chaddad, Michael J Kucharczyk, Tamim Niazi (2018) Multimodal Radiomic Features for the Predicting Gleason Score of Prostate Cancer. Cancers (Basel) 10(8):249. Chao Han, Shuai Ma, Xiang Liu, Yi Liu, Changxin Li, Yaofeng Zhang et al (2021) Radiomics Models Based on Apparent Diffusion Coefficient Maps for the Prediction of High-Grade Prostate Cancer at Radical Prostatectomy: Comparison With Preoperative Biopsy. J Magn Reson Imaging 54(6):1892–1901 Turkbey B, Pinto PA, Mani H, Bernardo M, Pang YX, McKinney YL, et al (2010) Prostate cancer: value of multiparametric MR imaging at 3T for detection–histopathologic correlation. Radiology 255(1):89–99. Donati OF, Mazaheri Y, Afaq A, Vargas HA, Zheng J, Moskowitz CS et al (2014) Prostate cancer aggressiveness: assessment with whole-lesion histogram analysis of the apparent diffusion coefficient. Radiology 271(1):143–152. Hambrock T, Hoeks C, Hulsbergen-van KC, Scheenen T, Fütterer J, Bouwense S et al (2012) Prospective assessment of prostate cancer aggressiveness using 3-T diffusion-weighted magnetic resonance imaging-guided biopsies versus a systematic 10-core transrectal ultrasound prostate biopsy cohort. Eur Urol 61(1):177–184. Felker ER, Margolis DJ, Nassiri N, Marks LS (2016) Prostate cancer risk stratification with magnetic resonance imaging. Urol Oncol 34(7):311–319. Yukihisa Takayama, Riwa Kishimoto, Shouhei Hanaoka, Hiroi Nonaka, Susumu Kandatsu, Hiroshi Tsuji et al (2008) ADC value and diffusion tensor imaging of prostate cancer: changes in carbon-ion radiotherapy. J Magn Reson Imaging 27(6):1331–5 Hiram Shaish, Stella K Kang, Andrew B Rosenkrantz (2017) The utility of quantitative ADC values for differentiating high-risk from low-risk prostate cancer: a systematic review and meta-analysis. Abdom Radiol (NY) 42(1):260–270. Algohary A, Shiradkar R, Pahwa S, Andrei P, Sadhna V, Daniel M et al (2020) Combination of Peri-Tumoral and Intra-Tumoral 4 Radiomic Features on Bi-Parametric MRI Accurately Stratifies Prostate Cancer Risk: A Multi-Site 5 Study. Cancers 12(8):2200. Wibmer A, Hricak H, Gondo T, Matsumoto K, Veeraraghavan H, Fehr D et al (2015) Haralick texture analysis of prostate MRI: Utility for differentiating non-cancerous prostate from prostate cancer and differentiating prostate cancers with different Gleason scores. Eur Radiol 25:2840–2850. Fehr D, Veeraraghavan H, Wibmer A, Gondo T, Matsumoto K, Vargas HA et al (2015) Automatic classification of prostate cancer Gleason scores from multiparametric magnetic resonance images. Proc Natl Acad Sci USA 112:6265–6273. Antonelli M, Johnston EW, Dikaios N, Cheung KK, Sidhu HS, Appayya MB et al (2019) Machine learning classifiers can predict Gleason pattern 4 prostate cancer with greater accuracy than experienced radiologists. Eur Radio 29:4754–4764. Chaddad A, Kucharczyk MJ, Niazi T (2018) Multimodal radiomic features for the predicting gleason score of prostate cancer. Cancers 10(8):249 Sarah Caulfield, Gerard Menezes, Laure Marignol, Claire Poole (2018) Nomograms are key decision-making tools in prostate cancer radiation therapy. Urol Oncol 36(6):283–292. McBee MP, Awan OA, Colucci AT, Ghobadi CW, Kadom N, Kansagra AP et al (2018) Deep learning in radiology. Acad Radiol 25:1472–1480. Khosravi P, Lysandrou M, Eljalby M, Li Q, Kazemi E, Zisimopoulos P et al (2021) A deep learning approach to diagnostic classification of prostate cancer using pathology-radiology fusion. J Magn Reson Imaging 54:462–471. Aldoj N, Lukas S, Dewey M, Penzkofer T (2020) Semi-automatic classification of prostate cancer on multi-parametric MR imaging using a multichannel 3D convolutional neural network. Eur Radiol 30:1243–1253. A Vignati, S Mazzetti, V Giannini, F Russo, E Bollito, F Porpigli et al (2015) Texture features on T2-weighted magnetic resonance imaging: new potential biomarkers for prostate cancer aggressiveness. Phys Med Biol 60:2685–2701. Khémara Gnep, Auréline Fargeas, Ricardo E Gutiérrez-Carvajal, Frédéric Commandeur, Romain Mathieu, Juan D Ospina et al (2017) Haralick textural features on T2-weighted MRI are associated with biochemical recurrence following radiotherapy for peripheral zone prostate cancer. J Magn. Reson Imaging 45:103–117. Evans SM, Patabendi Bandarage V, Kronborg C, Earnest A, Millar J, Clouston D (2016) Gleason group concordance between biopsy and radical prostatectomy specimens: a cohort study from prostate cancer outcome registry-Victoria. Prostate Int 4:145–151. Jurgen Peerlings, Henry C Woodruff, Jessica M Winfield, Abdalla Ibrahim, Bernard E Van Beers, Arend Heerschap et al (2019) Stability of radiomics features in apparent diffusion coefficient maps from a multi-centre test-retest trial. Sci Rep 9(1):4800. Shoshana B Ginsburg, Ahmad Algohary, Shivani Pahwa, Vikas Gulani, Lee Ponsky, Hannu J Aronen et al (2017) Radiomic features for prostate cancer detection on MRI differ between the transition and peripheral zones: preliminary findings from a multi-institutional study. J Magn Reson Imaging 46:184–193. Additional Declarations No competing interests reported. Supplementary Files Supplementarymaterial.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-4252741","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":292662138,"identity":"df633cd5-4483-4772-b76b-8d6d1cb4c202","order_by":0,"name":"Kexin Wang","email":"","orcid":"","institution":"Capital Medical University","correspondingAuthor":false,"prefix":"","firstName":"Kexin","middleName":"","lastName":"Wang","suffix":""},{"id":292662139,"identity":"a1f38d06-fdd9-4549-b3b7-28e13a6229b9","order_by":1,"name":"Ning Luo","email":"","orcid":"","institution":"the Second Affiliated Hospital of Dalian Medical University","correspondingAuthor":false,"prefix":"","firstName":"Ning","middleName":"","lastName":"Luo","suffix":""},{"id":292662140,"identity":"2387f835-32d4-4a6b-a09b-8ad1bc88ef4d","order_by":2,"name":"Zhaonan Sun","email":"","orcid":"","institution":"Peking University First Hospital","correspondingAuthor":false,"prefix":"","firstName":"Zhaonan","middleName":"","lastName":"Sun","suffix":""},{"id":292662141,"identity":"5f9e3655-397e-42c1-8c1d-3b3c5beb4f71","order_by":3,"name":"Xiangpeng Zhao","email":"","orcid":"","institution":"the Second Affiliated Hospital of Dalian Medical University","correspondingAuthor":false,"prefix":"","firstName":"Xiangpeng","middleName":"","lastName":"Zhao","suffix":""},{"id":292662142,"identity":"cd215a7e-8000-4804-9c26-8f1d15fd61b4","order_by":4,"name":"Lilan She","email":"","orcid":"","institution":"Fujian Medical University Union Hospital","correspondingAuthor":false,"prefix":"","firstName":"Lilan","middleName":"","lastName":"She","suffix":""},{"id":292662143,"identity":"95539b36-697e-4799-a12c-187f8ebb67a6","order_by":5,"name":"Zhangli Xing","email":"","orcid":"","institution":"Fujian Medical University Union Hospital","correspondingAuthor":false,"prefix":"","firstName":"Zhangli","middleName":"","lastName":"Xing","suffix":""},{"id":292662144,"identity":"6960473d-9ddd-43e6-a535-2f798ecff1f4","order_by":6,"name":"Yuntian Chen","email":"","orcid":"","institution":"Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Yuntian","middleName":"","lastName":"Chen","suffix":""},{"id":292662145,"identity":"f87b178d-95fa-480e-bfcd-745f18655069","order_by":7,"name":"Chunlei He","email":"","orcid":"","institution":"Sichuan University","correspondingAuthor":false,"prefix":"","firstName":"Chunlei","middleName":"","lastName":"He","suffix":""},{"id":292662146,"identity":"9b91fd12-65fb-40ae-9344-fd4e9e75450f","order_by":8,"name":"Pengsheng Wu","email":"","orcid":"","institution":"Beijing Smart Tree Medical Technology Co. Ltd","correspondingAuthor":false,"prefix":"","firstName":"Pengsheng","middleName":"","lastName":"Wu","suffix":""},{"id":292662147,"identity":"8cb58e3a-58dc-4999-8bf3-e40739b83c12","order_by":9,"name":"Xiangpeng Wang","email":"","orcid":"","institution":"Beijing Smart Tree Medical Technology Co. Ltd","correspondingAuthor":false,"prefix":"","firstName":"Xiangpeng","middleName":"","lastName":"Wang","suffix":""},{"id":292662148,"identity":"aff4d7d7-7f4d-4597-bd77-83a1cf0f4e55","order_by":10,"name":"ZiXuan Kong","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAwElEQVRIiWNgGAWjYDCCAwwMjA1Amp+Z+fAD0rRItrOlGZCmxeA8j4IEUTr4bh9/Jjmj4o7d5sM8DAYMNTbRBLVInktINtxw5lnytsO8Bx4wHEvLbSCkxeAMw8GHD9sOJ5sd5kswYGw4TIwWxoaDIC3GzTwGEkRqYWZ8uLHtsJ0BM7FaJM+wMRvOOHM4QeIwMJATiPEL3xn2Z5I9FYft+fsPH37wocaGsBYYSASrTCBWOQjYk6J4FIyCUTAKRhgAAEP6Q9fjZYOdAAAAAElFTkSuQmCC","orcid":"","institution":"the Second Affiliated Hospital of Dalian Medical University","correspondingAuthor":true,"prefix":"","firstName":"ZiXuan","middleName":"","lastName":"Kong","suffix":""}],"badges":[],"createdAt":"2024-04-11 13:27:18","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4252741/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4252741/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":55321428,"identity":"fc1bd1ef-f4f2-46e3-bff0-5a7e034f605d","added_by":"auto","created_at":"2024-04-25 16:21:57","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":96200,"visible":true,"origin":"","legend":"\u003cp\u003eData enrollment process\u003c/p\u003e","description":"","filename":"image1.png","url":"https://assets-eu.researchsquare.com/files/rs-4252741/v1/749566f7a81301434afeaf75.png"},{"id":55322712,"identity":"76efecf0-5940-43fe-859d-5fb81921b816","added_by":"auto","created_at":"2024-04-25 16:29:58","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":228145,"visible":true,"origin":"","legend":"\u003cp\u003eExtraction of image features and construction of prediction methods\u003c/p\u003e","description":"","filename":"image2.png","url":"https://assets-eu.researchsquare.com/files/rs-4252741/v1/47794605793b73d6f3959e63.png"},{"id":55321430,"identity":"52ef7722-d484-42dc-a816-5bb58dfbbd17","added_by":"auto","created_at":"2024-04-25 16:21:58","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":459838,"visible":true,"origin":"","legend":"\u003cp\u003eDistribution of the clinical variables among the different ISUP classes\u003c/p\u003e","description":"","filename":"image3.png","url":"https://assets-eu.researchsquare.com/files/rs-4252741/v1/2343357f75b6dcc07771fdb9.png"},{"id":55322713,"identity":"ae568728-ffaa-4bc2-9b5b-39f84ebe64ab","added_by":"auto","created_at":"2024-04-25 16:29:58","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":325287,"visible":true,"origin":"","legend":"\u003cp\u003eConfusion matrix of the prediction methods\u003c/p\u003e","description":"","filename":"image4.png","url":"https://assets-eu.researchsquare.com/files/rs-4252741/v1/74ae1aa2ebb7403cff50f196.png"},{"id":55321434,"identity":"68dd9632-7b0b-4369-893d-138764a5333b","added_by":"auto","created_at":"2024-04-25 16:21:58","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":170930,"visible":true,"origin":"","legend":"\u003cp\u003eROC curves of the prediction methods in the training and external validation datasets\u003c/p\u003e","description":"","filename":"image5.png","url":"https://assets-eu.researchsquare.com/files/rs-4252741/v1/b9cdcde6db5f132ff18d58fb.png"},{"id":55321431,"identity":"e071fbaa-c1c5-4e98-bcef-c30636e81e70","added_by":"auto","created_at":"2024-04-25 16:21:58","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":176669,"visible":true,"origin":"","legend":"\u003cp\u003eROC curves of the prediction methods among the hospitals in the external validation cohort\u003c/p\u003e","description":"","filename":"image6.png","url":"https://assets-eu.researchsquare.com/files/rs-4252741/v1/b7c34c2950ddaf3195b70bbc.png"},{"id":55712408,"identity":"9bfc0fcb-1371-482a-beb6-68eeae43ca69","added_by":"auto","created_at":"2024-05-02 06:45:52","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3285425,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4252741/v1/b21645d7-c050-4ec0-a8fa-b508a5e54879.pdf"},{"id":55321432,"identity":"799baa9b-615f-4bec-b2a9-f9a9a3802f83","added_by":"auto","created_at":"2024-04-25 16:21:58","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":2071095,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementarymaterial.docx","url":"https://assets-eu.researchsquare.com/files/rs-4252741/v1/ce80f531e1077689aec663c2.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Evaluating the Feasibility of AI-Predicted mpMRI Image Features for Predicting Prostate Cancer Aggressiveness: a Multicenter Study","fulltext":[{"header":"Introduction","content":"\u003cp\u003eProstate cancer (PCa) is the second most common cancer in men worldwide\u003csup\u003e[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]\u003c/sup\u003e. At present, how to avoid overtreatment of inert prostate cancer was a clinical challenge\u003csup\u003e[\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]\u003c/sup\u003e. Since the selection of a PCa treatment plan and its prognostic value are closely related to its biological behavior, early prediction and accurate judgment of whether the patient's lesion is invasive are clinically important for obtaining the best treatment plan and effect\u003csup\u003e[\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eGleason score (GS) is a commonly used and effective indicator of the invasiveness of PCa. In 2014, the International Society of Urology Pathology (ISUP) proposed the use of the Gleason score group (GG), which was divided into five groups according to the GS, which can better guide clinical work\u003csup\u003e[\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]\u003c/sup\u003e. Transrectal ultrasound (TRUS)-guided biopsy is often used to determine preoperative GS, but due to the randomness of biopsy sampling, it was not possible to accurately evaluate prostate cancer lesions and evaluate tumor heterogeneity. For the above reasons, GS obtained by preoperative biopsy is different from GS obtained by radical prostatectomy (RP), which may cause patients to be overtreated or undertreated\u003csup\u003e[\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]\u003c/sup\u003e. Advances in prostate multiparametric magnetic resonance imaging (mpMRI) and prostate imaging reporting and data system (PI-RADS v2.1) have improved the ability to detect and evaluate prostate cancer. Related research\u003csup\u003e[\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]\u003c/sup\u003e has shown that the apparent diffusion coefficient (ADC) of the PCa region is a better indicator of prostate cancer aggressiveness than is the TRUS biopsy GS. However, the analysis of prostate mpMRIs relies mainly on the experience and subjective judgment of radiologists, resulting in increased variability and decreased reliability of the results\u003csup\u003e[\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eRadiomics research involves the extraction of high-throughput and quantitative features from multimodal medical images and the use of machine learning algorithms to transform these features into high-dimensional mining information related to tumor pathophysiology, which may aid in clinical diagnosis and decision-making\u003csup\u003e[\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]\u003c/sup\u003e. Radiomic models based on mpMRI had been used for risk stratification of PCa patients\u003csup\u003e[\u003cspan additionalcitationids=\"CR12 CR13\" citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]\u003c/sup\u003e. Several studies have shown that the features extracted from T2WI and ADC are helpful for classifying the Gleason score\u003csup\u003e[\u003cspan additionalcitationids=\"CR16\" citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]\u003c/sup\u003e. At present, manual delineation of the region of interest (ROI) is still the main method of radiometric analysis and inevitably consumes certain human resources and time. Most related research reports\u003csup\u003e[\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e, \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]\u003c/sup\u003e at this stage are limited to single-center research. The generalization ability of relevant models and how to reach the level of clinical application still need to be discussed and verified.\u003c/p\u003e \u003cp\u003eIn this study, we used a pre-trained AI algorithm to automatically segment PCa lesions \u003csup\u003e[\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]\u003c/sup\u003e and extract mpMRI features. We developed various models and evaluated their feasibility for predicting PCa invasiveness before surgery. In addition, the clinical efficacy of the prediction model was evaluated and validated on a multicenter dataset.\u003c/p\u003e"},{"header":"Materials and methods","content":"\u003cp\u003e \u003cb\u003eData enrollment\u003c/b\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e This retrospective study received approval from the institutional review board (IRB number: 2021060), all methods were performed in accordance with the relevant guidelines.The Biomedical Research Ethics Committee of Peking University First Hospital granted a waiver for written patient informed consent. The data were collected retrospectively from four distinct hospitals. The study included patients who underwent prostate mpMRI followed by RP between November 2017 and December 2022. This encompassed gathering mpMRI images and pertinent clinical information, such as age, PSA levels, mpMRI results, biopsy pathology, and RP pathology reports. The exclusion criteria were as follows: 1) prior endocrine therapy, 2) benign prostate hyperplasia according to RP pathology, 3) missing PSA data, 4) incomplete biopsy pathology records, 5) incomplete MR images, and 6) evident image artifacts. The process of data enrollment is depicted in Fig.\u0026nbsp;1.\u003c/p\u003e \u003cp\u003e \u003cb\u003eFigure 1\u003c/b\u003e Data enrollment process\u003c/p\u003e \u003cp\u003eFlowchart depicting the process of data inclusion. The training dataset comprises data from Hospital_1, while the external validation dataset includes data from Hospital_2, Hospital_3, and Hospital_4. All the data adhered to the same inclusion and exclusion criteria. In total, the training dataset consisted of 345 included patients, and the external validation dataset included 533 patients, 18 from Hospital_2, 231 from Hospital_3, and 284 from Hospital_4.\u003c/p\u003e \u003cp\u003e \u003cb\u003eReference standard\u003c/b\u003e \u003c/p\u003e \u003cp\u003eAll patients underwent both biopsy and RP, and pathology samples were available for all patients. Pathology was assessed by experienced pathologists following the guidelines of the International Society of Urological Pathology (ISUP) group. The reference standard was established using the RP pathology results, categorizing the patients into three groups: ISUP_1, ISUP_2, and ISUP_3\u0026thinsp;~\u0026thinsp;5.\u003c/p\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eMR scanning protocols\u003c/h2\u003e \u003cp\u003eThe mpMRI images in this study were acquired with 13 MR scanners from five different vendors. Detailed information regarding the MR scanners and image acquisition protocols can be found in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eMR scanning protocols used in different hospitals\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"8\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eOverall\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eTraining dataset\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colspan=\"3\" nameend=\"c7\" namest=\"c5\"\u003e \u003cp\u003eExternal validation dataset (N\u0026thinsp;=\u0026thinsp;533)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003e\u003cem\u003eP\u003c/em\u003e value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;878)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHospital_1\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eHospital_2\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eHospital_3\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eHospital_4\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;345)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;18)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;231)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;284)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eMR scanner\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMagnetic field\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1.436 T\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5 (0.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e5 (1.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1.5 T\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e229 (26.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e19 (5.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e45 (19.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e165 (58.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e3.0 T\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e644 (73.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e326 (94.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e18 (100%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e186 (80.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e114 (40.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStation name\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGE-Scanner_1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e262 (29.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e262 (75.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGE-Scanner_2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e37 (4.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e37 (16.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGE-Scanner_3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e159 (18.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1 (5.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e44 (19.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e114 (40.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePHILIPS_Scanner_1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e38 (4.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e38 (11.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePHILIPS_Scanner_2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1 (0.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1 (0.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePHILIPS_Scanner_3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1 (0.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1 (0.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSIEMENS_Scanner_1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e14 (1.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e14 (77.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSIEMENS_Scanner_2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e43 (4.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e43 (18.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSIEMENS_Scanner_3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e64 (7.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e64 (27.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSIEMENS_Scanner_4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e19 (2.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e19 (5.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSIEMENS_Scanner_5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3 (0.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3 (16.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSIEMENS_Scanner_6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e208 (23.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e43 (18.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e165 (58.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUIH_Scanner_1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e29 (3.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e24 (7.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e5 (1.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eDWI/ADC protocol\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eB value (10^-6 s/mm2)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e800\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e8 (0.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e6 (1.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1 (5.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1 (0.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e288 (32.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2 (11.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e8 (3.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e278 (97.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1200\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e97 (11.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e96 (41.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1 (0.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1400\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e377 (42.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e285 (82.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1 (5.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e87 (37.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e4 (1.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e1500\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e52 (5.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e12 (66.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e40 (17.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e56 (6.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e54 (15.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2 (11.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRepetition time (ms)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian [Q1,Q3]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3000 [2630,3800]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2680 [2650,3000]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2500 [2500,3660]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e4310 [3000,4850]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e3000 [2090,3000]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEcho time (ms)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian [Q1,Q3]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e61.7 [59.9,70.0]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e61.0 [59.8,61.5]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e91.0 [82.8,91.0]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e70.0 [63.0,74.0]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e70.0 [59.7,70.0]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePixel bandwidth (MHz)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian [Q1,Q3]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1950 [1540,1950]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1950 [1950,1950]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1300 [1300,1500]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1630 [1540,1950]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1540 [1540,1950]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSlice thickness (mm)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian [Q1,Q3]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3.50 [3.00,4.00]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4.00 [4.00,4.00]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4.00 [4.00,4.00]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e3.00 [3.00,3.50]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e3.50 [3.00,3.50]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSlice spacing (mm)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian [Q1,Q3]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4.00 [3.50,4.00]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4.00 [4.00,4.00]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4.80 [4.80,4.80]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e3.60 [3.30,4.00]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e3.50 [3.50,4.00]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eReconstruction diameter (ms)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian [Q1,Q3]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e230 [200,240]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e240 [220,240]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e250 [250,250]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e221 [200,260]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e200 [200,340]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePixel spacing (ms)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian [Q1,Q3]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.25 [0.938,1.63]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.938 [0.938,0.938]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1.30 [1.30,1.30]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1.33 [0.877,1.63]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1.56 [1.33,1.79]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eT2WI protocol\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRepetition time (ms)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian [Q1,Q3]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3560 [3130,4470]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3130 [3020,3500]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4000 [4000,4000]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e3710 [3340,4120]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e5400 [3410,5540]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEcho time (ms)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian [Q1,Q3]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e105 [88.9,112]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e88.8 [87.3,93.0]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e100 [100,100]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e108 [96.0,112]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e112 [106,112]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePixel bandwidth (MHz)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian [Q1,Q3]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e163 [163,200]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e163 [163,163]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e203 [203,203]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e190 [160,200]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e200 [163,200]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSlice thickness (mm)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian [Q1,Q3]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3.50 [3.00,4.00]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4.00 [4.00,4.00]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3.00 [3.00,3.00]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e3.00 [3.00,3.50]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e3.50 [3.00,3.50]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSlice spacing (mm)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian [Q1,Q3]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4.00 [3.50,4.00]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4.00 [4.00,4.00]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3.60 [3.60,3.60]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e3.60 [3.50,4.00]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e3.50 [3.50,4.00]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eReconstruction diameter (ms)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian [Q1,Q3]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e200 [200,240]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e240 [240,240]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e178 [178,178]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e200 [200,220]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e200 [200,200]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePixel spacing (ms)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian [Q1,Q3]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.469 [0.391,0.625]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.469 [0.469,0.469]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.625 [0.625,0.625]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.391 [0.344,0.625]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.625 [0.391,0.625]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003eImage features extracted by AI algorithms\u003c/h2\u003e \u003cp\u003eFollowing anonymization, the DICOM files were converted to NIFTI format using dicom2nii.py (Python 3.5) and then input into pretrained AI software. For detailed technical information about the AI software, please refer to the supplementary material.\u003c/p\u003e \u003cp\u003eFor each enrolled patient, the largest lesion predicted by the AI algorithms on the mpMRI images was designated the index lesion for analysis. The lesion's location was classified as solely in the peripheral zone (PZ), solely in the transitional zone (TZ), or in both the peripheral and transitional zones (PZ\u0026thinsp;+\u0026thinsp;TZ). The following image features were computed and utilized in the prediction model: 1) lesion volume, 2) diameter of the lesion, signifying the largest diameter of the lesion, 3) mean ADC value of the lesion (ADClesion), 4) mean signal intensity of the lesion on DWI (DWIlesion), and 5) mean signal intensity of the lesion on T2WI (T2WIlesion) (as illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eStep 1 involves selecting and inputting anonymized mpMRI images, which encompass T2WI, DWI, and ADC maps. In Step 2, the pre-trained AI models are employed to identify suspicious areas indicative of PCa. Step 3 entails identifying the index lesion by selecting the largest lesion from the predicted labels. The index lesion was annotated as the red areas on the images. Step 4 encompasses extracting image features from the index lesion, encompassing 1) clinical image features such as lesion volume, lesion diameter, ADC value, DWI signal intensity, and T2WI intensity, 2) conventional radiomic features, and 3) deep radiomic features. In Step 5, various prediction methods are developed, including a) biopsy prediction, which forecasts ISUP grouping based on biopsy pathology; b) PIRADS prediction, which anticipates ISUP grouping based on the PIRADS category; c) a clinical model, which incorporates age, PSA, PIRADS category, and clinical image features to predict ISUP grouping; d) radiomics model, which forecasts ISUP grouping using conventional radiomic features obtained from the index lesion; and e) deep-radiomics model, which forecasts ISUP grouping using deep-radiomic features extracted from the index lesion.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003ePrediction methods\u003c/h2\u003e \u003cp\u003eTo predict the ISUP_1, ISUP_2, and ISUP_3\u0026thinsp;~\u0026thinsp;5 classes, five prediction methods were developed, namely, the PIRADS category, biopsy pathology, a clinical model, a radiomics model, and a deep-radiomics model. The steps involved in developing the prediction methods are illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e2\u003c/span\u003e. The training dataset utilized was sourced from Hospital_1, while the external validation dataset included data from Hospital_2, Hospital_3, and Hospital_4.\u003c/p\u003e \u003cp\u003eFor the construction of the clinical model, clinical variables were employed as covariates. These included age, PSA levels, PIRADS score, and image features of the suspected clinically significant prostate cancer (csPCa) lesion, which included position, volume, diameter, ADC value, signal intensity on DWI, and signal intensity on T2WI.\u003c/p\u003e \u003cp\u003eTo construct the radiomics model, image features were derived from regions of interest (ROIs) on the ADC maps using the PyRadiomics package in Python (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://pyradiomics.readthedocs.io/en/latest/changes.html\u003c/span\u003e\u003cspan address=\"https://pyradiomics.readthedocs.io/en/latest/changes.html\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). Z score normalization was applied to standardize the extracted features, and Pearson correlation coefficients (PCCs) were computed to identify highly correlated features. Features with a PCC value exceeding 0.99 were eliminated to mitigate multicollinearity. The Kruskal‒Wallis test was then employed to select the features for the final radiomics model. For the classifier, the least absolute shrinkage and selection operator (LASSO) algorithm was utilized.\u003c/p\u003e \u003cp\u003eTo construct the deep-radiomics model, image features were extracted using a deep learning algorithm. The construction process involved several key steps: 1) Preprocessing ADC maps: Initial preprocessing included normalizing the intensities of the ADC maps. 2) ROI resampling: ROIs were resampled to ensure a consistent voxel size. 3) Deep Feature Extraction: A pretrained deep learning model was used to extract features from the segmented ROIs. 4) Feature Dimension Reduction: The resulting channel feature maps underwent dimension reduction by filtering with the maximum value, resulting in a set of 512 one-dimensional features. 5) Model Construction: After extracting the deep features, the construction of the deep-radiomics model followed a procedure similar to that of the radiomics model, as previously described.\u003c/p\u003e \u003cp\u003eFor model tuning, 5-fold cross-validation was employed to select the optimal value of the hyperparameter α, which controls the strength of the L1 penalty. A grid search was performed over a range of α values, and the one yielding the highest cross-validated accuracy was chosen. This process identified a subset of significant predictors associated with ISUP classes.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003ePrediction efficacy evaluation\u003c/h2\u003e \u003cp\u003eThe external dataset was used to assess the prediction effectiveness of the methods via receiver operating characteristic (ROC) analysis. This evaluation was conducted by computing the area under the ROC curve (AUC).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003eStatistical analysis\u003c/h2\u003e \u003cp\u003eThe statistical analysis was conducted using R 4.1.3 software. Quantitative variables with a normal distribution are presented as the mean\u0026thinsp;\u0026plusmn;\u0026thinsp;standard deviation, while those with a nonnormal distribution are presented as the median [Q1, Q3]. Categorical variables are presented as frequencies. The normality of the variables was assessed using the Kolmogorov‒Smirnov test, associations between categorical variables were examined using the chi-square test, and differences between multiple groups were compared using the Kruskal‒Wallis test. The DeLong test was used to compare differences in the area under the curve (AUC) among the five types of prediction methods. A P value less than 0.05 was considered to indicate statistical significance.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003eClinical characteristics\u003c/h2\u003e \u003cp\u003eA total of 878 eligible patients participated in the study, with an average age of 68.7\u0026thinsp;\u0026plusmn;\u0026thinsp;6.6 years. Among these patients, 345 were recruited from Hospital_1, and 18, 231, and 284 were from Hospital_2, Hospital_3, and Hospital_4, respectively. All clinical variables exhibited statistically significant differences among the four hospitals (all P\u0026thinsp;\u0026lt;\u0026thinsp;0.001), except for age (P\u0026thinsp;=\u0026thinsp;0.878). Further details regarding the clinical variables among the different datasets can be found in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eClinical characteristics of patients in the training and external validation cohorts\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"8\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eOverall\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eTraining cohort\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colspan=\"3\" nameend=\"c7\" namest=\"c5\"\u003e \u003cp\u003eExternal validation cohort\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c8\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003e\u003cem\u003eP\u003c/em\u003e value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;878)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eHospital_1\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eHospital_2\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eHospital_3\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eHospital_4\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;345)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;18)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;231)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;284)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge (year)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e0.878\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMean(SD)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e68.7 (6.63)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e68.5 (6.58)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e69.4 (6.71)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e68.8 (6.73)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e68.7 (6.64)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePSA (ng/dl)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian [Q1,Q3]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e11.6 [7.90,19.1]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e10.9 [7.06,15.5]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e9.93 [6.35,14.0]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e12.8 [8.82,23.9]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e12.8 [8.20,22.5]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePIRADS\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePIRADS_1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2 (0.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1 (0.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1 (0.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePIRADS_2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e87 (9.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e12 (3.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4 (22.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e4 (1.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e67 (23.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePIRADS_3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e216 (24.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e45 (13.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e8 (44.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e57 (24.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e106 (37.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePIRADS_4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e286 (32.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e173 (50.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e57 (24.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e56 (19.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePIRADS_5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e287 (32.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e114 (33.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e6 (33.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e112 (48.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e55 (19.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePCa location\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNot measurable\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e137 (15.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e110 (31.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1 (5.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e12 (5.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e14 (4.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePZ\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e80 (9.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e38 (11.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3 (16.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e14 (6.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e25 (8.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePZ\u0026thinsp;+\u0026thinsp;TZ\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e491 (55.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e121 (35.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e10 (55.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e172 (74.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e188 (66.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTZ\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e170 (19.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e76 (22.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4 (22.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e33 (14.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e57 (20.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLesion volume\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian [Q1,Q3]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.60 [0.717,3.93]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.737 [0,1.55]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2.73 [0.976,4.68]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e3.25 [1.44,7.62]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e2.44 [1.16,5.25]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLesion diameter\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian [Q1,Q3]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e24.9 [17.8,33.5]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e18.2 [0,25.4]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e27.5 [23.7,37.2]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e31.1 [23.1,40.0]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e28.1 [21.3,36.2]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eADC\u003csub\u003elesion\u003c/sub\u003e (10^6 mm2/s)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian [Q1,Q3]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e802 [663,907]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e738 [0,865]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e848 [757,957]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e816 [714,909]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e840 [739,926]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDWI\u003csub\u003elesion\u003c/sub\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian [Q1,Q3]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e120 [46.3,404]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e52.2 [0,85.7]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e313 [176,398]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e266 [80.5,510]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e281 [102,506]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTWI\u003csub\u003elesion\u003c/sub\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian [Q1,Q3]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e324 [119,1150]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e127 [0,297]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e579 [436,1490]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e490 [160,1580]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e499 [274,1520]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBiopsy result\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo cancer\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e8 (0.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1 (0.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e7 (2.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eISUP_1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e132 (15.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e49 (14.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e14 (6.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e69 (24.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eISUP_2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e185 (21.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e128 (37.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e18 (7.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e39 (13.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eISUP_3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e331 (37.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e84 (24.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e18 (100%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e159 (68.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e70 (24.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eISUP_4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e133 (15.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e43 (12.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e26 (11.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e64 (22.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eISUP_5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e89 (10.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e41 (11.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e13 (5.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e35 (12.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePR result\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eISUP_1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e62 (7.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e10 (2.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e27 (11.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e25 (8.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eISUP_2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e300 (34.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e148 (42.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e6 (33.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e63 (27.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e83 (29.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eISUP_3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e241 (27.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e93 (27.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3 (16.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e66 (28.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e79 (27.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eISUP_4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e108 (12.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e40 (11.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4 (22.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e27 (11.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e37 (13.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eISUP_5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e167 (19.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e54 (15.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e5 (27.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e48 (20.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e60 (21.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eBased on the RP pathology results, 62 patients were categorized as ISUP_1, 300 as ISUP_2, and 516 as ISUP_3\u0026thinsp;~\u0026thinsp;5. Except for patient age distribution, which did not differ significantly among the three ISUP classes (P\u0026thinsp;=\u0026thinsp;0.563), all the other clinical variables demonstrated significant differences among the various ISUP classes (all P\u0026thinsp;\u0026lt;\u0026thinsp;0.05). Moreover, there was a statistically significant difference in the distribution of ISUP classes among the four hospitals (P\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Detailed information about the variables in different ISUP classes can be found in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e. The distributions of these variables across different ISUP classes are illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e3\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eClinical characteristics of patients in different ISUP classes\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eOverall\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eISUP_1\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eISUP_2\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eISUP_3\u0026thinsp;~\u0026thinsp;5\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e\u003cem\u003eP\u003c/em\u003e value\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;878)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;62)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;300)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003e(N\u0026thinsp;=\u0026thinsp;516)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge (year)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.563\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMean(SD)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e68.7(6.63)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e69.6(6.91)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e68.5(6.74)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e68.6(6.54)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePSA (ng/dl)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian [Q1,Q3]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e11.6 [7.90,19.1]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e11.9 [9.37,18.4]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e10.8 [6.99,15.8]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e12.3 [8.18,21.4]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePIRADS\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePIRADS_1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2 (0.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1 (0.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1 (0.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePIRADS_2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e87 (9.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e16 (25.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e29 (9.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e42 (8.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePIRADS_3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e216 (24.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e20 (32.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e81 (27.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e115 (22.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePIRADS_4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e286 (32.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e15 (24.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e119 (39.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e152 (29.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePIRADS_5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e287 (32.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e11 (17.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e70 (23.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e206 (39.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePCa location\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNot measurable\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e137 (15.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e7 (11.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e86 (28.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e44 (8.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePZ\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e80 (9.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e7 (11.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e39 (13.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e34 (6.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTZ\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e170 (19.4%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e14 (22.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e67 (22.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e89 (17.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePZ\u0026thinsp;+\u0026thinsp;TZ\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e491 (55.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e34 (54.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e108 (36.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e349 (67.6%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLesion volume\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian [Q1,Q3]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.60 [0.717,3.93]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.66 [0.869,2.84]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.887 [0,1.90]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2.37 [0.973,5.87]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLesion diameter\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian [Q1,Q3]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e24.9 [17.8,33.5]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e25.4 [19.1,31.1]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e19.6 [0,26.1]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e28.1 [20.6,38.1]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eADC\u003csub\u003elesion\u003c/sub\u003e (10^6 mm2/s)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian [Q1,Q3]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e802 [663,907]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e820 [732,915]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e773 [0,892]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e813 [697,907]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDWI\u003csub\u003elesion\u003c/sub\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian [Q1,Q3]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e120 [46.3,404]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e101 [54.5,282]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e68.6 [0,287]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e205 [61.4,489]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTWI\u003csub\u003elesion\u003c/sub\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMedian [Q1,Q3]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e324 [119,1150]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e303 [122,476]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e162 [0,576]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e434 [132,1390]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBiopsy result\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo cancer\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e8 (0.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2 (3.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1 (0.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e5 (1.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eISUP_1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e132 (15.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e33 (53.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e62 (20.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e37 (7.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eISUP_2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e185 (21.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3 (4.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e111 (37.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e71 (13.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eISUP_3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e331 (37.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e22 (35.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e99 (33.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e210 (40.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eISUP_4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e133 (15.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2 (3.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e21 (7.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e110 (21.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eISUP_5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e89 (10.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e6 (2.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e83 (16.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDataset\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTraining dataset\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e345 (39.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e10 (16.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e148 (49.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e187 (36.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTest dataset\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e533 (60.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e52 (83.9%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e152 (50.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e329 (63.8%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eInstitute name\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHospital_1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e345 (39.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e10 (16.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e148 (49.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e187 (36.2%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHospital_2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e18 (2.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0 (0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e6 (2.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e12 (2.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHospital_3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e231 (26.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e27 (43.5%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e63 (21.0%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e141 (27.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHospital_4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e284 (32.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e25 (40.3%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e83 (27.7%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e176 (34.1%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe PSA map displays the distribution of PSA levels. Columns 1 to 4 show the PSA values for lesions classified as \"not measurable,\" lesions in the PZ, lesions in the TZ, and lesions in both the PZ and TZ, respectively. Column 5 displays the PSA values for all lesions. Rows 1 to 5 display the PSA values for lesions categorized as PIRADS_1 to PIRADS_5, and row 6 shows the PSA values for all lesions.\u003c/p\u003e \u003cp\u003eThe distribution of measurements for lesion volume, lesion diameter, ADC value, DWI signal intensity, and T2WI signal intensity is presented in a similar manner. Columns 1 to 3 display the measurements for lesions in the PZ, lesions in the TZ, and lesions in both the PZ and TZ, respectively. Column 4 represents the measurements for all lesions. When observed by rows, rows 1 to 4 show the measurements for lesions categorized as PIRADS_2 to PIRADS_5, and row 5 displays the measurements for all lesions.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003eModel development metrics\u003c/h2\u003e \u003cp\u003eOrdinal logistic regression analysis results indicated that age, PSA level, lesion location, and the ADC value were statistically significant predictors (all P\u0026thinsp;\u0026lt;\u0026thinsp;0.001), while PIRADS, volume, diameter, DWI intensity, and T2WI intensity did not significantly differ (all P\u0026thinsp;\u0026gt;\u0026thinsp;0.05). The model exhibited a good fit to the data, as evidenced by the Brant test (P\u0026thinsp;\u0026gt;\u0026thinsp;0.05). Additional details of the ordinal logistic regression model can be found in the supplementary material (Table S 14).\u003c/p\u003e \u003cp\u003eThe radiomics model was constructed using the LASSO method, which identified 13 critical features (3 first-order features and 10 texture features) for inclusion in the final model. The deep-radiomics model incorporated 13 features extracted from the convolutional network. The development of the radiomics and deep-radiomics LASSO classifiers is depicted in the supplementary material (Figure S 8). The coefficients of the selected features for the radiomics model and the deep-radiomics model are presented in the supplementary material (Table S 15).\u003c/p\u003e \u003cp\u003eThe prediction results are visualized in the form of a confusion matrix, as illustrated in Fig.\u0026nbsp;4.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eModel evaluation\u003c/h2\u003e \u003cp\u003eThe performance of the clinical model, PIRADS category, radiomics model, deep-radiomics model, and biopsy pathology was assessed using ROC analysis, and the results are detailed in Figs.\u0026nbsp;5\u0026thinsp;~\u0026thinsp;6 and Tables\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e\u0026thinsp;~\u0026thinsp;5.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eEfficacy of the five prediction methods in the training and external validation datasets\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"7\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAUC\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eACC\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSEN\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eSPE\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003ePPV\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eNPV\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTraining dataset\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eISUP_1 (N\u0026thinsp;=\u0026thinsp;10)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eClinical model\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.815 (0.695, 0.934)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.736 (0.735, 0.737)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.900 (0.714, 1.000)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.731 (0.684, 0.779)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.091 (0.034, 0.148)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.996 (0.988, 1.004)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePIRADS category\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.507 (0.293, 0.722)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.838 (0.837, 0.838)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.300 (0.016, 0.584)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.854 (0.816, 0.892)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.058 (-0.006, 0.121)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.976 (0.959, 0.994)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRadiomics model\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.852 (0.741, 0.963)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.849 (0.849, 0.850)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.800 (0.552, 1.000)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.851 (0.813, 0.889)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.138 (0.049, 0.227)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.993 (0.983, 1.003)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDeep-radiomics model\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.882 (0.819, 0.946)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.739 (0.738, 0.740)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.000 (1.000, 1.000)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.731 (0.684, 0.779)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.100 (0.041, 0.159)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1.000 (1.000, 1.000)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBiopsy pathology\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.799 (0.653, 0.946)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.820 (0.819, 0.821)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.800 (0.552, 1.000)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.821 (0.780, 0.862)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.118 (0.041, 0.194)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.993 (0.983, 1.003)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eISUP_2 (N\u0026thinsp;=\u0026thinsp;148)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eClinical model\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.666 (0.609, 0.723)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.635 (0.633, 0.636)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.784 (0.717, 0.850)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.523 (0.453, 0.593)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.552 (0.485, 0.620)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.763 (0.691, 0.835)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePIRADS category\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.604 (0.556, 0.653)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.609 (0.607, 0.610)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.243 (0.174, 0.312)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.883 (0.838, 0.928)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.610 (0.486, 0.735)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.608 (0.552, 0.665)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRadiomics model\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.609 (0.549, 0.668)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.583 (0.581, 0.584)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.642 (0.565, 0.719)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.538 (0.468, 0.608)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.511 (0.439, 0.583)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.667 (0.593, 0.740)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDeep-radiomics model\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.791 (0.745, 0.837)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.713 (0.712, 0.714)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.878 (0.826, 0.931)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.589 (0.520, 0.658)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.616 (0.550, 0.682)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.866 (0.808, 0.923)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBiopsy pathology\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.565 (0.515, 0.616)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.597 (0.596, 0.598)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.385 (0.307, 0.464)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.756 (0.696, 0.816)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.543 (0.448, 0.638)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.621 (0.559, 0.682)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eISUP_3\u0026thinsp;~\u0026thinsp;5 (N\u0026thinsp;=\u0026thinsp;187)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eClinical model\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.699 (0.644, 0.754)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.661 (0.660, 0.662)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.561 (0.490, 0.633)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.778 (0.714, 0.843)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.750 (0.678, 0.822)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.600 (0.533, 0.667)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePIRADS category\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.602 (0.553, 0.651)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.597 (0.596, 0.598)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.893 (0.849, 0.937)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.247 (0.180, 0.314)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.584 (0.527, 0.641)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.661 (0.540, 0.782)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRadiomics model\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.649 (0.591, 0.707)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.635 (0.633, 0.636)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.786 (0.727, 0.845)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.456 (0.378, 0.533)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.631 (0.569, 0.693)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.643 (0.554, 0.732)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDeep-radiomics model\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.808 (0.764, 0.852)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.716 (0.715, 0.717)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.578 (0.507, 0.648)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.880 (0.829, 0.930)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.850 (0.788, 0.912)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.638 (0.574, 0.701)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBiopsy pathology\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.599 (0.549, 0.648)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.614 (0.613, 0.616)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.786 (0.727, 0.845)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.411 (0.335, 0.488)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.613 (0.551, 0.674)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.619 (0.526, 0.712)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e \u003cp\u003e\u003cb\u003eExternal validation dataset\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eISUP_1 (N\u0026thinsp;=\u0026thinsp;52)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eClinical model\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.718 (0.652, 0.785)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.717 (0.716, 0.717)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.635 (0.504, 0.765)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.726 (0.686, 0.765)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.200 (0.139, 0.261)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.948 (0.926, 0.971)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePIRADS category\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.554 (0.485, 0.623)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.475 (0.474, 0.476)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.673 (0.546, 0.801)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.453 (0.409, 0.498)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.117 (0.081, 0.154)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.928 (0.895, 0.961)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRadiomics model\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.632 (0.549, 0.714)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.829 (0.829, 0.830)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.365 (0.235, 0.496)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.879 (0.850, 0.909)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.247 (0.150, 0.343)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.928 (0.904, 0.951)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDeep-radiomics model\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.791 (0.728, 0.854)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.871 (0.870, 0.871)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.538 (0.403, 0.674)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.906 (0.880, 0.932)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.384 (0.272, 0.495)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.948 (0.928, 0.968)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBiopsy pathology\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.537 (0.492, 0.582)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.244 (0.243, 0.245)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.904 (0.824, 0.984)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.173 (0.139, 0.206)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.106 (0.077, 0.134)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.943 (0.895, 0.992)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eISUP_2 (N\u0026thinsp;=\u0026thinsp;152)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eClinical model\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.597 (0.547, 0.646)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.568 (0.568, 0.569)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.651 (0.576, 0.727)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.535 (0.485, 0.586)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.359 (0.302, 0.415)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.794 (0.744, 0.843)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePIRADS category\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.607 (0.561, 0.653)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.557 (0.556, 0.558)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.704 (0.631, 0.777)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.499 (0.448, 0.549)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.359 (0.305, 0.414)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.809 (0.758, 0.859)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRadiomics model\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.566 (0.511, 0.621)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.642 (0.641, 0.642)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.414 (0.336, 0.493)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.732 (0.688, 0.777)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.382 (0.308, 0.456)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.758 (0.714, 0.802)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDeep-radiomics model\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.700 (0.652, 0.748)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.595 (0.594, 0.596)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.789 (0.725, 0.854)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.517 (0.467, 0.567)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.395 (0.340, 0.450)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.860 (0.815, 0.905)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBiopsy pathology\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.578 (0.539, 0.618)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.707 (0.707, 0.708)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.276 (0.205, 0.347)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.879 (0.847, 0.912)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.477 (0.373, 0.582)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.753 (0.713, 0.793)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eISUP_3\u0026thinsp;~\u0026thinsp;5 (N\u0026thinsp;=\u0026thinsp;329)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eClinical model\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.668 (0.622, 0.713)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.598 (0.598, 0.599)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.498 (0.444, 0.553)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.760 (0.701, 0.818)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.770 (0.713, 0.826)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.484 (0.430, 0.539)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePIRADS category\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.613 (0.569, 0.656)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.591 (0.590, 0.592)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.526 (0.472, 0.580)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.696 (0.633, 0.759)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.736 (0.680, 0.793)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.477 (0.420, 0.533)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRadiomics model\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.602 (0.552, 0.652)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.649 (0.648, 0.650)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.845 (0.806, 0.884)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.333 (0.269, 0.398)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.671 (0.626, 0.717)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.571 (0.483, 0.660)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDeep-radiomics model\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.745 (0.703, 0.788)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.700 (0.699, 0.701)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.778 (0.733, 0.823)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.574 (0.506, 0.641)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.746 (0.700, 0.792)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.616 (0.547, 0.685)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eBiopsy pathology\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.554 (0.520, 0.588)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.629 (0.628, 0.629)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.875 (0.840, 0.911)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.230 (0.173, 0.288)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.647 (0.603, 0.692)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.534 (0.430, 0.638)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab5\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 5\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eEfficacy of the five prediction methods across the different hospitals in the external validation datasets\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"7\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eAUC\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eACC\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSEN\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eSPE\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003ePPV\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e \u003cp\u003eNPV\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eClinical model\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eISUP_2 (N\u0026thinsp;=\u0026thinsp;152)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHospital_2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.575 (0.285, 0.865)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.611 (0.585, 0.637)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.500 (0.190, 0.810)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.750 (0.450, 1.000)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.714 (0.380, 1.049)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.545 (0.251, 0.840)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHosptial_3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.638 (0.564, 0.711)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.623 (0.621, 0.625)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.692 (0.580, 0.805)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.596 (0.522, 0.671)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.402 (0.311, 0.493)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.832 (0.765, 0.899)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHosptial_4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.545 (0.476, 0.615)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.574 (0.572, 0.576)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.596 (0.506, 0.687)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.559 (0.484, 0.633)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.476 (0.394, 0.557)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.674 (0.596, 0.751)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eISUP_3\u0026thinsp;~\u0026thinsp;5 (N\u0026thinsp;=\u0026thinsp;329)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHospital_2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.575 (0.285, 0.865)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.611 (0.585, 0.637)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.750 (0.450, 1.000)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.500 (0.190, 0.810)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.545 (0.251, 0.840)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.714 (0.380, 1.049)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHosptial_3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.706 (0.639, 0.773)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.671 (0.669, 0.673)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.648 (0.571, 0.726)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.709 (0.613, 0.805)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.790 (0.717, 0.863)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.545 (0.452, 0.637)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHosptial_4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.579 (0.512, 0.645)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.581 (0.579, 0.583)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.573 (0.492, 0.654)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.589 (0.507, 0.670)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.586 (0.504, 0.667)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.576 (0.496, 0.657)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003ePIRADS category\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eISUP_2 (N\u0026thinsp;=\u0026thinsp;152)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHospital_2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.675 (0.429, 0.921)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.667 (0.642, 0.691)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.800 (0.552, 1.000)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.500 (0.154, 0.846)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.667 (0.400, 0.933)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.667 (0.289, 1.044)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHosptial_3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.616 (0.550, 0.682)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.597 (0.595, 0.599)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.566 (0.485, 0.646)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.651 (0.550, 0.752)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.732 (0.650, 0.814)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.471 (0.381, 0.560)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHosptial_4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.577 (0.515, 0.639)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.623 (0.622, 0.625)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.325 (0.239, 0.411)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.824 (0.766, 0.881)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.552 (0.433, 0.671)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.645 (0.581, 0.709)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eISUP_3\u0026thinsp;~\u0026thinsp;5 (N\u0026thinsp;=\u0026thinsp;329)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHospital_2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.675 (0.429, 0.921)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.667 (0.642, 0.691)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.500 (0.154, 0.846)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.800 (0.552, 1.000)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.667 (0.289, 1.044)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.667 (0.400, 0.933)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHosptial_3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.616 (0.550, 0.682)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.597 (0.595, 0.599)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.566 (0.485, 0.646)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.651 (0.550, 0.752)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.732 (0.650, 0.814)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.471 (0.381, 0.560)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHosptial_4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.567 (0.508, 0.626)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.556 (0.555, 0.558)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.818 (0.755, 0.881)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.291 (0.216, 0.366)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.539 (0.473, 0.605)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.612 (0.495, 0.729)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eRadiomics model\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eISUP_2 (N\u0026thinsp;=\u0026thinsp;152)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHospital_2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.525 (0.227, 0.823)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.667 (0.642, 0.691)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.900 (0.714, 1.000)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.375 (0.040, 0.710)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.643 (0.392, 0.894)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.750 (0.326, 1.174)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHosptial_3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.550 (0.471, 0.629)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.442 (0.439, 0.444)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.892 (0.817, 0.968)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.265 (0.198, 0.332)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.322 (0.254, 0.390)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.863 (0.768, 0.957)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHosptial_4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.519 (0.449, 0.588)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.556 (0.555, 0.558)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.482 (0.391, 0.574)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.606 (0.532, 0.679)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.451 (0.363, 0.539)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.636 (0.562, 0.710)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eISUP_3\u0026thinsp;~\u0026thinsp;5 (N\u0026thinsp;=\u0026thinsp;329)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHospital_2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.512 (0.214, 0.811)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.667 (0.642, 0.691)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.375 (0.040, 0.710)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.900 (0.714, 1.000)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.750 (0.326, 1.174)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.643 (0.392, 0.894)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHosptial_3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.617 (0.542, 0.691)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.558 (0.556, 0.561)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.455 (0.374, 0.536)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.733 (0.639, 0.826)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.742 (0.651, 0.833)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.444 (0.362, 0.525)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHosptial_4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.557 (0.490, 0.624)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.563 (0.562, 0.565)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.916 (0.871, 0.962)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.206 (0.139, 0.272)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.539 (0.476, 0.602)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.707 (0.568, 0.847)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eDeep-radiomics model\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eISUP_2 (N\u0026thinsp;=\u0026thinsp;152)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHospital_2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.725 (0.474, 0.976)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.722 (0.700, 0.744)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.900 (0.714, 1.000)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.500 (0.154, 0.846)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.692 (0.441, 0.943)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.800 (0.449, 1.151)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHosptial_3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.720 (0.651, 0.789)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.632 (0.630, 0.634)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.769 (0.667, 0.872)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.578 (0.503, 0.653)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.417 (0.328, 0.505)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.865 (0.801, 0.928)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHosptial_4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.696 (0.635, 0.756)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.616 (0.615, 0.618)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.746 (0.666, 0.826)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.529 (0.454, 0.604)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.515 (0.439, 0.591)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.756 (0.679, 0.833)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eISUP_3\u0026thinsp;~\u0026thinsp;5 (N\u0026thinsp;=\u0026thinsp;329)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHospital_2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.713 (0.461, 0.964)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.667 (0.642, 0.691)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.875 (0.646, 1.000)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.500 (0.190, 0.810)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.583 (0.304, 0.862)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.833 (0.535, 1.132)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHosptial_3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.770 (0.709, 0.831)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.654 (0.652, 0.656)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.545 (0.464, 0.626)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.837 (0.759, 0.915)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.849 (0.777, 0.922)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.522 (0.438, 0.605)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHosptial_4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.739 (0.683, 0.796)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.665 (0.664, 0.667)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.874 (0.820, 0.928)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.454 (0.372, 0.536)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.619 (0.552, 0.686)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.780 (0.691, 0.870)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eBiopsy pathology\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eISUP_2 (N\u0026thinsp;=\u0026thinsp;152)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHospital_2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.588 (0.394, 0.781)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.556 (0.529, 0.582)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.300 (0.016, 0.584)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.875 (0.646, 1.000)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.750 (0.326, 1.174)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.500 (0.238, 0.762)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHosptial_3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.527 (0.483, 0.571)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.706 (0.704, 0.707)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.123 (0.043, 0.203)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.934 (0.896, 0.972)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.421 (0.199, 0.643)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.731 (0.671, 0.791)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHosptial_4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.615 (0.555, 0.674)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.630 (0.629, 0.632)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.535 (0.444, 0.627)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.694 (0.625, 0.763)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.540 (0.448, 0.632)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.690 (0.621, 0.759)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eISUP_3\u0026thinsp;~\u0026thinsp;5 (N\u0026thinsp;=\u0026thinsp;329)\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHospital_2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.587 (0.394, 0.781)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.556 (0.529, 0.582)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.875 (0.646, 1.000)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.300 (0.016, 0.584)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.500 (0.238, 0.762)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.750 (0.326, 1.174)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHosptial_3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.536 (0.496, 0.576)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.641 (0.639, 0.643)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.945 (0.908, 0.982)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.128 (0.057, 0.198)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.646 (0.582, 0.711)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.579 (0.357, 0.801)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eHosptial_4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.598 (0.541, 0.656)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.592 (0.590, 0.593)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.692 (0.617, 0.768)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.489 (0.407, 0.572)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.579 (0.505, 0.653)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e \u003cp\u003e0.611 (0.521, 0.701)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \n\u003c/br\u003e\n\u003cp\u003eFigure 4a to Fig. 4e represent the confusion matrices of the training dataset, while Fig. 4f to Fig. 4j depict the confusion matrices of the external validation dataset. Figure 4a and Fig. 4f display the predictions of the clinical model. Figure 4b and Fig. 4g show the classification results based on the PIRADS category. Figure 4c and Fig. 4h show the predictions of the radiomics model. Figure 4d and Fig. 4i show the predictions of the deep radiomics model. Finally, Fig. 4e and Fig. 4j illustrate the classification outcomes according to the biopsy results.\u003c/p\u003e\n\u003cp\u003eWhen observed by columns, the first to three columns display the true values for ISUP_1, ISUP_2, and ISUP_3, respectively, while the fourth column represents the total true values. When observed by rows, from bottom to top, the first to three rows represent the predicted values for ISUP_1, ISUP_2, and ISUP_3, respectively, and the fourth row corresponds to the total predicted values.\u003c/p\u003e\n\u003cp\u003eFigure 5a to Fig. 5c represent the ROC curves of the training dataset, while Fig. 5d to Fig. 5e depict the ROC curves of the external validation dataset. Figure 5a display the ROC curves for various prediction methods targeting ISUP_1. Figure 5b and Fig. 5d illustrate the ROC curves for various prediction methods targeting ISUP_2. Figure 5c and Fig. 5e show the ROC curves for various prediction methods targeting ISUP_3\u0026thinsp;~\u0026thinsp;5.\u003c/p\u003e\n\u003cp\u003eFigure 6a displays the predictions of the clinical model across different hospitals. Figure\u0026nbsp;6b shows the classification results based on the PIRADS category across different hospitals. Figure\u0026nbsp;6c illustrates the predictions of the radiomics model across different hospitals. Figure\u0026nbsp;6d shows the predictions of the deep-radiomics model across different hospitals. Finally, Fig.\u0026nbsp;6e shows the classification results based on biopsy results across different hospitals.\u003c/p\u003e\n\u003cp\u003eIn Table \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003e, evaluation metrics, including AUC, accuracy (ACC), sensitivity (SEN), specificity (SPE), positive predictive value (PPV), negative predictive value (NPV), positive likelihood ratio (PLR), and negative likelihood ratio (NLR), are presented. In the external validation dataset, the AUCs for the prediction methods ranged from 0.597 to 0.718 for the clinical model, 0.554 to 0.613 for the PIRADS score, 0.566 to 0.632 for the radiomics model, 0.700 to 0.791 for the deep-radiomics model, and 0.537 to 0.578 for the biopsy. The ROC curves of the training dataset and the external validation dataset are depicted in Fig.\u0026nbsp;5.\u003c/p\u003e\n\u003cp\u003eFor the prediction of ISUP_1, the AUC of the deep-radiomics model was the highest, surpassing those of the radiomics model, PIRADS score, and biopsy pathology (all P\u0026thinsp;\u0026lt;\u0026thinsp;0.05); however, the deep-radiomics model did not significantly differ from the clinical model (P\u0026thinsp;=\u0026thinsp;0.120). For the classification of ISUP_2, the AUC of the deep-radiomics model was the highest compared to those of the other prediction methods (all P\u0026thinsp;\u0026lt;\u0026thinsp;0.05). For the classification of ISUP_3\u0026thinsp;~\u0026thinsp;5, the AUC of the deep-radiomics model also outperformed those of the other prediction methods (all P\u0026thinsp;\u0026lt;\u0026thinsp;0.05). The results of the DeLong test for the comparison of AUCs in discriminating ISUP classes in the external validation dataset are shown in the supplementary material (Table S 16).\u003c/p\u003e\n\u003cp\u003eTable \u003cspan class=\"InternalRef\"\u003e5\u003c/span\u003e displays the AUCs evaluated across the three hospitals in the external dataset. The AUCs were compared using the Delong test, and the results are outlined in the supplementary material (Table S 17). Since there was no ISUP_1 patient in the data from Hospital_2, a comparison of ISUP_1 could not be performed. For the prediction of ISUP_2, only the AUC of biopsy pathology was significantly different between Hospital_3 and Hospital_4 (P\u0026thinsp;=\u0026thinsp;0.021). For the prediction of ISUP_3\u0026thinsp;~\u0026thinsp;5, only the AUC of the clinical model showed a statistically significant difference between Hospital_3 and Hospital_4 (P\u0026thinsp;=\u0026thinsp;0.008). The PIRADS category, radiomics model, and deep-radiomics model did not significantly differ among the three hospitals in terms of the prediction of the ISUP_2 and ISUP_3\u0026thinsp;~\u0026thinsp;5 classes (all P\u0026thinsp;\u0026gt;\u0026thinsp;0.05). The ROC curves for the predictions are illustrated in Fig. 6.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003e In this study, according to the International Society of Urology (ISUP) grading system, we divided the data into the ISUP 1, ISUP 2 and ISUP 3\u0026ndash;5 groups, developed five methods to predict the grouping of PCa ISUP, and evaluated and verified the clinical efficacy of various prediction methods on external multicenter datasets. Our results showed that, compared with other methods, the radiomics model based on deep learning showed the best performance in predicting the invasiveness of PCa (the AUC ranged from 0.700 to 0.791), and there was no statistically significant difference among the other three external validation datasets (all P\u0026thinsp;\u0026gt;\u0026thinsp;0.05). Our study showed that the radiomics model based on deep learning is promising for providing an objective and noninvasive method for evaluating the invasiveness of PCa, which is helpful for further reducing overtreatment and avoiding unnecessary biopsy and has a certain generalizability.\u003c/p\u003e \u003cp\u003eMpMRI can provide important information about the invasiveness of PCa \u003csup\u003e[\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e, \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]\u003c/sup\u003e, and according to previous studies, ADC value had high value in predicting Gleason score \u003csup\u003e[\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e]\u003c/sup\u003e. ADC value is calculated on the basis of DWI, which uses the diffusion attenuation of water molecules to reflect the microstructure of living tissues and has good reliability and stability\u003csup\u003e[\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]\u003c/sup\u003e. The combination of radiomics and ADC in the diagnosis and evaluation of prostate cancer has also been a research hotspot. Several researchers\u003csup\u003e[\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e]\u003c/sup\u003e reported that the accuracy of feature prediction PCa of ADC sequence images can reach 97.39%, which means that better accuracy can be achieved only by extracting image features on ADC maps for prediction. Algohary\u003csup\u003e[\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]\u003c/sup\u003e et al. reported that intra-tumoral radiomics features on dual parameter MRI can predict the risk of prostate cancer, and the radiomics features are mainly derived from ADC maps.\u003c/p\u003e \u003cp\u003eIn this study, ADC maps were selected for feature extraction, and two models, traditional radiomics and deep learning radiomics, were developed to evaluate their clinical application value in predicting the invasiveness of PCa. The results of this study showed that the AUC of the traditional radiomics model in the training cohort was 0.609 to 0.882, and the AUC of the three external validation datasets was only 0.566 to 0.632. Several previous single center studies reported that\u003csup\u003e[\u003cspan additionalcitationids=\"CR29 CR30\" citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e]\u003c/sup\u003e the traditional radiomics model performed well in predicting the invasiveness of PCa (AUC value ranged from 0.70 to 0.93); however, this multi-centers study showed low accuracy. The reason may be that there is currently no unified machine learning algorithm for traditional radiomics. The systematic and strict machine learning framework includes feature extraction, classification, cross validation, algorithm selection and statistical analysis. The manual selection of these algorithm steps can affect the efficiency of the model to a certain extent\u003csup\u003e[\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e]\u003c/sup\u003e. Various algorithms have both advantages and disadvantages, and there is currently no unified standard for these algorithms. Improving the reproducibility of traditional radiomics models in different datasets still requires a large number of sample sets for research and validation.\u003c/p\u003e \u003cp\u003eMoreover, the deep learning radiomics model showed good efficacy in this study. The model included 13 features extracted from the convolutional network (Table\u0026nbsp;\u003cspan refid=\"Tab5\" class=\"InternalRef\"\u003e5\u003c/span\u003e), and its prediction effect ranged from 0.791 to 0.882 in the training cohort and from 0.700 to 0.791 in the external verification cohort. The use of deep learning algorithms is an important branch of machine learning. Various machine learning algorithms are used on neural networks to learn the internal laws of sample data. Convolutional neural networks (CNNs) are the most representative models in deep learning. Many neural networks with different structures, such as AlexNet, GoogLeNet, ResNet, and DeepLab, have been created. These algorithms can more fully extract data features, thus realizing the algorithm set of various tasks\u003csup\u003e[\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e]\u003c/sup\u003e. Our research results were consistent with previous work. Similarly, Khosravi et al.\u003csup\u003e[\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e]\u003c/sup\u003e developed a model based on GoogLeNet; their research showed that the AUC of the lesions that differentiated between the GG\u0026thinsp;=\u0026thinsp;1 group and the GG\u0026thinsp;\u0026ge;\u0026thinsp;2 group was 0.78, while that of lesions that differentiated between the GG\u0026thinsp;\u0026ge;\u0026thinsp;4 group and the GG\u0026thinsp;=\u0026thinsp;1 group was 0.86. Aldoj et al.\u003csup\u003e[\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e]\u003c/sup\u003e conducted an analysis based on a multi-channel 3D CNN and used the open source dataset ProstateX to establish a prostate cancer classifier. Research has shown that the sensitivity, specificity, and AUC of the lesions in the GG\u0026thinsp;\u0026ge;\u0026thinsp;2 group and GG\u0026thinsp;=\u0026thinsp;1 group were 81.9%, 86.1%, and 0.897, respectively. These results confirmed that deep learning technology can mine high-dimensional features of images, has obvious advantages in building prediction models, provides a new idea for predicting the invasiveness of PCa, and is expected to be applied in the clinical environment in the future.\u003c/p\u003e \u003cp\u003eIn recent years, although many studies have reported the broad prospects of the application of the radiomics model in prostate cancer\u003csup\u003e[\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e, \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]\u003c/sup\u003e, few studies based on multi-center data or focused on invasion have been conducted. Multiple centers of data research verification are highly important for training the applicability and stability of the radiomics model and can promote its clinical application. In this study, there was no statistically significant difference in the prediction of ISUP grouping of PCa lesions by the in-depth learning radiomics model among the three external hospitals (P\u0026thinsp;\u0026gt;\u0026thinsp;0.05), indicating that the model has strong generalizability among different hospitals, different MRI scanners, different field strengths and different scanning parameters. The reasons may be as follows: (1) This study used a pre-trained AI algorithm to process mpMRI images for focus segmentation and feature extraction, which made the selection of ROIs more accurate and stable and reduced human error. (2) The ISUP groups included in the study all had postoperative RP pathology. Compared with the results of biopsy pathology\u003csup\u003e[\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]\u003c/sup\u003e, the pathological information in this study was more accurate. Previous studies have shown that the consistency of biopsy pathology and total prostatectomy pathology is only 54%\u003csup\u003e[\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e]\u003c/sup\u003e. (3) In this study, the radiomics model selected lesions and extracted features based on ADC maps. According to previous reports\u003csup\u003e[\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e]\u003c/sup\u003e, there is no statistically significant difference (P\u0026thinsp;=\u0026thinsp;0.51) in the stability of features extracted from different vendors' scanners and ADC maps with different field strengths, which helps to develop a stable radiomics model and promote clinical application across multiple centers.\u003c/p\u003e \u003cp\u003eOur study has several limitations. Although this was a multi-center study, our cohort size was relatively small, and the distribution of patients and ISUP grouping data at each center were unbalanced, which may have affected our findings. Second, this study did not group the PCa in the TZ or PZ. Related studies have reported that\u003csup\u003e[\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e]\u003c/sup\u003e the effective radiomics features for PCa detection and evaluation differ between the TZ and peripheral zone. Future studies could group lesions in the TZ and peripheral zone, extract the radiomics features more accurately, and improve the diagnostic efficiency of the model.\u003c/p\u003e \u003cp\u003eThis multi-center study confirmed that the deep learning radiomics model using artificial intelligence to extract image features from mpMRI images can potentially predict the invasiveness of PCa and has strong external generalizability; additionally, this model is expected to provide an objective and noninvasive new method for evaluating PCa in the clinic.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cp\u003ePCa Prostate cancer\u003c/p\u003e\u003cp\u003eMpMRI Multiparametric magnetic resonance imaging\u003c/p\u003e\u003cp\u003eAI Artificial intelligence\u003c/p\u003e\u003cp\u003eRP Artificial intelligence\u003c/p\u003e\u003cp\u003eAUC Area under the curve\u003c/p\u003e\u003cp\u003eROC Receiver operating characteristic\u003c/p\u003e\u003cp\u003eROI Region of interest\u003c/p\u003e\u003cp\u003eGS Gleason score\u003c/p\u003e\u003cp\u003eGG Gleason score group\u003c/p\u003e\u003cp\u003eISUP International Society of Urology Pathology\u003c/p\u003e\u003cp\u003eADC Apparent diffusion coefficient\u003c/p\u003e\u003cp\u003eDWI Diffusion-weighted imaging\u003c/p\u003e\u003cp\u003eT2WI T2-weighted imaging\u003c/p\u003e\u003cp\u003ePZ Peripheral zone\u003c/p\u003e\u003cp\u003eTZ Transitional zone\u003c/p\u003e\u003cp\u003eCsPCa Clinically significant prostate cancer\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAuthor contributions\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eKexin Wang, Ning Luo wrote the main manuscript text.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eZhaonan Sun, Xiangpeng Zhao, Lilan She, Zhangli Xing, Yuntian Chen, Chunlei He, Pengsheng Wu, Xiangpeng Wang conducted data collection.\u003c/p\u003e\n\u003cp\u003eZiXuan Kong reviewed the manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData Availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets used and analysed during the current study available from the corresponding author on reasonable request.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eWHO. Cancer. 2018. Accessed on October 26, 2020. Available from: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.who.int/health-topics/cancer#tab=tab_1\u003c/span\u003e\u003cspan address=\"https://www.who.int/health-topics/cancer#tab=tab_1\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMottet N BJ, Briers E, Bolla M, Cornford P, De Santis M. European Association of Urology Prostate Cancer Guidelines. 2020. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://uroweb.org/wp-content/uploads/EAU-Guidelines-Prostate-Cancer-2015-v2\u003c/span\u003e\u003cspan address=\"https://uroweb.org/wp-content/uploads/EAU-Guidelines-Prostate-Cancer-2015-v2\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMatoso A, Epstein JI (2019) Defining clinically significant prostate cancer on the basis of pathological findings.Histopathology 74(1):135\u0026ndash;145\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEpstein JI, Egevad L, Amin MB, Delahunt B, Srigley JR, Humphrey PA et al (2016) The 2014 International Society of Urological Pathology (ISUP) Consensus Conference on Gleason Grading of Prostatic Carcinoma:Definition of Grading Patterns and Proposal for a New Grading System. Am J Surg Pathol 40(2):244\u0026ndash;252.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLizhi Shao, Ye Yan, Zhenyu Liu, Xiongjun Ye, Haizhui Xia, Xuehua Zhu et al (2020) Radiologist-like artificial intelligence for grade group prediction of radical prostatectomy for reducing upgrading and downgrading from biopsy. Theranostics 10(22):10200\u0026ndash;10212.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBoorjian SA, Karnes RJ, Crispen PL, Rangel LJ, Bergstralh EJ, Sebo TJ et al (2009) The impact of discordance between biopsy and pathological Gleason scores on survival after radical prostatectomy. J Urol 181: 95\u0026ndash;104.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChunmei Li, Min Chen, Jianye Wang, Xuan Wang, Wei Zhang, Chen Zhang (2017) Apparent diffusion coefficient values are superior to transrectal ultrasound-guided prostate biopsy for the assessment of prostate cancer aggressiveness. Acta Radiol 58(2):232\u0026ndash;239.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMidiri F, Vernuccio F, Purpura P, Pierpaolo A, Tommaso VB (2021) Multiparametric MRI and radiomics in prostate cancer: a review of the current literature. Diagnostics 11(10):1829.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eStanzione A, Gambardella M, Cuocolo R, Ponsiglione A, Romeo V, Imbriaco M (2020) Prostate MRI Radiomics: A Systematic Review and Radiomic Quality Score Assessment. Eur J Radiol 129:109095.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCastaldo R, Cavaliere C, Soricelli A, Salvatore M, Pecchia L, Franzese M (2021) Radiomic and Genomic Machine Learning Method Performance for Prostate Cancer Diagnosis: Systematic Literature Review. J Med Internet Res 23(4):e22394.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHectors SJ, Chen C, Chen J, Wang J, Sharon G, Miko Y et al (2021) Magnetic resonance imaging radiomics-based machine learning prediction of clinically significant prostate cancer in equivocal PI-RADS 3 lesions. J Magn Reson Imaging 54(5):1466\u0026ndash;1473.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen T, Zhang ZY, Tan SX, Zhang Y, Wei CG, Wang S et al (2022) MRI based radiomics compared with the PI-RADS v2.1 in the prediction of clinically significant prostate cancer: biparametric vs multiparametric MRI. Front Oncol 11:792456.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen T, Li M, Gu Y, Zhang Y, Yang S, Wei CG et al (2019) Prostate Cancer Differentiation and Aggressiveness: Assessment with a Radiomic-Based Model vs. PI-RADS v2. J Magn Reson Imaging 49(3):875\u0026ndash;884.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGabriel A Nketiah, Mattijs Elschot, Tom W Scheenen, Marnix C Maas, Tone F Bathen, Kirsten M Seln\u0026aelig;s et al (2021) Utility of T2-weighted MRI texture analysis in assessment of peripheral zone prostate cancer aggressiveness: a single-arm, multicenter study. Sci Rep 11(1):2085.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLeandro Donisi, Giuseppe Cesarelli, Anna Castaldo, Davide Raffaele De Lucia, Francesca Nessuno, Gaia Spadarella et al (2021) A Combined Radiomics and Machine Learning Approach to Distinguish Clinically Significant Prostate Lesions on a Publicly Available MRI Dataset. J Imaging 7(10):215.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhen Kang, Anhui Xu, Liang Wang (2021) Predictive role of T2WI and ADC-derived texture parameters in differentiating Gleason score 3\u0026thinsp;+\u0026thinsp;4 and 4\u0026thinsp;+\u0026thinsp;3 prostate cancer. J Xray Sci Technol 29(2):307\u0026ndash;315.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eElena Bertelli, Laura Mercatelli, Chiara Marzi, Eva Pachetti, Michela Baccini, Andrea Barucci et al (2022) Machine and Deep Learning Prediction Of Prostate Cancer Aggressiveness Using Multiparametric MRI. Front Oncol 11:802964\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDavid Bonekamp, Simon Kohl, Manuel Wiesenfarth, Patrick Schelb, Jan Philipp Radtke, Michael G\u0026ouml;tz et al (2018) Radiomic machine learning for characterization of prostate lesions with MRI: comparison to ADC values. Radiology 289:128\u0026ndash;137.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAhmad Chaddad, Michael J Kucharczyk, Tamim Niazi (2018) Multimodal Radiomic Features for the Predicting Gleason Score of Prostate Cancer. Cancers (Basel) 10(8):249.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChao Han, Shuai Ma, Xiang Liu, Yi Liu, Changxin Li, Yaofeng Zhang et al (2021) Radiomics Models Based on Apparent Diffusion Coefficient Maps for the Prediction of High-Grade Prostate Cancer at Radical Prostatectomy: Comparison With Preoperative Biopsy. J Magn Reson Imaging 54(6):1892\u0026ndash;1901\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTurkbey B, Pinto PA, Mani H, Bernardo M, Pang YX, McKinney YL, et al (2010) Prostate cancer: value of multiparametric MR imaging at 3T for detection\u0026ndash;histopathologic correlation. Radiology 255(1):89\u0026ndash;99.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDonati OF, Mazaheri Y, Afaq A, Vargas HA, Zheng J, Moskowitz CS et al (2014) Prostate cancer aggressiveness: assessment with whole-lesion histogram analysis of the apparent diffusion coefficient. Radiology 271(1):143\u0026ndash;152.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHambrock T, Hoeks C, Hulsbergen-van KC, Scheenen T, F\u0026uuml;tterer J, Bouwense S et al (2012) Prospective assessment of prostate cancer aggressiveness using 3-T diffusion-weighted magnetic resonance imaging-guided biopsies versus a systematic 10-core transrectal ultrasound prostate biopsy cohort. Eur Urol 61(1):177\u0026ndash;184.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFelker ER, Margolis DJ, Nassiri N, Marks LS (2016) Prostate cancer risk stratification with magnetic resonance imaging. Urol Oncol 34(7):311\u0026ndash;319.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYukihisa Takayama, Riwa Kishimoto, Shouhei Hanaoka, Hiroi Nonaka, Susumu Kandatsu, Hiroshi Tsuji et al (2008) ADC value and diffusion tensor imaging of prostate cancer: changes in carbon-ion radiotherapy. J Magn Reson Imaging 27(6):1331\u0026ndash;5\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHiram Shaish, Stella K Kang, Andrew B Rosenkrantz (2017) The utility of quantitative ADC values for differentiating high-risk from low-risk prostate cancer: a systematic review and meta-analysis. Abdom Radiol (NY) 42(1):260\u0026ndash;270.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAlgohary A, Shiradkar R, Pahwa S, Andrei P, Sadhna V, Daniel M et al (2020) Combination of Peri-Tumoral and Intra-Tumoral 4 Radiomic Features on Bi-Parametric MRI Accurately Stratifies Prostate Cancer Risk: A Multi-Site 5 Study. Cancers 12(8):2200.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWibmer A, Hricak H, Gondo T, Matsumoto K, Veeraraghavan H, Fehr D et al (2015) Haralick texture analysis of prostate MRI: Utility for differentiating non-cancerous prostate from prostate cancer and differentiating prostate cancers with different Gleason scores. Eur Radiol 25:2840\u0026ndash;2850.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFehr D, Veeraraghavan H, Wibmer A, Gondo T, Matsumoto K, Vargas HA et al (2015) Automatic classification of prostate cancer Gleason scores from multiparametric magnetic resonance images. Proc Natl Acad Sci USA 112:6265\u0026ndash;6273.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAntonelli M, Johnston EW, Dikaios N, Cheung KK, Sidhu HS, Appayya MB et al (2019) Machine learning classifiers can predict Gleason pattern 4 prostate cancer with greater accuracy than experienced radiologists. Eur Radio 29:4754\u0026ndash;4764.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChaddad A, Kucharczyk MJ, Niazi T (2018) Multimodal radiomic features for the predicting gleason score of prostate cancer. Cancers 10(8):249\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSarah Caulfield, Gerard Menezes, Laure Marignol, Claire Poole (2018) Nomograms are key decision-making tools in prostate cancer radiation therapy. Urol Oncol 36(6):283\u0026ndash;292.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMcBee MP, Awan OA, Colucci AT, Ghobadi CW, Kadom N, Kansagra AP et al (2018) Deep learning in radiology. Acad Radiol 25:1472\u0026ndash;1480.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKhosravi P, Lysandrou M, Eljalby M, Li Q, Kazemi E, Zisimopoulos P et al (2021) A deep learning approach to diagnostic classification of prostate cancer using pathology-radiology fusion. J Magn Reson Imaging 54:462\u0026ndash;471.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAldoj N, Lukas S, Dewey M, Penzkofer T (2020) Semi-automatic classification of prostate cancer on multi-parametric MR imaging using a multichannel 3D convolutional neural network. Eur Radiol 30:1243\u0026ndash;1253.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eA Vignati, S Mazzetti, V Giannini, F Russo, E Bollito, F Porpigli et al (2015) Texture features on T2-weighted magnetic resonance imaging: new potential biomarkers for prostate cancer aggressiveness. Phys Med Biol 60:2685\u0026ndash;2701.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKh\u0026eacute;mara Gnep, Aur\u0026eacute;line Fargeas, Ricardo E Guti\u0026eacute;rrez-Carvajal, Fr\u0026eacute;d\u0026eacute;ric Commandeur, Romain Mathieu, Juan D Ospina et al (2017) Haralick textural features on T2-weighted MRI are associated with biochemical recurrence following radiotherapy for peripheral zone prostate cancer. J Magn. Reson Imaging 45:103\u0026ndash;117.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEvans SM, Patabendi Bandarage V, Kronborg C, Earnest A, Millar J, Clouston D (2016) Gleason group concordance between biopsy and radical prostatectomy specimens: a cohort study from prostate cancer outcome registry-Victoria. Prostate Int 4:145\u0026ndash;151.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJurgen Peerlings, Henry C Woodruff, Jessica M Winfield, Abdalla Ibrahim, Bernard E Van Beers, Arend Heerschap et al (2019) Stability of radiomics features in apparent diffusion coefficient maps from a multi-centre test-retest trial. Sci Rep 9(1):4800.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShoshana B Ginsburg, Ahmad Algohary, Shivani Pahwa, Vikas Gulani, Lee Ponsky, Hannu J Aronen et al (2017) Radiomic features for prostate cancer detection on MRI differ between the transition and peripheral zones: preliminary findings from a multi-institutional study. J Magn Reson Imaging 46:184\u0026ndash;193.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Gleason score, prostate cancer, multiparametric magnetic resonance imaging, deep learning, radiomics","lastPublishedDoi":"10.21203/rs.3.rs-4252741/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4252741/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eObjective\u003c/h2\u003e \u003cp\u003eTo evaluate the feasibility of utilizing artificial intelligence (AI)-predicted multiparametric MRI (mpMRI) image features for predicting the aggressiveness of prostate cancer (PCa).\u003c/p\u003e\u003ch2\u003eMaterials and methods\u003c/h2\u003e \u003cp\u003eA total of 878 PCa patients from 4 hospitals were retrospectively collected, all of whom had pathological results after radical prostatectomy(RP).A pre-trained AI algorithm was used to select suspected PCa lesions and extract lesion features for model development. The study evaluated five prediction methods, including 1) A clinical model of clinical features and image features of suspected PCa lesions selected by AI algorithm, 2)the PIRADS category, 3)a conventional radiomics model, 4) a radiomics model based on deep learning, 5)biopsy pathology.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eIn the externally validated dataset, the deep learn-based radiomics model showed the highest area under the curve (AUC 0.700 to 0.791).It exceeded clinical model (AUC 0.597 to 0.718), traditional radiomic model (AUC 0.566 to 0.632), PIRADS score (AUC 0.554 to 0.613) and biopsy pathology (AUC 0.537 to 0.578). And the AUC predicted by the model did not show statistically significant difference among the three externally verified hospitals (P\u0026thinsp;\u0026gt;\u0026thinsp;0.05).\u003c/p\u003e\u003ch2\u003eConclusion\u003c/h2\u003e \u003cp\u003eDeep-radiomics models utilizing AI-extracted image features from mpMRI images can potentially be used to predict PCa aggressiveness, demonstrating a generalized ability for external validation.\u003c/p\u003e","manuscriptTitle":"Evaluating the Feasibility of AI-Predicted mpMRI Image Features for Predicting Prostate Cancer Aggressiveness: a Multicenter Study","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-04-25 16:21:53","doi":"10.21203/rs.3.rs-4252741/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"02d06f5b-33e4-473c-abff-d8f9e0bc14bd","owner":[],"postedDate":"April 25th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":30834300,"name":"Health sciences/Diseases/Cancer"},{"id":30834301,"name":"Health sciences/Diseases/Urogenital diseases"}],"tags":[],"updatedAt":"2024-05-02T06:37:39+00:00","versionOfRecord":[],"versionCreatedAt":"2024-04-25 16:21:53","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-4252741","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4252741","identity":"rs-4252741","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}