DWI and ADC Habitat Imaging in Predicting HER2 Expression Status in Bladder Cancer: A Retrospective Study

DWI and ADC Habitat Imaging in Predicting HER2 Expression Status in Bladder Cancer: A Retrospective 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 Research Article DWI and ADC Habitat Imaging in Predicting HER2 Expression Status in Bladder Cancer: A Retrospective Study Zeke Chen, Zhichang Fan, Xiaoyue Zhang, Wenxin Li, Yan Li, Bin Wang, and 3 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7710608/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 10 You are reading this latest preprint version Abstract Background: Human epidermal growth factor receptor 2 (HER2) antibody-coupled drugs have shown promising clinical benefits in patients with bladder cancer (BCa). HER2 expression status is generally detected clinically using postoperative pathological immunohistochemistry (IHC), but preoperative non-invasive detection of BCa HER2 expression status remains to be sought. The aim of this study was to investigate the value of diffusion-weighted imaging (DWI) and apparent diffusion coefficient (ADC) habitat imaging in predicting the expression of HER2 in BCa. Methods: This retrospective study included 232 BCa patients (November 2022–February 2024) with HER2 status confirmed by immunohistochemistry. The K-means clustering algorithm is used to re-segment the region of interest. Based on the spatial distribution of the habitat map, the histogram features of each subregion were extracted. Based on the Spearman correlation coefficient (> 0.7) feature screening results, a support vector machine (SVM) classification model was established to predict the expression of HER2 in BCa. The discrimination ability of the model was evaluated by plotting the receiver operating characteristic (ROC) curve and calculating the area under the curve (AUC), and the diagnostic performance of the model was comprehensively evaluated by combining the calibration curve and the decision curve. Results: Randomly divided patients into training cohort (N = 148, median age 68.66 years; 121men), validation (N = 47, median age 69.12 years; 39 men), and test cohort (N = 37, median age 67.92 years; 32men) according to the ratio of 6:2:2. Based on the contour coefficient, K = 2 is finally selected as the clustering parameter to cluster the DWI and ADC images into two subregions. A total of 80 features were extracted from the four sub-regions of the two sequences. After screening, an SVM prediction model was constructed from the remaining 17 features. In the SVM model, the AUC of the training set was 0.88 (95% CI: 0.82–0.93), the validation set was 0.85 (95% CI: 0.72–0.94), and the test set was 0.84 (95% CI: 0.88–0.94). Conclusion: MRI-based habitat analysis can help distinguish heterogeneous regions of BCa and effectively predict HER2 expression status of BCa. Clinical trial number: Not applicable. HER2 bladder cancer magnetic resonance imaging habitat image Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 1. Introduction Bladder cancer (BCa) ranks among the most prevalent malignancies of the urinary system and the 10th most common cancer globally, with its incidence steadily rising [1,2]. The human epidermal growth factor receptor 2 (HER2) has emerged as a critical biomarker for diagnosing, prognosing, and predicting therapeutic responses in urothelial carcinomas [3]. Previous studies have confirmed that HER2-targeted therapies such as RC48-ADC have demonstrated remarkable efficacy in HER2-overexpressing advanced BCa, achieving an objective response rate of 50% and a disease control rate of 100% [4]. Immunohistochemistry (IHC) is a low-cost and effective technique, which is widely used in the expression of HER2 protein in tissues and cells [5-7]. However, IHC has some limitations. First, it is an invasive method that is complex and time-consuming to operate. Second, different pathologists interpret the results differently. Finally, pathological samples are inevitably affected by surgery, and the limited samples can not fully reflect the expression of HER2 protein in the whole tumor [8]. Therefore, it is essential to find a preoperative non-invasive method to determine the status of HER2 in BCa. Diffusion weighted imaging (DWI) and apparent diffusion coefficient (ADC) are functional magnetic resonance imaging (MRI) techniques, which can reflect the Brownian motion of water molecules in different tissues and cells of the human body [9]. There is increasing evidence that DWI and ADC can be used as imaging biomarkers to characterize the pathophysiology of various types of malignant tumors [10, 11]. Habitat analysis, also known as habitat imaging, is an image analysis technology that divides the region of interest into several sub-regions with the same or similar heterogeneity according to the differences between the macro level and the micro level of the region of interest [12-14]. MRI habitat imaging, by combining quantitative information of images with tumor heterogeneity, captures subtle differences within tumors and can reflect tumor cellular and molecular heterogeneity non-invasively. Previous studies have shown that MRI-based imaging radiomics models can be used to predict the expression status of HER2 in BCa [15,16]. However, there are still some shortcomings in the prediction of HER2 expression in BCa by traditional radiomics, and the prediction performance of the model needs to be improved. Traditional imaging usually regards tumors as a whole with relatively uniform internal distribution, which can not truly express tumor heterogeneity. Therefore, this study aimed to explore the value of DWI and ADC habitat imaging in predicting HER2 expression status in BCa. 2. Materials and Methods This retrospective study (approval number: KYYJ-2023-064) was approved by the local institutional review board, and the need for written informed consent was waived. 2.1 Patients From November 2022 to February 2024, patients pathologically diagnosed with BCa were retrospectively selected. A total of 232 patients were included in this study. The patient enrollment pathway is presented in Fig. 1. The inclusion criteria comprised the following: (1) confirmed by postoperative pathological examination as BCa; (2) preoperative multiparameter bladder MRI examination; (3) the expression status of HER2 in BCa can be determined by IHC testing of postoperative specimens. The exclusion criteria were as follows: (1) Incomplete baseline data (n = 135); (2) Patients who received preoperative radiotherapy and chemotherapy (n = 67); (3) patients with lesions <5 mm (n = 18); (4) patients with nonurothelial carcinoma (n = 17); and (5) patients with poor image quality (n = 5). Fig. 1. Flow chart of patient screening. 2.2 MRI Image Acquisition To ensure that the bladder is properly filled, patients are required to avoid drinking or urinating before the completion of the scan and drink 500-1000 mL of water half an hour before the scan. Two 3.0T MR scanners from Siemens in Germany (MAGNETOM Vida, MAGNETOM Skyra) were used. The scanning range covers the pelvic cavity (bladder), and the scanning sequence includes DWI. The scanning parameters of axial DWI sequence are vida: repetition time (TR) = 4100msec, echo time (TE) = 65 msec, layer thickness, 6 mm, layer spacing, 1.2mm, field of view (FOV) = 280mmx245mm, 19 layers; Skyra: TR = 4000msec, TE = 76msec, layer thickness, 6mm, the layer spacing, 1.2mm, FOV = 280mmx245mm, and the number of layers is 19. Two b values (b0: 0 sec/mm2; b1:800 sec/mm2). ADC image was obtained by dual-exponential model transformation of the DWI image on an image post-processing workstation. 2.3 Histopathological Assessment The most representative morphological sections were selected from the postoperative pathological tissues of all patients, and the results were independently interpreted by two pathologists with more than 5 years of working experience. HER2 results were interpreted according to the clinical pathological expert consensus on HER2 testing in urothelial carcinoma in China [6]. The HER2 results were then categorized as follows: HER2 non-overexpressing (IHC score of 0 or 1+), HER2 overexpression (IHC score of 3+ or 2+). 2.4 Image Preprocessing and Delineation 2.4.1 Image Preprocessing Python 3.7.6 (https://www.python.org) software was used to construct the code for the N4-bias field correction algorithm to eliminate the variation of image signal strength caused by different scanners, noise, and many unknown problems. 2.4.2 Region of Interest Delineation Two independent radiologists (A and B, with 5 and 2 years of pelvic imaging experience, respectively) using ITK-SNAP 4.0 (http://www.ITK-SNAP.org/) manually delineated the entire tumor area layer by layer on DWI and ADC images to obtain the volume of interest (VOI). For patients with multiple tumor lesions, the largest lesion was selected for delineation. Use the inter-class correlation coefficient (ICC) to evaluate the consistency of selected regions among observers. After 30 days, 30 patients were randomly selected and their VOI was redrawn by physician A. The intra-class correlation coefficient (ICC) was used to evaluate the consistency of the selected areas by the same physician at different times. 2.5 Clustering to Generate Tumor Habitat The habitats analysis was implemented by an in-house software nnFAE, which was developed based on Python. The K-means clustering algorithm is used to automatically segment the tumor region to construct the habitat. Silhouette coefficient is an important indicator for measuring clustering performance, which can reflect the degree of separation between clusters. When the contour coefficient is high, it means that the similarity between different clusters is low, and the similarity within clusters is high. We calculated the contour coefficients of clustering results under different K values to select the optimal K value [14]. 2.6 Extract Features and Model Construction 2.6.1 Extract Habitat Features The internal software nnFAE developed based on Python, was used to extract the features of DWI and ADC subregions of each bladder cancer patient, including the volume of each subregion and the corresponding volume percentage of each subregion. Using the pyradiomics toolkit (https://github.com/Radiomics/pyradiomics), extract the histogram features of each subregion. Histogram features include maximum, minimum, skewness, kurtosis, entropy, etc. 2.6.2 Establishment of Habitat Prediction Model To improve the repeatability of the model, features with consistency greater than 0.75 intra- and inter-observer are selected. Then the Spearman correlation coefficient between features is calculated to evaluate the correlation between features. When the correlation coefficient is greater than 0.7, it is considered that there is a high correlation between features. To optimize feature selection, the features with a narrow data distribution range among the two highly correlated features are screened out, and the optimal features are finally selected. The support vector machine (SVM) model is constructed in the training set by using the optimal features, and the performance of the model is evaluated in the validation and test set. Receiver operating characteristic (ROC) was plotted and the area under the curve (AUC), accuracy, specificity, sensitivity, positive predictive value (PPV), and negative predictive value (NPV) were calculated. 2.7 Statistical analysis Statistical analysis was performed using R4.4.0 (http://www.r-project.org/ version) or SPSS 13.0. The mean ± standard deviation was used if the measurement data conformed to normality and homogeneity of variance, and the two independent samples t-test and variance test were used for comparison between groups. If it does not conform to the normality and or homogeneity of variance and is expressed by the median ± upper and lower quartiles, comparisons between groups were performed using Kruskal-Wallis or Mann-Whitney U tests. The count data were presented as numerical examples. χ2 test and Fisher's exact test were used for comparison between groups. ICC was used to evaluate the consistency between different physicians and the same physician at different times. When ICC ≥ 0.75, the consistency is considered to be high. AUC, Calibration curve, and decision curve analysis (DCA) are used to evaluate the performance of the SVM model. When p <0.05, the difference was considered statistically significant. The specific workflow is shown in Fig. 2. Fig. 2. Workflow of the habitat analysis. 3. Results 3.1 Patients ’ Characteristics In this study, 232 patients with bladder urothelial carcinoma were randomly divided into a train set (n=148), validation set (n=47), and test set (n=37) according to the ratio of 6:2:2. The clinical and pathological parameters including age, sex, smoking, pathological stage, etc were also collected and analyzed. Table 1 presents the clinical and pathological characteristics. There were no significant differences in clinical and pathological variables between cohorts (all p > 0.05). Table 1 . Clinical and Pathological Characteristics of Patients. Training Cohort（n=148） Validation Cohort （n=47） Test Cohort（n=37） p Age 68.66（59.33-73.67） 69.12 (65.49-76.20) 67.92 (60.28-75.10) 0.67 Gender 0.85 Male 121(81.8%) 39(83.0%) 32(86.5%) Female 27(18.2%) 8(17.0%) 5(13.5%) Smoking 0.19 Yes 57(38.5%) 25(53.2%) 17(45.9%) No 91(61.5%) 22(46.8%) 20(54.1%) Pathological stage 0.99 MIBC 55(37.2%) 18(38.3%) 14(37.8%) NMIBC 93(62.8%) 29(61.7%) 23(62.2%) Tumor level 0.54 High grade 77(52.0%) 25(53.2%) 23(62.2%) Low grade 71(48.0%) 22(46.8%) 14(37.8%) HER2 0.64 Overexpression 92(62.2%) 27(57.4%) 25(67.6%) Non-overexpression 56(37.8%) 20(42.6%) 12(32.4%) MIBC: muscle-invasive bladder cancer; NMIBC: non-muscle-invasive bladder cancer 3.2 Habitat Clustering and Feature Selection According to the silhouette coefficient, a K = 2 clustering parameter is selected to cluster DWI and ADC images into two sub-regions respectively. Fig. 3 and Fig. 4 show representative habitat imaging of two correctly classified lesions with different HER2 expressions. There was an obvious difference in the spatial distribution of the habitats for various HER2 expressions. Fig. 3: Case 1, representative case of a patient with HER2 non-overexpression BCa (HER2 1+). (A-D) The preoperative DWI and ADC images and their corresponding habitat images of the patients. (E-G) Microscopic image of hematoxylin-eosin staining (original magnification, *100) and postoperative pathological IHC image of the patient(original magnification, *100). Fig. 4: Case 2, representative case of a patient with HER2 overexpression BCa (HER2 2+). (A-D) The preoperative DWI and ADC images and their corresponding habitat images of the patients. (E-G) Microscopic image of hematoxylin-eosin staining (original magnification, *100) and postoperative pathological IHC image of the patient(original magnification, *100). Initially, a total of 80 features were extracted from four sub-regions of the two sequences, with 40 features each from ADC and DWI images. The ICC for each of these habitat features exceeds 0.75. When the Spearman correlation coefficient between two features is greater than 0.7, the features with a narrow data distribution range among the two highly correlated features are filtered out. In the final DWI image, habitat 1 and habitat 2 retain four features respectively, while in the ADC image, habitat 1 and habitat 2 retain four and five features respectively. The SVM model was constructed using these 17 habitat characteristics (Table 2 and Table 3). Table 2. Habitat features of the training cohort and validation cohort. Features Training and validation cohorts (n=195) Overexpression Non-overexpression p DWI_Habitat 1_Minimum * 93.29(1.86) 92.92(93.95) 0.003 DWI_Habitat1_Energy * 2243836.45 (10442044.48) 779561.18 (6014724.49) 0.22 DWI_Habitat1_Kurtosis * -0.40(1.33) 0(0.83) 0.07 DWI_Habitat1_Skewness * 0.50(0.80) 0.22(0.65) 0.20 DWI_Habitat2_Maximum * 90.39(2.20) 90.11(4.12) 0.13 DWI_Habitat2_Uniformity * 4321.51(1505.19) 3734.69(2054.14) 0.17 DWI_Habitat2_Kurtosis * -0.80(0.58) -0.82(0.58) 0.17 DWI_Habitat2_Variance & 261.48±109.68 249.45±116.69 0.55 ADC_Habitat1_Minimum * 94.10(0.31) 94.11(0.79) 0.04 ADC_Habitat1_Uniformity * 15339.71(3568.16) 15297.23(31433.09) 0.41 ADC_Habitat1_Energy * 2727387.71(7324436.08) 3317559.67(4105059.34) 0.16 ADC_Habitat1_Kurtosis * -0.27(0.91) -0.25(1.07) 0.26 ADC_Habitat2_Maximum * 93.90(0.26) 93.78(0.48) 0.02 ADC_Habitat2_Uniformity * 5302.16(1816.21) 5099.39(1938.89) 0.17 ADC_Habitat2_Kurtosis * -0.36(1.18) 0(2.04) 0.15 ADC_Habitat2_Skewness * -0.21(0.90) -0.07(1.12) 0.17 ADC_Habitat2_Variance * 138.94(122.76) 121.23(116.03) 0.34 Data are the mean ± standard deviation or median (inter-quartile range, IQR). p values were calculated with the t-test & or Mann–Whitney U test * . Table 3. Habitat features of the test cohort Features Test cohort (n=37) Overexpression Non-overexpression p DWI_Habitat1_Minimum * 93.23(1.07) 92.33(93.49) 0.01 DWI_Habitat1_Energy * 4230308.23(29832411.04) 4001832.78(8526170.1) 0.41 DWI_Habitat1_Kurtosis * -0.67(0.88) 0(0.51) 0.59 DWI_Habitat1_Skewness * 0.40(0.66) 0.39(0.80) 0.93 DWI_Habitat2_Maximum * 90.48(1.43) 90.21(4.15) 0.15 DWI_Habitat2_Uniformity * 4343.18(1142.08) 4387.63(2517.54) 0.10 DWI_Habitat2_Kurtosis * -0.79(0.31) -0.67(0.54) 0.49 DWI_Habitat2_Variance & 275.87±96.74 247.37±88.74 0.72 ADC_Habitat1_Minimum * 94.07 (0.20) 94.07(0.12) 0.97 ADC_Habitat1_Uniformity & 15625.22(2409.96) 16368.66(4330.80) 0.03 ADC_Habitat1_Energy * 4478899.05(5868083.84) 2601169.44(11336955.85) 0.29 ADC_Habitat1_Kurtosis * -0.02(1.04) -0.52(0.79) 0.37 ADC_Habitat2_Maximum * 93.86(0.53) 93.92(0.08) 0.80 ADC_Habitat2_Uniformity & 5404.46±1244.70 5595.93±1318.08 0.97 ADC_Habitat2_Kurtosis * -0.46(0.91) 0.55(2.39) 0.12 ADC_Habitat2_Skewness & -0.30±0.61 -0.66±0.98 0.05 ADC_Habitat2_Variance * 129.96(92.36) 177.04(130.30) 0.30 Data are the mean ± standard deviation or median (inter-quartile range, IQR). p values were calculated with the t-test & or Mann–Whitney U test * . 3.3 Construction and Evaluation of Habitat Prediction Model The tumor VOI was clustered based on ADC and DWI sequences, and the habitat characteristics were extracted and screened. SVM was used to construct a habitat model for predicting the expression of HER2 in BCa. AUC, 95% confidence interval (95% CI), accuracy, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were used to evaluate the performance of the prediction model. Table 4 and Fig. 5a show the performance of the SVM model in evaluating the HER2 expression status in the training, validation, and test sets. Within the training cohort, the SVM model achieved an AUC of 0.88 (95% CI: 0.82-0.93), accuracy of 0.84, sensitivity of 0.83, specificity of 0.86, PPV of 0.91, and NPV of 0.75. In the validation cohort, the SVM models achieved an AUC of 0.85 (95% CI: 0.72-0.94), accuracy of 0.81, sensitivity of 0.85, specificity of 0.75, PPV of 0.82, NPV of 0.79; and the test achieved an AUC of 0.84 (95% CI: 0.68-0.94), accuracy of 0.78, sensitivity of 0.72, specificity of 0.92, PPV of 0.95, NPV of 0.61. We also calculated the calibration curve (Fig. 5b-5d) and decision curve (Fig. 6). As shown in Fig. 5 and Fig. 6, our prediction model exhibits stability and considerable predictive ability. It shows that the model has no obvious difference between the predicted value and the actual observation value, has good consistency, and has good correction efficiency, which is considered to be one of the good clinical application values. Fig. 5. (a) SVM model receiver operating characteristic curves (ROC) in the training cohort, validation cohort, and test cohort. (b-d)The SVM model calibration curve of the training cohort、validation cohort, and test cohort. Fig. 6. The SVM model decision curve of the training cohort(red)、validation cohort(blue), and test cohort (green). Table 4. AUCs for the Performance of the SVM Models in All Cohorts Cohort AUC Accuracy Sensitivity Specificity PPV NPV Training 0.88(0.82-0.93) 0.84 0.83 0.86 0.91 0.75 Validation 0.85(0.72-0.94) 0.81 0.85 0.75 0.82 0.79 Test 0.84(0.68-0.94) 0.78 0.72 0.92 0.95 0.61 4. Discussion This study confirms that extracting habitat features based on DWI and ADC subregions and constructing an SVM model can be used to predict the expression status of HER2 in BCa. Good predictive performance was demonstrated in the training set, validation set, and test set. This research may reveal the feasibility and clinical value of MRI habitat in the preoperative non-invasive assessment of HER2 status in BCa. HER2 is a member of the epidermal growth factor receptor family, with tyrosine kinase activity, involved in signal transduction for cell growth and differentiation. HER2 protein plays an important role in cell proliferation, differentiation, and angiogenesis, and its high expression can promote cell division, proliferation and differentiation [17]. HER2 plays a key role in the development and progression of a variety of malignancies, including breast cancer and urothelial cancer, etc [18,19]. Previous studies have shown that HER2 overexpression is an independent predictor of BCa-related survival, and HER2 overexpression is significantly associated with poor prognosis [20]. Studies have found that anti-HER-2 antibody-drug conjugates show good efficacy and safety in the treatment of patients with locally advanced or metastatic urothelial carcinoma with overexpression of HER2, and can bring significant clinical benefit to this subset of patients [21,22]. Therefore, it is essential to develop a non-invasive and effective method to assess the expression status of HER2 in BCa. Previous studies have shown that MRI-based radiomics models can be used to predict the expression of HER2 in BCa [15,16]. However, traditional radiomics still has shortcomings in predicting HER2 expression in BCa, and its predictive efficacy needs to be improved. Traditional radiomics often considers the tumor as a homogeneous whole and fails to fully reflect the heterogeneity within the tumor. Habitat imaging can reflect the internal heterogeneity of tumors. In this study, the K-means clustering method [23-25] was used to cluster the VOI of BCa, divide the VOI into multiple different subregions, and perform feature extraction on different subregions, so as to improve the accuracy of feature extraction, an SVM model was constructed to predict the HER2 expression status of BCa. The results indicate that there were some differences in the habitat subregions of HER2-overexpressing and HER2 non-overexpressing BCa. The range of red subregions with higher ADC values in HER2 overexpressing BCa patients is larger than that in HER2 non-overexpressing BCa patients. Similarly, compared to HER2 non-overexpressing BCa, HER2 overexpressing BCa patients have a smaller range of green subregions with lower ADC values. This is consistent with previous studies on other tumors [26-28]. The reason may be that HER2 is highly expressed in luminal unstable (Lumu) BCa (39%, p<0.01), and papillary structures are more common in luminal BCa. This structural feature may provide a relatively large space for the diffusion movement of water molecules, which leads to more significant diffusion behavior of water molecules in this subtype. Therefore, higher ADC values were observed in HER2 overexpressing BCa [29]. In this study, radiomics features were extracted from different tumor subregions to explore the influence of radiomics features between different regions on the evaluation of HER2 expression status of BCa. Radiomics features derived from different subregions provide us with a wealth of information, showing the image information of BCa itself, which can reflect BCa heterogeneity [30,31]. An SVM model was constructed based on 17 radiomics features, including 4 from DWI habitat 1 and habitat 2 subregions, 4 from ADC habitat 1 subregion, and 5 from habitat 2 subregion. The above explanation shows that DWI and ADC sequences, as well as their subregions, contribute to the construction of the model. The model includes 7 first-order features: Minimum, Energy, Kurtosis, Skewness, Maximum, Uniformity, and Variance. By capturing the distribution characteristics of gray values within BCa, the heterogeneity of BCa can be accurately reflected, providing strong support for predicting the HER2 expression status of BCa [32,33]. It was also observed in the study that some lesions were not successfully divided into two subregions, which may be due to the insufficient sample size to fully reflect the real image information, so it is necessary to continue to enrich the research samples. In addition, in ADC images, the red subregions are often located at the edge of the lesion. Whether the edge of the lesion contains clinically significant imaging information is worth further study. Our study has several limitations. First, the sample size we analyzed is relatively small; Secondly, manually sketching the region of interest may lead to subjectivity of the data and deviation of the results; Finally, our study is a single-center retrospective study, which may lead to inevitable bias. Multi-center studies with larger sample sizes are needed to verify the stability and reproducibility of these results. 5. Conclusion MRI habitat imaging can effectively predict the expression status of HER2 in BCa, which provides a new method for the preoperative non-invasive diagnosis of bladder cancer and has important clinical value for the treatment decision of BCa patients. Abbreviations ADC Apparent Diffusion Coefficient AUC Area Under the Curve BCa Bladder Cancer DWI Diffusion Weighted Imaging HER2 Human Epidermal Growth Factor Receptor 2 ICC Intraclass Correlation Coefficient ICC Interclass Correlation Coefficient IHC Immunohistochemistry MIBC Muscle Invasive Bladder Cancer MRI Magnetic resonance imaging NMIBC Non-muscle Invasive Bladder Cancer NPV Negative Predictive Value PPV Positive Predictive Value ROC Receiver Operating Characteristic SVM Support Vector Machine VOI Volume of Interest 95% CI 95% Confidence Interval Declarations Ethics approval and consent to participate This retrospective study (No. KYYJ-2023-064) was approved by the institutional review board of the First Hospital of Shanxi Medical University, and the need for written informed consent was waived. The study was performed in compliance with the 2024 version of the Declaration of Helsinki-Ethical Principles for Medical Research Involving Human Participants. Consent for publication Not applicable. Availability of data and material The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request. Competing interests The authors declare that they have no competing interests. Funding This study was supported by grants from the Four “Batches” Innovation Project of Invigorating Medical through Science and Technology of Shanxi Province (2023XM011); from the China International Medical Foundation of China (z-2014-07-2301); from the 2024 Annual Shanxi Provincial Basic Research Program (Free Exploration Category) Second Batch (202403021222449). Authors' contributions ZC was primarily responsible for collecting the patient data, organizing the literature, designing the experimental approach, conducting statistical analysis, and writing the manuscript. ZF was responsible for designing the experimental approach, performing post-processing on the original images to obtain habitat maps, conducting statistical analysis, and writing all the codes used in the paper. XZ and WL collected the laboratory data. ZF analyzed the data, with statistical advice from YL and BW. YW and GY discussed the results and interpreted the data. XW was primarily responsible for project administration and providing resources, also served as the corresponding author, handling all communications with the journal and addressing any post-publication inquiries. All authors read and approved the final manuscript. Acknowledgments The authors thank the colleagues from First Hospital of Shanxi Medical University for their constructive suggestions in the conception and completion of this work. References Saginala K, Barsouk A, Aluru JS, Rawla P, Padala SA, Barsouk A. Epidemiol Bladder Cancer Med Sci (Basel). 2020;8(1):15. Siegel RL, Giaquinto AN, Jemal A, Cancer statistics. CA Cancer J Clin 2024 Jan-Feb. 2024;74(1):12–49. Sanguedolce F, Zanelli M, Palicelli A, et al. HER2 Expression in Bladder Cancer: A Focused View on Its Diagnostic, Prognostic, and Predictive Role. Int J Mol Sci. 2023;24(4):3720. Xu Y, Wang Y, Gong J, et al. Phase I study of the recombinant humanized anti-HER2 monoclonal antibody–MMAE conjugate RC48-ADC in patients with HER2-positive advanced solid tumors. Gastric Cancer. 2021;24(4):913–25. Fernandez AI, Liu M, Bellizzi A, et al. Examination of Low ERBB2 Protein Expression in Breast Cancer Tissue. JAMA Oncol. 2022;8(4):1–4. Committee of Tumor Pathology. Chinese anti-cancer association, Committee of Urothelial Carcinoma, Chinese society of Clinical Oncology. Chinese clinicopathological expert consensus on the detection of human epidermal growth factor receptor 2 in urothelial carcinoma. Chin J Oncol. 2021;43(10):1001–6. Wolff AC, Hammond MEH, Allison KH, et al. Human Epidermal Growth Factor Receptor 2 Testing in Breast Cancer: American Society of Clinical Oncology/College of American Pathologists Clinical Practice Guideline Focused Update. J Clin Oncol. 2018;36(20):2105–22. Glass B, Vandenberghe ME, Chavali ST, et al. Deployment of a Machine Learning Algorithm in a Real-World Cohort for Quality Control Monitoring of Human Epidermal Growth Factor-2-Stained Clinical Specimens in Breast Cancer. Arch Pathol Lab Med. 2025;149(8):751–60. Koh DM, Collins DJ. Diffusion-weighted MRI in the body: applications and challenges in oncology. AJR Am J Roentgenol. 2007;188(6):1622–35. Carr ME, Keenan KE, Beavan M, et al. Quantifying multi-institutional ADC measurement variability of 1.5 T MR-Linacs: A phantom and in vivo study. Med Phys. 2025;52(6):4120–33. He L, Li F, Qin Y, et al. Enhanced preoperative prediction of breast lesion pathology, prognostic biomarkers, and molecular subtypes using multiple models diffusion-weighted MR imaging. Sci Rep. 2025;15(1):4704. Gerlinger M, Rowan AJ, Horswell S, et al. Intratumor Heterogeneity and Branched Evolution Revealed by Multiregion Sequencing. N Engl J Med. 2012;366(10):883–92. Yuan Y. Spatial Heterogeneity in the Tumor Microenvironment. Cold Spring Harb Perspect Med. 2016;6(8):a026583. Wu H, Tong H, Du X, et al. Vascular habitat analysis based on dynamic susceptibility contrast perfusion MRI predicts IDH mutation status and prognosis in high-grade gliomas. Eur Radiol. 2020;30(6):3254–65. Yu R, Cai L, Gong Y, Sun X, Li K, Cao Q, Yang X, Lu Q. MRI-Based Machine Learning Radiomics for Preoperative Assessment of Human Epidermal Growth Factor Receptor 2 Status in Urothelial Bladder Carcinoma[J]. J Magn Reson Imaging. 2024;60(6):2694–704. Kong L, Ling J, Cao W, Wen Z, Lin Y, Cai Q, Chen Y, Guo Y, Chen J, Wang H. Multiparametric MR characterization for human epithelial growth factor receptor 2 expression in bladder cancer: an exploratory study[J]. Abdom Radiol (New York). 2024;49(7):2349–57. Arteaga CL, Engelman JA. ERBB receptors: from oncogene discovery to basic science to mechanism-based cancer therapeutics. Cancer Cell. 2014;25(3):282–303. Altriche N, Gallant S, Augustine TN, Xulu KR. Navigating the Intricacies of Tumor Heterogeneity: An Insight into Potential Prognostic Breast Cancer Biomarkers. Biomark Insights. 2024;19:11772719241256798. Sinha S, Choudhury S, Mishra J, Kundu G, Bera MK, Mondol PP. A study on immunohistochemical expression of HER2/Neu and p63 and its association with grade and invasiveness in case of bladder carcinoma. Urologia. 2024;91(2):284–8. Krüger S, Weitsch G, Büttner H, et al. Overexpression of c-erbB-2 oncoprotein in muscle-invasive bladder carcinoma: relationship with gene amplification, clinicopathological parameters and prognostic outcome. Int J Oncol. 2002;21(5):981–7. Vlachou E, Johnson BA, Hoffman-Censits J. The Role of Antibody-Drug Conjugates in Urothelial Cancer: A Review of Recent Advances in the Treatment of Locally Advanced and Metastatic Urothelial Cancer. Clin Med Insights Oncol. 2024;18:11795549241290787. Sheng X, Yan X, Wang L, et al. Open-label, Multicenter, Phase II Study of RC48-ADC, a HER2-Targeting Antibody-Drug Conjugate, in Patients with Locally Advanced or Metastatic Urothelial Carcinoma. Clin Cancer Res. 2021;27(1):43–51. Wang X, Xie Z, Wang X, et al. Preoperative prediction of IDH genotypes and prognosis in adult-type diffuse gliomas: intratumor heterogeneity habitat analysis using dynamic contrast-enhanced MRI and diffusion-weighted imaging. Cancer Imaging. 2025;25(1):11. Wang Y, Xie B, Wang K et al. Multi-parametric MRI Habitat Radiomics Based on Interpretable Machine Learning for Preoperative Assessment of Microsatellite Instability in Rectal Cancer. Acad Radiol 2025 Feb 26:S1076-6332(25)00111-4. Horvat-Menih I, Khan AS, McLean MA, et al. K-Means Clustering of Hyperpolarised 13C-MRI Identifies Intratumoral Perfusion/Metabolism Mismatch in Renal Cell Carcinoma as the Best Predictor of the Highest Grade. Cancers (Basel). 2025;17(4):569. He J, Shi H, Zhou Z, et al. Correlation between apparent diffusion coefficients and HER2 status in gastric cancers: pilot study. BMC Cancer. 2015;15:749. Martincich L, Deantoni V, Bertotto I, et al. Correlations between diffusion-weighted imaging and breast cancer biomarkers. Eur Radiol. 2012;22(7):1519–28. Choi BB, Kim SH, Kang BJ, et al. Diffusion-weighted imaging and FDG PET/CT: predicting the prognoses with apparent diffusion coefficient values and maximum standardized uptake values in patients with invasive ductal carcinoma. World J Surg Oncol. 2012;10:126. Kamoun A, de Reyniès A, Allory Y, et al. A Consensus Molecular Classification of Muscle-invasive Bladder Cancer. Eur Urol. 2020;77(4):420–33. El Saftawy E, Aboulhoda BE, Alghamdi MA, Abd Elkhalek MA, AlHariry NS. Heterogeneity of modulatory immune microenvironment in bladder cancer. Tissue Cell. 2025;93:102679. Dernbach G, Eich ML, Dragomir MP, et al. Spatial expression of HER2, NECTIN4, and TROP-2 in Muscle-Invasive Bladder Cancer and metastases: Implications for pathological and clinical management. Mod Pathol. 2025;38(7):100753. Xu C, Wang Z, Wang A, et al. Breast Cancer: Multi-b-Value Diffusion Weighted Habitat Imaging in Predicting Pathologic Complete Response to Neoadjuvant Chemotherapy. Acad Radiol. 2024;31(12):4733–42. Skwierawska D, Laun FB, Wenkel E, et al. Diffusion-Weighted Imaging for Skin Pathologies of the Breast-A Feasibility Study. Diagnostics (Basel). 2024;14(9):934. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 06 Mar, 2026 Reviews received at journal 13 Nov, 2025 Reviews received at journal 10 Nov, 2025 Reviewers agreed at journal 31 Oct, 2025 Reviewers agreed at journal 29 Oct, 2025 Reviewers invited by journal 29 Oct, 2025 Editor assigned by journal 28 Oct, 2025 Editor invited by journal 06 Oct, 2025 Submission checks completed at journal 06 Oct, 2025 First submitted to journal 06 Oct, 2025 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-7710608","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":541899905,"identity":"cc35a337-91d5-4c3a-9491-30d5098637b0","order_by":0,"name":"Zeke Chen","email":"","orcid":"","institution":"First Hospital of Shanxi Medical University","correspondingAuthor":false,"prefix":"","firstName":"Zeke","middleName":"","lastName":"Chen","suffix":""},{"id":541899906,"identity":"b1a87f7a-4590-4819-b2ea-6d17a9a63f01","order_by":1,"name":"Zhichang Fan","email":"","orcid":"","institution":"First Hospital of Shanxi Medical University","correspondingAuthor":false,"prefix":"","firstName":"Zhichang","middleName":"","lastName":"Fan","suffix":""},{"id":541899907,"identity":"fab5714f-5841-41a2-8ff8-d03303041f96","order_by":2,"name":"Xiaoyue Zhang","email":"","orcid":"","institution":"First Hospital of Shanxi Medical University","correspondingAuthor":false,"prefix":"","firstName":"Xiaoyue","middleName":"","lastName":"Zhang","suffix":""},{"id":541899908,"identity":"c9d8dc32-9d2d-479f-9f7d-51d8c9517f7a","order_by":3,"name":"Wenxin Li","email":"","orcid":"","institution":"First Hospital of Shanxi Medical University","correspondingAuthor":false,"prefix":"","firstName":"Wenxin","middleName":"","lastName":"Li","suffix":""},{"id":541899909,"identity":"1d51d0db-9e56-4b32-be02-20a644315c40","order_by":4,"name":"Yan Li","email":"","orcid":"","institution":"First Hospital of Shanxi Medical University","correspondingAuthor":false,"prefix":"","firstName":"Yan","middleName":"","lastName":"Li","suffix":""},{"id":541899910,"identity":"83eac8dd-73ca-405b-adc0-90c92d04d41e","order_by":5,"name":"Bin Wang","email":"","orcid":"","institution":"First Hospital of Shanxi Medical University","correspondingAuthor":false,"prefix":"","firstName":"Bin","middleName":"","lastName":"Wang","suffix":""},{"id":541899911,"identity":"0cf05fe7-a679-4519-93d8-b2d1ba68e804","order_by":6,"name":"Yongfang Wang","email":"","orcid":"","institution":"First Hospital of Shanxi Medical University","correspondingAuthor":false,"prefix":"","firstName":"Yongfang","middleName":"","lastName":"Wang","suffix":""},{"id":541899912,"identity":"7d2b2edd-46f9-4854-b8fc-3166d18d2c28","order_by":7,"name":"Guoqiang Yang","email":"","orcid":"","institution":"First Hospital of Shanxi Medical University","correspondingAuthor":false,"prefix":"","firstName":"Guoqiang","middleName":"","lastName":"Yang","suffix":""},{"id":541899913,"identity":"d5c1ee66-ac05-48af-aeb1-dafa51ed3518","order_by":8,"name":"Xiaochun Wang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA5klEQVRIie3PIQvCQBTA8TsOXDlcfUPwMzwYTMOYX8UxuLRgshjcGMziBzjwS5jM6qEWP8CCQRHMiwaDU7HJTpvh/vDghfcLjxCT6Q+zGU2WBKsNqinRD7TEmWTp8U2oHIhIS3C/y/C5VYTxck0TrSnCvMUHQdueTb2zj0tGLLWZ1wkqw9yRGLlw2HfcGA9NwoUo6giDMIcSWTiH2GvFeGEEuFdLGg/Sx/GLdFHRREc4X2VYonoR8g0BK01PEncuFGLoTFFEDd0vPWUdFb+N2raMFnC9+YFtqW0t+fDdb+cmk8lk+tQdIq1HI7RpTZQAAAAASUVORK5CYII=","orcid":"","institution":"First Hospital of Shanxi Medical University","correspondingAuthor":true,"prefix":"","firstName":"Xiaochun","middleName":"","lastName":"Wang","suffix":""}],"badges":[],"createdAt":"2025-09-25 08:38:22","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7710608/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7710608/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":95655528,"identity":"e55abb77-bf3f-4ee0-8958-4b645f0a1416","added_by":"auto","created_at":"2025-11-11 16:16:24","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":56523,"visible":true,"origin":"","legend":"","description":"","filename":"Maintextdocument2.docx","url":"https://assets-eu.researchsquare.com/files/rs-7710608/v1/8b0225e035830176de0e544c.docx"},{"id":95565764,"identity":"21e4099f-1523-4d70-b120-3832d9b5a80f","added_by":"auto","created_at":"2025-11-10 16:17:28","extension":"tiff","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2459974,"visible":true,"origin":"","legend":"","description":"","filename":"Fig.1.tiff","url":"https://assets-eu.researchsquare.com/files/rs-7710608/v1/cf104a5da86cb1ee25551215.tiff"},{"id":95655552,"identity":"f93f88dc-3b38-4d8f-bdda-81a656cf8e45","added_by":"auto","created_at":"2025-11-11 16:16:27","extension":"tiff","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2657994,"visible":true,"origin":"","legend":"","description":"","filename":"Fig.2.tiff","url":"https://assets-eu.researchsquare.com/files/rs-7710608/v1/68093e6af6da107df48a20bf.tiff"},{"id":95565756,"identity":"dd46596e-b501-4ace-b209-3c1ca1f0153a","added_by":"auto","created_at":"2025-11-10 16:17:27","extension":"tiff","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1683014,"visible":true,"origin":"","legend":"","description":"","filename":"Fig.3.tiff","url":"https://assets-eu.researchsquare.com/files/rs-7710608/v1/9baeacd41fd9ce46be858fca.tiff"},{"id":95655700,"identity":"c0db704d-8c85-4007-ae4e-b8b91889171a","added_by":"auto","created_at":"2025-11-11 16:16:44","extension":"tiff","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1679708,"visible":true,"origin":"","legend":"","description":"","filename":"Fig.4.tiff","url":"https://assets-eu.researchsquare.com/files/rs-7710608/v1/c809a9880404031155699cf3.tiff"},{"id":95565771,"identity":"67a8ad34-f437-4ff8-a545-23fe3804df27","added_by":"auto","created_at":"2025-11-10 16:17:28","extension":"tiff","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2381612,"visible":true,"origin":"","legend":"","description":"","filename":"Fig.5.tiff","url":"https://assets-eu.researchsquare.com/files/rs-7710608/v1/5e1aaeb274ddd3b93225d135.tiff"},{"id":95654924,"identity":"303f4101-5796-4b2f-a187-07b29370953f","added_by":"auto","created_at":"2025-11-11 16:13:50","extension":"tiff","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":975132,"visible":true,"origin":"","legend":"","description":"","filename":"Fig.6.tiff","url":"https://assets-eu.researchsquare.com/files/rs-7710608/v1/63dc6827b0c3db96868fa49d.tiff"},{"id":95654631,"identity":"712a1402-7c2b-4711-a223-37f6e498dfd6","added_by":"auto","created_at":"2025-11-11 16:12:37","extension":"json","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":10984,"visible":true,"origin":"","legend":"","description":"","filename":"5159709c9a7a4c93b7f254f17dd6aa33.json","url":"https://assets-eu.researchsquare.com/files/rs-7710608/v1/d11f575be32a292d76be5dce.json"},{"id":95654636,"identity":"79385bcb-88e2-435e-baef-7cff66cc4da1","added_by":"auto","created_at":"2025-11-11 16:12:37","extension":"xml","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":113221,"visible":true,"origin":"","legend":"","description":"","filename":"5159709c9a7a4c93b7f254f17dd6aa331enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7710608/v1/b3ae3a3e63e0a0605ac0071b.xml"},{"id":95565766,"identity":"6cc62099-65d4-4e08-bdff-b3e281eee0de","added_by":"auto","created_at":"2025-11-10 16:17:28","extension":"tiff","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2459974,"visible":true,"origin":"","legend":"","description":"","filename":"Fig.1.tiff","url":"https://assets-eu.researchsquare.com/files/rs-7710608/v1/89b08d788c0ead8b66092a22.tiff"},{"id":95655586,"identity":"578fc5d7-7ccf-4a36-b115-7b1ee1ce0832","added_by":"auto","created_at":"2025-11-11 16:16:32","extension":"tiff","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2657994,"visible":true,"origin":"","legend":"","description":"","filename":"Fig.2.tiff","url":"https://assets-eu.researchsquare.com/files/rs-7710608/v1/981f23ccda725b45f8fb0ee9.tiff"},{"id":95565768,"identity":"1393fa94-ca38-479e-973e-692cac6b6a48","added_by":"auto","created_at":"2025-11-10 16:17:28","extension":"tiff","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1683014,"visible":true,"origin":"","legend":"","description":"","filename":"Fig.3.tiff","url":"https://assets-eu.researchsquare.com/files/rs-7710608/v1/f22080a2b60b65cf56ed74c5.tiff"},{"id":95654887,"identity":"d6f489eb-7239-4ea5-b1d5-46cab2509c81","added_by":"auto","created_at":"2025-11-11 16:13:40","extension":"tiff","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1679708,"visible":true,"origin":"","legend":"","description":"","filename":"Fig.4.tiff","url":"https://assets-eu.researchsquare.com/files/rs-7710608/v1/17ae1e06278d9dbde25d1b00.tiff"},{"id":95565779,"identity":"a7f333aa-b06e-48f3-ae76-152d1a3c1dba","added_by":"auto","created_at":"2025-11-10 16:17:28","extension":"tiff","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2381612,"visible":true,"origin":"","legend":"","description":"","filename":"Fig.5.tiff","url":"https://assets-eu.researchsquare.com/files/rs-7710608/v1/7faca9a03ff877c16d4da949.tiff"},{"id":95565772,"identity":"6c83dc8a-0561-4454-9ba6-3946c5f30e5f","added_by":"auto","created_at":"2025-11-10 16:17:28","extension":"tiff","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":975132,"visible":true,"origin":"","legend":"","description":"","filename":"Fig.6.tiff","url":"https://assets-eu.researchsquare.com/files/rs-7710608/v1/0700191ab22985bbb991c20c.tiff"},{"id":95654607,"identity":"a0ea855c-a070-4342-8f83-cccd8763eb47","added_by":"auto","created_at":"2025-11-11 16:12:33","extension":"png","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":20950,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFig.1.png","url":"https://assets-eu.researchsquare.com/files/rs-7710608/v1/95b55a94618e41a17c9f7687.png"},{"id":95655921,"identity":"978b6a70-3c2d-4961-94bf-3716d2fd3cbe","added_by":"auto","created_at":"2025-11-11 16:17:18","extension":"png","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":87461,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFig.2.png","url":"https://assets-eu.researchsquare.com/files/rs-7710608/v1/7a1c89443071299cf1668775.png"},{"id":95565777,"identity":"40f4f025-9f98-462c-b66a-a3d524c89fa2","added_by":"auto","created_at":"2025-11-10 16:17:28","extension":"png","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":106932,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFig.3.png","url":"https://assets-eu.researchsquare.com/files/rs-7710608/v1/64f685d7fd10053e537409af.png"},{"id":95655671,"identity":"9d72d18f-49d7-4fdd-b848-edda0d29a60c","added_by":"auto","created_at":"2025-11-11 16:16:41","extension":"png","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":147094,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFig.4.png","url":"https://assets-eu.researchsquare.com/files/rs-7710608/v1/ab23b98c16234c178a69c6c3.png"},{"id":95565767,"identity":"30dcb2f8-dbc2-40b9-a8f5-b1d126e92975","added_by":"auto","created_at":"2025-11-10 16:17:28","extension":"png","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":41568,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFig.5.png","url":"https://assets-eu.researchsquare.com/files/rs-7710608/v1/084436da2ec541bbd94f5fe9.png"},{"id":95655049,"identity":"f2e71468-b6ae-494b-a7ad-44b78d836d37","added_by":"auto","created_at":"2025-11-11 16:14:02","extension":"png","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":13980,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFig.6.png","url":"https://assets-eu.researchsquare.com/files/rs-7710608/v1/ce883c184c86dfcf644dbe3c.png"},{"id":95565776,"identity":"825908ea-74ea-4c81-bfac-43d54a4a6fb6","added_by":"auto","created_at":"2025-11-10 16:17:28","extension":"xml","order_by":21,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":108644,"visible":true,"origin":"","legend":"","description":"","filename":"5159709c9a7a4c93b7f254f17dd6aa331structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7710608/v1/a198257f8f91d1bf8426d542.xml"},{"id":95655313,"identity":"0034d090-3315-413b-a5e1-3ee001c9abcf","added_by":"auto","created_at":"2025-11-11 16:15:19","extension":"html","order_by":22,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":119826,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7710608/v1/d1aeadbff1bd3f1690ab270a.html"},{"id":95565750,"identity":"fa4f585e-68bc-471e-9dc0-2f87efd7852f","added_by":"auto","created_at":"2025-11-10 16:17:27","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":573499,"visible":true,"origin":"","legend":"\u003cp\u003eFlow chart of patient screening.\u003c/p\u003e","description":"","filename":"Fig.1.png","url":"https://assets-eu.researchsquare.com/files/rs-7710608/v1/2915f0028c1ae4cb066bad11.png"},{"id":95565752,"identity":"533dc1c5-adb9-422b-a15f-ee67977fd182","added_by":"auto","created_at":"2025-11-10 16:17:27","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":3406140,"visible":true,"origin":"","legend":"\u003cp\u003eWorkflow of the habitat analysis.\u003c/p\u003e","description":"","filename":"Fig.2.png","url":"https://assets-eu.researchsquare.com/files/rs-7710608/v1/364a27efdba3a2384ca8b8f9.png"},{"id":95655273,"identity":"bf9dbf2b-9e33-4717-87c1-d12379ba02e7","added_by":"auto","created_at":"2025-11-11 16:15:03","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":3722420,"visible":true,"origin":"","legend":"\u003cp\u003eCase 1, representative case of a patient with HER2 non-overexpression BCa (HER2 1+). (A-D) The preoperative DWI and ADC images and their corresponding habitat images of the patients. (E-G) Microscopic image of hematoxylin-eosin staining (original magnification, *100) and postoperative pathological IHC image of the patient(original magnification, *100).\u003c/p\u003e","description":"","filename":"Fig.3.png","url":"https://assets-eu.researchsquare.com/files/rs-7710608/v1/c80883735bbaa636ab2fb0a0.png"},{"id":95655078,"identity":"eb2866c3-ec25-4a83-a10f-7fcfc8cfaa54","added_by":"auto","created_at":"2025-11-11 16:14:17","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":3997789,"visible":true,"origin":"","legend":"\u003cp\u003eCase 2, representative case of a patient with HER2 overexpression BCa (HER2 2+). (A-D) The preoperative DWI and ADC images and their corresponding habitat images of the patients. (E-G) Microscopic image of hematoxylin-eosin staining (original magnification, *100) and postoperative pathological IHC image of the patient(original magnification, *100).\u003c/p\u003e","description":"","filename":"Fig.4.png","url":"https://assets-eu.researchsquare.com/files/rs-7710608/v1/0644ac524111edd2e91d0265.png"},{"id":95565762,"identity":"707b6108-9308-468b-afbf-f675adb1b937","added_by":"auto","created_at":"2025-11-10 16:17:28","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":1124181,"visible":true,"origin":"","legend":"\u003cp\u003e(a) SVM model receiver operating characteristic curves (ROC) in the training cohort, validation cohort, and test cohort. (b-d)The SVM model calibration curve of the training cohort、validation cohort, and test cohort.\u003c/p\u003e","description":"","filename":"Fig.5.png","url":"https://assets-eu.researchsquare.com/files/rs-7710608/v1/6998a5e929b8246fc68125b9.png"},{"id":95565754,"identity":"3859f4d8-3922-4275-8d5c-0290825bff6d","added_by":"auto","created_at":"2025-11-10 16:17:27","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":378331,"visible":true,"origin":"","legend":"\u003cp\u003eThe SVM model decision curve of the training cohort(red)、validation cohort(blue), and test cohort (green).\u003c/p\u003e","description":"","filename":"Fig.6.png","url":"https://assets-eu.researchsquare.com/files/rs-7710608/v1/8e8cd42842526813b0df3c3b.png"},{"id":95797451,"identity":"08819704-54ce-4e9b-8af6-b6ac871f055d","added_by":"auto","created_at":"2025-11-13 08:05:26","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":12944053,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7710608/v1/e7b49f68-3b83-4abd-b112-5ca84eecd280.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"DWI and ADC Habitat Imaging in Predicting HER2 Expression Status in Bladder Cancer: A Retrospective Study","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eBladder cancer (BCa)\u0026nbsp;ranks among the most prevalent malignancies of the urinary system and the\u0026nbsp;10th most common cancer globally, with its incidence steadily rising [1,2]. The\u0026nbsp;human epidermal growth factor receptor 2 (HER2)\u0026nbsp;has emerged as a critical biomarker for diagnosing, prognosing, and predicting therapeutic responses in urothelial carcinomas [3]. Previous studies have confirmed that HER2-targeted therapies such as\u0026nbsp;RC48-ADC\u0026nbsp;have demonstrated remarkable efficacy in\u0026nbsp;HER2-overexpressing advanced BCa, achieving an\u0026nbsp;objective response rate of 50% and a disease control rate of 100%\u0026nbsp;[4].\u003c/p\u003e\n\u003cp\u003eImmunohistochemistry (IHC) is a low-cost and effective technique, which is widely used in the expression of HER2 protein in tissues and cells [5-7]. However, IHC has some limitations. First, it is an invasive method that is complex and time-consuming to operate. Second, different pathologists interpret the results differently. Finally, pathological samples are inevitably affected by surgery, and the limited samples can not fully reflect the expression of HER2 protein in the whole tumor [8]. Therefore, it is essential to find a preoperative non-invasive method to determine the status of HER2 in BCa.\u003c/p\u003e\n\u003cp\u003eDiffusion weighted imaging (DWI) and apparent diffusion coefficient (ADC) are functional magnetic resonance imaging (MRI) techniques, which can reflect the Brownian motion of water molecules in different tissues and cells of the human body [9]. There is increasing evidence that DWI and ADC can be used as imaging biomarkers to characterize the pathophysiology of various types of malignant tumors [10, 11].\u003c/p\u003e\n\u003cp\u003eHabitat analysis, also known as habitat imaging, is an image analysis technology that divides the region of interest into several sub-regions with the same or similar heterogeneity according to the differences between the macro level and the micro level of the region of interest [12-14]. MRI habitat imaging, by combining quantitative information of images with tumor heterogeneity, captures subtle differences within tumors and can reflect tumor cellular and molecular heterogeneity non-invasively. Previous studies have shown that MRI-based imaging radiomics models can be used to predict the expression status of HER2 in BCa [15,16]. However, there are still some shortcomings in the prediction of HER2 expression in BCa by traditional radiomics, and the prediction performance of the model needs to be improved. Traditional imaging usually regards tumors as a whole with relatively uniform internal distribution, which can not truly express tumor heterogeneity.\u003c/p\u003e\n\u003cp\u003eTherefore, this study aimed to explore the value of DWI and ADC habitat imaging in predicting HER2 expression status in BCa.\u003c/p\u003e"},{"header":"2. Materials and Methods","content":"\u003cp\u003eThis retrospective study (approval number: KYYJ-2023-064) was approved by the local institutional review board, and the need for written informed consent was waived.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.1 Patients\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFrom November 2022 to February 2024, patients pathologically diagnosed with BCa were retrospectively selected. A total of 232 patients were included in this study. The patient enrollment pathway is presented in Fig. 1. The inclusion criteria comprised the following: (1) confirmed by postoperative pathological examination as BCa; (2) preoperative multiparameter bladder MRI examination; (3) the expression status of HER2 in BCa can be determined by IHC testing of postoperative specimens. The exclusion criteria were as follows: (1) Incomplete baseline data (n = 135); (2) Patients who received preoperative radiotherapy and chemotherapy (n = 67); (3) patients with lesions \u0026lt;5 mm (n = 18); (4) patients with nonurothelial carcinoma (n = 17); and (5) patients with poor image quality (n = 5). \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFig. 1.\u0026nbsp;Flow chart of patient screening.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.2 MRI Image Acquisition\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo ensure that the bladder is properly filled, patients are required to avoid drinking or urinating before the completion of the scan and drink 500-1000 mL of water half an hour before the scan.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTwo 3.0T MR scanners from Siemens in Germany (MAGNETOM Vida, MAGNETOM Skyra) were used. The scanning range covers the pelvic cavity (bladder), and the scanning sequence includes DWI. The scanning parameters of axial DWI sequence are vida: repetition time (TR) = 4100msec, echo time (TE) = 65 msec, layer thickness, 6 mm, layer spacing, 1.2mm, field of view (FOV) = 280mmx245mm, 19 layers; Skyra: TR = 4000msec, TE = 76msec, layer thickness, 6mm, the layer spacing, 1.2mm, FOV = 280mmx245mm, and the number of layers is 19. Two b values (b0: 0 sec/mm2; b1:800 sec/mm2). ADC image was obtained by dual-exponential model transformation of the DWI image on an image post-processing workstation.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.3 Histopathological Assessment\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe most representative morphological sections were selected from the postoperative pathological tissues of all patients, and the results were independently interpreted by two pathologists with more than 5 years of working experience. HER2 results were interpreted according to the clinical pathological expert consensus on HER2 testing in urothelial carcinoma in China [6]. The HER2 results were then categorized as follows: HER2 non-overexpressing (IHC score of 0 or 1+), HER2 overexpression (IHC score of 3+ or 2+).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.4 Image Preprocessing and Delineation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.4.1 Image Preprocessing\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePython 3.7.6 (https://www.python.org) software was used to construct the code for the N4-bias field correction algorithm to eliminate the variation of image signal strength caused by different scanners, noise, and many unknown problems.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.4.2 Region of Interest Delineation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTwo independent radiologists (A and B, with 5 and 2 years of pelvic imaging experience, respectively) using ITK-SNAP 4.0 (http://www.ITK-SNAP.org/) manually delineated the entire tumor area layer by layer on DWI and ADC images to obtain the volume of interest (VOI). For patients with multiple tumor lesions, the largest lesion was selected for delineation. Use the inter-class correlation coefficient (ICC) to evaluate the consistency of selected regions among observers. After 30 days, 30 patients were randomly selected and their VOI was redrawn by physician A. The intra-class correlation coefficient (ICC) was used to evaluate the consistency of the selected areas by the same physician at different times.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.5 Clustering to Generate Tumor Habitat\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe habitats analysis was implemented by an in-house software nnFAE, which was developed based on Python. The K-means clustering algorithm is used to automatically segment the tumor region to construct the habitat. Silhouette coefficient is an important indicator for measuring clustering performance, which can reflect the degree of separation between clusters. When the contour coefficient is high, it means that the similarity between different clusters is low, and the similarity within clusters is high. We calculated the contour coefficients of clustering results under different K values to select the optimal K value [14].\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.6 Extract Features and Model Construction\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.6.1 Extract Habitat Features\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe internal software nnFAE developed based on Python, was used to extract the features of DWI and ADC subregions of each bladder cancer patient, including the volume of each subregion and the corresponding volume percentage of each subregion. Using the pyradiomics toolkit (https://github.com/Radiomics/pyradiomics), extract the histogram features of each subregion. Histogram features include maximum, minimum, skewness, kurtosis, entropy, etc.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.6.2 Establishment of Habitat Prediction Model\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTo improve the repeatability of the model, features with consistency greater than 0.75 intra- and inter-observer are selected. Then the Spearman correlation coefficient between features is calculated to evaluate the correlation between features. When the correlation coefficient is greater than 0.7, it is considered that there is a high correlation between features. To optimize feature selection, the features with a narrow data distribution range among the two highly correlated features are screened out, and the optimal features are finally selected.\u003c/p\u003e\n\u003cp\u003eThe support vector machine (SVM) model is constructed in the training set by using the optimal features, and the performance of the model is evaluated in the validation and test set. Receiver operating characteristic (ROC) was plotted and the area under the curve (AUC), accuracy, specificity, sensitivity, positive predictive value (PPV), and negative predictive value (NPV) were calculated.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.7 Statistical analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eStatistical analysis was performed using R4.4.0 (http://www.r-project.org/ version) or SPSS 13.0. The mean \u0026plusmn; standard deviation was used if the measurement data conformed to normality and homogeneity of variance, and the two independent samples t-test and variance test were used for comparison between groups. If it does not conform to the normality and or homogeneity of variance and is expressed by the median \u0026plusmn; upper and lower quartiles, comparisons between groups were performed using Kruskal-Wallis or Mann-Whitney U tests. The count data were presented as numerical examples. \u0026chi;2 test and Fisher\u0026apos;s exact test were used for comparison between groups. ICC was used to evaluate the consistency between different physicians and the same physician at different times. When ICC \u0026ge; 0.75, the consistency is considered to be high. AUC, Calibration curve, and decision curve analysis (DCA) are used to evaluate the performance of the SVM model. When \u003cem\u003ep\u003c/em\u003e\u0026lt;0.05, the difference was considered statistically significant. The specific workflow is shown in Fig. 2.\u003c/p\u003e\n\u003cp\u003eFig. 2. Workflow of the habitat analysis.\u003c/p\u003e"},{"header":"3. Results","content":"\u003cp\u003e\u003cstrong\u003e3.1 Patients\u003c/strong\u003e\u003cstrong\u003e\u0026rsquo;\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;Characteristics\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn this study, 232 patients with bladder urothelial carcinoma were randomly divided into a train set (n=148), validation set (n=47), and test set (n=37) according to the ratio of 6:2:2. The clinical and pathological parameters including age, sex, smoking, pathological stage, etc were also collected and analyzed. Table 1 presents the clinical and pathological characteristics. There were no significant differences in clinical and pathological variables between cohorts (all \u003cem\u003ep\u0026nbsp;\u003c/em\u003e\u0026gt; 0.05).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003e1\u003c/strong\u003e\u003cstrong\u003e.\u0026nbsp;\u003c/strong\u003eClinical and Pathological Characteristics of Patients.\u003c/p\u003e\n\u003cdiv\u003e\n \u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"586\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\"\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eTraining Cohort（n=148）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eValidation Cohort\u003c/p\u003e\n \u003cp\u003e（n=47）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eTest Cohort（n=37）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003ep\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\"\u003e\n \u003cp\u003eAge\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e68.66（59.33-73.67）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e69.12\u003c/p\u003e\n \u003cp\u003e(65.49-76.20)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e67.92\u003c/p\u003e\n \u003cp\u003e(60.28-75.10)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.67\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\"\u003e\n \u003cp\u003eGender\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.85\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eMale\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e121(81.8%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e39(83.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e32(86.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eFemale\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e27(18.2%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e8(17.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e5(13.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\"\u003e\n \u003cp\u003eSmoking\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.19\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e57(38.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e25(53.2%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e17(45.9%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eNo\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e91(61.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e22(46.8%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e20(54.1%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\"\u003e\n \u003cp\u003ePathological stage\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.99\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\"\u003e\n \u003cp\u003eMIBC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e55(37.2%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e18(38.3%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e14(37.8%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\"\u003e\n \u003cp\u003eNMIBC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e93(62.8%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e29(61.7%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e23(62.2%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\"\u003e\n \u003cp\u003eTumor level\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.54\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\"\u003e\n \u003cp\u003eHigh grade\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e77(52.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e25(53.2%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e23(62.2%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\"\u003e\n \u003cp\u003eLow grade\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e71(48.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e22(46.8%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e14(37.8%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\"\u003e\n \u003cp\u003eHER2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.64\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\"\u003e\n \u003cp\u003eOverexpression\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e92(62.2%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e27(57.4%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e25(67.6%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"2\"\u003e\n \u003cp\u003eNon-overexpression\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e56(37.8%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e20(42.6%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e12(32.4%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\u003cbr\u003e\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003eMIBC: muscle-invasive bladder cancer; NMIBC: non-muscle-invasive bladder cancer\u003cstrong\u003e\u003cbr\u003e\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.2 Habitat Clustering and Feature Selection\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAccording to the silhouette coefficient, a K = 2 clustering parameter is selected to cluster DWI and ADC images into two sub-regions respectively. Fig. 3 and Fig. 4 show representative habitat imaging of two correctly classified lesions with different HER2 expressions. There was an obvious difference in the spatial distribution of the habitats for various HER2 expressions.\u003c/p\u003e\n\u003cp\u003eFig. 3: Case 1, representative case of a patient with HER2 non-overexpression BCa (HER2 1+). (A-D) The preoperative DWI and ADC images and their corresponding habitat images of the patients. (E-G) Microscopic image of hematoxylin-eosin staining (original magnification, *100) and postoperative pathological IHC image of the patient(original magnification, *100).\u003c/p\u003e\n\u003cp\u003eFig. 4: Case 2, representative case of a patient with HER2 overexpression BCa (HER2 2+). (A-D) The preoperative DWI and ADC images and their corresponding habitat images of the patients. (E-G) Microscopic image of hematoxylin-eosin staining (original magnification, *100) and postoperative pathological IHC image of the patient(original magnification, *100).\u003c/p\u003e\n\u003cp\u003eInitially, a total of 80 features were extracted from four sub-regions of the two sequences, with 40 features each from ADC and DWI images. The ICC for each of these habitat features exceeds 0.75. When the Spearman correlation coefficient between two features is greater than 0.7, the features with a narrow data distribution range among the two highly correlated features are filtered out. In the final DWI image, habitat 1 and habitat 2 retain four features respectively, while in the ADC image, habitat 1 and habitat 2 retain four and five features respectively. The SVM model was constructed using these 17 habitat characteristics (Table 2 and Table 3).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 2.\u0026nbsp;\u003c/strong\u003eHabitat features of the training cohort and validation cohort.\u003c/p\u003e\n\u003cdiv\u003e\n \u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"580\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\"\u003e\n \u003cp\u003eFeatures\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\"\u003e\n \u003cp\u003eTraining and validation cohorts (n=195)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eOverexpression\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eNon-overexpression\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003ep\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eDWI_Habitat 1_Minimum\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e93.29(1.86)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e92.92(93.95)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.003\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eDWI_Habitat1_Energy\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e2243836.45\u003c/p\u003e\n \u003cp\u003e(10442044.48)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e779561.18\u003c/p\u003e\n \u003cp\u003e(6014724.49)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.22\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eDWI_Habitat1_Kurtosis\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e-0.40(1.33)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0(0.83)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.07\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eDWI_Habitat1_Skewness\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.50(0.80)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.22(0.65)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.20\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eDWI_Habitat2_Maximum\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e90.39(2.20)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e90.11(4.12)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.13\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eDWI_Habitat2_Uniformity\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e4321.51(1505.19)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e3734.69(2054.14)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.17\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eDWI_Habitat2_Kurtosis\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e-0.80(0.58)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e-0.82(0.58)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.17\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eDWI_Habitat2_Variance\u003csup\u003e\u0026amp;\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e261.48\u0026plusmn;109.68\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e249.45\u0026plusmn;116.69\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.55\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eADC_Habitat1_Minimum\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e94.10(0.31)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e94.11(0.79)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.04\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eADC_Habitat1_Uniformity\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e15339.71(3568.16)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e15297.23(31433.09)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.41\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eADC_Habitat1_Energy\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e2727387.71(7324436.08)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e3317559.67(4105059.34)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.16\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eADC_Habitat1_Kurtosis\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e-0.27(0.91)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e-0.25(1.07)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.26\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eADC_Habitat2_Maximum\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e93.90(0.26)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e93.78(0.48)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.02\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eADC_Habitat2_Uniformity\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e5302.16(1816.21)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e5099.39(1938.89)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.17\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eADC_Habitat2_Kurtosis\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e-0.36(1.18)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0(2.04)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.15\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eADC_Habitat2_Skewness\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e-0.21(0.90)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e-0.07(1.12)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.17\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eADC_Habitat2_Variance\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e138.94(122.76)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e121.23(116.03)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.34\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003eData are the mean \u0026plusmn; standard deviation or median (inter-quartile range, IQR). \u003cem\u003ep\u003c/em\u003e values were calculated with the t-test \u003csup\u003e\u0026amp;\u003c/sup\u003e or Mann\u0026ndash;Whitney U test\u003csup\u003e*\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 3.\u0026nbsp;\u003c/strong\u003eHabitat features of the test cohort\u003c/p\u003e\n\u003cdiv\u003e\n \u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"587\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\"\u003e\n \u003cp\u003eFeatures\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"3\"\u003e\n \u003cp\u003eTest cohort (n=37)\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eOverexpression\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eNon-overexpression\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e\u003cem\u003ep\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eDWI_Habitat1_Minimum\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e93.23(1.07)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e92.33(93.49)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eDWI_Habitat1_Energy\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e4230308.23(29832411.04)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e4001832.78(8526170.1)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.41\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eDWI_Habitat1_Kurtosis\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e-0.67(0.88)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0(0.51)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.59\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eDWI_Habitat1_Skewness\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.40(0.66)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.39(0.80)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.93\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eDWI_Habitat2_Maximum\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e90.48(1.43)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e90.21(4.15)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.15\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eDWI_Habitat2_Uniformity\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e4343.18(1142.08)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e4387.63(2517.54)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.10\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eDWI_Habitat2_Kurtosis\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e-0.79(0.31)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e-0.67(0.54)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.49\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eDWI_Habitat2_Variance\u003csup\u003e\u0026amp;\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e275.87\u0026plusmn;96.74\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e247.37\u0026plusmn;88.74\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.72\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eADC_Habitat1_Minimum\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e94.07 (0.20)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e94.07(0.12)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.97\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eADC_Habitat1_Uniformity\u003csup\u003e\u0026amp;\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e15625.22(2409.96)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e16368.66(4330.80)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.03\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eADC_Habitat1_Energy\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e4478899.05(5868083.84)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e2601169.44(11336955.85)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.29\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eADC_Habitat1_Kurtosis\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e-0.02(1.04)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e-0.52(0.79)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.37\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eADC_Habitat2_Maximum\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e93.86(0.53)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e93.92(0.08)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.80\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eADC_Habitat2_Uniformity\u003csup\u003e\u0026amp;\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e5404.46\u0026plusmn;1244.70\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e5595.93\u0026plusmn;1318.08\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.97\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eADC_Habitat2_Kurtosis\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e-0.46(0.91)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.55(2.39)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.12\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eADC_Habitat2_Skewness\u003csup\u003e\u0026amp;\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e-0.30\u0026plusmn;0.61\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e-0.66\u0026plusmn;0.98\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.05\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eADC_Habitat2_Variance\u003csup\u003e*\u003c/sup\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e129.96(92.36)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e177.04(130.30)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.30\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003eData are the mean \u0026plusmn; standard deviation or median (inter-quartile range, IQR). \u003cem\u003ep\u003c/em\u003e values were calculated with the t-test \u003csup\u003e\u0026amp;\u003c/sup\u003e or Mann\u0026ndash;Whitney U test\u003csup\u003e*\u003c/sup\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.3 Construction and Evaluation of Habitat Prediction Model\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe tumor VOI was clustered based on ADC and DWI sequences, and the habitat characteristics were extracted and screened. SVM was used to construct a habitat model for predicting the expression of HER2 in BCa. AUC, 95% confidence interval (95% CI), accuracy, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were used to evaluate the performance of the prediction model. Table 4 and Fig. \u0026nbsp;5a show the performance of the SVM model in evaluating the HER2 expression status in the training, validation, and test sets. Within the training cohort, the SVM model achieved an AUC of 0.88 (95% CI: 0.82-0.93), accuracy of 0.84, sensitivity of 0.83, specificity of 0.86, PPV of 0.91, and NPV of 0.75. In the validation cohort, the SVM models achieved an AUC of 0.85 (95% CI: 0.72-0.94), accuracy of 0.81, sensitivity of 0.85, specificity of 0.75, PPV of 0.82, NPV of 0.79; and the test achieved an AUC of 0.84 (95% CI: 0.68-0.94), \u0026nbsp;accuracy of 0.78, sensitivity of 0.72, specificity of 0.92, PPV of 0.95, NPV of 0.61. We also calculated the calibration curve (Fig. 5b-5d) and decision curve (Fig. 6). As shown in Fig. 5 and Fig. 6, our prediction model exhibits stability and considerable predictive ability. It shows that the model has no obvious difference between the predicted value and the actual observation value, has good consistency, and has good correction efficiency, which is considered to be one of the good clinical application values.\u003c/p\u003e\n\u003cp\u003eFig. 5. (a) SVM model receiver operating characteristic curves (ROC) in the training cohort, validation cohort, and test cohort. (b-d)The SVM model calibration curve of the training cohort、validation cohort, and test cohort.\u003c/p\u003e\n\u003cp\u003eFig. 6. The SVM model decision curve of the training cohort(red)、validation cohort(blue), and test cohort (green).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 4.\u003c/strong\u003e AUCs for the Performance of the SVM Models in All Cohorts\u003c/p\u003e\n\u003cdiv\u003e\n \u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\" width=\"583\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eCohort\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eAUC\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eAccuracy\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eSensitivity\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eSpecificity\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003ePPV\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003eNPV\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eTraining\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.88(0.82-0.93)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.84\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.83\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.86\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.91\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.75\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eValidation\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.85(0.72-0.94)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.81\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.85\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.75\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.82\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.79\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd\u003e\n \u003cp\u003eTest\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.84(0.68-0.94)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.78\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.72\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.92\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.95\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd\u003e\n \u003cp\u003e0.61\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eThis study confirms that extracting habitat features based on DWI and ADC subregions and constructing an SVM model can be used to predict the expression status of HER2 in BCa. Good predictive performance was demonstrated in the training set, validation set, and test set. This research may reveal the feasibility and clinical value of MRI habitat in the preoperative non-invasive assessment of HER2 status in BCa.\u003c/p\u003e\n\u003cp\u003eHER2 is a member of the epidermal growth factor receptor family, with tyrosine kinase activity, involved in signal transduction for cell growth and differentiation. HER2 protein plays an important role in cell proliferation, differentiation, and angiogenesis, and its high expression can promote cell division, proliferation and differentiation [17]. HER2 plays a key role in the development and progression of a variety of malignancies, including breast cancer and urothelial cancer, etc [18,19]. Previous studies have shown that HER2 overexpression is an independent predictor of BCa-related survival, and HER2 overexpression is significantly associated with poor prognosis [20]. Studies have found that anti-HER-2 antibody-drug conjugates show good efficacy and safety in the treatment of patients with locally advanced or metastatic urothelial carcinoma with overexpression of HER2, and can bring significant clinical benefit to this subset of patients [21,22]. Therefore, it is essential to develop a non-invasive and effective method to assess the expression status of HER2 in BCa.\u003c/p\u003e\n\u003cp\u003ePrevious studies have shown that MRI-based radiomics models can be used to predict the expression of HER2 in BCa [15,16]. However, traditional radiomics still has shortcomings in predicting HER2 expression in BCa, and its predictive efficacy needs to be improved. Traditional radiomics often considers the tumor as a homogeneous whole and fails to fully reflect the heterogeneity within the tumor.\u003c/p\u003e\n\u003cp\u003eHabitat imaging can reflect the internal heterogeneity of tumors. In this study, the K-means clustering method [23-25] was used to cluster the VOI of BCa, divide the VOI into multiple different subregions, and perform feature extraction on different subregions, so as to improve the accuracy of feature extraction, an SVM model was constructed to predict the HER2 expression status of BCa. The results indicate that there were some differences in the habitat subregions of HER2-overexpressing and HER2 non-overexpressing BCa. The range of red subregions with higher ADC values in HER2 overexpressing BCa patients is larger than that in HER2 non-overexpressing BCa patients. Similarly, compared to HER2 non-overexpressing BCa, HER2 overexpressing BCa patients have a smaller range of green subregions with lower ADC values. This is consistent with previous studies on other tumors [26-28]. The reason may be that HER2 is highly expressed in luminal unstable (Lumu) BCa (39%, p\u0026lt;0.01), and papillary structures are more common in luminal BCa. This structural feature may provide a relatively large space for the diffusion movement of water molecules, which leads to more significant diffusion behavior of water molecules in this subtype. Therefore, higher ADC values were observed in HER2 overexpressing BCa [29].\u003c/p\u003e\n\u003cp\u003eIn this study, radiomics features were extracted from different tumor subregions to explore the influence of radiomics features between different regions on the evaluation of HER2 expression status of BCa. Radiomics features derived from different subregions provide us with a wealth of information, showing the image information of BCa itself, which can reflect BCa heterogeneity [30,31]. An SVM model was constructed based on 17 radiomics features, including 4 from DWI \u0026nbsp;habitat 1 and habitat 2 subregions, 4 from ADC habitat 1 subregion, and 5 from habitat 2 subregion. The above explanation shows that DWI and ADC sequences, as well as their subregions, contribute to the construction of the model. The model includes 7 first-order features: Minimum, Energy, Kurtosis, Skewness, Maximum, \u0026nbsp; Uniformity, and Variance. By capturing the distribution characteristics of gray values within BCa, the heterogeneity of BCa can be accurately reflected, providing strong support for predicting the HER2 expression status of BCa [32,33].\u003c/p\u003e\n\u003cp\u003eIt was also observed in the study that some lesions were not successfully divided into two subregions, which may be due to the insufficient sample size to fully reflect the real image information, so it is necessary to continue to enrich the research samples. In addition, in ADC images, the red subregions are often located at the edge of the lesion. Whether the edge of the lesion contains clinically significant imaging information is worth further study.\u003c/p\u003e\n\u003cp\u003eOur study has several limitations. First, the sample size we analyzed is relatively small; Secondly, manually sketching the region of interest may lead to subjectivity of the data and deviation of the results; Finally, our study is a single-center retrospective study, which may lead to inevitable bias. Multi-center studies with larger sample sizes are needed to verify the stability and reproducibility of these results.\u0026nbsp;\u003c/p\u003e"},{"header":"5. Conclusion","content":"\u003cp\u003eMRI habitat imaging can effectively predict the expression status of HER2 in BCa, which provides a new method for the preoperative non-invasive diagnosis of bladder cancer and has important clinical value for the treatment decision of BCa patients.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cdiv class=\"DefinitionList\"\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eADC\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003eApparent Diffusion Coefficient\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eAUC\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003eArea Under the Curve\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eBCa\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003eBladder Cancer\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eDWI\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003eDiffusion Weighted Imaging\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eHER2\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003eHuman Epidermal Growth Factor Receptor 2\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eICC\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003eIntraclass Correlation Coefficient\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eICC\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003eInterclass Correlation Coefficient\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eIHC\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003eImmunohistochemistry\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eMIBC\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003eMuscle Invasive Bladder Cancer\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eMRI\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003eMagnetic resonance imaging\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eNMIBC\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003eNon-muscle Invasive Bladder Cancer\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eNPV\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003eNegative Predictive Value\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003ePPV\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003ePositive Predictive Value\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eROC\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003eReceiver Operating Characteristic\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eSVM\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003eSupport Vector Machine\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003eVOI\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003eVolume of Interest\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv class=\"DefinitionListEntry\"\u003e\u003cdiv class=\"Term\"\u003e95% CI\u003c/div\u003e\u003cdiv class=\"Description\"\u003e\u003cp\u003e95% Confidence Interval\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003cstrong\u003e\u003cu\u003e\u003cbr\u003e\u0026nbsp;\u003c/u\u003e\u003c/strong\u003eThis retrospective study (No. KYYJ-2023-064) was approved by the institutional review board of the First Hospital of Shanxi Medical University, and the need for written informed consent was waived. The study was performed in compliance with the 2024 version of the Declaration of Helsinki-Ethical Principles for Medical Research Involving Human Participants.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and material\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was supported by grants from the Four \u0026ldquo;Batches\u0026rdquo; Innovation Project of Invigorating Medical through Science and Technology of Shanxi Province (2023XM011); from the China International Medical Foundation of China (z-2014-07-2301); from the 2024 Annual Shanxi Provincial Basic Research Program (Free Exploration Category) Second Batch (202403021222449).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors\u0026apos; contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eZC was primarily responsible for collecting the patient data, organizing the literature, designing the experimental approach, conducting statistical analysis, and writing the manuscript. ZF was responsible for designing the experimental approach, performing post-processing on the original images to obtain habitat maps, conducting statistical analysis, and writing all the codes used in the paper. XZ and WL collected the laboratory data. ZF analyzed the data, with statistical advice from YL and BW. YW and GY discussed the results and interpreted the data. XW was primarily responsible for project administration and providing resources, also served as the corresponding author, handling all communications with the journal and addressing any post-publication inquiries. All authors read and approved the final manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgments\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors thank the colleagues from First Hospital of Shanxi Medical University \u0026nbsp;for their constructive suggestions in the conception and completion of this work.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eSaginala K, Barsouk A, Aluru JS, Rawla P, Padala SA, Barsouk A. Epidemiol Bladder Cancer Med Sci (Basel). 2020;8(1):15.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSiegel RL, Giaquinto AN, Jemal A, Cancer statistics. CA Cancer J Clin 2024 Jan-Feb. 2024;74(1):12\u0026ndash;49.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSanguedolce F, Zanelli M, Palicelli A, et al. HER2 Expression in Bladder Cancer: A Focused View on Its Diagnostic, Prognostic, and Predictive Role. Int J Mol Sci. 2023;24(4):3720.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eXu Y, Wang Y, Gong J, et al. Phase I study of the recombinant humanized anti-HER2 monoclonal antibody\u0026ndash;MMAE conjugate RC48-ADC in patients with HER2-positive advanced solid tumors. Gastric Cancer. 2021;24(4):913\u0026ndash;25.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFernandez AI, Liu M, Bellizzi A, et al. Examination of Low ERBB2 Protein Expression in Breast Cancer Tissue. JAMA Oncol. 2022;8(4):1\u0026ndash;4.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCommittee of Tumor Pathology. Chinese anti-cancer association, Committee of Urothelial Carcinoma, Chinese society of Clinical Oncology. Chinese clinicopathological expert consensus on the detection of human epidermal growth factor receptor 2 in urothelial carcinoma. Chin J Oncol. 2021;43(10):1001\u0026ndash;6.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWolff AC, Hammond MEH, Allison KH, et al. Human Epidermal Growth Factor Receptor 2 Testing in Breast Cancer: American Society of Clinical Oncology/College of American Pathologists Clinical Practice Guideline Focused Update. J Clin Oncol. 2018;36(20):2105\u0026ndash;22.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGlass B, Vandenberghe ME, Chavali ST, et al. Deployment of a Machine Learning Algorithm in a Real-World Cohort for Quality Control Monitoring of Human Epidermal Growth Factor-2-Stained Clinical Specimens in Breast Cancer. Arch Pathol Lab Med. 2025;149(8):751\u0026ndash;60.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKoh DM, Collins DJ. Diffusion-weighted MRI in the body: applications and challenges in oncology. AJR Am J Roentgenol. 2007;188(6):1622\u0026ndash;35.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCarr ME, Keenan KE, Beavan M, et al. Quantifying multi-institutional ADC measurement variability of 1.5 T MR-Linacs: A phantom and in vivo study. Med Phys. 2025;52(6):4120\u0026ndash;33.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHe L, Li F, Qin Y, et al. Enhanced preoperative prediction of breast lesion pathology, prognostic biomarkers, and molecular subtypes using multiple models diffusion-weighted MR imaging. Sci Rep. 2025;15(1):4704.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGerlinger M, Rowan AJ, Horswell S, et al. Intratumor Heterogeneity and Branched Evolution Revealed by Multiregion Sequencing. N Engl J Med. 2012;366(10):883\u0026ndash;92.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYuan Y. Spatial Heterogeneity in the Tumor Microenvironment. Cold Spring Harb Perspect Med. 2016;6(8):a026583.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWu H, Tong H, Du X, et al. Vascular habitat analysis based on dynamic susceptibility contrast perfusion MRI predicts IDH mutation status and prognosis in high-grade gliomas. Eur Radiol. 2020;30(6):3254\u0026ndash;65.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYu R, Cai L, Gong Y, Sun X, Li K, Cao Q, Yang X, Lu Q. MRI-Based Machine Learning Radiomics for Preoperative Assessment of Human Epidermal Growth Factor Receptor 2 Status in Urothelial Bladder Carcinoma[J]. J Magn Reson Imaging. 2024;60(6):2694\u0026ndash;704.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKong L, Ling J, Cao W, Wen Z, Lin Y, Cai Q, Chen Y, Guo Y, Chen J, Wang H. Multiparametric MR characterization for human epithelial growth factor receptor 2 expression in bladder cancer: an exploratory study[J]. Abdom Radiol (New York). 2024;49(7):2349\u0026ndash;57.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eArteaga CL, Engelman JA. ERBB receptors: from oncogene discovery to basic science to mechanism-based cancer therapeutics. Cancer Cell. 2014;25(3):282\u0026ndash;303.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAltriche N, Gallant S, Augustine TN, Xulu KR. Navigating the Intricacies of Tumor Heterogeneity: An Insight into Potential Prognostic Breast Cancer Biomarkers. Biomark Insights. 2024;19:11772719241256798.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSinha S, Choudhury S, Mishra J, Kundu G, Bera MK, Mondol PP. A study on immunohistochemical expression of HER2/Neu and p63 and its association with grade and invasiveness in case of bladder carcinoma. Urologia. 2024;91(2):284\u0026ndash;8.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKr\u0026uuml;ger S, Weitsch G, B\u0026uuml;ttner H, et al. Overexpression of c-erbB-2 oncoprotein in muscle-invasive bladder carcinoma: relationship with gene amplification, clinicopathological parameters and prognostic outcome. Int J Oncol. 2002;21(5):981\u0026ndash;7.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eVlachou E, Johnson BA, Hoffman-Censits J. The Role of Antibody-Drug Conjugates in Urothelial Cancer: A Review of Recent Advances in the Treatment of Locally Advanced and Metastatic Urothelial Cancer. Clin Med Insights Oncol. 2024;18:11795549241290787.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSheng X, Yan X, Wang L, et al. Open-label, Multicenter, Phase II Study of RC48-ADC, a HER2-Targeting Antibody-Drug Conjugate, in Patients with Locally Advanced or Metastatic Urothelial Carcinoma. Clin Cancer Res. 2021;27(1):43\u0026ndash;51.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWang X, Xie Z, Wang X, et al. Preoperative prediction of IDH genotypes and prognosis in adult-type diffuse gliomas: intratumor heterogeneity habitat analysis using dynamic contrast-enhanced MRI and diffusion-weighted imaging. Cancer Imaging. 2025;25(1):11.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWang Y, Xie B, Wang K et al. Multi-parametric MRI Habitat Radiomics Based on Interpretable Machine Learning for Preoperative Assessment of Microsatellite Instability in Rectal Cancer. Acad Radiol 2025 Feb 26:S1076-6332(25)00111-4.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHorvat-Menih I, Khan AS, McLean MA, et al. K-Means Clustering of Hyperpolarised 13C-MRI Identifies Intratumoral Perfusion/Metabolism Mismatch in Renal Cell Carcinoma as the Best Predictor of the Highest Grade. Cancers (Basel). 2025;17(4):569.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHe J, Shi H, Zhou Z, et al. Correlation between apparent diffusion coefficients and HER2 status in gastric cancers: pilot study. BMC Cancer. 2015;15:749.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMartincich L, Deantoni V, Bertotto I, et al. Correlations between diffusion-weighted imaging and breast cancer biomarkers. Eur Radiol. 2012;22(7):1519\u0026ndash;28.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChoi BB, Kim SH, Kang BJ, et al. Diffusion-weighted imaging and FDG PET/CT: predicting the prognoses with apparent diffusion coefficient values and maximum standardized uptake values in patients with invasive ductal carcinoma. World J Surg Oncol. 2012;10:126.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKamoun A, de Reyni\u0026egrave;s A, Allory Y, et al. A Consensus Molecular Classification of Muscle-invasive Bladder Cancer. Eur Urol. 2020;77(4):420\u0026ndash;33.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eEl Saftawy E, Aboulhoda BE, Alghamdi MA, Abd Elkhalek MA, AlHariry NS. Heterogeneity of modulatory immune microenvironment in bladder cancer. Tissue Cell. 2025;93:102679.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eDernbach G, Eich ML, Dragomir MP, et al. Spatial expression of HER2, NECTIN4, and TROP-2 in Muscle-Invasive Bladder Cancer and metastases: Implications for pathological and clinical management. Mod Pathol. 2025;38(7):100753.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eXu C, Wang Z, Wang A, et al. Breast Cancer: Multi-b-Value Diffusion Weighted Habitat Imaging in Predicting Pathologic Complete Response to Neoadjuvant Chemotherapy. Acad Radiol. 2024;31(12):4733\u0026ndash;42.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSkwierawska D, Laun FB, Wenkel E, et al. Diffusion-Weighted Imaging for Skin Pathologies of the Breast-A Feasibility Study. Diagnostics (Basel). 2024;14(9):934.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"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":"bmc-medical-imaging","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"bmim","sideBox":"Learn more about [BMC Medical Imaging](http://bmcmedimaging.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/bmim/default.aspx","title":"BMC Medical Imaging","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"HER2, bladder cancer, magnetic resonance imaging, habitat image","lastPublishedDoi":"10.21203/rs.3.rs-7710608/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7710608/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground:\u003c/h2\u003e\u003cp\u003eHuman epidermal growth factor receptor 2 (HER2) antibody-coupled drugs have shown promising clinical benefits in patients with bladder cancer (BCa). HER2 expression status is generally detected clinically using postoperative pathological immunohistochemistry (IHC), but preoperative non-invasive detection of BCa HER2 expression status remains to be sought. The aim of this study was to investigate the value of diffusion-weighted imaging (DWI) and apparent diffusion coefficient (ADC) habitat imaging in predicting the expression of HER2 in BCa.\u003c/p\u003e\u003ch2\u003eMethods:\u003c/h2\u003e\u003cp\u003eThis retrospective study included 232 BCa patients (November 2022\u0026ndash;February 2024) with HER2 status confirmed by immunohistochemistry. The K-means clustering algorithm is used to re-segment the region of interest. Based on the spatial distribution of the habitat map, the histogram features of each subregion were extracted. Based on the Spearman correlation coefficient (\u0026gt;\u0026thinsp;0.7) feature screening results, a support vector machine (SVM) classification model was established to predict the expression of HER2 in BCa. The discrimination ability of the model was evaluated by plotting the receiver operating characteristic (ROC) curve and calculating the area under the curve (AUC), and the diagnostic performance of the model was comprehensively evaluated by combining the calibration curve and the decision curve.\u003c/p\u003e\u003ch2\u003eResults:\u003c/h2\u003e\u003cp\u003eRandomly divided patients into training cohort (N\u0026thinsp;=\u0026thinsp;148, median age 68.66 years; 121men), validation (N\u0026thinsp;=\u0026thinsp;47, median age 69.12 years; 39 men), and test cohort (N\u0026thinsp;=\u0026thinsp;37, median age 67.92 years; 32men) according to the ratio of 6:2:2. Based on the contour coefficient, K\u0026thinsp;=\u0026thinsp;2 is finally selected as the clustering parameter to cluster the DWI and ADC images into two subregions. A total of 80 features were extracted from the four sub-regions of the two sequences. After screening, an SVM prediction model was constructed from the remaining 17 features. In the SVM model, the AUC of the training set was 0.88 (95% CI: 0.82\u0026ndash;0.93), the validation set was 0.85 (95% CI: 0.72\u0026ndash;0.94), and the test set was 0.84 (95% CI: 0.88\u0026ndash;0.94).\u003c/p\u003e\u003ch2\u003eConclusion:\u003c/h2\u003e\u003cp\u003eMRI-based habitat analysis can help distinguish heterogeneous regions of BCa and effectively predict HER2 expression status of BCa.\u003c/p\u003e\u003ch2\u003eClinical trial number:\u003c/h2\u003e\u003cp\u003eNot applicable.\u003c/p\u003e","manuscriptTitle":"DWI and ADC Habitat Imaging in Predicting HER2 Expression Status in Bladder Cancer: A Retrospective Study","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-11-10 16:17:23","doi":"10.21203/rs.3.rs-7710608/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-03-06T12:47:45+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-11-13T20:24:48+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-11-10T11:10:41+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"100581249053700654680718046443362979485","date":"2025-10-31T11:36:50+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"227589892590100425603120813835807456256","date":"2025-10-29T12:40:44+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-10-29T10:00:24+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-10-28T08:26:25+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-10-06T10:07:25+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-10-06T09:11:31+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Medical Imaging","date":"2025-10-06T09:07:59+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-medical-imaging","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"bmim","sideBox":"Learn more about [BMC Medical Imaging](http://bmcmedimaging.biomedcentral.com/)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/bmim/default.aspx","title":"BMC Medical Imaging","twitterHandle":"BMC_series","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"em","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"c7f4f522-9401-4293-97dd-cc04aed13ef7","owner":[],"postedDate":"November 10th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-03-19T15:08:21+00:00","versionOfRecord":[],"versionCreatedAt":"2025-11-10 16:17:23","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7710608","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7710608","identity":"rs-7710608","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}