CT-based Deep Learning Radiomics Nomogram for Differentiating pyelocaliceal Upper Tract Urothelial Carcinoma and Xanthogranulomatous Pyelonephritis

CT-based Deep Learning Radiomics Nomogram for Differentiating pyelocaliceal Upper Tract Urothelial Carcinoma and Xanthogranulomatous Pyelonephritis | 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 CT-based Deep Learning Radiomics Nomogram for Differentiating pyelocaliceal Upper Tract Urothelial Carcinoma and Xanthogranulomatous Pyelonephritis Yi Feng, Yiheng Cheng, Jian Zeng, Junyu Lin, Hao Qin, Jiawen Zhao, and 3 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7537981/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Objective This study aims to establish a CT-based clinical deep learning radiomics (CDLR) model for preoperative prediction of pyelocaliceal Upper Tract Urothelial Carcinoma (UTUC) and xanthogranulomatous pyelonephritis (XGP), providing scientific guidance for personalized treatment. Methods A retrospective analysis was conducted on 161 post-operative pathology-confirmed cases of pyelocaliceal UTUC and XGP patients, divided into training cohort (n = 112) and validation cohort (n = 49). Radiomics (Rad) and deep learning (LR) features were extracted from three-phase CT images, combined with clinical features, and after feature selection, a logistic regression (LR) classifier was used to construct clinical, radiomics (Rad), deep learning (LR), deep learning radiomics (DLR), and clinical deep learning radiomics (CDLR) models. The top-performing model was chosen utilizing receiver operating characteristic (ROC) curve analysis. Decision curve analysis (DCA) was used to evaluate the model's practicality. Results The evaluation of clinical, Rad, LR, DLR, and CDLR models revealed that the CDLR model exhibited superior diagnostic performance. The area under the receiver operating characteristic curve (AUC) of the CDLR model in the training and validation cohorts were 0.984 and 0.970, respectively, outperforming other models (clinical model, Rad model, DL model, DLR model). DCA results showed that the CDLR model provided a higher net benefit in preoperative prediction. Conclusion The CDLR model, combining clinical, Rad, and LR features, could serve as a non-invasive tool for differentiating pyelocaliceal UTUC and XGP, offering valuable guidance for clinical treatment. Upper Tract Urothelial Carcinoma Xanthogranulomatous Pyelonephritis Radiomics Deep Learning Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Introduction Upper Tract Urothelial Carcinoma (UTUC), which includes renal pelvis cancer and ureteral cancer, accounts for approximately 5%-10% of all urothelial cancers [ 1 ] . Xanthogranulomatous pyelonephritis (XGP) is a rare chronic granulomatous inflammation, often referred to as a pseudotumor because the kidney enlarges and presents characteristics similar to a tumor, potentially leading to local infiltration and destruction [ 2 ] . Radical nephroureterectomy combined with bladder cuff excision is considered the standard treatment, as the recurrence rate of the remaining distal ureter is high, ranging from 16% to 58% [ 3 – 5 ] . XGP, on the other hand, is a destructive bacterial infection typically requiring antibiotic treatment and nephrectomy [ 6 – 8 ] . This differs from the treatment for pyelocaliceal UTUC. UTUC patients generally have a poor prognosis, with the 5-year survival rate for advanced UTUC patients significantly lower than for other types of cancer, requiring close clinical follow-up [ 9 , 10 ] . However, XGP typically manifests as a unilateral disease and has a favorable prognosis with timely treatment [ 11 ] . Therefore, accurately distinguishing between pyelocaliceal UTUC and XGP preoperatively is crucial. Currently, the diagnosis of pyelocaliceal UTUC and XGP is dependent on the subjective experience of radiologists. Previous studies have shown that XGP is referred to as the "great imitator" because its clinical and radiologic presentations are very similar to other pathological entities, and about half of the cases are not correctly identified in preoperative imaging exams, making preoperative diagnosis difficult [ 12 , 13 ] . Furthermore, the radiologic features of pyelocaliceal UTUC and XGP share certain similarities [ 2 , 14 ] , which further complicates the accurate preoperative differentiation. Therefore, it is vital to find a simple, non-invasive, and accurate method for preoperatively differentiating pyelocaliceal UTUC and XGP. Radiomics is an image analysis method that can extract extensive lesion information from traditional medical images, deeply exploring and analyzing complex details within the images to reveal hidden patterns [ 15 – 17 ] . It mainly includes basic steps such as image acquisition and preprocessing, Region of Interest (ROI) delineation, feature extraction and selection, as well as model construction and prediction [ 18 ] . Early studies have explored the application of radiomics in UTUC post-operative prognosis prediction, differentiation from other renal tumors, pathological grading, etc [ 19 – 21 ] . In recent years, the practical value of deep learning (DL) algorithms based on convolutional neural networks (CNN) has been widely recognized and applied in medical imaging analysis [ 22 – 24 ] . CNN is a typical artificial neural network in deep learning, capable of advanced image and video recognition and segmentation [ 25 ] . However, to date, no study has reported the application of radiomics or deep learning in distinguishing between pyelocaliceal UTUC and XGP. Therefore, this study aims to develop and validate a clinical deep learning radiomics (CDLR) nomogram for accurately differentiating pyelocaliceal UTUC and XGP. Materials and methods Patients This study was approved by the ethics committee (Approval Number:2025-E0330 ). A retrospective analysis was conducted on patients diagnosed with pyelocaliceal UTUC and XGP, confirmed by postoperative pathology, between October 2018 and February 2024 at the First Affiliated Hospital of Guangxi Medical University. Inclusion criteria: (1) Patients with postoperative pathology reports confirming pyelocaliceal UTUC and XGP; (2) Patients who had complete preoperative non-enhanced and enhanced CT imaging data. Exclusion criteria: (1) Patients with poor quality preoperative CT images; (2) Patients with incomplete or missing clinical data; (3) Patients who received chemotherapy, radiotherapy, or immunotherapy before surgery; (4) Patients with a history of other malignant tumors. The patient recruitment pathway is shown in Fig. 1 . Clinical Data Collection The following clinical characteristics were extracted from the patient medical records: age, gender, fever, lumbago, basic illness (hypertension, diabetes, coronary heart disease), red blood cell count, and urinary protein. All CT images were analyzed by radiologists with 5 to 30 years of experience in abdominal radiology. The features of the CT images included the tumor's maximum diameter and location. All clinical data were retrospectively retrieved through the hospital's HIS system. CT Image Acquisition This study used two CT devices, namely the Siemens Somatom Definition Flash CT and the GE 256 Revolution CT. The scanning parameters were set as follows: tube voltage 120 kV, tube current 280 mA, slice thickness 1–8 mm. First, a non-enhanced scan was performed, followed by the injection of iodine contrast agent. Enhanced CT scans were then acquired at 30–40 seconds and 70–120 seconds to capture the corticomedullary phase (CMP) and nephrographic phase (NP), respectively. Image Segmentation and Feature Extraction We obtained the original CT images through the picture archiving and communication system (PACS) of the First Affiliated Hospital of Guangxi Medical University. To reduce the impact of different scanning protocols or devices, all CT images underwent standardized preprocessing. During preprocessing, the CT images were resampled to 1mm×1mm×1mm, with window width and window level set to 400 and 40 Hounsfield units (HU), respectively. Subsequently, the region of interest (ROI) of the lesions in the corticomedullary phase (CMP) was delineated using 3D Slicer software (version: 5.4.0). The ROI of the lesions in the nephrographic phase (NP) and precontrast phase (PCP) were delineated based on the contours from the CMP. The inter-observer and intra-observer reproducibility of the features were typically assessed using interclass and intraclass correlation coefficients (ICCs), with values greater than 0.75 indicating good consistency of the extracted features. Thirty patients were randomly selected, and image ROI delineation was performed by a radiologist with 5 years of experience (Reader 1) and another radiologist with 8 years of experience (Reader 2). After two weeks, Reader 1 repeated the image ROI delineation for the same 30 patients, and Reader 1 independently completed the ROI delineation for the remaining patients. Both radiologists were blinded to the clinical information and pathological results of the patients. We used the Pyradiomics package in Python (version: 3.7.12) to extract radiomics (Rad) features, obtaining a total of 5502 Rad features from the three-phase image ROI of each patient. This included (1) forty-two 2D shape-based features, (2) 1080 first-order features, (3) texture features, including features from gray level co-occurrence matrix (GLCM) (n = 1320), gray-level dependence matrix (GLDM) (n = 840), gray-level run length matrix (GLRLM) (n = 960), gray-level size zone matrix (GLSZM) (n = 960), and neighboring gray tone difference matrix (NGTDM) (n = 300). To obtain deep learning (DL) features, we selected the image with the largest lesion area to represent each patient and fine-tuned a pre-trained ResNet 50 network from the ImageNet database ( https://image-net.org/ ). We set the initial learning rate to 0.01, trained for 50 epochs, with a batch size of 32, using the stochastic gradient descent optimizer to adjust the model parameters. By fine-tuning the pre-trained ResNet 50 model, we obtained 6144 DL features from the three-phase image ROI of each patient. Feature Selection and Fusion For clinical features, univariate and multivariate logistic regression analysis were conducted in the training cohort, and variables with significant correlations were selected to build the clinical model. For Rad features, first, features with ICCs value greater than 0.75 were retained. Next, the remaining Rad features were normalized and selected using the Minimum Redundancy Maximum Relevance (mRMR) algorithm. Then, further selection of features with non-zero coefficients was performed using Least Absolute Shrinkage and Selection Operator (LASSO) regression with 5-fold cross-validation. For DL features, the DL features were first normalized and then selected using the mRMR algorithm. Features with non-zero coefficients were subsequently selected through LASSO regression with 5-fold cross-validation. Based on Rad features with ICCs value greater than 0.75 and the 6144 DL features, we constructed a deep learning radiomics (DLR) feature set. Feature selection followed the same path as for the DL features. Model Development and Evaluation In this study, we used a logistic regression (LR) classifier to build clinical model, Rad model, DL model, and DLR model by combining the best selected clinical features, Rad features, DL features, and DLR features, respectively. The CDLR model was developed by integrating clinical features and DLR features. Figure 2 illustrates the entire process of model construction. We compared all models based on area under the curve (AUC), accuracy, sensitivity, specificity, positive predictive value, and negative predictive value. The differences in AUC between models were evaluated using the Delong test, with a significance level set at p < 0.05. In addition, we further evaluated and compared the model performance using decision curve analysis (DCA), calibration curves, and confusion matrices. Statistical Analysis Statistical analysis was performed using SPSS software (version 23; IBM) and Python software (version 3.7.12). For continuous variables, t-test or Mann-Whitney U test was used based on their distribution characteristics. For categorical variables, chi-square test or Fisher’s exact test was applied. The comparison of AUC was performed using the Delong test. A p-value of < 0.05 was considered statistically significant. Results Clinical Information of Patients and Clinical Model Construction A total of 161 patients were included in this study, including 47 XGP patients (9 males, 38 females, mean age 53.43 ± 11.88 years) and 114 pyelocaliceal UTUC patients (77 males, 37 females, mean age 66.67 ± 10.72 years). The patients were randomly divided into a training cohort (n = 112) and a validation cohort (n = 49) at a ratio of 7:3. The clinical baseline data of the training and validation cohorts are shown in Table 1 . Univariate logistic regression analysis in the training cohort revealed significant differences in age, gender, fever, and maximum diameter between pyelocaliceal UTUC and XGP (p < 0.05). After conducting multivariate logistic regression analysis on these significantly different clinical factors, the results indicated that age, gender, fever, and maximum diameter were independent discriminative factors in distinguishing pyelocaliceal UTUC from XGP (p < 0.05) (see Table 2 ). The clinical model was constructed based on these four independent clinical factors. As shown in Table 3 , the AUC for the training cohort was 0.912 (95% CI: 0.8542–0.9701), and the AUC for the validation cohort was 0.799 (95% CI: 0.6250–0.9733). Table 1 The clinical features of UTUC and XGP patients in the training cohort and validation cohort. Variable Training cohort(n = 112) Validation cohort(n = 49) XGP UTUC P -value XGP UTUC P -value Age (years) 52.68 ± 12.32 66.86 ± 10.96 < 0.001 55.38 ± 10.87 66.25 ± 10.30 0.004 Red blood cell count 9.42 ± 5.74 7.73 ± 4.01 0.046 11.00 ± 5.70 7.32 ± 2.63 0.003 Max-diameter(cm) 2.52 ± 2.42 3.79 ± 2.83 0.006 4.22 ± 5.60 3.71 ± 2.03 0.303 Gender < 0.001 0.002 Male 7 52 2 25 Female 27 26 11 11 Location 0.386 0247 Right 19 35 7 11 Left 15 43 6 25 Basic illness 1.0 0.719 Without 23 53 9 21 With 11 25 4 15 Fever < 0.001 0.342 Without 23 76 11 35 With 11 2 2 1 Lumbago 0.176 0.207 Without 12 40 5 23 With 22 38 8 13 Urinary protein 0.449 0.853 feminine 9 28 4 14 masculine 25 50 9 22 Table 2 Univariate and multivariate logistic regression analysis of the clinical features of UTUC and XGP patients in the training cohort. Variable Univeriate analysis Multivariate analysis OR (95% CI ) P -value OR (95% CI ) P -value Age (years) 1.018 (1.013, 1.022) < 0.001 1.014 (1.010,1.018) < 0.001 Red blood cell count 0.983 (0.969, 0.999) 0.076 Max-diameter(cm) 1.036 (1.010, 1.063) 0.024 1.031(1.011,1.050) 0.010 Gender 0.677 (0.593, 0.772) < 0.001 0.773(0.694,0.861) < 0.001 Location 1.098(0.949, 1.269) 0.288 Basic illness 0.997 (0.853, 1.165) 0.975 Fever 0.541 (0.441, 0.664) < 0.001 0.606(0.514,0.715) < 0.001 Lumbago 0.873 (0.756, 1.008) 0.121 Urinary protein 0.914 (0.783, 1.066) 0.334 OR, odds ratio;CI, confidence interval. Table 3 Diagnostic metrics of different models in the training cohort and validation cohort. Model Cohort AUC(95%CI) Accuracy Sensitivity Specificity PPV NPV Clinic Training 0.912(0.8542–0.9701) 0.902 0.974 0.735 0.894 0.926 Validation 0.779(0.6250–0.9733) 0.878 0.972 0.615 0.875 0.889 Rad Training 0.893(0.8215–0.9636) 0.848 0.859 0.824 0.918 0.718 Validation 0.779(0.6348–0.9635) 0.857 0.944 0.615 0.872 0.800 DL Training 0.936(0.8747–0.9963) 0.911 0.897 0.941 0.972 0.800 Validation 0.870(0.7412–0.9982) 0.898 0.944 0.769 0.919 0.833 DLR Training 0.940(0.8854–0.9947) 0.929 0.962 0.853 0.937 0.906 Validation 0.891(0.7739–1.0000) 0.878 0.917 0.769 0.917 0.769 CDLR Training 0.984(0.9642–1.0000) 0.938 0.923 0.971 0.986 0.846 Validation 0.970(0.9292 -1.0000) 0.857 0.806 1.000 1.000 0.650 AUC, area under the receiver operating characteristic curve; CI, confidence interval; PPV, positive predictive value; NPV, negative predictive value; Rad, radiomics model; DL, deep learning model; DLR, deep learning radiomics model; CDLR, clinical deep learning radiomics nomogram. Construction of Rad, DL, and DLR Models After evaluating the inter-observer and intra-observer reproducibility of Rad features (ICCs > 0.75), 2080 Rad features were considered reliable. After feature selection, the remaining Rad features were used to construct the Rad model. After feature selection of 6144 DL features, the remaining DL features were used to construct the DL model. Based on 2080 Rad features with ICCs > 0.75 and 6144 DL features, after feature selection, we obtained 2 Rad features and 7 DL features to construct the DLR model (see Fig. 3 ). As shown in Table 3 , the AUC of the Rad model in the training cohort was 0.893 (95% CI: 0.8215–0.9636), and the AUC in the validation cohort was 0.799 (95% CI: 0.6348–0.9635). The AUC of the DL model in the training cohort was 0.936 (95% CI: 0.8747–0.9963), and the AUC in the validation cohort was 0.870 (95% CI: 0.7412–0.9982). The DLR model was the best model between the Rad model and the DL model. The AUC of the DLR model in the training cohort was 0.940 (95% CI: 0.8854–0.9947), and the AUC in the validation cohort was 0.891 (95% CI: 0.7739–1.0000). Establishment of the CDLR Model and Evaluation of Performance Across Different Models Given the superior performance of the DLR model compared to other models, we combined clinical features with DLR features to develop the CDLR model. Table 3 shows all models used to predict pyelocaliceal UTUC and XGP. Figure 4 displays the AUC values for the training and validation cohorts. To compare the performance of the clinical model, Rad model, DL model, DLR model, and CDLR model, we conducted the Delong test (see Table 4 ). The calibration curve (see Fig. 5 ) shows good consistency between the predicted values of each model and the actual observed values. Decision curve analysis (DCA) indicates that the CDLR model provided higher net benefits for patients (see Fig. 6 ). Based on clinical features and DLR features, we constructed a nomogram to predict pyelocaliceal UTUC and XGP (see Fig. 7 ). The confusion matrix results indicate that the CDLR model performed best in the validation cohort (see Fig. 8 ). Figure 9 displays the activation maps of a convolutional neural network used to identify pyelocaliceal UTUC and XGP. Table 4 Delong test. Cohort CDRL vs Clinic CDRL vs Rad CDRL vs DL CDRL vs DLR Training 0.0093 0.0105 0.1036 0.0778 Validation 0.0715 0.0459 0.0473 0.0611 Discussion In this study, we developed and independently validated five diagnostic models: a clinical model based on patient age, gender, fever, and maximum diameter; a radiomics model (Rad model); a deep learning model (DL model); a DLR model combining radiomics features and deep learning features; and a CDLR model combining clinical features with DLR features. In both the training and validation cohorts, the CDLR model demonstrated the best predictive performance. The AUC of the CDLR model in the training cohort was 0.984 (95% CI: 0.9642–1.0000), and in the validation cohort, it was 0.970 (95% CI: 0.9292–1.0000). Accurately distinguishing between pyelocaliceal UTUC and XGP is of significant clinical importance, as the surgical approaches and prognoses for the two are completely different. For pyelocaliceal UTUC patients, radical nephroureterectomy combined with bladder cuff resection is required [ 3 – 5 ] , while XGP patients are typically treated with antibiotics and, if necessary, nephrectomy [ 6 – 8 ] , avoiding unnecessary surgical resection. In this study, through univariate and multivariate logistic regression analysis, we identified patient age, gender, fever, and maximum diameter as independent predictors for distinguishing between pyelocaliceal UTUC and XGP. Our findings indicate that pyelocaliceal UTUC patients have a significantly older average age of onset compared to XGP patients, and XGP patients are more commonly female, a group that warrants special attention. Previous studies have shown that the average diagnostic age of XGP patients ranges from 45 to 55.2 years, with the majority being female, and nearly all patients have symptoms, with fever being one of the most common symptoms [ 26 – 28 ] . Our results align with previous studies. The clinical model had an AUC of 0.912 and 0.799 in the training and validation cohorts, respectively, demonstrating good predictive ability. Radiomics and deep learning, as emerging interdisciplinary fields, combine medical imaging and computer science to extract vast amounts of quantitative information from medical images, showing great potential in aiding clinical diagnosis and treatment [ 18 , 29 ] . Radiomics and deep learning have been widely applied in kidney diseases, especially in distinguishing between benign and malignant lesions [ 30 – 32 ] . Additionally, radiomics has been widely used in UTUC postoperative prognosis prediction, distinguishing it from other renal tumors, and predicting pathological grading [ 19 – 21 ] , achieving good predictive performance in all of these areas. In this study, we applied CNN-based deep learning algorithms and successfully distinguished pyelocaliceal UTUC from XGP. Among the 9 features selected by the CDLR model, 2 are Rad features and 7 are DL features, with DL features having the highest weight, indicating that deep learning technology can effectively extract quantitative information to distinguish pyelocaliceal UTUC from XGP. Our results show that the diagnostic ability of the DLR model outperforms both the Rad and DL models, and the DL model based on DL features is superior to the Rad model based on Rad features, further proving the application value of deep learning in this field. To our knowledge, this is the first study that combines clinical, radiomics, and deep learning features for preoperative prediction of pyelocaliceal UTUC and XGP. It is worth noting that there are some limitations in this study. First, the sample size is relatively small. Second, since this is a retrospective study, selection bias is inevitable. Finally, all subjects in the study came from a single center, lacking validation from multi-center data. Therefore, we plan to expand the sample size in the future and seek external validation from multiple centers to enhance the reliability and generalizability of the study. Conclusion In conclusion, the CDLR model based on clinical, radiomics, and deep learning features has significant clinical value in accurately distinguishing pyelocaliceal UTUC from XGP. Compared to individual clinical, radiomics, deep learning models, and the DLR model combining radiomics and deep learning features, the CDLR model significantly improves classification performance. Therefore, the CDLR model can help clinicians more accurately distinguish pyelocaliceal UTUC from XGP and promote more effective treatment guidance. Abbreviations Rad radiomics DL deep learning DLR deep learning radiomics CDLR clinical deep learning radiomics ROC receiver operating characteristic DCA decision curve analysis ROI region of interest CNN convolutional neural network CMP corticomedullary phase NP nephrographic phase PCP precontrast phase HU hounsfield units PACS picture archiving and communication system ICCs interclass and intraclass correlation coefficients GLCM gray level co-occurrence matrix GLDM gray-level dependence matrix GLRLM gray-level run length matrix GLSZM gray-level size zone matrix NGTDM neighboring gray tone difference matrix LASSO least absolute shrinkage and selection operator mRMR minimum redundancy maximum relevance LR logistic regression AUC area under the receiver operating characteristic curve PPV positive predictive value NPV negative predictive value CI confidence interval OR odds ratio Declarations Conflict of Interest The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Ethical Statement The study protocol was approved by the Institutional Review Board. Because the study was retrospective, the institutional review board waived the patient's informed consent. Funding This work was supported by grants from the National Natural Science Foundation of China (82460156,82400898). Author Contribution C.Y.L and F.X.F: Conceptualization, funding acquisition and methodology.Y.F and Y.H.C: formal analysis, software, and writing – original draft. J.Z, J.Y L, H.Q, J.W.Z and B.W: Data curation. Data Availability The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding authors. References Rouprêt M, Babjuk M, Compérat E, et al. European guidelines on upper tract urothelial carcinomas: 2013 update [J]. European urology, 2013, 63: 1059-71.10.1016/j.eururo.2013.03.032 Kang Y, Lee T W, Bae E, et al. Xanthogranulomatous pyelonephritis in a patient with polycystic kidney disease without underlying risk factors: a case report [J]. Frontiers in medicine, 2024, 11: 1419965.10.3389/fmed.2024.1419965 Colin P, Ouzzane A, Pignot G, et al. Comparison of oncological outcomes after segmental ureterectomy or radical nephroureterectomy in urothelial carcinomas of the upper urinary tract: results from a large French multicentre study [J]. BJU international, 2012, 110: 1134-41.10.1111/j.1464-410X.2012.10960.x Lughezzani G, Jeldres C, Isbarn H, et al. Nephroureterectomy and segmental ureterectomy in the treatment of invasive upper tract urothelial carcinoma: a population-based study of 2299 patients [J]. European journal of cancer (Oxford, England : 1990), 2009, 45: 3291-7.10.1016/j.ejca.2009.06.016 Oosterlinck W, Solsona E, Van Der Meijden A P, et al. EAU guidelines on diagnosis and treatment of upper urinary tract transitional cell carcinoma [J]. European urology, 2004, 46: 147-54.10.1016/j.eururo.2004.04.011 Burbano M A, Nati-Castillo H A, Castaño-Giraldo N, et al. Fatal nephrobronchial fistula arising from xanthogranulomatous pyelonephritis: a case report [J]. Frontiers in medicine, 2024, 11: 1374043.10.3389/fmed.2024.1374043 Xie L, Tapiero S, Flores A R, et al. Long-Term Antibiotic Treatment Prior to Laparoscopic Nephrectomy for Xanthogranulomatous Pyelonephritis Improves Postoperative Outcomes: Results from a Multicenter Study [J]. The Journal of urology, 2021, 205: 820-5.10.1097/ju.0000000000001429 Bercowsky E, Shalhav A L, Portis A, et al. Is the laparoscopic approach justified in patients with xanthogranulomatous pyelonephritis? [J]. Urology, 1999, 54: 437-42; discussion 42-3.10.1016/s0090-4295(99)00261-7 Margulis V, Shariat S F, Matin S F, et al. Outcomes of radical nephroureterectomy: a series from the Upper Tract Urothelial Carcinoma Collaboration [J]. Cancer, 2009, 115: 1224-33.10.1002/cncr.24135 Abouassaly R, Alibhai S M, Shah N, et al. Troubling outcomes from population-level analysis of surgery for upper tract urothelial carcinoma [J]. Urology, 2010, 76: 895-901.10.1016/j.urology.2010.04.020 Malek R S, Elder J S. Xanthogranulomatous pyelonephritis: a critical analysis of 26 cases and of the literature [J]. The Journal of urology, 1978, 119: 589-93.10.1016/s0022-5347(17)57559-x Chlif M, Chakroun M, Ben Rhouma S, et al. Xanthogranulomatous pyelonephritis presenting as a pseudotumour [J]. Canadian Urological Association journal = Journal de l'Association des urologues du Canada, 2016, 10: E36-40.10.5489/cuaj.3225 Chandrashekhar S, Kumar S, Jambunathan S, et al. Xanthogranulomatous Pyelonephritis Mimicking a Complex Renal Cyst: A Report of a Rare Case [J]. Cureus, 2024, 16: e69233.10.7759/cureus.69233 Xie X, Wang N, Wang Y, et al. Non-invasive papillary urothelial carcinoma, low-grade of the renal pelvis mimicking a xanthogranulomatous pyelonephritis in a male patient: A case report and review of literature [J]. International journal of immunopathology and pharmacology, 2020, 34: 2058738420925720.10.1177/2058738420925720 Lafata K J, Wang Y, Konkel B, et al. Radiomics: a primer on high-throughput image phenotyping [J]. Abdominal radiology (New York), 2022, 47: 2986-3002.10.1007/s00261-021-03254-x Gillies R J, Kinahan P E, Hricak H. Radiomics: Images Are More than Pictures, They Are Data [J]. Radiology, 2016, 278: 563-77.10.1148/radiol.2015151169 Lambin P, Rios-Velazquez E, Leijenaar R, et al. Radiomics: extracting more information from medical images using advanced feature analysis [J]. European journal of cancer (Oxford, England : 1990), 2012, 48: 441-6.10.1016/j.ejca.2011.11.036 Zhang X, Zhang Y, Zhang G, et al. Deep Learning With Radiomics for Disease Diagnosis and Treatment: Challenges and Potential [J]. Frontiers in oncology, 2022, 12: 773840.10.3389/fonc.2022.773840 Al Mopti A, Alqahtani A, Alshehri A H D, et al. Perirenal Fat CT Radiomics-Based Survival Model for Upper Tract Urothelial Carcinoma: Integrating Texture Features with Clinical Predictors [J]. Cancers, 2024, 16.10.3390/cancers16223772 Zhai X, Sun P, Yu X, et al. CT-based radiomics signature for differentiating pyelocaliceal upper urinary tract urothelial carcinoma from infiltrative renal cell carcinoma [J]. Frontiers in oncology, 2023, 13: 1244585.10.3389/fonc.2023.1244585 Zheng Y, Shi H, Fu S, et al. Development and validation of a radiomics-based nomogram for predicting pathological grade of upper urinary tract urothelial carcinoma [J]. BMC cancer, 2024, 24: 1546.10.1186/s12885-024-13325-z Lotter W, Diab A R, Haslam B, et al. Robust breast cancer detection in mammography and digital breast tomosynthesis using an annotation-efficient deep learning approach [J]. Nature medicine, 2021, 27: 244-9.10.1038/s41591-020-01174-9 Song Z, Liu T, Shi L, et al. The deep learning model combining CT image and clinicopathological information for predicting ALK fusion status and response to ALK-TKI therapy in non-small cell lung cancer patients [J]. European journal of nuclear medicine and molecular imaging, 2021, 48: 361-71.10.1007/s00259-020-04986-6 Esteva A, Kuprel B, Novoa R A, et al. Dermatologist-level classification of skin cancer with deep neural networks [J]. Nature, 2017, 542: 115-8.10.1038/nature21056 Zhang K, Zuo W, Chen Y, et al. Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoising [J]. IEEE transactions on image processing : a publication of the IEEE Signal Processing Society, 2017, 26: 3142-55.10.1109/tip.2017.2662206 Li L, Parwani A V. Xanthogranulomatous pyelonephritis [J]. Archives of pathology & laboratory medicine, 2011, 135: 671-4.10.5858/2009-0769-rsr.1 Schneeberger C, Holleman F, Geerlings S E. Febrile urinary tract infections: pyelonephritis and urosepsis [J]. Current opinion in infectious diseases, 2016, 29: 80-5.10.1097/qco.0000000000000227 Goldman S M, Hartman D S, Fishman E K, et al. CT of xanthogranulomatous pyelonephritis: radiologic-pathologic correlation [J]. AJR American journal of roentgenology, 1984, 142: 963-9.10.2214/ajr.142.5.963 Liu Y X, Liu Q H, Hu Q H, et al. Ultrasound-Based Deep Learning Radiomics Nomogram for Tumor and Axillary Lymph Node Status Prediction After Neoadjuvant Chemotherapy [J]. Academic radiology, 2025, 32: 12-23.10.1016/j.acra.2024.07.036 Han J H, Kim B W, Kim T M, et al. Fully automated segmentation and classification of renal tumors on CT scans via machine learning [J]. BMC cancer, 2025, 25: 173.10.1186/s12885-025-13582-6 Wu Y, Cao F, Lei H, et al. Interpretable multiphasic CT-based radiomic analysis for preoperatively differentiating benign and malignant solid renal tumors: a multicenter study [J]. Abdominal radiology (New York), 2024, 49: 3096-106.10.1007/s00261-024-04351-3 Yu S, Hao R, Cui J, et al. Integration of radiomic and deep features to reliably differentiate benign renal lesions from renal cell carcinoma [J]. European journal of radiology, 2025, 184: 111989.10.1016/j.ejrad.2025.111989 Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7537981","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":516925428,"identity":"83cfaed3-6e8a-45c9-a0a8-f88a56b18bd9","order_by":0,"name":"Yi Feng","email":"","orcid":"","institution":"the First Affiliated Hospital of Guangxi Medical University","correspondingAuthor":false,"prefix":"","firstName":"Yi","middleName":"","lastName":"Feng","suffix":""},{"id":516925429,"identity":"98760b12-d3f0-4d20-a02a-06a62a7ce7f4","order_by":1,"name":"Yiheng Cheng","email":"","orcid":"","institution":"the First Affiliated Hospital of Guangxi Medical University","correspondingAuthor":false,"prefix":"","firstName":"Yiheng","middleName":"","lastName":"Cheng","suffix":""},{"id":516925430,"identity":"a1d1ce27-c119-4fda-b5c0-45bfa88d475d","order_by":2,"name":"Jian Zeng","email":"","orcid":"","institution":"the First Affiliated Hospital of Guangxi Medical University","correspondingAuthor":false,"prefix":"","firstName":"Jian","middleName":"","lastName":"Zeng","suffix":""},{"id":516925431,"identity":"0e0a25f8-98c5-4beb-bfa9-fa509dd88190","order_by":3,"name":"Junyu Lin","email":"","orcid":"","institution":"Guangxi Medical University","correspondingAuthor":false,"prefix":"","firstName":"Junyu","middleName":"","lastName":"Lin","suffix":""},{"id":516925432,"identity":"6362d724-f71a-4392-8494-823e1366327c","order_by":4,"name":"Hao Qin","email":"","orcid":"","institution":"Guangxi Medical University","correspondingAuthor":false,"prefix":"","firstName":"Hao","middleName":"","lastName":"Qin","suffix":""},{"id":516925433,"identity":"f98e93c8-822d-414c-804d-dfd615c9c267","order_by":5,"name":"Jiawen Zhao","email":"","orcid":"","institution":"the First Affiliated Hospital of Guangxi Medical University","correspondingAuthor":false,"prefix":"","firstName":"Jiawen","middleName":"","lastName":"Zhao","suffix":""},{"id":516925439,"identity":"37c65376-4b2e-4352-a0de-43d9e518ca5a","order_by":6,"name":"Bo Wu","email":"","orcid":"","institution":"the First Affiliated Hospital of Guangxi Medical University","correspondingAuthor":false,"prefix":"","firstName":"Bo","middleName":"","lastName":"Wu","suffix":""},{"id":516925441,"identity":"54f25633-af3e-4426-a48b-eb193f35fc12","order_by":7,"name":"Fuxiang Fang","email":"","orcid":"","institution":"the First Affiliated Hospital of Guangxi Medical University","correspondingAuthor":false,"prefix":"","firstName":"Fuxiang","middleName":"","lastName":"Fang","suffix":""},{"id":516925443,"identity":"a74eca31-b2fa-495e-83d0-9ee1a7bf101a","order_by":8,"name":"Chengyang Li","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAzklEQVRIiWNgGAWjYBACPigtA8SMDxIqaghrYYPSPEDMbPDgzDHStLBJPmxhJkIL+xkDZt49Njz8/GvMKhIb2Bj427sT8GvhyTFg5nmWxiM5443ZjcQdMgwSZ85uwK9Fggeo5cBhHoMbZ4BazrAxGEjkkqClILGNmRQt53vMGIjTwpNWwDjnAMgvbMUSCWeO8RD0Cz/74Q0Mbw7YyPHzH9748UdFjRx/ey9+LQwMHOY/wLREhgGI4iGgHATYH0DtO/6ACNWjYBSMglEwEgEA9t1AFUHWD/YAAAAASUVORK5CYII=","orcid":"","institution":"the First Affiliated Hospital of Guangxi Medical University","correspondingAuthor":true,"prefix":"","firstName":"Chengyang","middleName":"","lastName":"Li","suffix":""}],"badges":[],"createdAt":"2025-09-04 16:23:21","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7537981/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7537981/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":91937476,"identity":"93a723d8-39f2-4e6e-8dff-24419a9c8135","added_by":"auto","created_at":"2025-09-23 02:57:03","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":521937,"visible":true,"origin":"","legend":"","description":"","filename":"Manuscript.docx","url":"https://assets-eu.researchsquare.com/files/rs-7537981/v1/d07b3df5a96c6d48ddb5f5cd.docx"},{"id":91934683,"identity":"9cd256f5-a586-4733-aa80-8835e79cd4ae","added_by":"auto","created_at":"2025-09-23 02:41:02","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":9631,"visible":true,"origin":"","legend":"","description":"","filename":"101a66e5d16e43539bd30ccde7e6f96e.json","url":"https://assets-eu.researchsquare.com/files/rs-7537981/v1/e726b1cfe31b978abb140ec9.json"},{"id":91936407,"identity":"cbe589ef-3a7e-4dff-a67a-164a925c0458","added_by":"auto","created_at":"2025-09-23 02:49:02","extension":"xml","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":93693,"visible":true,"origin":"","legend":"","description":"","filename":"101a66e5d16e43539bd30ccde7e6f96e1enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7537981/v1/daedaa0cb7467c8b83cb1eef.xml"},{"id":91937475,"identity":"f1b08077-0093-40a8-bdb6-a8784ea4e739","added_by":"auto","created_at":"2025-09-23 02:57:03","extension":"jpeg","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":221549,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7537981/v1/7a5f2933ad887fec00db3753.jpeg"},{"id":91934709,"identity":"d0ea7c5d-8671-4e1f-88fa-3153fd432dd5","added_by":"auto","created_at":"2025-09-23 02:41:03","extension":"jpeg","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":80242,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7537981/v1/1e0dcfa8dd92d1062da6de1c.jpeg"},{"id":91936410,"identity":"4d733369-87d7-42e3-8e31-5a36f7892622","added_by":"auto","created_at":"2025-09-23 02:49:03","extension":"jpeg","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":313461,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7537981/v1/a79e901ab6b648febbfcfdd7.jpeg"},{"id":91936416,"identity":"3b55b865-5417-46d6-845d-77b6cb46c944","added_by":"auto","created_at":"2025-09-23 02:49:03","extension":"jpeg","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":193294,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7537981/v1/75c71b8c5947200ad7057662.jpeg"},{"id":91934692,"identity":"fe833ead-fb3e-4a4c-a543-763633ad0e06","added_by":"auto","created_at":"2025-09-23 02:41:03","extension":"jpeg","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":197681,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage5.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7537981/v1/66c7000983f4f1a69158217f.jpeg"},{"id":91936412,"identity":"4c182198-fc6a-4bab-a3fc-82bc1832289f","added_by":"auto","created_at":"2025-09-23 02:49:03","extension":"jpeg","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":229401,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage6.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7537981/v1/4e74257fa3747c8763120794.jpeg"},{"id":91937477,"identity":"651616f0-e753-4f46-b916-d2e60d7b9494","added_by":"auto","created_at":"2025-09-23 02:57:03","extension":"jpeg","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":81576,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage7.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7537981/v1/50dadda0e075f18f552c12c2.jpeg"},{"id":91936419,"identity":"2fc942ff-c751-4878-ac07-b40fd53663a7","added_by":"auto","created_at":"2025-09-23 02:49:03","extension":"jpeg","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":164142,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage8.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7537981/v1/e6baa3ceefdcd7725fc928a5.jpeg"},{"id":91936414,"identity":"21ac63ce-4a23-4ebe-bf32-3191fa96cc9e","added_by":"auto","created_at":"2025-09-23 02:49:03","extension":"jpeg","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":432802,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage9.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7537981/v1/7f292af82ac07b32107b3682.jpeg"},{"id":91934698,"identity":"bc230490-bd31-43a7-8b27-f9e938380dc9","added_by":"auto","created_at":"2025-09-23 02:41:03","extension":"png","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":52104,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7537981/v1/167ff4fd9f508e90d1268e65.png"},{"id":91936418,"identity":"b58db69b-a032-4c4c-b2fc-eceadd3bc7b0","added_by":"auto","created_at":"2025-09-23 02:49:03","extension":"png","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":49162,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7537981/v1/363a52a0b1f405c7ec3f838e.png"},{"id":91934711,"identity":"f2da7722-6ded-4a9e-9819-183928c91185","added_by":"auto","created_at":"2025-09-23 02:41:03","extension":"png","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":50811,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7537981/v1/937698eca4b71ae8413ca5bc.png"},{"id":91936421,"identity":"a6af8053-66da-4eda-9d0a-b81982fb2616","added_by":"auto","created_at":"2025-09-23 02:49:03","extension":"png","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":35220,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7537981/v1/0a86827f17fff30563de36ad.png"},{"id":91934710,"identity":"460cdd87-8c96-450d-8e2d-f3d981421332","added_by":"auto","created_at":"2025-09-23 02:41:03","extension":"png","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":49746,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7537981/v1/a12ec139a10cd2e0194b0fc3.png"},{"id":91934701,"identity":"e0623319-b6cf-4731-b7d7-7e45e46b37a0","added_by":"auto","created_at":"2025-09-23 02:41:03","extension":"png","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":53029,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7537981/v1/12be88ed964f7207c08eeb42.png"},{"id":91938060,"identity":"ef36e5ce-34bf-435f-a6ce-37e994adfb63","added_by":"auto","created_at":"2025-09-23 03:05:03","extension":"png","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":16209,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-7537981/v1/86f4e70179ec85d24de36cfd.png"},{"id":91934696,"identity":"a0539d6e-569d-4d74-a309-c9b2826a22cf","added_by":"auto","created_at":"2025-09-23 02:41:03","extension":"png","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":32035,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-7537981/v1/447f57f802b9188be68675e0.png"},{"id":91934716,"identity":"54992853-6fd5-42be-9311-1f1cc35c04a6","added_by":"auto","created_at":"2025-09-23 02:41:03","extension":"png","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":111145,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage9.png","url":"https://assets-eu.researchsquare.com/files/rs-7537981/v1/1995e993db89e4d29177be50.png"},{"id":91934713,"identity":"07725574-de04-4218-8c87-ae6cfe1d6271","added_by":"auto","created_at":"2025-09-23 02:41:03","extension":"xml","order_by":21,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":91655,"visible":true,"origin":"","legend":"","description":"","filename":"101a66e5d16e43539bd30ccde7e6f96e1structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7537981/v1/8b129f1a38102b99a29ff5b5.xml"},{"id":91934715,"identity":"6adb50db-3abf-4ff2-bfe4-8c736d0e2fec","added_by":"auto","created_at":"2025-09-23 02:41:03","extension":"html","order_by":22,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":100017,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7537981/v1/a777439ce7406877d7ba7ab2.html"},{"id":91934686,"identity":"61b0e67e-1f93-4461-9d21-0b436812e113","added_by":"auto","created_at":"2025-09-23 02:41:02","extension":"jpeg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":221549,"visible":true,"origin":"","legend":"\u003cp\u003ePatient recruitment pathway.\u003c/p\u003e","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7537981/v1/c5b7fa3e1d1eeb9eccfa1178.jpeg"},{"id":91936406,"identity":"97e8e5f5-aa9e-4c3c-943a-4bdada8e39bd","added_by":"auto","created_at":"2025-09-23 02:49:02","extension":"jpeg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":80242,"visible":true,"origin":"","legend":"\u003cp\u003eThe process of model construction.\u003c/p\u003e","description":"","filename":"floatimage2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7537981/v1/2f77c857fdbff4b7c3b57c7c.jpeg"},{"id":91934688,"identity":"becd8c9c-a8e2-46b6-8828-8b20b7bbf309","added_by":"auto","created_at":"2025-09-23 02:41:03","extension":"jpeg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":313461,"visible":true,"origin":"","legend":"\u003cp\u003eFeature selection of DLR featuresusing the LASSO algorithm. (a) The coefficient profile plot (b) The cross-validation plot (c) Feature weight histogram of the features.\u003c/p\u003e","description":"","filename":"floatimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7537981/v1/dafe8bfc8c8ced88698345b2.jpeg"},{"id":91934685,"identity":"5381ac40-f553-4c7f-b842-2a2cf44a236d","added_by":"auto","created_at":"2025-09-23 02:41:02","extension":"jpeg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":193294,"visible":true,"origin":"","legend":"\u003cp\u003eReceiver operating characteristic (ROC) curves of the clinical model, Rad model, DL model, DLR model, and CDLR model. (a) In the training cohort; (b) In the validation cohort.\u003c/p\u003e","description":"","filename":"floatimage4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7537981/v1/3436758901fdafed111e4f6a.jpeg"},{"id":91934689,"identity":"bc8b265e-e8d8-4886-9fb0-a052c23a51f2","added_by":"auto","created_at":"2025-09-23 02:41:03","extension":"jpeg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":197681,"visible":true,"origin":"","legend":"\u003cp\u003eCalibration curves of the clinical model, Rad model, DL model, DLR model, and CDRL model. (a) In the training cohort; (b) In the validation cohort.\u003c/p\u003e","description":"","filename":"floatimage5.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7537981/v1/de358457256ae6f14bbaaaf9.jpeg"},{"id":91934695,"identity":"90c00b3f-8555-4407-ac10-412d828a0d47","added_by":"auto","created_at":"2025-09-23 02:41:03","extension":"jpeg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":229401,"visible":true,"origin":"","legend":"\u003cp\u003eDecision curve analysis (DCA) of the clinical model, Rad model, DL model, DLR model, and CDLR model. (a) In the training cohort; (b) In the validation cohort.\u003c/p\u003e","description":"","filename":"floatimage6.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7537981/v1/d3271442a3fefca88f7886db.jpeg"},{"id":91936408,"identity":"25bf83cb-720e-4807-8a8e-6c908631b3fd","added_by":"auto","created_at":"2025-09-23 02:49:03","extension":"jpeg","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":81576,"visible":true,"origin":"","legend":"\u003cp\u003eNomogram of the CDLR model.\u003c/p\u003e","description":"","filename":"floatimage7.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7537981/v1/3001cdfee37bda9d513ec4d6.jpeg"},{"id":91937479,"identity":"4e9f4385-61da-4682-8b63-44fdb2164050","added_by":"auto","created_at":"2025-09-23 02:57:03","extension":"jpeg","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":164142,"visible":true,"origin":"","legend":"\u003cp\u003eConfusion matrices of the clinical, DLR, and CDLR models in the validation cohort. (a) Confusion matrix of the clinical model; (b) Confusion matrix of the DLR model; (c) Confusion matrix of the CDLR model.\u003c/p\u003e","description":"","filename":"floatimage8.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7537981/v1/13b805ccb43e422c8f8d8587.jpeg"},{"id":91934707,"identity":"b2a9fab2-d434-4456-84d4-919e8755031e","added_by":"auto","created_at":"2025-09-23 02:41:03","extension":"jpeg","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":432802,"visible":true,"origin":"","legend":"\u003cp\u003eConvolutional neural network activation maps used to identify UTUC and XGP.These maps were constructed using data from the precontrast phase, corticomedullary phase, and nephrographic phase.The red regions on these maps highlight areas that correlate with the nature of the mass.\u003c/p\u003e","description":"","filename":"floatimage9.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7537981/v1/765835c446055199884fd367.jpeg"},{"id":93140270,"identity":"f4424043-fda9-4ba7-9c92-0265c320dc34","added_by":"auto","created_at":"2025-10-09 12:54:08","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2837945,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7537981/v1/66e298e7-7658-43d6-b166-10bb3910e337.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"CT-based Deep Learning Radiomics Nomogram for Differentiating pyelocaliceal Upper Tract Urothelial Carcinoma and Xanthogranulomatous Pyelonephritis","fulltext":[{"header":"Introduction","content":"\u003cp\u003eUpper Tract Urothelial Carcinoma (UTUC), which includes renal pelvis cancer and ureteral cancer, accounts for approximately 5%-10% of all urothelial cancers \u003csup\u003e[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]\u003c/sup\u003e. Xanthogranulomatous pyelonephritis (XGP) is a rare chronic granulomatous inflammation, often referred to as a pseudotumor because the kidney enlarges and presents characteristics similar to a tumor, potentially leading to local infiltration and destruction \u003csup\u003e[\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]\u003c/sup\u003e. Radical nephroureterectomy combined with bladder cuff excision is considered the standard treatment, as the recurrence rate of the remaining distal ureter is high, ranging from 16% to 58% \u003csup\u003e[\u003cspan additionalcitationids=\"CR4\" citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]\u003c/sup\u003e. XGP, on the other hand, is a destructive bacterial infection typically requiring antibiotic treatment and nephrectomy \u003csup\u003e[\u003cspan additionalcitationids=\"CR7\" citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]\u003c/sup\u003e. This differs from the treatment for pyelocaliceal UTUC. UTUC patients generally have a poor prognosis, with the 5-year survival rate for advanced UTUC patients significantly lower than for other types of cancer, requiring close clinical follow-up \u003csup\u003e[\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]\u003c/sup\u003e. However, XGP typically manifests as a unilateral disease and has a favorable prognosis with timely treatment \u003csup\u003e[\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]\u003c/sup\u003e. Therefore, accurately distinguishing between pyelocaliceal UTUC and XGP preoperatively is crucial. Currently, the diagnosis of pyelocaliceal UTUC and XGP is dependent on the subjective experience of radiologists. Previous studies have shown that XGP is referred to as the \"great imitator\" because its clinical and radiologic presentations are very similar to other pathological entities, and about half of the cases are not correctly identified in preoperative imaging exams, making preoperative diagnosis difficult \u003csup\u003e[\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]\u003c/sup\u003e. Furthermore, the radiologic features of pyelocaliceal UTUC and XGP share certain similarities \u003csup\u003e[\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]\u003c/sup\u003e, which further complicates the accurate preoperative differentiation. Therefore, it is vital to find a simple, non-invasive, and accurate method for preoperatively differentiating pyelocaliceal UTUC and XGP.\u003c/p\u003e\u003cp\u003eRadiomics is an image analysis method that can extract extensive lesion information from traditional medical images, deeply exploring and analyzing complex details within the images to reveal hidden patterns \u003csup\u003e[\u003cspan additionalcitationids=\"CR16\" citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]\u003c/sup\u003e. It mainly includes basic steps such as image acquisition and preprocessing, Region of Interest (ROI) delineation, feature extraction and selection, as well as model construction and prediction \u003csup\u003e[\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]\u003c/sup\u003e. Early studies have explored the application of radiomics in UTUC post-operative prognosis prediction, differentiation from other renal tumors, pathological grading, etc \u003csup\u003e[\u003cspan additionalcitationids=\"CR20\" citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]\u003c/sup\u003e. In recent years, the practical value of deep learning (DL) algorithms based on convolutional neural networks (CNN) has been widely recognized and applied in medical imaging analysis \u003csup\u003e[\u003cspan additionalcitationids=\"CR23\" citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e]\u003c/sup\u003e. CNN is a typical artificial neural network in deep learning, capable of advanced image and video recognition and segmentation \u003csup\u003e[\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]\u003c/sup\u003e. However, to date, no study has reported the application of radiomics or deep learning in distinguishing between pyelocaliceal UTUC and XGP.\u003c/p\u003e\u003cp\u003eTherefore, this study aims to develop and validate a clinical deep learning radiomics (CDLR) nomogram for accurately differentiating pyelocaliceal UTUC and XGP.\u003c/p\u003e"},{"header":"Materials and methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003ePatients\u003c/h2\u003e\u003cp\u003e This study was approved by the ethics committee (Approval Number:2025-E0330 ). A retrospective analysis was conducted on patients diagnosed with pyelocaliceal UTUC and XGP, confirmed by postoperative pathology, between October 2018 and February 2024 at the First Affiliated Hospital of Guangxi Medical University. Inclusion criteria: (1) Patients with postoperative pathology reports confirming pyelocaliceal UTUC and XGP; (2) Patients who had complete preoperative non-enhanced and enhanced CT imaging data. Exclusion criteria: (1) Patients with poor quality preoperative CT images; (2) Patients with incomplete or missing clinical data; (3) Patients who received chemotherapy, radiotherapy, or immunotherapy before surgery; (4) Patients with a history of other malignant tumors. The patient recruitment pathway is shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eClinical Data Collection\u003c/h3\u003e\n\u003cp\u003eThe following clinical characteristics were extracted from the patient medical records: age, gender, fever, lumbago, basic illness (hypertension, diabetes, coronary heart disease), red blood cell count, and urinary protein. All CT images were analyzed by radiologists with 5 to 30 years of experience in abdominal radiology. The features of the CT images included the tumor's maximum diameter and location. All clinical data were retrospectively retrieved through the hospital's HIS system.\u003c/p\u003e\n\u003ch3\u003eCT Image Acquisition\u003c/h3\u003e\n\u003cp\u003eThis study used two CT devices, namely the Siemens Somatom Definition Flash CT and the GE 256 Revolution CT. The scanning parameters were set as follows: tube voltage 120 kV, tube current 280 mA, slice thickness 1\u0026ndash;8 mm. First, a non-enhanced scan was performed, followed by the injection of iodine contrast agent. Enhanced CT scans were then acquired at 30\u0026ndash;40 seconds and 70\u0026ndash;120 seconds to capture the corticomedullary phase (CMP) and nephrographic phase (NP), respectively.\u003c/p\u003e\n\u003ch3\u003eImage Segmentation and Feature Extraction\u003c/h3\u003e\n\u003cp\u003eWe obtained the original CT images through the picture archiving and communication system (PACS) of the First Affiliated Hospital of Guangxi Medical University. To reduce the impact of different scanning protocols or devices, all CT images underwent standardized preprocessing. During preprocessing, the CT images were resampled to 1mm\u0026times;1mm\u0026times;1mm, with window width and window level set to 400 and 40 Hounsfield units (HU), respectively. Subsequently, the region of interest (ROI) of the lesions in the corticomedullary phase (CMP) was delineated using 3D Slicer software (version: 5.4.0). The ROI of the lesions in the nephrographic phase (NP) and precontrast phase (PCP) were delineated based on the contours from the CMP. The inter-observer and intra-observer reproducibility of the features were typically assessed using interclass and intraclass correlation coefficients (ICCs), with values greater than 0.75 indicating good consistency of the extracted features. Thirty patients were randomly selected, and image ROI delineation was performed by a radiologist with 5 years of experience (Reader 1) and another radiologist with 8 years of experience (Reader 2). After two weeks, Reader 1 repeated the image ROI delineation for the same 30 patients, and Reader 1 independently completed the ROI delineation for the remaining patients. Both radiologists were blinded to the clinical information and pathological results of the patients.\u003c/p\u003e\u003cp\u003eWe used the Pyradiomics package in Python (version: 3.7.12) to extract radiomics (Rad) features, obtaining a total of 5502 Rad features from the three-phase image ROI of each patient. This included (1) forty-two 2D shape-based features, (2) 1080 first-order features, (3) texture features, including features from gray level co-occurrence matrix (GLCM) (n\u0026thinsp;=\u0026thinsp;1320), gray-level dependence matrix (GLDM) (n\u0026thinsp;=\u0026thinsp;840), gray-level run length matrix (GLRLM) (n\u0026thinsp;=\u0026thinsp;960), gray-level size zone matrix (GLSZM) (n\u0026thinsp;=\u0026thinsp;960), and neighboring gray tone difference matrix (NGTDM) (n\u0026thinsp;=\u0026thinsp;300). To obtain deep learning (DL) features, we selected the image with the largest lesion area to represent each patient and fine-tuned a pre-trained ResNet 50 network from the ImageNet database (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://image-net.org/\u003c/span\u003e\u003cspan address=\"https://image-net.org/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). We set the initial learning rate to 0.01, trained for 50 epochs, with a batch size of 32, using the stochastic gradient descent optimizer to adjust the model parameters. By fine-tuning the pre-trained ResNet 50 model, we obtained 6144 DL features from the three-phase image ROI of each patient.\u003c/p\u003e\n\u003ch3\u003eFeature Selection and Fusion\u003c/h3\u003e\n\u003cp\u003eFor clinical features, univariate and multivariate logistic regression analysis were conducted in the training cohort, and variables with significant correlations were selected to build the clinical model.\u003c/p\u003e\u003cp\u003eFor Rad features, first, features with ICCs value greater than 0.75 were retained. Next, the remaining Rad features were normalized and selected using the Minimum Redundancy Maximum Relevance (mRMR) algorithm. Then, further selection of features with non-zero coefficients was performed using Least Absolute Shrinkage and Selection Operator (LASSO) regression with 5-fold cross-validation.\u003c/p\u003e\u003cp\u003eFor DL features, the DL features were first normalized and then selected using the mRMR algorithm. Features with non-zero coefficients were subsequently selected through LASSO regression with 5-fold cross-validation.\u003c/p\u003e\u003cp\u003eBased on Rad features with ICCs value greater than 0.75 and the 6144 DL features, we constructed a deep learning radiomics (DLR) feature set. Feature selection followed the same path as for the DL features.\u003c/p\u003e\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\u003ch2\u003eModel Development and Evaluation\u003c/h2\u003e\u003cp\u003eIn this study, we used a logistic regression (LR) classifier to build clinical model, Rad model, DL model, and DLR model by combining the best selected clinical features, Rad features, DL features, and DLR features, respectively. The CDLR model was developed by integrating clinical features and DLR features. Figure\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e illustrates the entire process of model construction. We compared all models based on area under the curve (AUC), accuracy, sensitivity, specificity, positive predictive value, and negative predictive value. The differences in AUC between models were evaluated using the Delong test, with a significance level set at p\u0026thinsp;\u0026lt;\u0026thinsp;0.05. In addition, we further evaluated and compared the model performance using decision curve analysis (DCA), calibration curves, and confusion matrices.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e\u003ch2\u003eStatistical Analysis\u003c/h2\u003e\u003cp\u003eStatistical analysis was performed using SPSS software (version 23; IBM) and Python software (version 3.7.12). For continuous variables, t-test or Mann-Whitney U test was used based on their distribution characteristics. For categorical variables, chi-square test or Fisher\u0026rsquo;s exact test was applied. The comparison of AUC was performed using the Delong test. A p-value of \u0026lt;\u0026thinsp;0.05 was considered statistically significant.\u003c/p\u003e\u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\u003ch2\u003eClinical Information of Patients and Clinical Model Construction\u003c/h2\u003e\u003cp\u003eA total of 161 patients were included in this study, including 47 XGP patients (9 males, 38 females, mean age 53.43\u0026thinsp;\u0026plusmn;\u0026thinsp;11.88 years) and 114 pyelocaliceal UTUC patients (77 males, 37 females, mean age 66.67\u0026thinsp;\u0026plusmn;\u0026thinsp;10.72 years). The patients were randomly divided into a training cohort (n\u0026thinsp;=\u0026thinsp;112) and a validation cohort (n\u0026thinsp;=\u0026thinsp;49) at a ratio of 7:3. The clinical baseline data of the training and validation cohorts are shown in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. Univariate logistic regression analysis in the training cohort revealed significant differences in age, gender, fever, and maximum diameter between pyelocaliceal UTUC and XGP (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05). After conducting multivariate logistic regression analysis on these significantly different clinical factors, the results indicated that age, gender, fever, and maximum diameter were independent discriminative factors in distinguishing pyelocaliceal UTUC from XGP (p\u0026thinsp;\u0026lt;\u0026thinsp;0.05) (see Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). The clinical model was constructed based on these four independent clinical factors. As shown in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, the AUC for the training cohort was 0.912 (95% CI: 0.8542\u0026ndash;0.9701), and the AUC for the validation cohort was 0.799 (95% CI: 0.6250\u0026ndash;0.9733).\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eThe clinical features of UTUC and XGP patients in the training cohort and validation cohort.\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"7\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eVariable\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u003cp\u003eTraining cohort(n\u0026thinsp;=\u0026thinsp;112)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colspan=\"4\" nameend=\"c7\" namest=\"c4\"\u003e\u003cp\u003eValidation cohort(n\u0026thinsp;=\u0026thinsp;49)\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eXGP\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eUTUC\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003e\u003cem\u003eP\u003c/em\u003e-value\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eXGP\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003eUTUC\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c7\"\u003e\u003cp\u003e\u003cem\u003eP\u003c/em\u003e-value\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eAge (years)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e52.68\u0026thinsp;\u0026plusmn;\u0026thinsp;12.32\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e66.86\u0026thinsp;\u0026plusmn;\u0026thinsp;10.96\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e55.38\u0026thinsp;\u0026plusmn;\u0026thinsp;10.87\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e66.25\u0026thinsp;\u0026plusmn;\u0026thinsp;10.30\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.004\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eRed blood cell count\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e9.42\u0026thinsp;\u0026plusmn;\u0026thinsp;5.74\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e7.73\u0026thinsp;\u0026plusmn;\u0026thinsp;4.01\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.046\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e11.00\u0026thinsp;\u0026plusmn;\u0026thinsp;5.70\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e7.32\u0026thinsp;\u0026plusmn;\u0026thinsp;2.63\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.003\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMax-diameter(cm)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e2.52\u0026thinsp;\u0026plusmn;\u0026thinsp;2.42\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e3.79\u0026thinsp;\u0026plusmn;\u0026thinsp;2.83\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.006\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e4.22\u0026thinsp;\u0026plusmn;\u0026thinsp;5.60\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e3.71\u0026thinsp;\u0026plusmn;\u0026thinsp;2.03\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.303\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eGender\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.002\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMale\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e7\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e52\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e25\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eFemale\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e27\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e26\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e11\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e11\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLocation\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.386\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0247\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eRight\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e19\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e35\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e7\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e11\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLeft\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e15\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e43\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e6\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e25\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eBasic illness\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e1.0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.719\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWithout\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e23\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e53\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e9\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e21\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWith\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e11\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e25\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e15\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eFever\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.342\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWithout\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e23\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e76\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e11\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e35\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWith\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e11\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLumbago\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.176\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.207\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWithout\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e12\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e40\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e23\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWith\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e22\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e38\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e8\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e13\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eUrinary protein\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.449\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.853\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003efeminine\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e9\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e28\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e14\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003emasculine\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e25\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e50\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e9\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e22\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eUnivariate and multivariate logistic regression analysis of the clinical features of UTUC and XGP patients in the training cohort.\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"5\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eVariable\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u003cp\u003eUniveriate analysis\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colspan=\"2\" nameend=\"c5\" namest=\"c4\"\u003e\u003cp\u003eMultivariate analysis\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eOR (95%\u003cem\u003eCI\u003c/em\u003e)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003cem\u003eP\u003c/em\u003e-value\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eOR (95%\u003cem\u003eCI\u003c/em\u003e)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u003cem\u003eP\u003c/em\u003e-value\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eAge (years)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e1.018 (1.013, 1.022)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e1.014 (1.010,1.018)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eRed blood cell count\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.983 (0.969, 0.999)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.076\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMax-diameter(cm)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e1.036 (1.010, 1.063)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.024\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e1.031(1.011,1.050)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.010\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eGender\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.677 (0.593, 0.772)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.773(0.694,0.861)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLocation\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e1.098(0.949, 1.269)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.288\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eBasic illness\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.997 (0.853, 1.165)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.975\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eFever\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.541 (0.441, 0.664)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.606(0.514,0.715)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eLumbago\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.873 (0.756, 1.008)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.121\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eUrinary protein\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.914 (0.783, 1.066)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.334\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003ctfoot\u003e\u003ctr\u003e\u003ctd colspan=\"5\"\u003eOR, odds ratio;CI, confidence interval.\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eDiagnostic metrics of different models in the training cohort and validation cohort.\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"8\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eModel\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eCohort\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eAUC(95%CI)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eAccuracy\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eSensitivity\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003eSpecificity\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c7\"\u003e\u003cp\u003ePPV\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c8\"\u003e\u003cp\u003eNPV\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eClinic\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eTraining\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.912(0.8542\u0026ndash;0.9701)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.902\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.974\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e0.735\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e0.894\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e\u003cp\u003e0.926\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eValidation\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.779(0.6250\u0026ndash;0.9733)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.878\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.972\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e0.615\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e0.875\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e\u003cp\u003e0.889\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eRad\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eTraining\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.893(0.8215\u0026ndash;0.9636)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.848\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.859\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e0.824\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e0.918\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e\u003cp\u003e0.718\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eValidation\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.779(0.6348\u0026ndash;0.9635)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.857\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.944\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e0.615\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e0.872\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e\u003cp\u003e0.800\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eDL\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eTraining\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.936(0.8747\u0026ndash;0.9963)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.911\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.897\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e0.941\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e0.972\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e\u003cp\u003e0.800\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eValidation\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.870(0.7412\u0026ndash;0.9982)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.898\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.944\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e0.769\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e0.919\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e\u003cp\u003e0.833\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eDLR\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eTraining\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.940(0.8854\u0026ndash;0.9947)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.929\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.962\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e0.853\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e0.937\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e\u003cp\u003e0.906\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eValidation\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.891(0.7739\u0026ndash;1.0000)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.878\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.917\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e0.769\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e0.917\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e\u003cp\u003e0.769\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCDLR\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eTraining\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.984(0.9642\u0026ndash;1.0000)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.938\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.923\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e0.971\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e0.986\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e\u003cp\u003e0.846\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eValidation\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.970(0.9292 -1.0000)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.857\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.806\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e1.000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e1.000\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e\u003cp\u003e0.650\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003ctfoot\u003e\u003ctr\u003e\u003ctd colspan=\"8\"\u003eAUC, area under the receiver operating characteristic curve; CI, confidence interval; PPV, positive predictive value; NPV, negative predictive value; Rad, radiomics model; DL, deep learning model; DLR, deep learning radiomics model; CDLR, clinical deep learning radiomics nomogram.\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e\u003ch2\u003eConstruction of Rad, DL, and DLR Models\u003c/h2\u003e\u003cp\u003eAfter evaluating the inter-observer and intra-observer reproducibility of Rad features (ICCs\u0026thinsp;\u0026gt;\u0026thinsp;0.75), 2080 Rad features were considered reliable. After feature selection, the remaining Rad features were used to construct the Rad model. After feature selection of 6144 DL features, the remaining DL features were used to construct the DL model. Based on 2080 Rad features with ICCs\u0026thinsp;\u0026gt;\u0026thinsp;0.75 and 6144 DL features, after feature selection, we obtained 2 Rad features and 7 DL features to construct the DLR model (see Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). As shown in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, the AUC of the Rad model in the training cohort was 0.893 (95% CI: 0.8215\u0026ndash;0.9636), and the AUC in the validation cohort was 0.799 (95% CI: 0.6348\u0026ndash;0.9635). The AUC of the DL model in the training cohort was 0.936 (95% CI: 0.8747\u0026ndash;0.9963), and the AUC in the validation cohort was 0.870 (95% CI: 0.7412\u0026ndash;0.9982). The DLR model was the best model between the Rad model and the DL model. The AUC of the DLR model in the training cohort was 0.940 (95% CI: 0.8854\u0026ndash;0.9947), and the AUC in the validation cohort was 0.891 (95% CI: 0.7739\u0026ndash;1.0000).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec13\" class=\"Section2\"\u003e\u003ch2\u003eEstablishment of the CDLR Model and Evaluation of Performance Across Different Models\u003c/h2\u003e\u003cp\u003eGiven the superior performance of the DLR model compared to other models, we combined clinical features with DLR features to develop the CDLR model. Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e shows all models used to predict pyelocaliceal UTUC and XGP. Figure\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e displays the AUC values for the training and validation cohorts. To compare the performance of the clinical model, Rad model, DL model, DLR model, and CDLR model, we conducted the Delong test (see Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). The calibration curve (see Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e) shows good consistency between the predicted values of each model and the actual observed values. Decision curve analysis (DCA) indicates that the CDLR model provided higher net benefits for patients (see Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e). Based on clinical features and DLR features, we constructed a nomogram to predict pyelocaliceal UTUC and XGP (see Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e). The confusion matrix results indicate that the CDLR model performed best in the validation cohort (see Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003e). Figure\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003e displays the activation maps of a convolutional neural network used to identify pyelocaliceal UTUC and XGP.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eDelong test.\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"5\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCohort\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eCDRL vs Clinic\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eCDRL vs Rad\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eCDRL vs DL\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eCDRL vs DLR\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTraining\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.0093\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.0105\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.1036\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.0778\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eValidation\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.0715\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.0459\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.0473\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.0611\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eIn this study, we developed and independently validated five diagnostic models: a clinical model based on patient age, gender, fever, and maximum diameter; a radiomics model (Rad model); a deep learning model (DL model); a DLR model combining radiomics features and deep learning features; and a CDLR model combining clinical features with DLR features. In both the training and validation cohorts, the CDLR model demonstrated the best predictive performance. The AUC of the CDLR model in the training cohort was 0.984 (95% CI: 0.9642\u0026ndash;1.0000), and in the validation cohort, it was 0.970 (95% CI: 0.9292\u0026ndash;1.0000).\u003c/p\u003e\u003cp\u003eAccurately distinguishing between pyelocaliceal UTUC and XGP is of significant clinical importance, as the surgical approaches and prognoses for the two are completely different. For pyelocaliceal UTUC patients, radical nephroureterectomy combined with bladder cuff resection is required \u003csup\u003e[\u003cspan additionalcitationids=\"CR4\" citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]\u003c/sup\u003e, while XGP patients are typically treated with antibiotics and, if necessary, nephrectomy \u003csup\u003e[\u003cspan additionalcitationids=\"CR7\" citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]\u003c/sup\u003e, avoiding unnecessary surgical resection.\u003c/p\u003e\u003cp\u003eIn this study, through univariate and multivariate logistic regression analysis, we identified patient age, gender, fever, and maximum diameter as independent predictors for distinguishing between pyelocaliceal UTUC and XGP. Our findings indicate that pyelocaliceal UTUC patients have a significantly older average age of onset compared to XGP patients, and XGP patients are more commonly female, a group that warrants special attention. Previous studies have shown that the average diagnostic age of XGP patients ranges from 45 to 55.2 years, with the majority being female, and nearly all patients have symptoms, with fever being one of the most common symptoms \u003csup\u003e[\u003cspan additionalcitationids=\"CR27\" citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e]\u003c/sup\u003e. Our results align with previous studies. The clinical model had an AUC of 0.912 and 0.799 in the training and validation cohorts, respectively, demonstrating good predictive ability.\u003c/p\u003e\u003cp\u003eRadiomics and deep learning, as emerging interdisciplinary fields, combine medical imaging and computer science to extract vast amounts of quantitative information from medical images, showing great potential in aiding clinical diagnosis and treatment \u003csup\u003e[\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]\u003c/sup\u003e. Radiomics and deep learning have been widely applied in kidney diseases, especially in distinguishing between benign and malignant lesions \u003csup\u003e[\u003cspan additionalcitationids=\"CR31\" citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e]\u003c/sup\u003e. Additionally, radiomics has been widely used in UTUC postoperative prognosis prediction, distinguishing it from other renal tumors, and predicting pathological grading \u003csup\u003e[\u003cspan additionalcitationids=\"CR20\" citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]\u003c/sup\u003e, achieving good predictive performance in all of these areas. In this study, we applied CNN-based deep learning algorithms and successfully distinguished pyelocaliceal UTUC from XGP. Among the 9 features selected by the CDLR model, 2 are Rad features and 7 are DL features, with DL features having the highest weight, indicating that deep learning technology can effectively extract quantitative information to distinguish pyelocaliceal UTUC from XGP. Our results show that the diagnostic ability of the DLR model outperforms both the Rad and DL models, and the DL model based on DL features is superior to the Rad model based on Rad features, further proving the application value of deep learning in this field. To our knowledge, this is the first study that combines clinical, radiomics, and deep learning features for preoperative prediction of pyelocaliceal UTUC and XGP.\u003c/p\u003e\u003cp\u003eIt is worth noting that there are some limitations in this study. First, the sample size is relatively small. Second, since this is a retrospective study, selection bias is inevitable. Finally, all subjects in the study came from a single center, lacking validation from multi-center data. Therefore, we plan to expand the sample size in the future and seek external validation from multiple centers to enhance the reliability and generalizability of the study.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eIn conclusion, the CDLR model based on clinical, radiomics, and deep learning features has significant clinical value in accurately distinguishing pyelocaliceal UTUC from XGP. Compared to individual clinical, radiomics, deep learning models, and the DLR model combining radiomics and deep learning features, the CDLR model significantly improves classification performance. Therefore, the CDLR model can help clinicians more accurately distinguish pyelocaliceal UTUC from XGP and promote more effective treatment guidance.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cp\u003eRad radiomics\u003c/p\u003e\n\u003cp\u003eDL deep learning\u003c/p\u003e\n\u003cp\u003eDLR deep learning radiomics\u003c/p\u003e\n\u003cp\u003eCDLR clinical deep learning radiomics\u003c/p\u003e\n\u003cp\u003eROC receiver operating characteristic\u003c/p\u003e\n\u003cp\u003eDCA decision curve analysis\u003c/p\u003e\n\u003cp\u003eROI region of interest\u003c/p\u003e\n\u003cp\u003eCNN convolutional neural network\u003c/p\u003e\n\u003cp\u003eCMP corticomedullary phase\u003c/p\u003e\n\u003cp\u003eNP nephrographic phase\u003c/p\u003e\n\u003cp\u003ePCP precontrast phase\u003c/p\u003e\n\u003cp\u003eHU hounsfield units\u003c/p\u003e\n\u003cp\u003ePACS picture archiving and communication system\u003c/p\u003e\n\u003cp\u003eICCs interclass and intraclass correlation coefficients\u003c/p\u003e\n\u003cp\u003eGLCM gray level co-occurrence matrix\u003c/p\u003e\n\u003cp\u003eGLDM gray-level dependence matrix\u003c/p\u003e\n\u003cp\u003eGLRLM gray-level run length matrix\u003c/p\u003e\n\u003cp\u003eGLSZM gray-level size zone matrix\u003c/p\u003e\n\u003cp\u003eNGTDM neighboring gray tone difference matrix\u003c/p\u003e\n\u003cp\u003eLASSO least absolute shrinkage and selection operator\u003c/p\u003e\n\u003cp\u003emRMR\u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp; \u0026nbsp;\u0026nbsp;minimum redundancy maximum relevance\u003c/p\u003e\n\u003cp\u003eLR logistic regression\u003c/p\u003e\n\u003cp\u003eAUC area under the receiver operating characteristic curve\u003c/p\u003e\n\u003cp\u003ePPV positive predictive value\u003c/p\u003e\n\u003cp\u003eNPV negative predictive value\u003c/p\u003e\n\u003cp\u003eCI confidence interval\u003c/p\u003e\n\u003cp\u003eOR odds ratio\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003ch2\u003eConflict of Interest\u003c/h2\u003e\u003cp\u003eThe authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.\u003c/p\u003e\u003c/p\u003e\u003cp\u003e\u003ch2\u003eEthical Statement\u003c/h2\u003e\u003cp\u003e The study protocol was approved by the Institutional Review Board. Because the study was retrospective, the institutional review board waived the patient's informed consent.\u003c/p\u003e\u003c/p\u003e\u003ch2\u003eFunding\u003c/h2\u003e\u003cp\u003eThis work was supported by grants from the National Natural Science Foundation of China (82460156,82400898).\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eC.Y.L and F.X.F: Conceptualization, funding acquisition and methodology.Y.F and Y.H.C: formal analysis, software, and writing \u0026ndash; original draft. J.Z, J.Y L, H.Q, J.W.Z and B.W: Data curation.\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eThe original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding authors.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eRoupr\u0026ecirc;t M, Babjuk M, Comp\u0026eacute;rat E, et al. European guidelines on upper tract urothelial carcinomas: 2013 update [J]. European urology, 2013, 63: 1059-71.10.1016/j.eururo.2013.03.032\u003c/li\u003e\n\u003cli\u003eKang Y, Lee T W, Bae E, et al. Xanthogranulomatous pyelonephritis in a patient with polycystic kidney disease without underlying risk factors: a case report [J]. Frontiers in medicine, 2024, 11: 1419965.10.3389/fmed.2024.1419965\u003c/li\u003e\n\u003cli\u003eColin P, Ouzzane A, Pignot G, et al. Comparison of oncological outcomes after segmental ureterectomy or radical nephroureterectomy in urothelial carcinomas of the upper urinary tract: results from a large French multicentre study [J]. BJU international, 2012, 110: 1134-41.10.1111/j.1464-410X.2012.10960.x\u003c/li\u003e\n\u003cli\u003eLughezzani G, Jeldres C, Isbarn H, et al. Nephroureterectomy and segmental ureterectomy in the treatment of invasive upper tract urothelial carcinoma: a population-based study of 2299 patients [J]. European journal of cancer (Oxford, England : 1990), 2009, 45: 3291-7.10.1016/j.ejca.2009.06.016\u003c/li\u003e\n\u003cli\u003eOosterlinck W, Solsona E, Van Der Meijden A P, et al. EAU guidelines on diagnosis and treatment of upper urinary tract transitional cell carcinoma [J]. European urology, 2004, 46: 147-54.10.1016/j.eururo.2004.04.011\u003c/li\u003e\n\u003cli\u003eBurbano M A, Nati-Castillo H A, Casta\u0026ntilde;o-Giraldo N, et al. Fatal nephrobronchial fistula arising from xanthogranulomatous pyelonephritis: a case report [J]. Frontiers in medicine, 2024, 11: 1374043.10.3389/fmed.2024.1374043\u003c/li\u003e\n\u003cli\u003eXie L, Tapiero S, Flores A R, et al. Long-Term Antibiotic Treatment Prior to Laparoscopic Nephrectomy for Xanthogranulomatous Pyelonephritis Improves Postoperative Outcomes: Results from a Multicenter Study [J]. The Journal of urology, 2021, 205: 820-5.10.1097/ju.0000000000001429\u003c/li\u003e\n\u003cli\u003eBercowsky E, Shalhav A L, Portis A, et al. Is the laparoscopic approach justified in patients with xanthogranulomatous pyelonephritis? [J]. Urology, 1999, 54: 437-42; discussion 42-3.10.1016/s0090-4295(99)00261-7\u003c/li\u003e\n\u003cli\u003eMargulis V, Shariat S F, Matin S F, et al. Outcomes of radical nephroureterectomy: a series from the Upper Tract Urothelial Carcinoma Collaboration [J]. Cancer, 2009, 115: 1224-33.10.1002/cncr.24135\u003c/li\u003e\n\u003cli\u003eAbouassaly R, Alibhai S M, Shah N, et al. Troubling outcomes from population-level analysis of surgery for upper tract urothelial carcinoma [J]. Urology, 2010, 76: 895-901.10.1016/j.urology.2010.04.020\u003c/li\u003e\n\u003cli\u003eMalek R S, Elder J S. Xanthogranulomatous pyelonephritis: a critical analysis of 26 cases and of the literature [J]. The Journal of urology, 1978, 119: 589-93.10.1016/s0022-5347(17)57559-x\u003c/li\u003e\n\u003cli\u003eChlif M, Chakroun M, Ben Rhouma S, et al. Xanthogranulomatous pyelonephritis presenting as a pseudotumour [J]. Canadian Urological Association journal = Journal de l\u0026apos;Association des urologues du Canada, 2016, 10: E36-40.10.5489/cuaj.3225\u003c/li\u003e\n\u003cli\u003eChandrashekhar S, Kumar S, Jambunathan S, et al. Xanthogranulomatous Pyelonephritis Mimicking a Complex Renal Cyst: A Report of a Rare Case [J]. Cureus, 2024, 16: e69233.10.7759/cureus.69233\u003c/li\u003e\n\u003cli\u003eXie X, Wang N, Wang Y, et al. Non-invasive papillary urothelial carcinoma, low-grade of the renal pelvis mimicking a xanthogranulomatous pyelonephritis in a male patient: A case report and review of literature [J]. International journal of immunopathology and pharmacology, 2020, 34: 2058738420925720.10.1177/2058738420925720\u003c/li\u003e\n\u003cli\u003eLafata K J, Wang Y, Konkel B, et al. Radiomics: a primer on high-throughput image phenotyping [J]. Abdominal radiology (New York), 2022, 47: 2986-3002.10.1007/s00261-021-03254-x\u003c/li\u003e\n\u003cli\u003eGillies R J, Kinahan P E, Hricak H. Radiomics: Images Are More than Pictures, They Are Data [J]. Radiology, 2016, 278: 563-77.10.1148/radiol.2015151169\u003c/li\u003e\n\u003cli\u003eLambin P, Rios-Velazquez E, Leijenaar R, et al. Radiomics: extracting more information from medical images using advanced feature analysis [J]. European journal of cancer (Oxford, England : 1990), 2012, 48: 441-6.10.1016/j.ejca.2011.11.036\u003c/li\u003e\n\u003cli\u003eZhang X, Zhang Y, Zhang G, et al. Deep Learning With Radiomics for Disease Diagnosis and Treatment: Challenges and Potential [J]. Frontiers in oncology, 2022, 12: 773840.10.3389/fonc.2022.773840\u003c/li\u003e\n\u003cli\u003eAl Mopti A, Alqahtani A, Alshehri A H D, et al. Perirenal Fat CT Radiomics-Based Survival Model for Upper Tract Urothelial Carcinoma: Integrating Texture Features with Clinical Predictors [J]. Cancers, 2024, 16.10.3390/cancers16223772\u003c/li\u003e\n\u003cli\u003eZhai X, Sun P, Yu X, et al. CT-based radiomics signature for differentiating pyelocaliceal upper urinary tract urothelial carcinoma from infiltrative renal cell carcinoma [J]. Frontiers in oncology, 2023, 13: 1244585.10.3389/fonc.2023.1244585\u003c/li\u003e\n\u003cli\u003eZheng Y, Shi H, Fu S, et al. Development and validation of a radiomics-based nomogram for predicting pathological grade of upper urinary tract urothelial carcinoma [J]. BMC cancer, 2024, 24: 1546.10.1186/s12885-024-13325-z\u003c/li\u003e\n\u003cli\u003eLotter W, Diab A R, Haslam B, et al. Robust breast cancer detection in mammography and digital breast tomosynthesis using an annotation-efficient deep learning approach [J]. Nature medicine, 2021, 27: 244-9.10.1038/s41591-020-01174-9\u003c/li\u003e\n\u003cli\u003eSong Z, Liu T, Shi L, et al. The deep learning model combining CT image and clinicopathological information for predicting ALK fusion status and response to ALK-TKI therapy in non-small cell lung cancer patients [J]. European journal of nuclear medicine and molecular imaging, 2021, 48: 361-71.10.1007/s00259-020-04986-6\u003c/li\u003e\n\u003cli\u003eEsteva A, Kuprel B, Novoa R A, et al. Dermatologist-level classification of skin cancer with deep neural networks [J]. Nature, 2017, 542: 115-8.10.1038/nature21056\u003c/li\u003e\n\u003cli\u003eZhang K, Zuo W, Chen Y, et al. Beyond a Gaussian Denoiser: Residual Learning of Deep CNN for Image Denoising [J]. IEEE transactions on image processing : a publication of the IEEE Signal Processing Society, 2017, 26: 3142-55.10.1109/tip.2017.2662206\u003c/li\u003e\n\u003cli\u003eLi L, Parwani A V. Xanthogranulomatous pyelonephritis [J]. Archives of pathology \u0026amp; laboratory medicine, 2011, 135: 671-4.10.5858/2009-0769-rsr.1\u003c/li\u003e\n\u003cli\u003eSchneeberger C, Holleman F, Geerlings S E. Febrile urinary tract infections: pyelonephritis and urosepsis [J]. Current opinion in infectious diseases, 2016, 29: 80-5.10.1097/qco.0000000000000227\u003c/li\u003e\n\u003cli\u003eGoldman S M, Hartman D S, Fishman E K, et al. CT of xanthogranulomatous pyelonephritis: radiologic-pathologic correlation [J]. AJR American journal of roentgenology, 1984, 142: 963-9.10.2214/ajr.142.5.963\u003c/li\u003e\n\u003cli\u003eLiu Y X, Liu Q H, Hu Q H, et al. Ultrasound-Based Deep Learning Radiomics Nomogram for Tumor and Axillary Lymph Node Status Prediction After Neoadjuvant Chemotherapy [J]. Academic radiology, 2025, 32: 12-23.10.1016/j.acra.2024.07.036\u003c/li\u003e\n\u003cli\u003eHan J H, Kim B W, Kim T M, et al. Fully automated segmentation and classification of renal tumors on CT scans via machine learning [J]. BMC cancer, 2025, 25: 173.10.1186/s12885-025-13582-6\u003c/li\u003e\n\u003cli\u003eWu Y, Cao F, Lei H, et al. Interpretable multiphasic CT-based radiomic analysis for preoperatively differentiating benign and malignant solid renal tumors: a multicenter study [J]. Abdominal radiology (New York), 2024, 49: 3096-106.10.1007/s00261-024-04351-3\u003c/li\u003e\n\u003cli\u003eYu S, Hao R, Cui J, et al. Integration of radiomic and deep features to reliably differentiate benign renal lesions from renal cell carcinoma [J]. European journal of radiology, 2025, 184: 111989.10.1016/j.ejrad.2025.111989\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Upper Tract Urothelial Carcinoma, Xanthogranulomatous Pyelonephritis, Radiomics, Deep Learning","lastPublishedDoi":"10.21203/rs.3.rs-7537981/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7537981/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eObjective\u003c/h2\u003e\u003cp\u003eThis study aims to establish a CT-based clinical deep learning radiomics (CDLR) model for preoperative prediction of pyelocaliceal Upper Tract Urothelial Carcinoma (UTUC) and xanthogranulomatous pyelonephritis (XGP), providing scientific guidance for personalized treatment.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e\u003cp\u003eA retrospective analysis was conducted on 161 post-operative pathology-confirmed cases of pyelocaliceal UTUC and XGP patients, divided into training cohort (n\u0026thinsp;=\u0026thinsp;112) and validation cohort (n\u0026thinsp;=\u0026thinsp;49). Radiomics (Rad) and deep learning (LR) features were extracted from three-phase CT images, combined with clinical features, and after feature selection, a logistic regression (LR) classifier was used to construct clinical, radiomics (Rad), deep learning (LR), deep learning radiomics (DLR), and clinical deep learning radiomics (CDLR) models. The top-performing model was chosen utilizing receiver operating characteristic (ROC) curve analysis. Decision curve analysis (DCA) was used to evaluate the model's practicality.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e\u003cp\u003eThe evaluation of clinical, Rad, LR, DLR, and CDLR models revealed that the CDLR model exhibited superior diagnostic performance. The area under the receiver operating characteristic curve (AUC) of the CDLR model in the training and validation cohorts were 0.984 and 0.970, respectively, outperforming other models (clinical model, Rad model, DL model, DLR model). DCA results showed that the CDLR model provided a higher net benefit in preoperative prediction.\u003c/p\u003e\u003ch2\u003eConclusion\u003c/h2\u003e\u003cp\u003eThe CDLR model, combining clinical, Rad, and LR features, could serve as a non-invasive tool for differentiating pyelocaliceal UTUC and XGP, offering valuable guidance for clinical treatment.\u003c/p\u003e","manuscriptTitle":"CT-based Deep Learning Radiomics Nomogram for Differentiating pyelocaliceal Upper Tract Urothelial Carcinoma and Xanthogranulomatous Pyelonephritis","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-09-23 02:40:58","doi":"10.21203/rs.3.rs-7537981/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"751ac558-f3c5-4765-90a4-b62157a8f3f5","owner":[],"postedDate":"September 23rd, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-10-21T09:23:38+00:00","versionOfRecord":[],"versionCreatedAt":"2025-09-23 02:40:58","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7537981","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7537981","identity":"rs-7537981","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}