CT–Based AI Score Predicts Perioperative Outcomes in Nephron–Sparing Surgery for Renal Cell Carcinoma

CT–Based AI Score Predicts Perioperative Outcomes in Nephron–Sparing Surgery for Renal Cell Carcinoma | 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 AI Score Predicts Perioperative Outcomes in Nephron–Sparing Surgery for Renal Cell Carcinoma Shengfa Lin, Liqing Su, Shu Chen, Huijian Chen, Yuying Lin, Zijie Lin, and 5 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7676212/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 29 Dec, 2025 Read the published version in Cancer Imaging → Version 1 posted 13 You are reading this latest preprint version Abstract Background To develop and validate a CT–based artificial intelligence (AI) score model integrating the R.E.N.A.L. nephrometry and contact surface area (CSA) for efficient, accurate prediction of perioperative outcomes in renal cell carcinoma (RCC) patients undergoing nephron–sparing surgery (NSS), addressing the subjectivity and inefficiency of manual score. Methods Retrospective data from two RCC cohorts were analyzed. Ninety percent of the n1 cohort was randomly allocated to develop and validate AI–driven kidney/tumor segmentation models and derive AI–calculated R.E.N.L. (The “A” score was ignored) and AI–calculated CSA scores. The remaining 10% of Cohort n1, combined with Cohort n2, were used for risk stratification prediction. Manual image annotation/scoring was conducted by experienced radiologists and urologists. Interrater consistency was evaluated via weighted kappa coefficients; risk stratification was performed viaKruskal–Wallis tests and Mann–Whitney U tests. Results A total of 550 patients were included in this study (median age, 56 [IQR: 46–66] years; 341 males), with n1=500 and n2=50. Automatic segmentation achieved high accuracy (Dice similarity coefficients: kidney 0.95, tumor 0.80). The R, E, N, L, R.E.N.L., and CSA score models had good consistency compared with the manual score, and the kappa coefficients were 0.82, 0.49, 0.63, 0.60, 0.65, and 0.69, respectively (all P < 0.01). Risk stratification by AI score significantly predicted warm ischemia time, surgical duration, intraoperative blood loss, serum creatinine changes, pathological T stage, and nuclear grade (all P < 0.05). Conclusions This study establishes a CT–based AI framework that integrates R.E.N.L. and CSA metrics, enabling standardized, objective preoperative risk assessment for NSS in RCC. Renal cell carcinoma Computed tomography Artificial intelligence Nephron-sparing surgery R.E.N.A.L. nephrometry contact surface area Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 1 Introduction Renal cell carcinoma (RCC) accounts for 2%–3% of adult malignant tumors, with a rising global incidence (average annual growth rate: 0.7%–2% over the past decade) [ 1 – 4 ] . Advances in imaging techniques (ultrasound, CT/MRI) have significantly improved early RCC detection [ 5 ] . Clinical guidelines designate nephron–sparing surgery (NSS) as the preferred treatment for T1a RCC and a viable option for selecting T1b tumors [ 6 , 7 ] . Tumor anatomical characteristics, such as size, location, relationship to the renal vasculature/collecting system, and contact surface area (CSA), are pivotal determinants of NSS outcomes. Widely used scoring systems, including the R.E.N.A.L. nephrometry score [ 8 ] , PADUA classification [ 9 ] , and renal tumor CSA quantify these features from CT/MRI to predict perioperative outcomes such as warm ischemia time (WIT), operative duration, blood loss, pathological stage, and nuclear grade [ 10 ] . Among them, the R.E.N.A.L. score includes five parameters: maximum tumor diameter (R), exophytic ratio (E), nearest distance to the renal sinus (N), anterior‒posterior location (A), and relationship to the polar line (L). CSA represents the tumor contact surface area. The meta–analyses confirmed a strong correlation between R.E.N.A.L. and PADUA (r = 0.8607) and their efficacy in outcome prediction [ 11 – 16 ] . The CSA score has unique advantages because it can assess residual renal function after surgery [ 17 – 21 ] . All score systems rely on manual clinician–based measurements and are plagued by inefficiency, subjectivity, poor reproducibility, suboptimal accuracy, and limited clinical utility. Simplified newer alternatives created by scholars have failed to address these limitations [ 22 – 28 ] . Artificial intelligence (AI), which leverages automation, efficiency, and precision, has emerged as a transformative tool in RCC management, demonstrating promise in diagnosis, treatment planning, and prognosis [ 29 – 38 ] . At present, the exploration of R.E.N.A.L. CSA scores generated by AI have been reported only individually [ 39 , 40 ] , and its efficacy needs to be further improved. However, the ability of the NSS to predict perioperative outcomes has not yet been reported. This study aimed to develop and validate a CT-based AI score framework for RCC to predict perioperative outcomes in NSS. Notably, the "A" parameter was excluded from R.E.N.A.L. nephrometry score because it yields only qualitative classifications (A, P, X) with no numerical score. By integrating R.E.N.L. With CSA parameters via deep learning (convolutional neural networks), the model seeks to establish an objective, efficient system to overcome the subjectivity and inefficiency of manual assessment. Drawing on large–scale datasets from Fuzhou University Affiliated Provincial Hospital and Fujian Cancer Hospital and leveraging advanced medical image processing and AI expertise from Fuzhou University’s College of Physics and Information Engineering, this project endeavors to create a standardized, evidence–based tool for preoperative risk stratification. The ultimate goal is to enhance clinical decision–making by providing surgeons with objective, data–driven insights into NSS feasibility and potential outcomes for RCC patients. 2 Methods 2.1 Ethical A pproval and Data Collection This study was approved by the Institutional Ethics Committee of Fuzhou University Affiliated Provincial Hospital and Fujian Cancer Hospital, with approval numbers K2021–12–018 and SQ2023–107, respectively. Retrospectively, RCC patients who underwent NSS were collected from two centers: Fuzhou University Affiliated Provincial Hospital (n1 cohort: January 2016–January 2025) and Fujian Cancer Hospital (n2 cohort: January 2023–January 2025). The inclusion criteria were as follows: 1) postoperative pathologically confirmed RCC; 2) complete contrast–enhanced arterial–phase CT imaging of the kidney and tumor; and 3) comprehensive clinical records available for analysis. The exclusion criteria were as follows: 1) CT images containing artifacts that interfere with score or model development and 2) the presence of multiple renal tumors (in a single or bilateral kidney). 2.2 Manual Score Imaging data were evaluated by a team of radiologists and urologists with 7–12 years of clinical experience in RCC diagnosis. Disputed cases underwent collective decision–making by the entire team. The R.E.N.A nephrometry score followed the classic algorithm developed by Kutikov et al [8] , but this protocol included only five parameters (R, E, N, L, and R.E.N.L.) while excluding the A parameter, as the A parameter yields only qualitative classifications (A, P, X) without numerical scores. The manual CSA was calculated via a spherical crown–like surface area formula: , where R represents the tumor diameter and h denotes the depth of the tumor within the renal parenchyma. 2.3 Developing and Validating AI Score Models 2.3.1 Datasets and Preprocessing This study integrated two subsets (n1 and n2 cohorts), with all CT scans including complete arterial–phase images at a resolution of 512×512. For the n1 cohort, 90% of the cases were randomly allocated to training, internal validation, and test sets at a 7:1:2 ratio, whereas the remaining 10% were combined with the n2 cohort to form a mixed validation set. The data used to establish the models were manually annotated by radiologists and urologists with 7–12 years of clinical experience via Mimics Research 21.0 software. Raw DICOM files were converted to NIfTI format via Python scripts (using the dicom2nifti library), with voxel sizes standardized and background noise cropped to ensure uniform preprocessing. 2.3.2 Development and Validation of Automated Kidney and Tumor Segmentation Models Figure 1 shows the data distribution flowchart. Randomly, 50 cases from the n1 cohort were manually annotated, and an additional 265 cases were annotated via semi–supervised learning, yielding 315 cases as "ground truths" for deep learning. An internal validation set (45 cases) and a test set (90 cases) were separately reserved. The segmentation model was trained via the nnU–Net framework. Preprocessed single–channel CT images served as model inputs, generating binary segmentation masks for kidneys and tumors as outputs. Prior to training, the data underwent three preprocessing steps: 1) spatial normalization—images were resampled to the median voxel size via trilinear interpolation, with labels resampled via nearest–neighbor interpolation to preserve anatomical boundaries; 2) intensity standardization—after adjusting the window width/level to 400 HU/30 HU, pixel intensities were normalized via Z–score transformation to standardize brightness across scans; and 3) data augmentation—random rotations (±15°), elastic deformations, and scaling (0.8–1.2×)—were applied to enhance model generalization and reduce overfitting. The training was iterated for a total of 505 rounds, and overfitting was avoided through the early stop method (monitored by the internal validation set). Model performance for renal and tumor segmentation was evaluated via the Dice similarity coefficient (DSC, a metric quantifying the spatial overlap between the predicted and ground–truth segmentation masks). 2.3.3 Development and Validation of the R . E . N . L . Score Model Based on the segmentation results, the R.E.N.L. score model parameters were defined as follows: 1) R (maximum tumor diameter): calculated via 3D connected component analysis and the minimum bounding box method to measure the longest tumor dimension; 2) E (exophytic ratio): the renal parenchymal boundary was reconstructed to quantify the proportion of tumor volume protruding outside the kidney, defined as the exophytic volume divided by total tumor volume; and 3) N (nearest distance to the renal sinus): the renal sinus was segmented using a HU threshold (-110 to 0), and the minimum Euclidean distance between the tumor and sinus region was computed. 4) L (relationship to the polar line): Renal poles were localized via morphological analysis, and the proportion of tumor voxels crossing the polar line (connecting the upper and lower renal poles) was determined. 5) R.E.N.L. total score: Sum of the R, E, N, and L parameters. Notably, the "A" parameter was excluded from the score. 2.3.4 Development and Validation of the CSA Score Model The process of the CSA score model mainly includes the following steps: 1) Contact surface extraction: the marching cubes algorithm was applied to generate a triangular mesh representing the tumor–renal parenchyma contact surface from segmentation masks; 2) area calculation: on the basis of the vertex coordinates of the triangular mesh and voxel physical dimensions, the area of each triangle was computed and summed to obtain the total contact surface area; and 3) unit conversion: the result was converted from mm² to cm² by dividing by 100. 2.3.5 AI–R.E.N.L. and AI–CSA enable risk stratification of NSS -related perioperative outcomes R.E.N.L. score risk stratification: low complexity (4–6 points), medium complexity (7–9 points) and high complexity (10–12 points). CSA score risk stratification was as follows: low-CSA group (CSA value < 20 cm²) and high-CSA group (CSA value ≥20 cm²). The evaluated outcomes included WIT, surgical duration, intraoperative blood loss, serum creatinine change, time to extubation, postoperative hospital stay, pathological T staging, and nuclear grade. WIT was defined as the duration from renal artery clamping to reperfusion. Surgical duration was calculated as the time from anesthesia induction to anesthesia termination. Postoperative serum creatinine changes were measured as the difference between day 1 postoperative and preoperative creatinine levels. The nuclear classification is based on the WHO/ISUP standard (Fifth Edition). 2.4 Statistical Analysis Statistical analyses were performed via SPSS 27.0 and GraphPad Prism 9.5. Continuous variables are summarized as medians (interquartile ranges, IQRs), and categorical variables are summarized as frequencies (percentages). Data normality was tested via the Shapiro–Wilk test. Agreement between automated and manual scores was evaluated via the weighted Kappa coefficient. For risk stratification of NSS perioperative outcomes, the Kruskal–Wallis test and Mann–Whitney U test were used for group comparisons, with post hoc Dunn’s multiple comparisons test used to identify pairwise differences. Statistical significance was set at P < 0.05. 3 Results 3.1 Cohort C haracteristics and Flowchart On the basis of the inclusion and exclusion criteria, a total of N = 550 patients were ultimately included in this study, consisting of 500 patients from Fuzhou University Affiliated Provincial Hospital (n1 cohort) and 50 patients from Fujian Cancer Hospital (n2 cohort). Figure 1 shows the data distribution flowchart. Figure 2 shows the flowchart of the technique of this research. This real–world dataset included a cohort with a median age of 56 (46–66) years, comprising 341 males (62.0%) and 209 females (38%). The majority of cases (471, 85.64%) were detected incidentally during routine physical examinations. Hypertension and diabetes were present in 199 (36.18%) and 78 (18.18%) patients, respectively. The median values for preoperative serum creatinine, postoperative day 1 serum creatinine, and creatinine changes were 73 (60–82) μmol/L, 80 (66–97) μmol/L, and 9 (1–19) μmol/L, respectively. Preoperative prophylactic antibiotic use was recorded in 268 patients (48.73%). Pathological analysis revealed clear cell carcinoma as the predominant histotype (449 cases, 81.64%), with most tumors staged as T1a (428 cases, 77.82%) or nuclear grade 1/2 (82.55%). The remaining demographic and clinical characteristics are detailed in Table 1. 3.2 Automated Kidney and Tumor Segmentation Models The kidney and tumor segmentation models achieved DSCs of 0.95 and 0.80, respectively. Figures 3 and 4 depict schematic visualizations of the segmentation results for two representative cases. Figure 3 illustrates a low–complexity case: both manual–calculated and AI–calculated R.E.N.L. (AI–R.E.N.L.) scores were , with manually–calculated CSA (M–CSA) of 14.83 cm² and an AI–calculated CSA (AI–CSA) of 14.26 cm². Figure 4 shows a high–complexity case (irregular morphology with extensive necrosis and liquefaction): the manually calculated R.E.N.L. score (M–R.E.N.L.) was , whereas the AI–R.E.N.L. ; the M–CSA and AI–CSA values were 67.86 cm² and 59.07 cm², respectively. Both cases demonstrated the excellent segmentation performance of the AI model, particularly in capturing anatomical details across varying levels of tumor complexity. 3.3 Agreement Validation between AI–calculated R.E.N.L./CSA Scores and Manual Scores Agreement for six parameters–R, E, N, L, R.E.N.L. total score and CSA were assessed via weighted kappa coefficients, and the detailed results are presented in Table 2. The median (IQR) of M–CSA was 18.56 (10.30–27.00) cm², whereas it was 14.17 (7.76–21.10) cm² for AI–CSA. 3.4 AI–R.E.N.L. and AI–CSA enable risk stratification of NSS -related perioperative outcomes On the basis of the previous findings, both AI–R.E.N.L. and AI–CSA scores showed substantial agreement with the manual assessments. Thus, AI–R.E.N.L. was directly applied to stratify risks into low-, moderate-, and high-complexity groups, predicting intergroup differences in warm ischemia time, operative duration, intraoperative blood loss, postoperative serum creatinine change, time to extubation, postoperative hospital stay, pathological stage, and nuclear grade (detailed results in Table 3 and Figure 5). Similarly, AI–CSA scores were directly used for risk stratification into low- and high-score groups, and the results are presented in Table 4. 4 Discussion First, we successfully developed an automated CT–based segmentation model for kidneys and tumors, which demonstrated favorable performance compared with manual segmentation. On the basis of this model, we subsequently constructed AI–R.E.N.L. and the AI–CSA score models, both of which strongly agreed with the manual score. Finally, risk stratification of NSS perioperative outcomes via these models revealed that AI scores effectively predict WIT, surgical duration, intraoperative blood loss, postoperative serum creatinine change, pathological stage, and nuclear grade. To enhance generalizability, this study integrated 550 real–world, multicenter (two–center) imaging datasets, with 450 cases used for developing/validating the segmentation and score models and 100 cases in a mixed dataset for perioperative outcome prediction. By combining manual annotation with semi–supervised learning, we efficiently created high–accuracy "ground truth" data for the segmentation model. The model achieved DSCs for kidney and tumor segmentation comparable to the KITS challenge average [ 41 ] , confirming its reliability. For the AI–R.E.N.L. model, weighted Kappa coefficients for parameters R, E, N, L, and total R.E.N.L. scores were 0.82, 0.49, 0.63, 0.60, and 0.65, respectively, indicating almost perfect agreement for R and substantial agreement for N, L and total R.E.N.L. score and moderate agreement for E. These results exceed the clinical utility reported by Heller et al. [ 39 ] . The AI–CSA model showed strong agreement with M–CSA (Kappa = 0.69), which was comparable to the results of Wood et al. [ 40 ] . Notably, AI–CSA directly calculates the tumor–kidney contact surface via pixel–level analysis, avoiding the systematic bias of manual methods, which assume spherical tumor geometry and often overestimate M–CSA (median M–CSA: 18.56 [10.30–27.00] cm² vs. AI–CSA: 14.17 [7.76–21.10] cm²). Using the AI–R.E.N.L. and AI–CSA score models for risk stratification of NSS perioperative outcomes, the AI–R.E.N.L. score analysis revealed the following: ① the median WIT was 25 min (low complexity) < 28 min (moderate complexity) < 30 min (high complexity), with statistically significant differences ( P < 0.05), and post–hoc pairwise comparisons revealed that both the moderate- and high-complexity groups required longer ischemia than the low-complexity group, which is consistent with prior reports [ 42 , 43 ] , prompting clinicians to monitor ischemia duration; ② the median surgical duration was 141 min (low) < 170 min (moderate) < 200 min (high), which is statistically significant ( P < 0.05), with longer durations in the moderate/high- vs. low-complexity groups, aligning with Wang et al.’s findings [ 44 ] ; ③ the median intraoperative blood loss was 20 ml (low) < 50 ml (moderate/high), which is statistically significant ( P < 0.05), with greater loss in the moderate/high-complexity groups, which is consistent with Styopushkin et al.’s findings of complexity and bleeding risk [ 45 ] ; and ④ the median serum creatinine changes were 8 µmol/L (low), 3 µmol/L (moderate), and 20 µmol/L (high), which is statistically significant ( P < 0.05), with marked elevation in high vs. low/moderate complexity groups, supporting literature on complexity–correlated creatinine increases [ 46 ] ; ⑤ pathological T1a stage proportions were 90.48% (low), 76.32% (moderate), and 46.67% (high), which is statistically significant ( P < 0.05), with fewer low–stage cases in high complexity; ⑥ nuclear grade 1/2 proportions were 87.80% (low), 86.49% (moderate), and 57.14% (high), which is statistically significant ( P < 0.05), with higher grades in high complexity, indicating increased invasiveness and recurrence risk, consistent with Sun et al.’s validation of R.E.N.A.L. predictive value for nuclear grade [ 47 ] ; ⑦ no statistical differences in postoperative hospital stay or extubation time, possibly due to uniform laparoscopic NSS reducing variability. For AI–CSA stratification: ① median WIT was 25 min (low CSA) < 30 min (high CSA, P < 0.05), ② median operative duration 142 min (low) < 185 min (high, P < 0.05), ③ median blood loss 30 ml (low) < 50 ml (high, P < 0.05), ④ T1a stage proportion 88.89% (low) vs. 56.25% (high, P < 0.05), indicating higher T stages in high CSA; ⑤ no differences were observed in creatinine change, extubation time, hospital stay, or nuclear grade. These findings align with Leslie et al.’s report of positive correlations between CSA score and WIT, surgical duration, blood loss, creatinine change, and pathological stage [ 48 ] . While this study achieved meaningful achievements, several limitations should be acknowledged. First, to avoid confounding segmentation and scoring of multiple tumors, the cohort excluded patients with multiple renal tumors, and future studies could include these populations to expand generalizability. Second, a small proportion of tumors smaller than 1 cm were under–segmented, and a minor subset of renal cysts were misclassified as tumors and erroneously segmented; future studies should further improve segmentation precision. Third, the analysis was limited to nephron–sparing surgery (NSS) data, and future research may explore perioperative outcomes in patients undergoing radical nephrectomy. Fourth, the study relied primarily on CT imaging and lacked investigations of magnetic resonance imaging (MRI) or other modalities, which could be incorporated to enhance multi–imaging applicability. 5 Conclusions In summary, this study successfully developed a CT-based AI score model for RCC, which effectively performs risk stratification for NSS-related perioperative outcomes. The model demonstrates high efficiency and generalizability, featuring both robust automated kidney/tumor segmentation and precise prediction of key metrics, including WIT, surgical duration, intraoperative blood loss, postoperative serum creatinine change, pathological T stage, and nuclear grade. These capabilities have significantly increased the objectivity and accuracy of preoperative assessment, as well as the efficiency of clinical decision–making. They have reduced the previously required several hours of manual labor to just approximately ten minutes, highlighting their potential for clinical application. With advancements in AI technology, future research may explore more complex scenarios, such as radical nephrectomy and multifocal renal cancer, to enable AI to comprehensively support the entire workflow of renal cancer diagnosis and treatment. Abbreviations AI: Artificial intelligence CSA: Contact surface area NSS: Nephron–sparing surgery RCC: Renal cell carcinoma WIT: Warm ischemia time DSC: Dice similarity coefficient IQR: Interquartile range M–CSA: Manually–calculated CSA AI–CSA: AI–calculated CSA M–R.E.N.L. : Manually–calculated R.E.N.L. AI–R.E.N.L. : AI–calculated R.E.N.L. Declarations Ethics approval and consent to participate This study has been approved by the Institutional Review Boards of Fuzhou University Affiliated Provincial Hospital and Fujian Cancer Hospital, with approval numbers K2021-12-018 (approval date: 2021/12/08) and SQ2023-107 (approval date: 2023/12/12), respectively. Consent for publication Not applicable. Availability of data and material The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request. Competing interests The authors declare that they have no competing interests. Funding This study was supported by the Guiding Project of Fujian Provincial Department of Science and Technology (No. 2022Y0052, Shengfa Lin), the Startup Fund for Scientific Research, Fujian Medical University (No. 2023QH1174, Liqing Su), the Startup Fund for Scientific Research and Fujian Medical University (2021QH1283, Minxiong Hu), the National Key R&D Program of China (No. 2023YFC2413500, Zhuting Fang), the Joint Funds for the Innovation of Science and Technology, Fujian Province (No. 2023Y9320, Zhuting Fang), the Fujian Province Natural Science Fund Project (No. 2024J011105, Zhuting Fang), and the Major Project of Fujian Provincial Health Commission (No. 2024ZD01004, Zhuting Fang). Authors' contributions SFL was responsible for data collection, organization, analysis, and manuscript preparation; LQS was in charge of data collection and organization at Fujian Provincial Cancer Hospital; HJC, YYL, ZJL, and YFX were tasked with image scoring; SC was responsible for AI model development; ZTF, MPM, and MXH provided project guidance and support. Acknowledgements The authors are very grateful to Dr. Wang Xuefei and Dr. Wu Yangbiao for helping to collect and organize the data. References Wein AJ, Kavoussi LR, Partin AW, Peters CA (Eds). (2016). Campbell-Walsh urology (11th ed, 4-volume set). Elsevier. Miller KD, Nogueira L, Devasia T et al (2022) Cancer treatment and survivorship statistics, 2022. CA-CANCER J CLIN 72:409–436. doi: 10.3322/caac.21731 Han B, Zheng R, Zeng H et al (2024) Cancer incidence and mortality in China, 2022. J Natl Cancer Cent 4(1):47-53. doi: 10.1016/j.jncc.2024.01.006 Ferlay J, Colombet M, Soerjomataram I et al (2018) Cancer incidence and mortality patterns in Europe: Estimates for 40 countries and 25 major cancers in 2018. EUR J CANCER 103:356-387. doi: 10.1016/j.ejca.2018.07.005 Escudier B, Porta C, Schmidinger M, et al (2019) Renal cell carcinoma: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up†. ANN ONCOL 30(5):706-720. doi: 10.1093/annonc/mdz056 Huang J, Zhang X (Eds).(2022). Guidelines for the diagnosis and treatment of urologic and andrologic diseases in China (2022 ed). Science Press. European Association of Urology (EAU). (2025). European Association of Urology Guidelines on Renal Cell Carcinoma: The 2025 Update. https://uroweb.org/guideline/renal-cell-carcinoma/ Kutikov A, Uzzo RG (2009) The R.E.N.A.L. nephrometry score: a comprehensive standardized system for quantitating renal tumor size, location and depth. J UROLOGY 182(3): 844-53. doi: 10.1016/j.juro.2009.05.035 Ficarra V, Novara G, Secco S et al (2009) Preoperative aspects and dimensions used for an anatomical (PADUA) classification of renal tumours in patients who are candidates for nephron-sparing surgery. EUR UROL 56(5): 786-93. doi: 10.1016/j.eururo.2009.07.040 Veccia A, Antonelli A, Uzzo RG et al (2019) Predictive Value of Nephrometry Scores in Nephron-sparing Surgery: A Systematic Review and Meta-analysis. Eur Urol Focus 6(3):490-504. doi: 10.1016/j.euf.2019.11.004 Schiavina, R, Novara, G, Borghesi, M, et al (2017) PADUA and R.E.N.A.L. nephrometry scores correlate with perioperative outcomes of robot-assisted partial nephrectomy: analysis of the Vattikuti Global Quality Initiative in Robotic Urologic Surgery (GQI-RUS) database. BJU INT 119(3):456-463. doi: 10.1111/bju.13628 Bertolo R, Pozzi L (2019) From PADUA to R.E.N.A.L. Score and Vice Versa: Development and validation of a mathematical converter. J UROLOGY 201 (4): 674-675. doi: 10.1016/j.juro.2018.10.020 Zhou HJ, Yan Y, Zhang JZ et al (2017) Role of R.E.N.A.L. Nephrometry Score in Laparoscopic Partial Nephrectomy. CHINESE MED J-PEKING 130(18):2170-2175. doi: 10.4103/0366-6999.213973 Zinssius D, Jünemann KP, Geiger F et al_(2022) Evaluation of the Padua and R.E.N.A.L. scores regarding their validity and implication in the perioperative management during partial nephrectomy. AKTUEL UROL 53(5):423-430. doi: 10.1055/a-0888-7234 Tanaka H, Wang Y, Suk-Ouichai C et al (2018) Can We Predict Functional Outcomes after Partial Nephrectomy?. J Urol 201(4):693-701. doi:10.1016/j.juro.2018.09.055 Kwon KJ, Ryu H, Kim M, et al (2021) Personalised three-dimensional printed transparent kidney model for robot-assisted partial nephrectomy in patients with complex renal tumours (R.E.N.A.L. nephrometry score ≥7): a prospective case-matched study. BJU Int 127(5):567-574. doi:10.1111/bju.15275 Leslie S, Gill IS, de Castro Abreu AL, et al (2014) Renal tumor contact surface area: a novel parameter for predicting complexity and outcomes of partial nephrectomy. Eur Urol 66(5):884-893. doi:10.1016/j.eururo.2014.03.010 Takagi T, Yoshida K, Kondo T, et al (2019) Association between tumor contact surface area and parenchymal volume change in robot-assisted laparoscopic partial nephrectomy carried out using the enucleation technique. Int J Urol 26(7):745-751. doi:10.1111/iju.14004 Kahn AE, Shumate AM, Galler IJ, Ball CT, Thiel DD (2020) Contact surface area and its association with outcomes in robotic-assisted partial nephrectomy. Int J Med Robot 16(1):e2069. doi:10.1002/rcs.2069 Lee CH, Ku JY, Park YJ, Seo WI, Ha HK (2019) The superiority of contact surface area as a predictor of renal cortical volume change after partial nephrectomy compared to RENAL, PADUA and C-index: an approach using computed tomography-based renal volumetry. Scand J Urol 53(2-3):129-133. doi:10.1080/21681805.2019.1614663 Hsieh PF, Wang YD, Huang CP, et al (2016) A Mathematical Method to Calculate Tumor Contact Surface Area: An Effective Parameter to Predict Renal Function after Partial Nephrectomy. J Urol 196(1):33-40. doi:10.1016/j.juro.2016.01.092 Benadiba S, Verin AL, Pignot G, et al (2015) Are urologists and radiologists equally effective in determining the RENAL Nephrometry score?. Ann Surg Oncol 22(5):1618-1624. doi:10.1245/s10434-014-4152-1 Nisen H, Ruutu M, Glücker E, Visapää H, Taari K (2014) Renal tumour invasion index as a novel anatomical classification predicting urological complications after partial nephrectomy. Scand J Urol 48(1):41-51. doi:10.3109/21681805.2013.797491 Shin TY, Komninos C, Kim DW, et al (2015) A novel mathematical model to predict the severity of postoperative functional reduction before partial nephrectomy: the importance of calculating resected and ischemic volume. J Urol 193(2):423-429. doi:10.1016/j.juro.2014.07.084 Spaliviero M, Poon BY, Karlo CA, et al (2015) An Arterial Based Complexity (ABC) Scoring System to Assess the Morbidity Profile of Partial Nephrectomy. Eur Urol 69(1):72-79. doi:10.1016/j.eururo.2015.08.008 Zhang R, Wu G, Huang J, et al (2017) Peritumoral Artery Scoring System: a Novel Scoring System to Predict Renal Function Outcome after Laparoscopic Partial Nephrectomy. Sci Rep 7(1):2853. doi:10.1038/s41598-017-03135-8 Li Y, Zhou L, Bian T, et al (2017) The zero ischemia index (ZII): a novel criterion for predicting complexity and outcomes of off-clamp partial nephrectomy. World J Urol 35(7):1095-1102. doi:10.1007/s00345-016-1975-3 Ficarra V, Porpiglia F, Crestani A, et al (2019) The Simplified PADUA REnal (SPARE) nephrometry system: a novel classification of parenchymal renal tumours suitable for partial nephrectomy. BJU Int 124(4):621-628. doi:10.1111/bju.14772 Carlier M, Lareyre F, Lê CD, et al (2021) A pilot study investigating the feasibility of using a fully automatic software to assess the RENAL and PADUA score. Prog Urol 32(8-9):558-566. doi:10.1016/j.purol.2022.04.001 Khene ZE, Bigot P, Doumerc N, et al (2023) Application of Machine Learning Models to Predict Recurrence After Surgical Resection of Nonmetastatic Renal Cell Carcinoma. Eur Urol Oncol 6(3):323-330. doi:10.1016/j.euo.2022.07.007 Zhou T, Guan J, Feng B, et al (2023) Distinguishing common renal cell carcinomas from benign renal tumors based on machine learning: comparing various CT imaging phases, slices, tumor sizes, and ROI segmentation strategies. Eur Radiol 33(6):4323-4332. doi:10.1007/s00330-022-09384-0 Cui E, Li Z, Ma C, et al (2020) Predicting the ISUP grade of clear cell renal cell carcinoma with multiparametric MR and multiphase CT radiomics. Eur Radiol 30(5):2912-2921. doi:10.1007/s00330-019-06601-1 Nassiri N, Maas M, Cacciamani G, et al (2022) A Radiomic-based Machine Learning Algorithm to Reliably Differentiate Benign Renal Masses from Renal Cell Carcinoma. Eur Urol Focus 8(4):988-994. doi:10.1016/j.euf.2021.09.004 Ohe C, Yoshida T, Amin MB, et al (2023) Deep learning-based predictions of clear and eosinophilic phenotypes in clear cell renal cell carcinoma. Hum Pathol 131:68-78. doi:10.1016/j.humpath.2022.11.004 Chen S, Song D, Chen L, et al (2023) Artificial intelligence-based non-invasive tumor segmentation, grade stratification and prognosis prediction for clear-cell renal-cell carcinoma. Precis Clin Med 6(3):pbad019. Published 2023 Aug 17. doi:10.1093/pcmedi/pbad019 Peng Q, Shen Y, Fu K, et al (2021) Artificial intelligence prediction model for overall survival of clear cell renal cell carcinoma based on a 21-gene molecular prognostic score system. Aging (Albany NY) 13(5):7361-7381. doi:10.18632/aging.202594 Barkan E, Porta C, Rabinovici-Cohen S, Tibollo V, Quaglini S, Rizzo M (2023) Artificial intelligence-based prediction of overall survival in metastatic renal cell carcinoma. Front Oncol 13:1021684. doi:10.3389/fonc.2023.1021684 Schulz S, Woerl AC, Jungmann F, et al. Multimodal Deep Learning for Prognosis Prediction in Renal Cancer. Front Oncol 11:788740. doi:10.3389/fonc.2021.788740 Heller N, Tejpaul R, Isensee F, et al (2022) Computer-Generated R.E.N.A.L. Nephrometry Scores Yield Comparable Predictive Results to Those of Human-Expert Scores in Predicting Oncologic and Perioperative Outcomes. J Urol 207(5):1105-1115. doi:10.1097/JU.0000000000002390 Wood AM, Abdallah N, Heller N, et al (2024) Fully Automated Versions of Clinically Validated Nephrometry Scores Demonstrate Superior Predictive Utility versus Human Scores. BJU Int 133(6):690-698. doi:10.1111/bju.16276 Heller, N, Isensee, F, Maier-Hein, KH, et al (2021) The state of the art in kidney and kidney tumor segmentation in contrast-enhanced CT imaging: Results of the KiTS19 challenge. MED IMAGE ANAL 67:101821. doi: 10.1016/j.media.2020.101821 Hu C, Sun J, Zhang Z, et al (2021) Parallel comparison of R.E.N.A.L., PADUA, and C-index scoring systems in predicting outcomes after partial nephrectomy: A systematic review and meta-analysis. Cancer Med 10(15):5062-5077. doi:10.1002/cam4.4047 Dubeux VT, Zanier JFC, Gabrich PN, Carrerette FB, Milfont JCA, Damião R (2022) Practical evaluation of the R.E.N.A.L. score system in 150 laparoscopic nephron-sparing surgeries. Int Braz J Urol 48(1):110-119. doi:10.1590/S1677-5538.IBJU.2021.0424 Wang Q, Qian B, Li Q, Ni Z, Li Y, Wang X (2015) Application of modified R.E.N.A.L. nephrometry score system in evaluating the retroperitoneal partial nephrectomy for T1 renal cell carcinoma. Int J Clin Exp Med 8(4):6482-6488. Styopushkin S, Chaikovskyi V, Chernylovskyi V, Sokolenk R, Bondarenko D (2021) POSTOPERATIVE HEMORRHAGE AS A COMPLICATION OF A PARTIAL NEPHRECTOMY: FREQUENCY, FEATURES AND MANAGEMENT. Georgian Med News (313):12-20. Dahlkamp L, Haeuser L, Winnekendonk G, et al (2019) Interdisciplinary Comparison of PADUA and R.E.N.A.L. Scoring Systems for Prediction of Conversion to Nephrectomy in Patients with Renal Mass Scheduled for Nephron Sparing Surgery. J Urol 202(5):890-898. doi:10.1097/JU.0000000000000361 Sun R, Zhao S, Jiang H, et al (2021) Imaging Tool for Predicting Renal Clear Cell Carcinoma Fuhrman Grade: Comparing R.E.N.A.L. Nephrometry Score and CT Texture Analysis. Biomed Res Int 2021:1821876. doi:10.1155/2021/1821876 Leslie S, Gill IS, de Castro Abreu AL, et al (2014) Renal tumor contact surface area: a novel parameter for predicting complexity and outcomes of partial nephrectomy. Eur Urol 66(5):884-893. doi:10.1016/j.eururo.2014.03.010 Tables Table 1 Cohort Characteristics Demographic and clinical characteristics N=550 Gender - n（%） Male 341（62.0%） Female 209（38.0%） Age（yrs.）- Median（IQR） - 56（46-66） BMI（Kg/m²）- Median（IQR） - 24.11（22.03-26.10） Symptom - n（%） Incidentally detected 471（85.64%） Pain 27（4.91%） Hematuria 34（6.18%） Others 18（3.27%） Hypertension - n（%） Yes 199（36.18%） No 351（63.82%） Diabetes - n（%） Yes 78（18.18%） No 472（85.82%） Preoperative serum creatinine（μmol/L）- Median（IQR） - 73（60-82） Day 1 postoperative serum creatinine（μmol/L）- Median（IQR） - 80（66-97） Serum creatinine change（μmol/L） - Median（IQR） - 9（1-19） Prophylactic use of antibiotics before the operation - n（%） Yes 268（48.73%） No 282（51.27%） Pathological types - n（%） Clear cell carcinoma 449（81.64%） Papillary cell carcinoma 17（3.09%） Chromophobe cell carcinoma 34（6.18%） Others 50（9.09%） T staging - n (%) T1a 428（77.82%） Stage T1b and above 122（22.18%） nuclear classification - n (%) 1/2 454（82.55%） 3/4 96（17.45%） IQR: Interquartile range Table 2 Consistency Verification of AI-R.E.N.L. Score /AI-CSA Score and Manual Score Manual VS AI （n=90） Weighted Kappa coefficients Standard error 95% CI P Value Agreement R 0.82 0.07 0.68-0.96 ＜0.01 Almost perfect E 0.49 0.06 0.38-0.61 ＜0.01 Moderate N 0.63 0.06 0.51-0.74 ＜0.01 Substantial L 0.60 0.06 0.48-0.71 ＜0.01 Substantial R.E.N.L. 0.65 0.04 0.57-0.72 ＜0.01 Substantial CSA 0.69 0.03 0.63-0.75 ＜0.01 Substantial AI : Artificial intelligence CSA : Contact surface area AI-R.E.N.L. : AI–calculated R.E.N.L. AI-CSA: AI–calculated CSA Table 3 Risk stratification of perioperative outcomes for NSS by AI-R.E.N.L. AI-R.E.N.L. （N=100） low complexity（4-6） moderate complexity（7-9） high complexity（10-12） Kruskal-Wallis statistic P n（%） 46（46%） 39（39%） 15（15%） WIT ( min ) - Median（ IQR ） 25（20-28） 28（25-30） 30（28-30） 16.63 ＜0.01 Surgical duration ( min ) - Median （ IQR ） 141（109-175） 170（140-200） 200（139-290） 10.87 ＜0.01 Intraoperative blood loss ( ml ) - Median（ IQR ） 20（20-47） 50（20-100） 50（50-100） 18.30 ＜0.01 Postoperative serum creatinine change ( μmol/L ) - M（ IQR ） 8（4-18） 3（-2-13） 20（12 - 33） 13.86 ＜0.01 Time to extubation - Median（ IQR ） 5（4-6） 6（4-7） 5（4-6） 2.57 0.28 Postoperative hospital stay - Median（ IQR ） 8（4-11） 8（5-10） 6（5-6） 4.51 0.10 T staging - Median T1a T1a T1b 11.12 ＜0.01 Nuclear classification 2（2-2） 2（1.5-2） 2（2-3） 7.40 0.02 NSS : Nephron–sparing surgery AI : Artificial intelligence AI-R.E.N.L. : AI–calculated R.E.N.L. WIT : Warm ischemia time IQR : Interquartile range Table 4 Risk stratification of perioperative outcomes for NSS by AI-CSA AI-CSA （N=100） Low score group（＜20cm²） High score group（≥20cm²） Mann-Whitney U test-U value P N（%） 66（66%） 34（34%） WIT ( min ) - Median（ IQR ） 25（20-28） 30（26-30） 516.5 ＜0.01 Surgical duration ( min ) - Median （ IQR ） 142（124-175） 185（148-243） 647.5 ＜0.01 Intraoperative blood loss ( ml ) - Median（ IQR ） 30（20-50） 50（30-100） 680.0 ＜0.01 Serum creatinine change ( μmol/L ) - M（ IQR ） 7(0-17) 11（1-25） 720.5 0.16 Time to extubation - Median（ IQR ） 5（4-6） 5（4-6.25） 1025.0 0.47 Postoperative hospital stay - Median（ IQR ） 8（4.75-11） 6.5（5-8） 1250.0 0.35 T staging - Median T1a T1a 682.5 ＜0.01 Nuclear classification 2 2 915.0 0.77 NSS : Nephron–sparing surgery AI : Artificial intelligence CSA : Contact surface area AI–CSA : AI–calculated CSA WIT : Warm ischemia time IQR : Interquartile range Additional Declarations No competing interests reported. Cite Share Download PDF Status: Published Journal Publication published 29 Dec, 2025 Read the published version in Cancer Imaging → Version 1 posted Editorial decision: Revision requested 22 Oct, 2025 Reviews received at journal 18 Oct, 2025 Reviews received at journal 16 Oct, 2025 Reviews received at journal 12 Oct, 2025 Reviewers agreed at journal 12 Oct, 2025 Reviews received at journal 10 Oct, 2025 Reviewers agreed at journal 10 Oct, 2025 Reviewers agreed at journal 09 Oct, 2025 Reviewers agreed at journal 03 Oct, 2025 Reviewers invited by journal 02 Oct, 2025 Editor assigned by journal 25 Sep, 2025 Submission checks completed at journal 24 Sep, 2025 First submitted to journal 22 Sep, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7676212","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":528426158,"identity":"2ce63f86-501e-4e78-911e-40222138c60a","order_by":0,"name":"Shengfa Lin","email":"","orcid":"","institution":"Shengli Clinical College of Fujian Medical University, Fujian Provincial Hospital, Fuzhou University Affiliated Provincial Hospital","correspondingAuthor":false,"prefix":"","firstName":"Shengfa","middleName":"","lastName":"Lin","suffix":""},{"id":528426159,"identity":"dbe02378-ec56-40ed-9a8a-ec790e104fa3","order_by":1,"name":"Liqing Su","email":"","orcid":"","institution":"Clinical Oncology School of Fujian Medical University, Fujian Cancer Hospital","correspondingAuthor":false,"prefix":"","firstName":"Liqing","middleName":"","lastName":"Su","suffix":""},{"id":528426160,"identity":"667ea75c-c107-4f6a-bc6c-18c5e6af390a","order_by":2,"name":"Shu Chen","email":"","orcid":"","institution":"Fuzhou University","correspondingAuthor":false,"prefix":"","firstName":"Shu","middleName":"","lastName":"Chen","suffix":""},{"id":528426161,"identity":"b2574e16-3dfb-4719-95b5-fc169c631488","order_by":3,"name":"Huijian Chen","email":"","orcid":"","institution":"Shengli Clinical College of Fujian Medical University, Fujian Provincial Hospital, Fuzhou University Affiliated Provincial Hospital","correspondingAuthor":false,"prefix":"","firstName":"Huijian","middleName":"","lastName":"Chen","suffix":""},{"id":528426162,"identity":"901f7c8d-e9a9-4a1a-a3c6-b15fe7015668","order_by":4,"name":"Yuying Lin","email":"","orcid":"","institution":"Shengli Clinical College of Fujian Medical University, Fujian Provincial Hospital, Fuzhou University Affiliated Provincial Hospital","correspondingAuthor":false,"prefix":"","firstName":"Yuying","middleName":"","lastName":"Lin","suffix":""},{"id":528426163,"identity":"f3b64885-909b-4b6a-a68a-3ae78781eb77","order_by":5,"name":"Zijie Lin","email":"","orcid":"","institution":"Shengli Clinical College of Fujian Medical University, Fujian Provincial Hospital, Fuzhou University Affiliated Provincial Hospital","correspondingAuthor":false,"prefix":"","firstName":"Zijie","middleName":"","lastName":"Lin","suffix":""},{"id":528426164,"identity":"f38ce5b0-ddd1-466e-8b6b-94e2e459c39f","order_by":6,"name":"Yinfeng Xia","email":"","orcid":"","institution":"Shengli Clinical College of Fujian Medical University, Fujian Provincial Hospital, Fuzhou University Affiliated Provincial Hospital","correspondingAuthor":false,"prefix":"","firstName":"Yinfeng","middleName":"","lastName":"Xia","suffix":""},{"id":528426165,"identity":"76d7d9b8-9167-4ff2-88f5-ba9528d50cde","order_by":7,"name":"Qianwen Li","email":"","orcid":"","institution":"Fuzhou University Affiliated Provincial Hospital, Provincial Key Laboratory of Medical Big Data Engineering","correspondingAuthor":false,"prefix":"","firstName":"Qianwen","middleName":"","lastName":"Li","suffix":""},{"id":528426166,"identity":"e682807b-8286-427c-8f73-ac1cba4e2e93","order_by":8,"name":"Zhuting Fang","email":"","orcid":"","institution":"Clinical Oncology School of Fujian Medical University, Fujian Cancer Hospital","correspondingAuthor":false,"prefix":"","firstName":"Zhuting","middleName":"","lastName":"Fang","suffix":""},{"id":528426167,"identity":"0d3383c6-eebe-47b5-b3b7-28580cdcace8","order_by":9,"name":"Mingping Ma","email":"","orcid":"","institution":"Shengli Clinical College of Fujian Medical University, Fujian Provincial Hospital, Fuzhou University Affiliated Provincial Hospital","correspondingAuthor":false,"prefix":"","firstName":"Mingping","middleName":"","lastName":"Ma","suffix":""},{"id":528426168,"identity":"c747c895-6bcb-446d-9f83-f311485340a5","order_by":10,"name":"Minxiong Hu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAyElEQVRIiWNgGAWjYBACefmHDQcS/7HJ2bc3EKnFsCG58cEHNj5jA54DxFpzIL3ZcAabXOIGiQQidTA2HGyT5uExMzaXfLzxBkONTTRBLeyMjUAtEmlylrPTii0YjqXlNhC0pZkRqMXgmDHD7RwzCcaGw4S1MBwDaUn4n9hw8wyxWs4wAr1/gC1xww0eIrUYzmBsfPCxgc1YsgfolwRi/CIvwf7gQGIDmxw/++GNNz7U2BDhMCRgQHTUIGkhVccoGAWjYBSMDAAAKA1ADmFwJVUAAAAASUVORK5CYII=","orcid":"","institution":"Shengli Clinical College of Fujian Medical University, Fuzhou University Affiliated Provincial Hospital","correspondingAuthor":true,"prefix":"","firstName":"Minxiong","middleName":"","lastName":"Hu","suffix":""}],"badges":[],"createdAt":"2025-09-22 13:08:48","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7676212/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7676212/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1186/s40644-025-00961-2","type":"published","date":"2025-12-29T15:56:51+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":93728627,"identity":"ce67950e-e9b0-4b73-9ee5-ded2f086968d","added_by":"auto","created_at":"2025-10-17 02:12:31","extension":"tif","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":197532,"visible":true,"origin":"","legend":"","description":"","filename":"Figure1.tif","url":"https://assets-eu.researchsquare.com/files/rs-7676212/v1/23a11cc0964d046863b1d022.tif"},{"id":93726657,"identity":"a525ea6e-36eb-44f2-b2a3-98eea63d7084","added_by":"auto","created_at":"2025-10-17 02:04:31","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":57509,"visible":true,"origin":"","legend":"","description":"","filename":"ManuscriptcancerimagingNEW.docx","url":"https://assets-eu.researchsquare.com/files/rs-7676212/v1/70484bd6192c69f0e6f528ea.docx"},{"id":93726656,"identity":"453e855e-6a12-4866-a2a5-54a0356669b1","added_by":"auto","created_at":"2025-10-17 02:04:30","extension":"docx","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":27514,"visible":true,"origin":"","legend":"","description":"","filename":"Tablecancerimaging.docx","url":"https://assets-eu.researchsquare.com/files/rs-7676212/v1/5ac3bd8a979a658b49f505ff.docx"},{"id":93726659,"identity":"a7d910df-c89e-4f4f-91db-3153da322b7a","added_by":"auto","created_at":"2025-10-17 02:04:31","extension":"tif","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":280614,"visible":true,"origin":"","legend":"","description":"","filename":"Figure2.tif","url":"https://assets-eu.researchsquare.com/files/rs-7676212/v1/6a840074d062ac42d876aae2.tif"},{"id":93726667,"identity":"0bcbcb61-40f8-49f3-9423-47bd0a67a797","added_by":"auto","created_at":"2025-10-17 02:04:31","extension":"tif","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":4121640,"visible":true,"origin":"","legend":"","description":"","filename":"Figure3.tif","url":"https://assets-eu.researchsquare.com/files/rs-7676212/v1/fd8ce82c9199c8e1aeb2bed7.tif"},{"id":93728633,"identity":"2ed7e6e9-ca56-4b26-8e51-33b531b75394","added_by":"auto","created_at":"2025-10-17 02:12:31","extension":"tif","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":5018816,"visible":true,"origin":"","legend":"","description":"","filename":"Figure4.tif","url":"https://assets-eu.researchsquare.com/files/rs-7676212/v1/472fed85d705a940fd83cbbb.tif"},{"id":93728632,"identity":"5f0cb00e-0a97-4a66-9cc2-14a96a82e833","added_by":"auto","created_at":"2025-10-17 02:12:31","extension":"tif","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2624696,"visible":true,"origin":"","legend":"","description":"","filename":"Figure5.tif","url":"https://assets-eu.researchsquare.com/files/rs-7676212/v1/a29a39665dd28dec289304f6.tif"},{"id":93728628,"identity":"53ce78ae-ba6e-4804-8294-dd884c01dceb","added_by":"auto","created_at":"2025-10-17 02:12:31","extension":"json","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":11436,"visible":true,"origin":"","legend":"","description":"","filename":"5678de1d1a91412a9d706ad9a50fd73f.json","url":"https://assets-eu.researchsquare.com/files/rs-7676212/v1/17cf1a20369dafff54ea6cfd.json"},{"id":93726675,"identity":"c772d091-226c-4007-928d-aba12a9045db","added_by":"auto","created_at":"2025-10-17 02:04:31","extension":"xml","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":140356,"visible":true,"origin":"","legend":"","description":"","filename":"5678de1d1a91412a9d706ad9a50fd73f1enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7676212/v1/b0d77f3e26d37eb4326f1a0f.xml"},{"id":93726677,"identity":"9c6a9290-e0db-42cd-ac39-0efa3df4da20","added_by":"auto","created_at":"2025-10-17 02:04:32","extension":"tif","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":197532,"visible":true,"origin":"","legend":"","description":"","filename":"Figure1.tif","url":"https://assets-eu.researchsquare.com/files/rs-7676212/v1/2c0e6cf0455b4a652cdc2860.tif"},{"id":93726664,"identity":"7feab319-60e0-4bd4-972c-edbf0fdabcef","added_by":"auto","created_at":"2025-10-17 02:04:31","extension":"tif","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":280614,"visible":true,"origin":"","legend":"","description":"","filename":"Figure2.tif","url":"https://assets-eu.researchsquare.com/files/rs-7676212/v1/451c066798ed0d169e159901.tif"},{"id":93726679,"identity":"32111757-3765-43d9-98f9-790fdea430c6","added_by":"auto","created_at":"2025-10-17 02:04:32","extension":"tif","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":4121640,"visible":true,"origin":"","legend":"","description":"","filename":"Figure3.tif","url":"https://assets-eu.researchsquare.com/files/rs-7676212/v1/9cf50870c5b449ac77b2270e.tif"},{"id":93726669,"identity":"1a01d6ba-809a-4437-98fe-468f5d5aa3d7","added_by":"auto","created_at":"2025-10-17 02:04:31","extension":"tif","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":5018816,"visible":true,"origin":"","legend":"","description":"","filename":"Figure4.tif","url":"https://assets-eu.researchsquare.com/files/rs-7676212/v1/bdc0e67b8244aae31dc73d88.tif"},{"id":93730407,"identity":"083ca6b1-033f-49fe-90f9-740a96967fe7","added_by":"auto","created_at":"2025-10-17 02:20:31","extension":"tif","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2624696,"visible":true,"origin":"","legend":"","description":"","filename":"Figure5.tif","url":"https://assets-eu.researchsquare.com/files/rs-7676212/v1/ca84253d18a047fffde87a11.tif"},{"id":93726674,"identity":"bca4121d-6378-4dee-af61-cc8ac570ac14","added_by":"auto","created_at":"2025-10-17 02:04:31","extension":"png","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":63651,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFigure1.png","url":"https://assets-eu.researchsquare.com/files/rs-7676212/v1/8e8773d800e9bdc99a9e8d54.png"},{"id":93726666,"identity":"ff5b9c4d-1cba-40bd-b56f-c8d01fe41c01","added_by":"auto","created_at":"2025-10-17 02:04:31","extension":"png","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":94498,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFigure2.png","url":"https://assets-eu.researchsquare.com/files/rs-7676212/v1/af606eee69f7cf12044896b9.png"},{"id":93728629,"identity":"d8741f83-0689-4243-8ae0-e4318af3fb2a","added_by":"auto","created_at":"2025-10-17 02:12:31","extension":"png","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":409969,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFigure3.png","url":"https://assets-eu.researchsquare.com/files/rs-7676212/v1/e7de8b946903eecb44653033.png"},{"id":93728634,"identity":"75d2aa5d-2595-47b8-a393-94d5bbca7de5","added_by":"auto","created_at":"2025-10-17 02:12:32","extension":"png","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":502359,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFigure4.png","url":"https://assets-eu.researchsquare.com/files/rs-7676212/v1/9bfbe8cb69efaf64774d5cf1.png"},{"id":93726672,"identity":"35c49b14-4c72-45ed-9612-af0c56ebd45e","added_by":"auto","created_at":"2025-10-17 02:04:31","extension":"png","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":310333,"visible":true,"origin":"","legend":"","description":"","filename":"OnlineFigure5.png","url":"https://assets-eu.researchsquare.com/files/rs-7676212/v1/fd9e57a8c3ff24a058eab0f7.png"},{"id":93728630,"identity":"94e50bdf-3fcd-4127-8161-5e01a27c4e2b","added_by":"auto","created_at":"2025-10-17 02:12:31","extension":"xml","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":140765,"visible":true,"origin":"","legend":"","description":"","filename":"5678de1d1a91412a9d706ad9a50fd73f1structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7676212/v1/c1630860df9c7b804b9836db.xml"},{"id":93726676,"identity":"0d15beba-a8c7-49ca-ba03-fb5f237ada59","added_by":"auto","created_at":"2025-10-17 02:04:31","extension":"html","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":153660,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7676212/v1/fa618b87724b0d3604ade317.html"},{"id":93726654,"identity":"ae8786c3-14cc-406c-b8be-bf83c200c5c8","added_by":"auto","created_at":"2025-10-17 02:04:30","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":324341,"visible":true,"origin":"","legend":"\u003cp\u003eData distribution flowchart\u003c/p\u003e","description":"","filename":"Figure1.png","url":"https://assets-eu.researchsquare.com/files/rs-7676212/v1/17190d7967cde93c394cdcfc.png"},{"id":93728626,"identity":"87a58542-661c-4398-904d-fb132ee59ba9","added_by":"auto","created_at":"2025-10-17 02:12:30","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":449343,"visible":true,"origin":"","legend":"\u003cp\u003eFlowchart of the technical route of this research\u003c/p\u003e","description":"","filename":"Figure2.png","url":"https://assets-eu.researchsquare.com/files/rs-7676212/v1/3c8822786807f2c17c4893ed.png"},{"id":93726661,"identity":"e7179985-a7e0-4852-af5f-48526913a9fe","added_by":"auto","created_at":"2025-10-17 02:04:31","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":9311365,"visible":true,"origin":"","legend":"\u003cp\u003eA 31-year-old male with incidentally detected left renal clear cell carcinoma (pathological stage T1a, grade 2). Panels \u003cstrong\u003ea\u003c/strong\u003e, \u003cstrong\u003eb\u003c/strong\u003e, and \u003cstrong\u003ec\u003c/strong\u003e show original axial, coronal, and sagittal CT arterial-phase images, respectively (red arrows: tumor). Panels \u003cstrong\u003ed\u003c/strong\u003e, \u003cstrong\u003ee\u003c/strong\u003e, and \u003cstrong\u003ef\u003c/strong\u003edepict corresponding axial, coronal, and sagittal views of AI-annotated kidney and tumor contours (red arrows: tumor). Panels \u003cstrong\u003eg\u003c/strong\u003e, \u003cstrong\u003eh\u003c/strong\u003e, and \u003cstrong\u003ei\u003c/strong\u003epresent 3D reconstructed segmentation labels for the kidney, tumor, and their combined overlay, respectively.\u003c/p\u003e","description":"","filename":"Figure3.png","url":"https://assets-eu.researchsquare.com/files/rs-7676212/v1/8a67ca465cb36498308bbb21.png"},{"id":93726668,"identity":"3007ffac-f78e-4bd3-9b9a-f383354df758","added_by":"auto","created_at":"2025-10-17 02:04:31","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":11577506,"visible":true,"origin":"","legend":"\u003cp\u003eA 70-year-old male with incidentally detected right renal clear cell carcinoma (pathological stage T1b, grade 3). Panels \u003cstrong\u003ea\u003c/strong\u003e, \u003cstrong\u003eb\u003c/strong\u003e, and \u003cstrong\u003ec\u003c/strong\u003e display original axial, coronal, and sagittal CT arterial-phase images, respectively (red arrows: tumor). Panels \u003cstrong\u003ed\u003c/strong\u003e, \u003cstrong\u003ee\u003c/strong\u003e, and \u003cstrong\u003ef\u003c/strong\u003e show corresponding axial, coronal, and sagittal views of the AI-annotated kidney and tumor contours (red arrows: tumor). Panels \u003cstrong\u003eg\u003c/strong\u003e, \u003cstrong\u003eh\u003c/strong\u003e, and \u003cstrong\u003ei\u003c/strong\u003epresent 3D reconstructed segmentation labels for the kidney, tumor, and their combined overlay, respectively.\u003c/p\u003e","description":"","filename":"Figure4.png","url":"https://assets-eu.researchsquare.com/files/rs-7676212/v1/a1daec5835eb2642ac9af6f8.png"},{"id":93726663,"identity":"c7340b56-5f43-4151-ba2c-1c87b20c47d9","added_by":"auto","created_at":"2025-10-17 02:04:31","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":4217527,"visible":true,"origin":"","legend":"\u003cp\u003eStatistical differences across risk stratification groups were evaluated via Dunn’s multiple comparisons test: (\u003cstrong\u003ea\u003c/strong\u003e) WIT, (\u003cstrong\u003eb\u003c/strong\u003e) surgical duration, (\u003cstrong\u003ec\u003c/strong\u003e) intraoperative blood loss, (\u003cstrong\u003ed\u003c/strong\u003e) serum creatinine change, (\u003cstrong\u003ee\u003c/strong\u003e) T - staging, (\u003cstrong\u003ef\u003c/strong\u003e) nuclear classification.\u003c/p\u003e","description":"","filename":"Figure5.png","url":"https://assets-eu.researchsquare.com/files/rs-7676212/v1/7d6da58acfddeebe23f41748.png"},{"id":99545160,"identity":"4a864b97-6440-4105-a6a3-89d7357d1292","added_by":"auto","created_at":"2026-01-05 15:59:18","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":25614063,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7676212/v1/42a0d2ec-801f-444d-8a27-2c52a514bfeb.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"CT–Based AI Score Predicts Perioperative Outcomes in Nephron–Sparing Surgery for Renal Cell Carcinoma","fulltext":[{"header":"1 Introduction","content":"\u003cp\u003eRenal cell carcinoma (RCC) accounts for 2%\u0026ndash;3% of adult malignant tumors, with a rising global incidence (average annual growth rate: 0.7%\u0026ndash;2% over the past decade) \u003csup\u003e[\u003cspan additionalcitationids=\"CR2 CR3\" citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]\u003c/sup\u003e. Advances in imaging techniques (ultrasound, CT/MRI) have significantly improved early RCC detection \u003csup\u003e[\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]\u003c/sup\u003e. Clinical guidelines designate nephron\u0026ndash;sparing surgery (NSS) as the preferred treatment for T1a RCC and a viable option for selecting T1b tumors \u003csup\u003e[\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]\u003c/sup\u003e. Tumor anatomical characteristics, such as size, location, relationship to the renal vasculature/collecting system, and contact surface area (CSA), are pivotal determinants of NSS outcomes. Widely used scoring systems, including the R.E.N.A.L. nephrometry score \u003csup\u003e[\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]\u003c/sup\u003e, PADUA classification \u003csup\u003e[\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]\u003c/sup\u003e, and renal tumor CSA quantify these features from CT/MRI to predict perioperative outcomes such as warm ischemia time (WIT), operative duration, blood loss, pathological stage, and nuclear grade \u003csup\u003e[\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]\u003c/sup\u003e. Among them, the R.E.N.A.L. score includes five parameters: maximum tumor diameter (R), exophytic ratio (E), nearest distance to the renal sinus (N), anterior‒posterior location (A), and relationship to the polar line (L). CSA represents the tumor contact surface area. The meta\u0026ndash;analyses confirmed a strong correlation between R.E.N.A.L. and PADUA (r\u0026thinsp;=\u0026thinsp;0.8607) and their efficacy in outcome prediction \u003csup\u003e[\u003cspan additionalcitationids=\"CR12 CR13 CR14 CR15\" citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]\u003c/sup\u003e. The CSA score has unique advantages because it can assess residual renal function after surgery \u003csup\u003e[\u003cspan additionalcitationids=\"CR18 CR19 CR20\" citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]\u003c/sup\u003e. All score systems rely on manual clinician\u0026ndash;based measurements and are plagued by inefficiency, subjectivity, poor reproducibility, suboptimal accuracy, and limited clinical utility. Simplified newer alternatives created by scholars have failed to address these limitations \u003csup\u003e[\u003cspan additionalcitationids=\"CR23 CR24 CR25 CR26 CR27\" citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e]\u003c/sup\u003e. Artificial intelligence (AI), which leverages automation, efficiency, and precision, has emerged as a transformative tool in RCC management, demonstrating promise in diagnosis, treatment planning, and prognosis \u003csup\u003e[\u003cspan additionalcitationids=\"CR30 CR31 CR32 CR33 CR34 CR35 CR36 CR37\" citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e]\u003c/sup\u003e. At present, the exploration of R.E.N.A.L. CSA scores generated by AI have been reported only individually \u003csup\u003e[\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e, \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e]\u003c/sup\u003e, and its efficacy needs to be further improved. However, the ability of the NSS to predict perioperative outcomes has not yet been reported.\u003c/p\u003e\u003cp\u003eThis study aimed to develop and validate a CT-based AI score framework for RCC to predict perioperative outcomes in NSS. Notably, the \"A\" parameter was excluded from R.E.N.A.L. nephrometry score because it yields only qualitative classifications (A, P, X) with no numerical score. By integrating R.E.N.L. With CSA parameters via deep learning (convolutional neural networks), the model seeks to establish an objective, efficient system to overcome the subjectivity and inefficiency of manual assessment. Drawing on large\u0026ndash;scale datasets from Fuzhou University Affiliated Provincial Hospital and Fujian Cancer Hospital and leveraging advanced medical image processing and AI expertise from Fuzhou University\u0026rsquo;s College of Physics and Information Engineering, this project endeavors to create a standardized, evidence\u0026ndash;based tool for preoperative risk stratification. The ultimate goal is to enhance clinical decision\u0026ndash;making by providing surgeons with objective, data\u0026ndash;driven insights into NSS feasibility and potential outcomes for RCC patients.\u003c/p\u003e"},{"header":"2 Methods","content":"\u003cp\u003e\u003cstrong\u003e2.1 Ethical A\u003c/strong\u003e\u003cstrong\u003epproval\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;and\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003eData Collection\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was approved by the Institutional Ethics Committee of Fuzhou University Affiliated Provincial Hospital and Fujian Cancer Hospital, with approval numbers K2021\u0026ndash;12\u0026ndash;018 and SQ2023\u0026ndash;107, respectively. Retrospectively, RCC patients who underwent NSS were collected from two centers: Fuzhou University Affiliated Provincial Hospital (n1 cohort: January 2016\u0026ndash;January 2025) and Fujian Cancer Hospital (n2 cohort: January 2023\u0026ndash;January 2025). The inclusion criteria were as follows: 1) postoperative pathologically confirmed RCC; 2) complete contrast\u0026ndash;enhanced arterial\u0026ndash;phase CT imaging of the kidney and tumor; and 3) comprehensive clinical records available for analysis. The exclusion criteria were as follows: 1) CT images containing artifacts that interfere with score or model development and 2) the presence of multiple renal tumors (in a single or bilateral kidney).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.2 Manual Score\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eImaging data were evaluated by a team of radiologists\u0026nbsp;and urologists\u0026nbsp;with 7\u0026ndash;12 years of clinical experience in RCC diagnosis. Disputed cases underwent collective decision\u0026ndash;making by the entire team. The R.E.N.A nephrometry score followed the classic algorithm developed by Kutikov et al \u003csup\u003e[8]\u003c/sup\u003e, but this protocol included only five parameters (R, E, N, L, and R.E.N.L.) while excluding the A parameter, as the A parameter yields only qualitative classifications (A, P, X) without numerical scores. The manual CSA was calculated via a spherical crown\u0026ndash;like surface area formula:\u0026nbsp;\u003cimg width=\"56\" height=\"19\" src=\"https://myfiles.space/user_files/58895_8739fc6c57c1c19a/58895_custom_files/img1760602941.gif\" alt=\"image\"\u003e,\u0026nbsp;where R represents\u0026nbsp;the\u0026nbsp;tumor diameter and h denotes\u0026nbsp;the\u0026nbsp;depth\u0026nbsp;of the tumor\u0026nbsp;within the renal parenchyma.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.3\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003eDeveloping\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;and\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003eValidating\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;AI Score Models\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.3.1 Datasets and Preprocessing\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study integrated two subsets (n1 and n2 cohorts), with all CT scans including complete arterial\u0026ndash;phase images at a resolution of 512\u0026times;512. For the n1 cohort, 90% of\u0026nbsp;the cases were randomly allocated to training, internal validation, and test sets at a 7:1:2 ratio, whereas the remaining 10% were combined with the n2 cohort to form a mixed validation set. The data used to establish the models were manually annotated by radiologists and urologists with 7\u0026ndash;12 years of clinical experience via Mimics Research 21.0 software. Raw DICOM files were converted to NIfTI format via Python scripts (using the dicom2nifti library), with voxel sizes standardized and background noise cropped to ensure uniform preprocessing.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.3.2 Development and Validation of Automated Kidney and Tumor Segmentation Models\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFigure 1\u0026nbsp;shows the data distribution flowchart. Randomly, 50 cases from the n1 cohort were manually annotated, and an additional 265 cases were annotated via semi\u0026ndash;supervised learning, yielding 315 cases as \u0026quot;ground truths\u0026quot; for deep learning. An internal validation set (45 cases) and a test set (90 cases) were separately reserved. The segmentation model was trained via the nnU\u0026ndash;Net framework. Preprocessed single\u0026ndash;channel CT images served as model inputs, generating binary segmentation masks for kidneys and tumors as outputs. Prior to training, the data underwent three preprocessing steps: 1) spatial normalization\u0026mdash;images were resampled to the median voxel size via trilinear interpolation, with labels resampled via nearest\u0026ndash;neighbor interpolation to preserve anatomical boundaries; 2) intensity standardization\u0026mdash;after adjusting the window width/level to 400 HU/30 HU, pixel intensities were normalized via Z\u0026ndash;score transformation to standardize brightness across scans; and 3) data augmentation\u0026mdash;random rotations (\u0026plusmn;15\u0026deg;), elastic deformations, and scaling (0.8\u0026ndash;1.2\u0026times;)\u0026mdash;were applied to enhance model generalization and reduce overfitting. The training was iterated for a total of 505 rounds, and overfitting was avoided through the early stop method (monitored by the internal validation set). Model performance for renal and tumor segmentation was evaluated via the Dice similarity coefficient (DSC, a metric quantifying the spatial overlap between the predicted and ground\u0026ndash;truth segmentation masks).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.3.3 Development and Validation of the\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003eR\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e\u003cstrong\u003eE\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e\u003cstrong\u003eN\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e\u003cstrong\u003eL\u003c/strong\u003e\u003cstrong\u003e.\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;Score Model\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBased on the segmentation results, the R.E.N.L. score model parameters were defined as follows: 1) R (maximum tumor diameter):\u0026nbsp;calculated via 3D connected component analysis and the minimum bounding box method to measure the longest tumor dimension; 2) E (exophytic ratio): the renal parenchymal boundary was reconstructed to quantify the proportion of tumor volume protruding outside the kidney, defined as the exophytic volume divided by total tumor volume; and 3) N (nearest distance to the renal sinus): the renal sinus was segmented using a HU threshold (-110 to 0), and the minimum Euclidean distance between the tumor and sinus region was computed. 4) L (relationship to the polar line): Renal poles were localized via morphological analysis, and the proportion of tumor voxels crossing the polar line (connecting the upper and lower renal poles) was determined. 5) R.E.N.L. total score: Sum of the R, E, N, and L parameters. Notably, the \u0026quot;A\u0026quot; parameter was excluded from the score.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.3.4 Development and Validation of\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003ethe CSA Score Model\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe process of the CSA score model mainly includes\u0026nbsp;the following steps: 1) Contact surface extraction: the marching cubes algorithm was applied to generate a triangular mesh representing the tumor\u0026ndash;renal parenchyma contact surface from segmentation masks; 2) area calculation: on the basis of the vertex coordinates of the triangular mesh and voxel physical dimensions, the area of each triangle was computed and summed to obtain the total contact surface area; and 3) unit conversion: the result was converted from mm\u0026sup2; to cm\u0026sup2; by dividing by 100.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.3.5 AI\u0026ndash;R.E.N.L. and AI\u0026ndash;CSA enable risk stratification of NSS\u003c/strong\u003e\u003cstrong\u003e-related\u0026nbsp;perioperative outcomes\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eR.E.N.L. score risk stratification: low complexity (4\u0026ndash;6 points), medium complexity (7\u0026ndash;9 points) and high complexity (10\u0026ndash;12 points). CSA score risk stratification\u0026nbsp;was as follows: low-CSA group (CSA value \u0026lt; 20 cm\u0026sup2;) and high-CSA group (CSA value \u0026ge;20 cm\u0026sup2;). The evaluated outcomes included WIT, surgical duration, intraoperative blood loss, serum creatinine change, time to extubation, postoperative hospital stay, pathological T staging, and nuclear grade. WIT was defined as the duration from renal artery clamping to reperfusion. Surgical duration was calculated as the time from anesthesia induction to anesthesia termination. Postoperative serum creatinine changes were measured as the difference between day 1 postoperative and preoperative creatinine levels. The nuclear classification is based on the WHO/ISUP standard (Fifth Edition).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e2.4 Statistical Analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eStatistical analyses were performed\u0026nbsp;via SPSS 27.0 and GraphPad Prism 9.5. Continuous variables are summarized as medians (interquartile ranges, IQRs), and categorical variables are summarized as frequencies (percentages). Data normality was tested via the Shapiro\u0026ndash;Wilk test. Agreement between automated and manual scores was evaluated via the weighted Kappa coefficient. For risk stratification of NSS perioperative outcomes, the Kruskal\u0026ndash;Wallis test and Mann\u0026ndash;Whitney U test were used for group comparisons, with post hoc Dunn\u0026rsquo;s multiple comparisons test used to identify pairwise differences. Statistical significance was set at \u003cem\u003eP\u003c/em\u003e \u0026lt; 0.05.\u003c/p\u003e"},{"header":"3 Results","content":"\u003cp\u003e\u003cstrong\u003e3.1 Cohort C\u003c/strong\u003e\u003cstrong\u003eharacteristics\u0026nbsp;and Flowchart\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eOn the basis of\u0026nbsp;the inclusion and exclusion criteria, a total of N = 550\u0026nbsp;patients\u0026nbsp;were ultimately included in this study, consisting of 500\u0026nbsp;patients\u0026nbsp;from Fuzhou University Affiliated Provincial Hospital (n1 cohort) and 50\u0026nbsp;patients\u0026nbsp;from Fujian Cancer Hospital (n2 cohort). Figure 1\u0026nbsp;shows\u0026nbsp;the data distribution flowchart. Figure 2\u0026nbsp;shows\u0026nbsp;the flowchart of the\u0026nbsp;technique\u0026nbsp;of this research. This real\u0026ndash;world dataset included a cohort with a median age of 56 (46\u0026ndash;66) years, comprising 341 males (62.0%) and 209 females (38%). The majority of cases (471, 85.64%) were detected incidentally during routine physical examinations. Hypertension and diabetes were present in 199 (36.18%) and 78 (18.18%) patients, respectively.\u0026nbsp;The median\u0026nbsp;values for preoperative serum creatinine, postoperative day 1 serum creatinine, and creatinine\u0026nbsp;changes\u0026nbsp;were 73 (60\u0026ndash;82) \u0026mu;mol/L, 80 (66\u0026ndash;97) \u0026mu;mol/L, and 9 (1\u0026ndash;19) \u0026mu;mol/L, respectively. Preoperative prophylactic antibiotic use was recorded in 268\u0026nbsp;patients\u0026nbsp;(48.73%). Pathological analysis revealed clear cell carcinoma as the predominant histotype (449 cases, 81.64%), with most tumors staged as T1a (428 cases, 77.82%)\u0026nbsp;or\u0026nbsp;nuclear grade 1/2 (82.55%).\u0026nbsp;The remaining\u0026nbsp;demographic and clinical characteristics are detailed in Table 1.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.2 Automated Kidney and Tumor Segmentation Models\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe kidney and tumor segmentation models achieved\u0026nbsp;DSCs of 0.95 and 0.80, respectively. Figures 3 and 4 depict schematic visualizations of the segmentation results for two representative cases. Figure 3 illustrates a low\u0026ndash;complexity case: both manual\u0026ndash;calculated and AI\u0026ndash;calculated R.E.N.L. (AI\u0026ndash;R.E.N.L.) scores were \u003cimg width=\"122\" height=\"19\" src=\"https://myfiles.space/user_files/58895_8739fc6c57c1c19a/58895_custom_files/img1760603094.gif\" alt=\"image\"\u003e, with manually\u0026ndash;calculated CSA (M\u0026ndash;CSA)\u0026nbsp;of 14.83 cm\u0026sup2; and an AI\u0026ndash;calculated CSA (AI\u0026ndash;CSA) of 14.26 cm\u0026sup2;. Figure 4 shows a high\u0026ndash;complexity case (irregular morphology with extensive necrosis and liquefaction): the manually calculated R.E.N.L. score (M\u0026ndash;R.E.N.L.) was \u003cimg width=\"131\" height=\"19\" src=\"https://myfiles.space/user_files/58895_8739fc6c57c1c19a/58895_custom_files/img1760603094.gif\" alt=\"image\"\u003e,\u0026nbsp;whereas the AI\u0026ndash;R.E.N.L. \u003cimg width=\"131\" height=\"19\" src=\"https://myfiles.space/user_files/58895_8739fc6c57c1c19a/58895_custom_files/img1760603095.gif\" alt=\"image\"\u003e;\u0026nbsp;the M\u0026ndash;CSA and AI\u0026ndash;CSA values were 67.86 cm\u0026sup2; and 59.07 cm\u0026sup2;, respectively. Both cases demonstrated the excellent segmentation performance of the AI model, particularly in capturing anatomical details across varying levels of tumor complexity.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.3 Agreement Validation\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003ebetween\u0026nbsp;AI\u0026ndash;calculated R.E.N.L./CSA Scores and Manual Scores\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAgreement for six parameters\u0026ndash;R, E, N, L, R.E.N.L. total score and\u0026nbsp;CSA were assessed via weighted kappa coefficients, and the detailed results are presented in Table 2. The median (IQR) of M\u0026ndash;CSA was 18.56 (10.30\u0026ndash;27.00) cm\u0026sup2;, whereas it was 14.17 (7.76\u0026ndash;21.10) cm\u0026sup2; for AI\u0026ndash;CSA.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.4 AI\u0026ndash;R.E.N.L. and AI\u0026ndash;CSA enable risk stratification of NSS\u003c/strong\u003e\u003cstrong\u003e-related\u0026nbsp;perioperative outcomes\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eOn the basis of the previous findings, both AI\u0026ndash;R.E.N.L. and AI\u0026ndash;CSA scores showed substantial agreement with the manual assessments. Thus, AI\u0026ndash;R.E.N.L. was directly applied to stratify risks into low-, moderate-, and high-complexity groups, predicting intergroup differences in warm ischemia time, operative duration, intraoperative blood loss, postoperative serum creatinine change, time to extubation, postoperative hospital stay, pathological stage, and nuclear grade (detailed results in Table 3 and Figure 5). Similarly, AI\u0026ndash;CSA scores were directly used for risk stratification into low- and high-score groups, and the results are presented in Table 4.\u003c/p\u003e"},{"header":"4 Discussion","content":"\u003cp\u003eFirst, we successfully developed an automated CT\u0026ndash;based segmentation model for kidneys and tumors, which demonstrated favorable performance compared with manual segmentation. On the basis of this model, we subsequently constructed AI\u0026ndash;R.E.N.L. and the AI\u0026ndash;CSA score models, both of which strongly agreed with the manual score. Finally, risk stratification of NSS perioperative outcomes via these models revealed that AI scores effectively predict WIT, surgical duration, intraoperative blood loss, postoperative serum creatinine change, pathological stage, and nuclear grade.\u003c/p\u003e\u003cp\u003eTo enhance generalizability, this study integrated 550 real\u0026ndash;world, multicenter (two\u0026ndash;center) imaging datasets, with 450 cases used for developing/validating the segmentation and score models and 100 cases in a mixed dataset for perioperative outcome prediction. By combining manual annotation with semi\u0026ndash;supervised learning, we efficiently created high\u0026ndash;accuracy \"ground truth\" data for the segmentation model. The model achieved DSCs for kidney and tumor segmentation comparable to the KITS challenge average \u003csup\u003e[\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e]\u003c/sup\u003e, confirming its reliability.\u003c/p\u003e\u003cp\u003eFor the AI\u0026ndash;R.E.N.L. model, weighted Kappa coefficients for parameters R, E, N, L, and total R.E.N.L. scores were 0.82, 0.49, 0.63, 0.60, and 0.65, respectively, indicating almost perfect agreement for R and substantial agreement for N, L and total R.E.N.L. score and moderate agreement for E. These results exceed the clinical utility reported by Heller et al. \u003csup\u003e[\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e]\u003c/sup\u003e. The AI\u0026ndash;CSA model showed strong agreement with M\u0026ndash;CSA (Kappa\u0026thinsp;=\u0026thinsp;0.69), which was comparable to the results of Wood et al. \u003csup\u003e[\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e]\u003c/sup\u003e. Notably, AI\u0026ndash;CSA directly calculates the tumor\u0026ndash;kidney contact surface via pixel\u0026ndash;level analysis, avoiding the systematic bias of manual methods, which assume spherical tumor geometry and often overestimate M\u0026ndash;CSA (median M\u0026ndash;CSA: 18.56 [10.30\u0026ndash;27.00] cm\u0026sup2; vs. AI\u0026ndash;CSA: 14.17 [7.76\u0026ndash;21.10] cm\u0026sup2;).\u003c/p\u003e\u003cp\u003eUsing the AI\u0026ndash;R.E.N.L. and AI\u0026ndash;CSA score models for risk stratification of NSS perioperative outcomes, the AI\u0026ndash;R.E.N.L. score analysis revealed the following: ① the median WIT was 25 min (low complexity)\u0026thinsp;\u0026lt;\u0026thinsp;28 min (moderate complexity)\u0026thinsp;\u0026lt;\u0026thinsp;30 min (high complexity), with statistically significant differences (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05), and post\u0026ndash;hoc pairwise comparisons revealed that both the moderate- and high-complexity groups required longer ischemia than the low-complexity group, which is consistent with prior reports \u003csup\u003e[\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e, \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e]\u003c/sup\u003e, prompting clinicians to monitor ischemia duration; ② the median surgical duration was 141 min (low)\u0026thinsp;\u0026lt;\u0026thinsp;170 min (moderate)\u0026thinsp;\u0026lt;\u0026thinsp;200 min (high), which is statistically significant (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05), with longer durations in the moderate/high- vs. low-complexity groups, aligning with Wang et al.\u0026rsquo;s findings \u003csup\u003e[\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e]\u003c/sup\u003e; ③ the median intraoperative blood loss was 20 ml (low)\u0026thinsp;\u0026lt;\u0026thinsp;50 ml (moderate/high), which is statistically significant (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05), with greater loss in the moderate/high-complexity groups, which is consistent with Styopushkin et al.\u0026rsquo;s findings of complexity and bleeding risk \u003csup\u003e[\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e]\u003c/sup\u003e; and ④ the median serum creatinine changes were 8 \u0026micro;mol/L (low), 3 \u0026micro;mol/L (moderate), and 20 \u0026micro;mol/L (high), which is statistically significant (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05), with marked elevation in high vs. low/moderate complexity groups, supporting literature on complexity\u0026ndash;correlated creatinine increases \u003csup\u003e[\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e]\u003c/sup\u003e; ⑤ pathological T1a stage proportions were 90.48% (low), 76.32% (moderate), and 46.67% (high), which is statistically significant (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05), with fewer low\u0026ndash;stage cases in high complexity; ⑥ nuclear grade 1/2 proportions were 87.80% (low), 86.49% (moderate), and 57.14% (high), which is statistically significant (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05), with higher grades in high complexity, indicating increased invasiveness and recurrence risk, consistent with Sun et al.\u0026rsquo;s validation of R.E.N.A.L. predictive value for nuclear grade \u003csup\u003e[\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e]\u003c/sup\u003e; ⑦ no statistical differences in postoperative hospital stay or extubation time, possibly due to uniform laparoscopic NSS reducing variability. For AI\u0026ndash;CSA stratification: ① median WIT was 25 min (low CSA)\u0026thinsp;\u0026lt;\u0026thinsp;30 min (high CSA, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05), ② median operative duration 142 min (low)\u0026thinsp;\u0026lt;\u0026thinsp;185 min (high, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05), ③ median blood loss 30 ml (low)\u0026thinsp;\u0026lt;\u0026thinsp;50 ml (high, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05), ④ T1a stage proportion 88.89% (low) vs. 56.25% (high, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05), indicating higher T stages in high CSA; ⑤ no differences were observed in creatinine change, extubation time, hospital stay, or nuclear grade. These findings align with Leslie et al.\u0026rsquo;s report of positive correlations between CSA score and WIT, surgical duration, blood loss, creatinine change, and pathological stage \u003csup\u003e[\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e]\u003c/sup\u003e.\u003c/p\u003e\u003cp\u003eWhile this study achieved meaningful achievements, several limitations should be acknowledged. First, to avoid confounding segmentation and scoring of multiple tumors, the cohort excluded patients with multiple renal tumors, and future studies could include these populations to expand generalizability. Second, a small proportion of tumors smaller than 1 cm were under\u0026ndash;segmented, and a minor subset of renal cysts were misclassified as tumors and erroneously segmented; future studies should further improve segmentation precision. Third, the analysis was limited to nephron\u0026ndash;sparing surgery (NSS) data, and future research may explore perioperative outcomes in patients undergoing radical nephrectomy. Fourth, the study relied primarily on CT imaging and lacked investigations of magnetic resonance imaging (MRI) or other modalities, which could be incorporated to enhance multi\u0026ndash;imaging applicability.\u003c/p\u003e"},{"header":"5 Conclusions","content":"\u003cp\u003eIn summary, this study successfully developed a CT-based AI score model for RCC, which effectively performs risk stratification for NSS-related perioperative outcomes. The model demonstrates high efficiency and generalizability, featuring both robust automated kidney/tumor segmentation and precise prediction of key metrics, including WIT, surgical duration, intraoperative blood loss, postoperative serum creatinine change, pathological T stage, and nuclear grade. These capabilities have significantly increased the objectivity and accuracy of preoperative assessment, as well as the efficiency of clinical decision\u0026ndash;making. They have reduced the previously required several hours of manual labor to just approximately ten minutes, highlighting their potential for clinical application. With advancements in AI technology, future research may explore more complex scenarios, such as radical nephrectomy and multifocal renal cancer, to enable AI to comprehensively support the entire workflow of renal cancer diagnosis and treatment.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cp\u003eAI: Artificial intelligence\u003c/p\u003e\n\u003cp\u003eCSA: Contact surface area\u003c/p\u003e\n\u003cp\u003eNSS: Nephron\u0026ndash;sparing surgery\u003c/p\u003e\n\u003cp\u003eRCC: Renal cell carcinoma\u003c/p\u003e\n\u003cp\u003eWIT: Warm ischemia time\u003c/p\u003e\n\u003cp\u003eDSC: Dice similarity coefficient\u003c/p\u003e\n\u003cp\u003eIQR: Interquartile range\u003c/p\u003e\n\u003cp\u003eM\u0026ndash;CSA: Manually\u0026ndash;calculated CSA\u003c/p\u003e\n\u003cp\u003eAI\u0026ndash;CSA: AI\u0026ndash;calculated CSA\u003c/p\u003e\n\u003cp\u003eM\u0026ndash;R.E.N.L. : Manually\u0026ndash;calculated R.E.N.L.\u003c/p\u003e\n\u003cp\u003eAI\u0026ndash;R.E.N.L. : AI\u0026ndash;calculated R.E.N.L.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study has been approved by the Institutional Review Boards of Fuzhou University Affiliated Provincial Hospital and Fujian Cancer Hospital, with approval numbers K2021-12-018 (approval date: 2021/12/08) and SQ2023-107 (approval date: 2023/12/12), respectively.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and material\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis study was supported by the Guiding Project of Fujian Provincial Department of Science and Technology (No. 2022Y0052, Shengfa Lin), the Startup Fund for Scientific Research, Fujian Medical University (No. 2023QH1174, Liqing Su),\u0026nbsp;the Startup Fund for Scientific Research and Fujian Medical University (2021QH1283, Minxiong Hu),\u0026nbsp;the National Key R\u0026amp;D Program of China (No. 2023YFC2413500, Zhuting Fang), the Joint Funds for the Innovation of Science and Technology, Fujian Province (No. 2023Y9320, Zhuting Fang), the Fujian Province Natural Science Fund Project (No. 2024J011105, Zhuting Fang), and the Major Project of Fujian Provincial Health Commission (No. 2024ZD01004, Zhuting Fang).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthors\u0026apos; contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSFL was responsible for data collection, organization, analysis, and manuscript preparation; LQS was in charge of data collection and organization at Fujian Provincial Cancer Hospital; HJC, YYL, ZJL, and YFX were tasked with image scoring; SC was responsible for AI model development; ZTF, MPM, and MXH provided project guidance and support.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors are very grateful to Dr. Wang Xuefei and Dr. Wu Yangbiao for helping to collect and organize the data.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eWein AJ, Kavoussi LR, Partin AW, Peters CA (Eds). (2016). Campbell-Walsh urology (11th ed, 4-volume set). Elsevier.\u003c/li\u003e\n\u003cli\u003eMiller KD, Nogueira L, Devasia T et al (2022) Cancer treatment and survivorship statistics, 2022. CA-CANCER J CLIN 72:409\u0026ndash;436. doi: 10.3322/caac.21731\u003c/li\u003e\n\u003cli\u003eHan B, Zheng R, Zeng H et al (2024) Cancer incidence and mortality in China, 2022. J Natl Cancer Cent 4(1):47-53. doi: 10.1016/j.jncc.2024.01.006\u003c/li\u003e\n\u003cli\u003eFerlay J, Colombet M, Soerjomataram I et al (2018) Cancer incidence and mortality patterns in Europe: Estimates for 40 countries and 25 major cancers in 2018. EUR J CANCER 103:356-387. doi: 10.1016/j.ejca.2018.07.005\u003c/li\u003e\n\u003cli\u003eEscudier B, Porta C, Schmidinger M, et al (2019) Renal cell carcinoma: ESMO Clinical Practice Guidelines for diagnosis, treatment and follow-up\u0026dagger;. ANN ONCOL 30(5):706-720. doi: 10.1093/annonc/mdz056\u003c/li\u003e\n\u003cli\u003eHuang J, Zhang X (Eds).(2022). Guidelines for the diagnosis and treatment of urologic and andrologic diseases in China (2022 ed). Science Press.\u003c/li\u003e\n\u003cli\u003eEuropean Association of Urology (EAU). (2025). European Association of Urology Guidelines on Renal Cell Carcinoma: The 2025 Update. https://uroweb.org/guideline/renal-cell-carcinoma/\u003c/li\u003e\n\u003cli\u003eKutikov A, Uzzo RG (2009) The R.E.N.A.L. nephrometry score: a comprehensive standardized system for quantitating renal tumor size, location and depth. J UROLOGY 182(3): 844-53. doi: 10.1016/j.juro.2009.05.035\u003c/li\u003e\n\u003cli\u003eFicarra V, Novara G, Secco S et al (2009) Preoperative aspects and dimensions used for an anatomical (PADUA) classification of renal tumours in patients who are candidates for nephron-sparing surgery. EUR UROL 56(5): 786-93. doi: 10.1016/j.eururo.2009.07.040\u003c/li\u003e\n\u003cli\u003eVeccia A, Antonelli A, Uzzo RG et al (2019) Predictive Value of Nephrometry Scores in Nephron-sparing Surgery: A Systematic Review and Meta-analysis. Eur Urol Focus 6(3):490-504. doi: 10.1016/j.euf.2019.11.004\u003c/li\u003e\n\u003cli\u003eSchiavina, R, Novara, G, Borghesi, M, et al (2017) PADUA and R.E.N.A.L. nephrometry scores correlate with perioperative outcomes of robot-assisted partial nephrectomy: analysis of the Vattikuti Global Quality Initiative in Robotic Urologic Surgery (GQI-RUS) database. BJU INT 119(3):456-463. doi: 10.1111/bju.13628\u003c/li\u003e\n\u003cli\u003eBertolo R, Pozzi L (2019) From PADUA to R.E.N.A.L. Score and Vice Versa: Development and validation of a mathematical converter. J UROLOGY 201 (4): 674-675. doi: 10.1016/j.juro.2018.10.020\u003c/li\u003e\n\u003cli\u003eZhou HJ, Yan Y, Zhang JZ et al (2017) Role of R.E.N.A.L. Nephrometry Score in Laparoscopic Partial Nephrectomy. CHINESE MED J-PEKING 130(18):2170-2175. doi: 10.4103/0366-6999.213973\u003c/li\u003e\n\u003cli\u003eZinssius D, J\u0026uuml;nemann KP, Geiger F et al_(2022) Evaluation of the Padua and R.E.N.A.L. scores regarding their validity and implication in the perioperative management during partial nephrectomy. AKTUEL UROL 53(5):423-430. doi: 10.1055/a-0888-7234\u003c/li\u003e\n\u003cli\u003eTanaka H, Wang Y, Suk-Ouichai C et al (2018) Can We Predict Functional Outcomes after Partial Nephrectomy?. J Urol 201(4):693-701. doi:10.1016/j.juro.2018.09.055\u003c/li\u003e\n\u003cli\u003eKwon KJ, Ryu H, Kim M, et al (2021) Personalised three-dimensional printed transparent kidney model for robot-assisted partial nephrectomy in patients with complex renal tumours (R.E.N.A.L. nephrometry score \u0026ge;7): a prospective case-matched study. BJU Int 127(5):567-574. doi:10.1111/bju.15275\u003c/li\u003e\n\u003cli\u003eLeslie S, Gill IS, de Castro Abreu AL, et al (2014) Renal tumor contact surface area: a novel parameter for predicting complexity and outcomes of partial nephrectomy. Eur Urol 66(5):884-893. doi:10.1016/j.eururo.2014.03.010\u003c/li\u003e\n\u003cli\u003eTakagi T, Yoshida K, Kondo T, et al (2019) Association between tumor contact surface area and parenchymal volume change in robot-assisted laparoscopic partial nephrectomy carried out using the enucleation technique. Int J Urol 26(7):745-751. doi:10.1111/iju.14004\u003c/li\u003e\n\u003cli\u003eKahn AE, Shumate AM, Galler IJ, Ball CT, Thiel DD (2020) Contact surface area and its association with outcomes in robotic-assisted partial nephrectomy. Int J Med Robot 16(1):e2069. doi:10.1002/rcs.2069\u003c/li\u003e\n\u003cli\u003eLee CH, Ku JY, Park YJ, Seo WI, Ha HK (2019) The superiority of contact surface area as a predictor of renal cortical volume change after partial nephrectomy compared to RENAL, PADUA and C-index: an approach using computed tomography-based renal volumetry. Scand J Urol 53(2-3):129-133. doi:10.1080/21681805.2019.1614663\u003c/li\u003e\n\u003cli\u003eHsieh PF, Wang YD, Huang CP, et al (2016) A Mathematical Method to Calculate Tumor Contact Surface Area: An Effective Parameter to Predict Renal Function after Partial Nephrectomy. J Urol 196(1):33-40. doi:10.1016/j.juro.2016.01.092\u003c/li\u003e\n\u003cli\u003eBenadiba S, Verin AL, Pignot G, et al (2015) Are urologists and radiologists equally effective in determining the RENAL Nephrometry score?. Ann Surg Oncol 22(5):1618-1624. doi:10.1245/s10434-014-4152-1\u003c/li\u003e\n\u003cli\u003eNisen H, Ruutu M, Gl\u0026uuml;cker E, Visap\u0026auml;\u0026auml; H, Taari K (2014) Renal tumour invasion index as a novel anatomical classification predicting urological complications after partial nephrectomy. Scand J Urol 48(1):41-51. doi:10.3109/21681805.2013.797491\u003c/li\u003e\n\u003cli\u003eShin TY, Komninos C, Kim DW, et al (2015) A novel mathematical model to predict the severity of postoperative functional reduction before partial nephrectomy: the importance of calculating resected and ischemic volume. J Urol 193(2):423-429. doi:10.1016/j.juro.2014.07.084\u003c/li\u003e\n\u003cli\u003eSpaliviero M, Poon BY, Karlo CA, et al (2015) An Arterial Based Complexity (ABC) Scoring System to Assess the Morbidity Profile of Partial Nephrectomy. Eur Urol 69(1):72-79. doi:10.1016/j.eururo.2015.08.008\u003c/li\u003e\n\u003cli\u003eZhang R, Wu G, Huang J, et al (2017) Peritumoral Artery Scoring System: a Novel Scoring System to Predict Renal Function Outcome after Laparoscopic Partial Nephrectomy. Sci Rep 7(1):2853. doi:10.1038/s41598-017-03135-8\u003c/li\u003e\n\u003cli\u003eLi Y, Zhou L, Bian T, et al (2017) The zero ischemia index (ZII): a novel criterion for predicting complexity and outcomes of off-clamp partial nephrectomy. World J Urol 35(7):1095-1102. doi:10.1007/s00345-016-1975-3\u003c/li\u003e\n\u003cli\u003eFicarra V, Porpiglia F, Crestani A, et al (2019) The Simplified PADUA REnal (SPARE) nephrometry system: a novel classification of parenchymal renal tumours suitable for partial nephrectomy. BJU Int 124(4):621-628. doi:10.1111/bju.14772\u003c/li\u003e\n\u003cli\u003eCarlier M, Lareyre F, L\u0026ecirc; CD, et al (2021) A pilot study investigating the feasibility of using a fully automatic software to assess the RENAL and PADUA score. Prog Urol 32(8-9):558-566. doi:10.1016/j.purol.2022.04.001\u003c/li\u003e\n\u003cli\u003eKhene ZE, Bigot P, Doumerc N, et al (2023) Application of Machine Learning Models to Predict Recurrence After Surgical Resection of Nonmetastatic Renal Cell Carcinoma. Eur Urol Oncol 6(3):323-330. doi:10.1016/j.euo.2022.07.007\u003c/li\u003e\n\u003cli\u003eZhou T, Guan J, Feng B, et al (2023) Distinguishing common renal cell carcinomas from benign renal tumors based on machine learning: comparing various CT imaging phases, slices, tumor sizes, and ROI segmentation strategies. Eur Radiol 33(6):4323-4332. doi:10.1007/s00330-022-09384-0\u003c/li\u003e\n\u003cli\u003eCui E, Li Z, Ma C, et al (2020) Predicting the ISUP grade of clear cell renal cell carcinoma with multiparametric MR and multiphase CT radiomics. Eur Radiol 30(5):2912-2921. doi:10.1007/s00330-019-06601-1\u003c/li\u003e\n\u003cli\u003eNassiri N, Maas M, Cacciamani G, et al (2022) A Radiomic-based Machine Learning Algorithm to Reliably Differentiate Benign Renal Masses from Renal Cell Carcinoma. Eur Urol Focus 8(4):988-994. doi:10.1016/j.euf.2021.09.004\u003c/li\u003e\n\u003cli\u003eOhe C, Yoshida T, Amin MB, et al (2023) Deep learning-based predictions of clear and eosinophilic phenotypes in clear cell renal cell carcinoma. Hum Pathol 131:68-78. doi:10.1016/j.humpath.2022.11.004\u003c/li\u003e\n\u003cli\u003eChen S, Song D, Chen L, et al (2023) Artificial intelligence-based non-invasive tumor segmentation, grade stratification and prognosis prediction for clear-cell renal-cell carcinoma. Precis Clin Med 6(3):pbad019. Published 2023 Aug 17. doi:10.1093/pcmedi/pbad019\u003c/li\u003e\n\u003cli\u003ePeng Q, Shen Y, Fu K, et al (2021) Artificial intelligence prediction model for overall survival of clear cell renal cell carcinoma based on a 21-gene molecular prognostic score system. Aging (Albany NY) 13(5):7361-7381. doi:10.18632/aging.202594\u003c/li\u003e\n\u003cli\u003eBarkan E, Porta C, Rabinovici-Cohen S, Tibollo V, Quaglini S, Rizzo M (2023) Artificial intelligence-based prediction of overall survival in metastatic renal cell carcinoma. Front Oncol 13:1021684. doi:10.3389/fonc.2023.1021684\u003c/li\u003e\n\u003cli\u003eSchulz S, Woerl AC, Jungmann F, et al. Multimodal Deep Learning for Prognosis Prediction in Renal Cancer. Front Oncol 11:788740. doi:10.3389/fonc.2021.788740\u003c/li\u003e\n\u003cli\u003eHeller N, Tejpaul R, Isensee F, et al (2022) Computer-Generated R.E.N.A.L. Nephrometry Scores Yield Comparable Predictive Results to Those of Human-Expert Scores in Predicting Oncologic and Perioperative Outcomes. J Urol 207(5):1105-1115. doi:10.1097/JU.0000000000002390\u003c/li\u003e\n\u003cli\u003eWood AM, Abdallah N, Heller N, et al (2024) Fully Automated Versions of Clinically Validated Nephrometry Scores Demonstrate Superior Predictive Utility versus Human Scores. BJU Int 133(6):690-698. doi:10.1111/bju.16276\u003c/li\u003e\n\u003cli\u003eHeller, N, Isensee, F, Maier-Hein, KH, et al (2021) The state of the art in kidney and kidney tumor segmentation in contrast-enhanced CT imaging: Results of the KiTS19 challenge. MED IMAGE ANAL 67:101821. doi: 10.1016/j.media.2020.101821\u003c/li\u003e\n\u003cli\u003eHu C, Sun J, Zhang Z, et al (2021) Parallel comparison of R.E.N.A.L., PADUA, and C-index scoring systems in predicting outcomes after partial nephrectomy: A systematic review and meta-analysis. Cancer Med 10(15):5062-5077. doi:10.1002/cam4.4047\u003c/li\u003e\n\u003cli\u003eDubeux VT, Zanier JFC, Gabrich PN, Carrerette FB, Milfont JCA, Dami\u0026atilde;o R (2022) Practical evaluation of the R.E.N.A.L. score system in 150 laparoscopic nephron-sparing surgeries. Int Braz J Urol 48(1):110-119. doi:10.1590/S1677-5538.IBJU.2021.0424\u003c/li\u003e\n\u003cli\u003eWang Q, Qian B, Li Q, Ni Z, Li Y, Wang X (2015) Application of modified R.E.N.A.L. nephrometry score system in evaluating the retroperitoneal partial nephrectomy for T1 renal cell carcinoma. Int J Clin Exp Med 8(4):6482-6488.\u003c/li\u003e\n\u003cli\u003eStyopushkin S, Chaikovskyi V, Chernylovskyi V, Sokolenk R, Bondarenko D (2021) POSTOPERATIVE HEMORRHAGE AS A COMPLICATION OF A PARTIAL NEPHRECTOMY: FREQUENCY, FEATURES AND MANAGEMENT. Georgian Med News (313):12-20.\u003c/li\u003e\n\u003cli\u003eDahlkamp L, Haeuser L, Winnekendonk G, et al (2019) Interdisciplinary Comparison of PADUA and R.E.N.A.L. Scoring Systems for Prediction of Conversion to Nephrectomy in Patients with Renal Mass Scheduled for Nephron Sparing Surgery. J Urol 202(5):890-898. doi:10.1097/JU.0000000000000361\u003c/li\u003e\n\u003cli\u003eSun R, Zhao S, Jiang H, et al (2021) Imaging Tool for Predicting Renal Clear Cell Carcinoma Fuhrman Grade: Comparing R.E.N.A.L. Nephrometry Score and CT Texture Analysis. Biomed Res Int 2021:1821876. doi:10.1155/2021/1821876\u003c/li\u003e\n\u003cli\u003eLeslie S, Gill IS, de Castro Abreu AL, et al (2014) Renal tumor contact surface area: a novel parameter for predicting complexity and outcomes of partial nephrectomy. Eur Urol 66(5):884-893. doi:10.1016/j.eururo.2014.03.010\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003e\u003cstrong\u003eTable 1 Cohort Characteristics\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 256px;\"\u003e\n \u003cp\u003eDemographic and clinical characteristics\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 181px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 131px;\"\u003e\n \u003cp\u003eN=550\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" style=\"width: 256px;\"\u003e\n \u003cp\u003eGender - n（%）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 181px;\"\u003e\n \u003cp\u003eMale\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 131px;\"\u003e\n \u003cp\u003e341（62.0%）\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 181px;\"\u003e\n \u003cp\u003eFemale\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 131px;\"\u003e\n \u003cp\u003e209（38.0%）\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 256px;\"\u003e\n \u003cp\u003eAge（yrs.）- Median（IQR）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 181px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 131px;\"\u003e\n \u003cp\u003e56（46-66）\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 256px;\"\u003e\n \u003cp\u003eBMI（Kg/m\u0026sup2;）- Median（IQR）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 181px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 131px;\"\u003e\n \u003cp\u003e24.11（22.03-26.10）\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"4\" style=\"width: 256px;\"\u003e\n \u003cp\u003eSymptom - n（%）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 181px;\"\u003e\n \u003cp\u003eIncidentally detected\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd style=\"width: 131px;\"\u003e\n \u003cp\u003e471（85.64%）\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 181px;\"\u003e\n \u003cp\u003ePain\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 131px;\"\u003e\n \u003cp\u003e27（4.91%）\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 181px;\"\u003e\n \u003cp\u003eHematuria\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 131px;\"\u003e\n \u003cp\u003e34（6.18%）\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 181px;\"\u003e\n \u003cp\u003eOthers\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 131px;\"\u003e\n \u003cp\u003e18（3.27%）\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" style=\"width: 256px;\"\u003e\n \u003cp\u003eHypertension - n（%）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 181px;\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 131px;\"\u003e\n \u003cp\u003e199（36.18%）\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 181px;\"\u003e\n \u003cp\u003eNo\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 131px;\"\u003e\n \u003cp\u003e351（63.82%）\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" style=\"width: 256px;\"\u003e\n \u003cp\u003eDiabetes - n（%）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 181px;\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 131px;\"\u003e\n \u003cp\u003e78（18.18%）\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 181px;\"\u003e\n \u003cp\u003eNo\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 131px;\"\u003e\n \u003cp\u003e472（85.82%）\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 256px;\"\u003e\n \u003cp\u003ePreoperative serum creatinine（\u0026mu;mol/L）- Median（IQR）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 181px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 131px;\"\u003e\n \u003cp\u003e73（60-82）\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 256px;\"\u003e\n \u003cp\u003eDay 1 postoperative serum creatinine（\u0026mu;mol/L）- Median（IQR）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 181px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 131px;\"\u003e\n \u003cp\u003e80（66-97）\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 256px;\"\u003e\n \u003cp\u003eSerum creatinine change（\u0026mu;mol/L）\u0026nbsp;-\u0026nbsp;Median（IQR）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 181px;\"\u003e\n \u003cp\u003e-\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 131px;\"\u003e\n \u003cp\u003e9（1-19）\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" style=\"width: 256px;\"\u003e\n \u003cp\u003eProphylactic use of antibiotics before the operation - n（%）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 181px;\"\u003e\n \u003cp\u003eYes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 131px;\"\u003e\n \u003cp\u003e268（48.73%）\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 181px;\"\u003e\n \u003cp\u003eNo\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 131px;\"\u003e\n \u003cp\u003e282（51.27%）\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"4\" style=\"width: 256px;\"\u003e\n \u003cp\u003ePathological types - n（%）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 181px;\"\u003e\n \u003cp\u003eClear cell carcinoma\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 131px;\"\u003e\n \u003cp\u003e449（81.64%）\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 181px;\"\u003e\n \u003cp\u003ePapillary cell carcinoma\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 131px;\"\u003e\n \u003cp\u003e17（3.09%）\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 181px;\"\u003e\n \u003cp\u003eChromophobe cell carcinoma\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 131px;\"\u003e\n \u003cp\u003e34（6.18%）\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 181px;\"\u003e\n \u003cp\u003eOthers\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 131px;\"\u003e\n \u003cp\u003e50（9.09%）\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" style=\"width: 256px;\"\u003e\n \u003cp\u003eT staging - n (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 181px;\"\u003e\n \u003cp\u003eT1a\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 131px;\"\u003e\n \u003cp\u003e428（77.82%）\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 181px;\"\u003e\n \u003cp\u003eStage T1b and above\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 131px;\"\u003e\n \u003cp\u003e122（22.18%）\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd rowspan=\"2\" style=\"width: 256px;\"\u003e\n \u003cp\u003enuclear classification\u0026nbsp;- n (%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 181px;\"\u003e\n \u003cp\u003e1/2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 131px;\"\u003e\n \u003cp\u003e454（82.55%）\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 181px;\"\u003e\n \u003cp\u003e3/4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 131px;\"\u003e\n \u003cp\u003e96（17.45%）\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eIQR: Interquartile range\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 2 Consistency Verification of AI-R.E.N.L. Score /AI-CSA Score and Manual Score\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003eManual\u0026nbsp;VS AI\u003c/p\u003e\n \u003cp\u003e（n=90）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 115px;\"\u003e\n \u003cp\u003eWeighted Kappa coefficients\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003eStandard error\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 76px;\"\u003e\n \u003cp\u003e95% CI\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 77px;\"\u003e\n \u003cp\u003e\u003cem\u003eP\u0026nbsp;\u003c/em\u003eValue\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eAgreement\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003eR\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 115px;\"\u003e\n \u003cp\u003e0.82\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e0.07\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 76px;\"\u003e\n \u003cp\u003e0.68-0.96\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 77px;\"\u003e\n \u003cp\u003e＜0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003eAlmost perfect\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003eE\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 115px;\"\u003e\n \u003cp\u003e0.49\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e0.06\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 76px;\"\u003e\n \u003cp\u003e\u0026nbsp;0.38-0.61\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 77px;\"\u003e\n \u003cp\u003e＜0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eModerate\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003eN\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 115px;\"\u003e\n \u003cp\u003e0.63\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e0.06\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 76px;\"\u003e\n \u003cp\u003e0.51-0.74\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 77px;\"\u003e\n \u003cp\u003e＜0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSubstantial\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003eL\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 115px;\"\u003e\n \u003cp\u003e0.60\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e0.06\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 76px;\"\u003e\n \u003cp\u003e0.48-0.71\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 77px;\"\u003e\n \u003cp\u003e＜0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSubstantial\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003eR.E.N.L.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 115px;\"\u003e\n \u003cp\u003e0.65\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e0.04\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 76px;\"\u003e\n \u003cp\u003e0.57-0.72\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 77px;\"\u003e\n \u003cp\u003e＜0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSubstantial\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 104px;\"\u003e\n \u003cp\u003eCSA\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 115px;\"\u003e\n \u003cp\u003e0.69\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 78px;\"\u003e\n \u003cp\u003e0.03\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 76px;\"\u003e\n \u003cp\u003e0.63-0.75\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 77px;\"\u003e\n \u003cp\u003e＜0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 118px;\"\u003e\n \u003cp\u003e\u003cstrong\u003eSubstantial\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eAI : Artificial intelligence \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eCSA :\u0026nbsp;Contact surface area\u003c/p\u003e\n\u003cp\u003eAI-R.E.N.L. : AI\u0026ndash;calculated R.E.N.L.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAI-CSA: AI\u0026ndash;calculated CSA\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 3 Risk stratification of perioperative outcomes for NSS by AI-R.E.N.L.\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"1\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 206px;\"\u003e\n \u003cp\u003eAI-R.E.N.L.\u003c/p\u003e\n \u003cp\u003e（N=100）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003elow complexity（4-6）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003emoderate \u0026nbsp;complexity（7-9）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003ehigh complexity（10-12）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003eKruskal-Wallis statistic\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 53px;\"\u003e\n \u003cp\u003e\u003cem\u003eP\u003c/em\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 206px;\"\u003e\n \u003cp\u003en（%）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003e46（46%）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003e39（39%）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e15（15%）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd colspan=\"2\" valign=\"top\" style=\"width: 112px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 206px;\"\u003e\n \u003cp\u003eWIT ( min ) - Median（\u0026nbsp;IQR\u0026nbsp;）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003e25（20-28）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003e28（25-30）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e30（28-30）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e16.63\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 53px;\"\u003e\n \u003cp\u003e＜0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 206px;\"\u003e\n \u003cp\u003eSurgical duration ( min )\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e- Median\u0026nbsp;（\u0026nbsp;IQR\u0026nbsp;）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003e141（109-175）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003e170（140-200）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e200（139-290）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e10.87\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 53px;\"\u003e\n \u003cp\u003e＜0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 206px;\"\u003e\n \u003cp\u003eIntraoperative blood loss (\u0026nbsp;ml )\u003c/p\u003e\n \u003cp\u003e- Median（\u0026nbsp;IQR\u0026nbsp;）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003e20（20-47）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003e50（20-100）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e50（50-100）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e18.30\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 53px;\"\u003e\n \u003cp\u003e＜0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 206px;\"\u003e\n \u003cp\u003ePostoperative serum creatinine change ( \u0026mu;mol/L ) - M（\u0026nbsp;IQR\u0026nbsp;）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003e8（4-18）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003e3（-2-13）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e20（12 - 33）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e13.86\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 53px;\"\u003e\n \u003cp\u003e＜0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 206px;\"\u003e\n \u003cp\u003eTime to extubation\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e- Median（\u0026nbsp;IQR\u0026nbsp;）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003e5（4-6）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003e6（4-7）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e5（4-6）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e2.57\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 53px;\"\u003e\n \u003cp\u003e0.28\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 206px;\"\u003e\n \u003cp\u003ePostoperative hospital stay\u003c/p\u003e\n \u003cp\u003e- Median（\u0026nbsp;IQR\u0026nbsp;）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003e8（4-11）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003e8（5-10）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e6（5-6）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e4.51\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 53px;\"\u003e\n \u003cp\u003e0.10\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 206px;\"\u003e\n \u003cp\u003eT staging - Median\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003eT1a\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003eT1a\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003eT1b\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e11.12\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 53px;\"\u003e\n \u003cp\u003e＜0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 206px;\"\u003e\n \u003cp\u003eNuclear classification\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003e2（2-2）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 83px;\"\u003e\n \u003cp\u003e2（1.5-2）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 84px;\"\u003e\n \u003cp\u003e2（2-3）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e7.40\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 53px;\"\u003e\n \u003cp\u003e0.02\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eNSS : Nephron\u0026ndash;sparing surgery\u003c/p\u003e\n\u003cp\u003eAI : Artificial intelligence \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eAI-R.E.N.L. : AI\u0026ndash;calculated R.E.N.L.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eWIT : Warm ischemia time\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIQR : Interquartile range\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTable 4 Risk stratification of perioperative outcomes for NSS by AI-CSA\u003c/strong\u003e\u003c/p\u003e\n\u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\" align=\"\" width=\"577\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003eAI-CSA\u003c/p\u003e\n \u003cp\u003e（N=100）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 121px;\"\u003e\n \u003cp\u003eLow score group（＜20cm\u0026sup2;）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 116px;\"\u003e\n \u003cp\u003eHigh score group（\u0026ge;20cm\u0026sup2;）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 107px;\"\u003e\n \u003cp\u003eMann-Whitney U test-U value\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003eP\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003eN（%）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 121px;\"\u003e\n \u003cp\u003e66（66%）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 116px;\"\u003e\n \u003cp\u003e34（34%）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 107px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e\u0026nbsp;\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003eWIT ( min ) - Median（\u0026nbsp;IQR\u0026nbsp;）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 121px;\"\u003e\n \u003cp\u003e25（20-28）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 116px;\"\u003e\n \u003cp\u003e30（26-30）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 107px;\"\u003e\n \u003cp\u003e516.5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e＜0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003eSurgical duration ( min )\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e- Median\u0026nbsp;（\u0026nbsp;IQR\u0026nbsp;）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 121px;\"\u003e\n \u003cp\u003e142（124-175）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 116px;\"\u003e\n \u003cp\u003e185（148-243）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 107px;\"\u003e\n \u003cp\u003e647.5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e＜0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003eIntraoperative blood loss ( ml ) - Median（\u0026nbsp;IQR\u0026nbsp;）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 121px;\"\u003e\n \u003cp\u003e30（20-50）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 116px;\"\u003e\n \u003cp\u003e50（30-100）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 107px;\"\u003e\n \u003cp\u003e680.0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e＜0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003eSerum creatinine change ( \u0026mu;mol/L ) - M（\u0026nbsp;IQR\u0026nbsp;）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 121px;\"\u003e\n \u003cp\u003e7(0-17)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 116px;\"\u003e\n \u003cp\u003e11（1-25）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 107px;\"\u003e\n \u003cp\u003e720.5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e0.16\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003eTime to extubation\u0026nbsp;\u003c/p\u003e\n \u003cp\u003e- Median（\u0026nbsp;IQR\u0026nbsp;）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 121px;\"\u003e\n \u003cp\u003e5（4-6）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 116px;\"\u003e\n \u003cp\u003e5（4-6.25）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 107px;\"\u003e\n \u003cp\u003e1025.0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e0.47\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003ePostoperative hospital stay\u003c/p\u003e\n \u003cp\u003e- Median（\u0026nbsp;IQR\u0026nbsp;）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 121px;\"\u003e\n \u003cp\u003e8（4.75-11）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 116px;\"\u003e\n \u003cp\u003e6.5（5-8）\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 107px;\"\u003e\n \u003cp\u003e1250.0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e0.35\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003eT staging - Median\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 121px;\"\u003e\n \u003cp\u003eT1a\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 116px;\"\u003e\n \u003cp\u003eT1a\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 107px;\"\u003e\n \u003cp\u003e682.5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e＜0.01\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd valign=\"top\" style=\"width: 174px;\"\u003e\n \u003cp\u003eNuclear classification\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 121px;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 116px;\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 107px;\"\u003e\n \u003cp\u003e915.0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd valign=\"top\" style=\"width: 59px;\"\u003e\n \u003cp\u003e0.77\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003eNSS : Nephron\u0026ndash;sparing surgery\u003c/p\u003e\n\u003cp\u003eAI : Artificial intelligence \u0026nbsp;\u003c/p\u003e\n\u003cp\u003eCSA : Contact surface area\u003c/p\u003e\n\u003cp\u003eAI\u0026ndash;CSA : AI\u0026ndash;calculated CSA\u003c/p\u003e\n\u003cp\u003eWIT : Warm ischemia time\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eIQR : Interquartile range\u0026nbsp;\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"cancer-imaging","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"caig","sideBox":"Learn more about [Cancer Imaging](https://cancerimagingjournal.biomedcentral.com/)","snPcode":"40644","submissionUrl":"https://submission.nature.com/new-submission/40644/3","title":"Cancer Imaging","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Renal cell carcinoma, Computed tomography, Artificial intelligence, Nephron-sparing surgery, R.E.N.A.L. nephrometry, contact surface area","lastPublishedDoi":"10.21203/rs.3.rs-7676212/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7676212/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground \u003c/strong\u003eTo develop and validate a CT–based artificial intelligence (AI) score model integrating the R.E.N.A.L. nephrometry and contact surface area (CSA) for efficient, accurate prediction of perioperative outcomes in renal cell carcinoma (RCC) patients undergoing nephron–sparing surgery (NSS), addressing the subjectivity and inefficiency of manual score.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods \u003c/strong\u003eRetrospective data from two RCC cohorts were analyzed. Ninety percent of the n1 cohort was randomly allocated to develop and validate AI–driven kidney/tumor segmentation models and derive AI–calculated R.E.N.L. (The “A” score was ignored) and AI–calculated CSA scores. The remaining 10% of Cohort n1, combined with Cohort n2, were used for risk stratification prediction. Manual image annotation/scoring was conducted by experienced radiologists and urologists. Interrater consistency was evaluated via weighted kappa coefficients; risk stratification was performed viaKruskal–Wallis tests and Mann–Whitney U tests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults \u003c/strong\u003eA total of 550 patients were included in this study (median age, 56 [IQR: 46–66] years; 341 males), with n1=500 and n2=50. Automatic segmentation achieved high accuracy (Dice similarity coefficients: kidney 0.95, tumor 0.80). The R, E, N, L, R.E.N.L., and CSA score models had good consistency compared with the manual score, and the kappa coefficients were 0.82, 0.49, 0.63, 0.60, 0.65, and 0.69, respectively (all \u003cem\u003eP\u003c/em\u003e \u0026lt; 0.01). Risk stratification by AI score significantly predicted warm ischemia time, surgical duration, intraoperative blood loss, serum creatinine changes, pathological T stage, and nuclear grade (all \u003cem\u003eP\u003c/em\u003e\u0026lt; 0.05).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusions \u003c/strong\u003eThis study establishes a CT–based AI framework that integrates R.E.N.L. and CSA metrics, enabling standardized, objective preoperative risk assessment for NSS in RCC.\u003c/p\u003e","manuscriptTitle":"CT–Based AI Score Predicts Perioperative Outcomes in Nephron–Sparing Surgery for Renal Cell Carcinoma","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-10-17 02:04:26","doi":"10.21203/rs.3.rs-7676212/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-10-22T19:25:19+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-10-18T05:37:52+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-10-16T06:46:08+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-10-12T09:26:30+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"306857870613731661078098863458090182917","date":"2025-10-12T09:25:27+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-10-10T15:13:20+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"100793980137562303396706303844435631303","date":"2025-10-10T09:31:24+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"286902648419386984625151446744890511536","date":"2025-10-09T17:30:41+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"319176723866770162220736117284958409254","date":"2025-10-03T05:06:05+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-10-02T21:10:44+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-09-25T05:11:36+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-09-25T02:23:40+00:00","index":"","fulltext":""},{"type":"submitted","content":"Cancer Imaging","date":"2025-09-22T10:13:48+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"cancer-imaging","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"caig","sideBox":"Learn more about [Cancer Imaging](https://cancerimagingjournal.biomedcentral.com/)","snPcode":"40644","submissionUrl":"https://submission.nature.com/new-submission/40644/3","title":"Cancer Imaging","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"f15713b1-4369-48d3-a53e-58933a8155f1","owner":[],"postedDate":"October 17th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2026-01-05T15:58:41+00:00","versionOfRecord":{"articleIdentity":"rs-7676212","link":"https://doi.org/10.1186/s40644-025-00961-2","journal":{"identity":"cancer-imaging","isVorOnly":false,"title":"Cancer Imaging"},"publishedOn":"2025-12-29 15:56:51","publishedOnDateReadable":"December 29th, 2025"},"versionCreatedAt":"2025-10-17 02:04:26","video":"","vorDoi":"10.1186/s40644-025-00961-2","vorDoiUrl":"https://doi.org/10.1186/s40644-025-00961-2","workflowStages":[]},"version":"v1","identity":"rs-7676212","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7676212","identity":"rs-7676212","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}