中华外科杂志最新文献

{"title":"[Application research and evolutionary history of risk prediction models for postoperative pancreatic fistula after pancreaticoduodenectomy].","authors":"B Q Li, T Zhao, H R Liu, S S Tong, L Li, C H Yuan","doi":"10.3760/cma.j.cn112139-20250803-00390","DOIUrl":"10.3760/cma.j.cn112139-20250803-00390","url":null,"abstract":"<p><p>Postoperative pancreatic fistula (POPF) is a severe complication following pancreaticoduodenectomy,among which clinically relevant postoperative pancreatic fistula has garnered substantial attention due to its high incidence (10% to 25%) and fatal consequences. This article reviews the evolutionary trajectory of risk prediction models for POPF after pancreaticoduodenectomy,progressing from single-center risk prediction models dominated by intraoperative factors, to weighted risk prediction models and their modified versions,and further to novel models integrating radiomics with machine learning. This evolutionary process reflects the tendency of the prediction system toward preoperativeness,simplification,and objectivity. Early models were limited by their dependence on intraoperative factors and insufficient external validation. Modified models enhanced preoperative applicability by streamlining variables and underwent adequate multicenter external validation;however,they remain constrained by the lack of effective preoperative methods for assessing pancreatic texture. In recent years,machine learning models have demonstrated superior predictive efficacy by leveraging multimodal data;however, they still rely on manual delineation,and challenges persist in terms of interpretability and clinical translation. Future studies should achieve in-depth integration of preoperative imaging and artificial intelligence,focus on multicenter validation,and develop fully automated prediction systems to provide a basis for individualized risk stratification and precise intervention more efficiently.</p>","PeriodicalId":60685,"journal":{"name":"中华外科杂志","volume":"64 7","pages":"829-833"},"PeriodicalIF":0.0,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148044718","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"[Ethical challenges and prospective strategies in orthopedic robotic-assisted surgery].","authors":"D Z Zhou, H Zhou, H J Chen, L Liang","doi":"10.3760/cma.j.cn112139-20251129-00554","DOIUrl":"10.3760/cma.j.cn112139-20251129-00554","url":null,"abstract":"<p><p>The ethical challenges of orthopedic robotic-assisted surgery encompass the ambiguity of the acting agent during surgery and uncertainties in patient benefits, the transfer of decision-making authority alongside dilemmas in intelligent model explainability, and blurred lines of accountability for surgical risks. To address these challenges, the review proposes prospective strategies from the three aspects of medical policy and governance, technological iteration and clinical integration, and psychological role adaptation of surgeons, and aims to provide theoretical insights for the standardized development and ethical governance of orthopedic robotic-assisted surgery.</p>","PeriodicalId":60685,"journal":{"name":"中华外科杂志","volume":"64 7","pages":"735-739"},"PeriodicalIF":0.0,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148044666","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"[Prognostic impact of lymphovascular invasion and efficacy of adjuvant therapy in patients with non-metastatic clear cell renal cell carcinoma: a single-center retrospective cohort study].","authors":"B Guo, X P Zou, J Li, L B Xiong, Y L Peng, W S Wei, S J Guo, P Dong, H Han, F J Zhou, Z H Zhou, Z L Zhang","doi":"10.3760/cma.j.cn112139-20260321-00121","DOIUrl":"10.3760/cma.j.cn112139-20260321-00121","url":null,"abstract":"&lt;p&gt;&lt;p&gt;&lt;b&gt;Objectives:&lt;/b&gt; To investigate the prognostic impact of lymphovascular invasion (LVI) on outcomes in patients with non-metastatic clear cell renal cell carcinoma (ccRCC) and to evaluate the efficacy of adjuvant therapy in LVI-positive patients. &lt;b&gt;Methods:&lt;/b&gt; This is a retrospective cohort study. Clinical and pathological data of 342 ccRCC patients who underwent surgical treatment at Sun Yat-sen University Cancer Center and Sun Yat-sen University Cancer Hospital Gansu Hospital between January 2018 and December 2024 were collected. There were 239 males and 103 females, with an age of (&lt;i&gt;M&lt;/i&gt; (IQR)) of 57 (15) years (range: 13 to 91 years). Patients were classified into LVI-positive and LVI-negative groups based on the postoperative pathological evidence of tumor cell invasion into lymphatic and(or) microvascular lumens. The comparison of data between groups was conducted using Mann-Whitney &lt;i&gt;U&lt;/i&gt; test&lt;i&gt;, χ&lt;/i&gt;&lt;sup&gt;2&lt;/sup&gt; test or Fisher's exact probability method. Kaplan-Meier survival analysis and Cox proportional hazards regression models were used to evaluate the impact of LVI on disease-free survival (DFS) and overall survival (OS). Three separate analyses were performed: full-cohort analysis, subgroup analysis after excluding patients who received adjuvant therapy, and efficacy analysis of adjuvant therapy within LVI-positive patients. &lt;b&gt;Results:&lt;/b&gt; Among the 342 patients, 128 (37.4%) were LVI-positive and 214 (62.6%) were LVI-negative. No statistically significant differences were observed between the two groups in age, sex, type of surgery, maximum tumor diameter, cT stage, or cN stage (all &lt;i&gt;P&lt;/i&gt;&gt;0.05). The proportion of minimally invasive surgery was lower in the LVI-positive group than in the LVI-negative group (66.4% (85/128)&lt;i&gt;vs.&lt;/i&gt; 77.6% (166/214), &lt;i&gt;χ&lt;/i&gt;²=4.556, &lt;i&gt;P&lt;/i&gt;=0.033). Regarding pathological features, the LVI-positive group had significantly higher rates of tumor necrosis (56.3%(72/128) &lt;i&gt;vs.&lt;/i&gt; 32.2%(69/214), &lt;i&gt;χ&lt;/i&gt;²=18.073, &lt;i&gt;P&lt;/i&gt;&lt;0.01), sarcomatoid differentiation (18.8%(24/128) &lt;i&gt;vs&lt;/i&gt;. 7.9%(17/214), &lt;i&gt;χ&lt;/i&gt;²=7.870, &lt;i&gt;P&lt;/i&gt;=0.005), and Fuhrman grade≥3 (63.3%(81/128) &lt;i&gt;vs.&lt;/i&gt; 40.2%(86/214), &lt;i&gt;χ&lt;/i&gt;²=18.256, &lt;i&gt;P&lt;/i&gt;&lt;0.01) than the LVI-negative group. In the full-cohort analysis, no statistically significant differences in DFS or OS were observed between the LVI-positive and LVI-negative groups (DFS: &lt;i&gt;P&lt;/i&gt;=0.127; OS: &lt;i&gt;P&lt;/i&gt;=0.627). Postoperative adjuvant therapy was administered to 41 LVI-positive and 44 LVI-negative patients. After excluding patients who received adjuvant therapy, LVI-positive patients demonstrated worse DFS than LVI-negative patients (&lt;i&gt;HR&lt;/i&gt;=1.885, &lt;i&gt;P&lt;/i&gt;=0.009). Multivariable Cox regression analysis showed that type of surgery, cN stage, Fuhrman grade, and tumor necrosis were independent prognostic factors for DFS (all &lt;i&gt;P&lt;/i&gt;&lt;0.05), while age, cN stage, and Fuhrman grade were independent prognostic factors for OS (all &lt;i&gt;P&lt;/i&gt;&lt;0.05). LVI was not an indepe","PeriodicalId":60685,"journal":{"name":"中华外科杂志","volume":"64 7","pages":"760-767"},"PeriodicalIF":0.0,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148044760","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"[Technology-driven management of early-onset scoliosis:new technologies and concepts].","authors":"X H Xue, W Z Li","doi":"10.3760/cma.j.cn112139-20251221-00600","DOIUrl":"10.3760/cma.j.cn112139-20251221-00600","url":null,"abstract":"<p><p>Early-onset scoliosis (EOS) refers to spinal deformity occurring in children younger than 10 years of age. The main goals of EOS treatment are not only to correct spinal curvature, but also to preserve spinal and thoracic growth potential. In recent years, the rapid development and clinical application of robotic technology and artificial intelligence (AI) have driven a paradigm shift in EOS management from experience-based practice toward intelligent and data-driven care. Emerging technologies, including robotic navigation, virtual and augmented reality (AR), AI-based imaging analysis, wearable devices and smart braces, digital twin modeling, and AI-driven integrated decision support systems, have been progressively introduced into different stages of EOS management. Together, these technologies contribute to a comprehensive framework encompassing preoperative planning, intraoperative navigation, and postoperative monitoring. Robotic navigation and AR-assisted systems have demonstrated improved accuracy and safety in pedicle screw insertion; AI algorithms have shown high concordance with conventional methods in automated Cobb angle measurement and scoliosis classification; wearable devices and smart braces enhance treatment adherence and enable continuous monitoring; and AI-driven decision support systems facilitate integrated and intelligent clinical management through multimodal data analysis. Nevertheless, current clinical evidence supporting the application of these technologies specifically in EOS remains limited, and challenges related to data security, cost-effectiveness, and ethical considerations persist. Looking forward, with continued technological integration and the accumulation of high-quality clinical evidence, EOS management is expected to evolve toward a closed-loop care model characterized by precise preoperative planning, accurate intraoperative execution, continuous postoperative surveillance, and long-term outcome prediction.</p>","PeriodicalId":60685,"journal":{"name":"中华外科杂志","volume":"64 7","pages":"708-714"},"PeriodicalIF":0.0,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148044722","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"[Liquid biopsy empowers precision medicine across the full course of bladder cancer].","authors":"W L Zhong, H R Zheng, T X Lin","doi":"10.3760/cma.j.cn112139-20260320-00118","DOIUrl":"10.3760/cma.j.cn112139-20260320-00118","url":null,"abstract":"<p><p>Bladder cancer management encompasses the entire clinical course, including screening, diagnosis, treatment decision-making, and recurrence monitoring. Although conventional approaches based on cystoscopy, imaging, and histopathology remain the foundation of clinical practice, they have limitations in noninvasive dynamic monitoring and molecular stratification. By analyzing tumor-derived cells and molecular information in urine or blood, liquid biopsy provides an important complementary approach for precision management of bladder cancer across the disease course. Owing to the direct contact between bladder tumors and urine, urine-based liquid biopsy is more suitable for screening, adjunctive diagnosis, and intravesical recurrence monitoring. In contrast, blood-based liquid biopsy better reflects systemic tumor biology and has potential value in perioperative adjuvant treatment decision-making and treatment response monitoring. These two approaches have distinct focuses and complementary roles across different disease stages and clinical scenarios. Nevertheless, the clinical translation of liquid biopsy still faces challenges, including insufficient sample standardization, heterogeneous detection systems, and relatively high costs. Future efforts should further define its complementary role within existing guideline-based pathways, optimize biomarker panels, standardize detection procedures, and promote its integration into precision management of bladder cancer through prospective studies.</p>","PeriodicalId":60685,"journal":{"name":"中华外科杂志","volume":"64 7","pages":"695-700"},"PeriodicalIF":0.0,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148044677","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"[The role of hemodynamic assessment in the treatment of lower extremity varicose veins based on the CHIVA strategy].","authors":"C Wang, W P Zou, L Y Zhou, Y F Zhu","doi":"10.3760/cma.j.cn112139-20251215-00591","DOIUrl":"10.3760/cma.j.cn112139-20251215-00591","url":null,"abstract":"<p><p>Hemodynamic assessment in patients with lower limb varicose veins should be performed with the patient standing, the limb examined in a non-weight-bearing position, and reflux elicited by appropriate provocative maneuvers. Duplex ultrasound is used to identify reflux and localize the anatomical reflux pathways, with attention to pelvic and nonsaphenous venous pathways. Quantitative parameters, including peak reflux velocity, reflux volume or volume flow, ambulatory venous pressure, venous filling index, and ejection fraction, help assess reflux burden, calf muscle pump function, and venous hypertension. The Cure Conservatrice et Hémodynamique de l'Insuffisance Veineuse en Ambulatoire (CHIVA) is a hemodynamic, vein sparing strategy for the outpatient treatment of lower limb varicose veins. It achieves hemodynamic correction by interrupting reflux entry points, preserving or remodeling re-entry points, and maintaining the saphenous trunk. Short-, mid-term, and long-term outcomes of CHIVA are not inferior to, and may in some settings be better than, those of conventional stripping or endovenous thermal ablation. Postoperative pain and ecchymosis are less pronounced, and the need for adjunctive sclerotherapy is lower. In selected patients with limited disease and only mild truncal change, ambulatory selective varices ablation under local anesthesia may be considered. Follow up and outcome assessment should move beyond closure alone. Imaging and functional endpoints should be evaluated together. A standardized postoperative surveillance strategy that integrates hemodynamic parameters with varicose veins scales may help reduce recurrence and improve quality of life.</p>","PeriodicalId":60685,"journal":{"name":"中华外科杂志","volume":"64 7","pages":"728-734"},"PeriodicalIF":0.0,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148044791","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"[The application of breast cancer organoids in drug screening and gene function research].","authors":"Y Q Zhang, S H Wu, J W Fang, L Q Zhang, B Xiao, L H Li","doi":"10.3760/cma.j.cn112139-20250707-00334","DOIUrl":"10.3760/cma.j.cn112139-20250707-00334","url":null,"abstract":"<p><p>Organoids refer to cell clusters with certain spatial structures formed by human adult stem cells or pluripotent stem cells through three-dimensional culture in vitro, which can be stably passaged. Organoids can simulate real organs in both structure and function, and to the greatest extent reproduce the structure and function of living organs or tissues. They have three-dimensional structural features, retain cellular heterogeneity, and can self-renew and self-organize. In the field of breast cancer research, breast cancer organoid models can faithfully reproduce the heterogeneity, tissue structure, and drug response characteristics of patients' tumors, providing a powerful platform for in-depth understanding of breast cancer biology and achieving individualized treatment. Currently, significant progress has been made in drug sensitivity screening and drug resistance mechanism research using breast cancer organoids, and they have shown unique advantages in elucidating the functions of key genes. This review will introduce organoid technology, especially the application of breast cancer organoids in drug screening and gene function research, and look forward to their application prospects in promoting the development of precision medicine for breast cancer.</p>","PeriodicalId":60685,"journal":{"name":"中华外科杂志","volume":"64 7","pages":"824-829"},"PeriodicalIF":0.0,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148044747","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"[Investigating prognostic factors for T1 stage high-grade bladder cancer and developing a multi-marker risk prediction model using immunohistochemistry].","authors":"N Zhang, H Liang, L Wang, T Yang, L Zhang, B Yang, J H Fan","doi":"10.3760/cma.j.cn112139-20260320-00119","DOIUrl":"10.3760/cma.j.cn112139-20260320-00119","url":null,"abstract":"<p><p><b>Objectives:</b> To identify prognostic factors for T1 high-grade bladder cancer (T1HG BCa), select pertinent biomarkers, and to construct an immunohistochemistry-based multi-biomarker risk prediction model. <b>Methods:</b> A retrospective case series study was performed involving 181 T1HG BCa patients who underwent transurethral resection of bladder tumor at the First Affiliated Hospital of Xi'an Jiaotong University between January 2014 and December 2018. The cohort comprised 136 males (75.1%) and 45 females (24.9%), with an age of (<i>M</i>(IQR)) 69(13) years (range: 36 to 88 years). Using simple random sampling, subjects were allocated to a training set (127 cases) and a validation set (54 cases) in a 7∶3 ratio. Differences between independent groups were assessed using chi-square or corrected chi-square tests. Immunohistochemical staining quantified the expression of nine biomarkers; least absolute shrinkage and selection operator (LASSO) regression identified key biomarkers to derive a risk score. The prognostic value of this score was evaluated via Kaplan-Meier survival analysis and receiver operating characteristic curve testing. Significant prognostic factors for T1HG BCa were identified through univariate and multivariate Cox regression analyses, leading to the construction of a Nomogram model. <b>Results:</b> During follow-up of 51(41) months (range:3 to 95 months), tumor recurrence occurred in 45 patients (24.9%), progression in 28 patients (15.5%), and death in 9 patients (5.0%). LASSO regression on the training set identified six significant biomarkers: CD44, Ki67, Her2, CK20, GATA3, and CK5/6. Patients were stratified into low-risk and high-risk groups based on the training set's median risk score (-1.416). Kaplan-Meier analysis within the training set revealed significantly higher rates of tumor recurrence and progression in the high-risk group compared to the low-risk group (all <i>P<</i>0.01). Validation set results confirmed these findings. Univariate and multivariate Cox regression models for tumor recurrence and progression, developed using the training set, demonstrated that bladder tumor count (<i>HR</i>=2.154, 95<i>%CI:</i> 1.021 to 4.503, <i>P=</i>0.043) and the risk score (<i>HR=</i>6.172, 95<i>%CI</i>: 3.250 to 11.719, <i>P<</i>0.01) were independent predictors of recurrence. The risk score alone (<i>HR=</i>3.975, 95<i>%CI</i>: 1.858 to 8.505, <i>P<</i>0.01) emerged as an independent predictor of progression. A Nomogram model integrating the risk score with clinical parameters achieved a C-index of 0.914 (95<i>%CI</i>: 0.845 to 0.984) for recurrence prediction and 0.811 (95<i>%CI</i>: 0.696 to 0.926) for progression prediction. <b>Conclusion:</b> By integrating the risk score with clinical characteristics of T1HG BCa patients, a robust risk prediction model capable of assessing the probability of tumor recurrence and progression is established, demonstrating significant prognostic predictive value.</p>","PeriodicalId":60685,"journal":{"name":"中华外科杂志","volume":"64 7","pages":"768-777"},"PeriodicalIF":0.0,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148044694","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"[The surgical strategy for all-inside arthroscopic procedures in chronic lateral ankle instability].","authors":"S M Feng","doi":"10.3760/cma.j.cn112139-20260118-00028","DOIUrl":"10.3760/cma.j.cn112139-20260118-00028","url":null,"abstract":"<p><p>With the continuous advancement of arthroscopic techniques and the accumulation of clinical experience among arthroscopic surgeons, all-inside arthroscopic procedures have been increasingly applied in the management of chronic lateral ankle instability. Existing evidence has demonstrated that, compared with traditional open surgery, all-inside arthroscopic techniques offer significant technical and clinical advantages, including improved clinical outcomes, reduced surgical trauma, and a shortened rehabilitation period. In 2025, the Evidence-Based Clinical Practice Guidelines for the all-inside arthroscopic techniques in the treatment of chronic lateral ankle instability provided more standardized and systematic guidance for their clinical application. These guidelines comprehensively addressed key issues such as surgical indications, portal selection, anchor insertion angles, suture configurations, management of talar osteochondral lesions, procedure selection, and postoperative rehabilitation, and offered specific evidence-based recommendations. This article interprets these guidelines in conjunction with the clinical experience of the authors' team in the application of all-inside arthroscopic techniques for chronic lateral ankle instability, with the aim of providing a reference for clinical practice and future research.</p>","PeriodicalId":60685,"journal":{"name":"中华外科杂志","volume":"64 7","pages":"715-721"},"PeriodicalIF":0.0,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148044758","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"[MRI-based deep learning model for preoperative prediction of urothelial carcinoma with variant histology of bladder: a retrospective, multicenter study].","authors":"L K Cai, X Yang, Q Cao, W Tian, X T Liu, B Liang, M H Jiang, G C Wang, Q Shao, H L Que, X P Jiang, J Lyu","doi":"10.3760/cma.j.cn112139-20260327-00132","DOIUrl":"10.3760/cma.j.cn112139-20260327-00132","url":null,"abstract":"&lt;p&gt;&lt;p&gt;&lt;b&gt;Objectives:&lt;/b&gt; To develop a deep learning model based on magnetic resonance imaging (MRI) for the preoperative prediction of urothelial carcinoma with variant histology (VUC), and to evaluate its predictive performance and prognostic stratification value. &lt;b&gt;Methods:&lt;/b&gt; This is a retrospective cohort study. Data from 1 221 patients with bladder cancer across seven centers of China between June 2013 and May 2025 were retrospectively analyzed. There were 1 032 male cases and 189 female cases, with an age (&lt;i&gt;M&lt;/i&gt;(IQR)) of 68.0 (15.0) years (range: 19 to 96 years). Data from Jiangsu Provincial People's Hospital (&lt;i&gt;n=&lt;/i&gt;959) were randomly assigned to a training set (&lt;i&gt;n=&lt;/i&gt;767) and a validation set (&lt;i&gt;n=&lt;/i&gt;192) in an 8∶2 ratio using stratified random sampling method, while data from the other six centers (&lt;i&gt;n=&lt;/i&gt;262) served as an external test set. Patients were categorized into pure urothelial carcinoma (PUC) group and VUC group. A deep learning prediction model for VUC (DeepVUC) was developed using a 3D ResNet50-based framework, integrating MRI T2WI and diffusion-weighted imaging (DWI) through a self-attention mechanism. The comparison of data between groups was conducted using Mann-Whitney &lt;i&gt;U, χ&lt;/i&gt;&lt;sup&gt;2&lt;/sup&gt; test or Fisher's exact probability method. Pairwise comparisons among multiple groups were corrected using the Bonferroni method. Diagnostic performance was evaluated using the area under the receiver operating characteristic curve (AUC), and the prognostic value was assessed through Kaplan-Meier analysis. &lt;b&gt;Results:&lt;/b&gt; Among the 1 221 patients with bladder cancer, 1 078 cases were PUC and 143 cases were VUC, with a proportion of VUC at 11.7%. Comparative analysis revealed that relative to the PUC group, the proportions of lymph node metastasis, high-grade urothelial carcinoma, carcinoma in situ, and lymphovascular invasion were significantly higher in the VUC group (all &lt;i&gt;P&lt;&lt;/i&gt;0.01). Bonferroni pairwise comparison showed that the proportion of VUC in pT3 stage patients was significantly higher than that in pT2 stage (&lt;i&gt;χ&lt;/i&gt;&lt;sup&gt;2&lt;/sup&gt;=14.30, &lt;i&gt;P&lt;&lt;/i&gt;0.01), pT1 stage (&lt;i&gt;χ&lt;/i&gt;&lt;sup&gt;2&lt;/sup&gt;=56.76, &lt;i&gt;P&lt;&lt;/i&gt;0.01) and pTa stage (&lt;i&gt;χ&lt;/i&gt;&lt;sup&gt;2&lt;/sup&gt;=174.02, &lt;i&gt;P&lt;&lt;/i&gt;0.01). The VUC group also exhibited significantly higher vesical imaging-reporting and data system(VI-RADS) scores (&lt;i&gt;χ&lt;/i&gt;&lt;sup&gt;2&lt;/sup&gt;=141.48, &lt;i&gt;P&lt;&lt;/i&gt;0.01). Specifically, the percentage of VUC in patients with VI-RADS scores of 1 to 5 was 3.5% (4/113),4.2% (22/524), 9.1% (25/275), 25.0% (37/148) and 34.2% (55/161), respectively. Squamous differentiation was the most common variant, accounting for 51.7% (74/143) of all VUC patients, with the highest frequency observed in patients with a VI-RADS score of 5 (69.1%, 38/55). The DeepVUC model achieved AUCs of 0.972 (95&lt;i&gt;%CI&lt;/i&gt;: 0.955 to 0.989), 0.940 (95&lt;i&gt;%CI&lt;/i&gt;: 0.902 to 0.977), and 0.925 (95&lt;i&gt;%CI&lt;/i&gt;: 0.889 to 0.961) in the training, validation, and external test sets, respectively. Accuracies wer","PeriodicalId":60685,"journal":{"name":"中华外科杂志","volume":"64 7","pages":"752-759"},"PeriodicalIF":0.0,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148044754","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}