丁国庆教授“AI分身”之尿路上皮癌多模态高质量数据集

该数据为浙江大学医学院附属邵逸夫医院在泌尿外科尿路上皮癌（包括膀胱癌、肾盂癌、输尿管癌）诊疗中产生，凝聚了丁国庆教授在该领域的核心知识与经验。数据集囊括了全景多模态数据（如诊断、手术、放化疗、影像检查记录、临床记录等）。数据形态包括病情描述与医学影像报告、结构化治疗信息与临床数据等，用于构建丁国庆教授尿路上皮癌AI分身。业务场景聚焦于血尿病因诊断、肿瘤早期发现、手术方式选择（保膀胱或根治性切除）以及术后灌注化疗/免疫治疗方案的制定。该数据集可支撑AI分身辅助进行肿瘤分期与风险分层、为保留器官的综合治疗提供决策支持，并优化术后随访策略，助力实现尿路上皮癌的规范化、微创化与个体化治疗。

This dataset was generated during the diagnosis and treatment of urothelial carcinoma (including bladder cancer, renal pelvic carcinoma, and ureteral carcinoma) in the Department of Urology at Sir Run Run Shaw Hospital Affiliated to Zhejiang University School of Medicine, and embodies the core knowledge and clinical experience of Professor Ding Guoqing in this field. It covers comprehensive multimodal data, such as records of diagnosis, surgery, chemoradiotherapy, imaging examinations, and clinical notes. The data formats include condition descriptions, medical imaging reports, structured treatment information, and clinical data, and is used to build Professor Ding Guoqing's urothelial carcinoma AI digital twin. The application scenarios focus on etiological diagnosis of hematuria, early tumor detection, selection of surgical approaches (bladder-sparing or radical resection), and formulation of postoperative perfusion chemotherapy/immunotherapy regimens. This dataset can support the AI digital twin in assisting with tumor staging and risk stratification, provide decision support for organ-preserving comprehensive treatment, optimize postoperative follow-up strategies, and facilitate the realization of standardized, minimally invasive, and individualized treatment for urothelial carcinoma.

搜集汇总

背景与挑战

背景概述

该数据集由浙江大学医学院附属邵逸夫医院提供，包含尿路上皮癌诊疗的全景多模态临床数据，用于构建AI分身以辅助肿瘤分期、治疗决策和术后随访。数据集覆盖诊断、手术、影像等核心环节，支持血尿病因分析、手术方案选择等关键临床场景。