胸外科专科评估集

GAPS胸外科专科评估数据集是一个基于临床指南和专家知识的AI评估基准，包含92个非小细胞肺癌（NSCLC）分期与治疗规划的真实临床案例。数据集采用多维度正负评分系统（A级加分，S级扣分），评估AI模型在术前评估、诊断、分期、治疗决策等方面的临床准确性。提供标准化Excel格式，涵盖多个主流AI模型（如GPT-4、Gemini、Claude）的回答及详细评分标准，支持AI临床决策支持系统的性能 benchmark、医学教育与质量评估。所有病例经多学科团队验证，确保高质量与临床相关性。The GAPS Thoracic Surgery Specialist Assessment Dataset is an AI evaluation benchmark based on clinical guidelines and expert knowledge, encompassing 92 real clinical cases of non-small cell lung cancer (NSCLC) staging and treatment planning. It adopts a multi-dimensional positive and negative scoring system (Grade A for bonus points, Grade S for deduction points) to assess the clinical accuracy of AI models in preoperative assessment, diagnosis, staging, treatment decision-making and other clinical scenarios. Provided in standardized Excel format, the dataset includes responses and detailed scoring criteria from multiple mainstream AI models such as GPT-4, Gemini and Claude, supporting performance benchmarking of AI clinical decision support systems, medical education and quality assessment. All cases have been validated by a multidisciplinary team to ensure high quality and clinical relevance.

搜集汇总

背景与挑战

背景概述

该数据集是GAPS医学AI评估数据集，专门用于评估AI模型在临床场景中的表现，包含92个肺癌专科临床案例，并采用基于医疗指南的多维评分系统。它覆盖胸外科的术前评估、诊断、分期和治疗规划等关键领域，提供标准化格式以支持AI临床决策和模型比较。