arXivDaily arXiv每日学术速递 周一至周五更新
​通知:联邦节假日调休:哥伦布日 / 原住民日(Columbus Day / Indigenous Peoples' Day),今日(13日)无更新。
arXiv 2508.11628cs.CVcs.AI

ChatGPT-5 是否已准备好用于乳腺 X 线影像视觉问答?

Is ChatGPT-5 Ready for Mammogram VQA?

Department of Radiation Oncology, Winship Cancer Institute, Emory University School of Medicine(放射肿瘤科、Winship癌症研究所、埃默里大学医学院)

Qiang Li, Shansong Wang, Mingzhe Hu, Mojtaba Safari, Zachary Eidex, Xiaofeng Yang

AI总结:

本文系统评估 GPT-5 与 GPT-4o 在四个乳腺 X 线数据集上的 VQA 表现,发现 GPT-5 虽优于其他 GPT 版本但仍不及人类专家和专用模型,显示通用 LLM 在乳腺影像筛查中的潜力与不足。

摘要:

乳腺 X 线影像视觉问答(VQA)将图像解读与临床推理相结合,具有支持乳腺癌筛查的潜力。我们在四个公开乳腺影像数据集(EMBED、InBreast、CMMD、CBIS-DDSM)上系统评估了 GPT-5 系列和 GPT-4o 模型在 BI-RADS 评估、异常检测和恶性分类任务上的表现。GPT-5 始终是表现最好的模型,但仍落后于人类专家和领域专用的微调模型。在 EMBED 上,GPT-5 在 GPT 各版本中取得了密度(56.8%)、结构扭曲(52.5%)、肿块(64.5%)、钙化(63.5%)和恶性(52.8%)分类的最高分。在 InBreast 上,其 BI-RADS 准确率为 36.9%,异常检测为 45.9%,恶性分类为 35.0%。在 CMMD 上,GPT-5 的异常检测达到 32.3%,恶性准确率为 55.0%。在 CBIS-DDSM 上,其 BI-RADS 准确率为 69.3%,异常检测为 66.0%,恶性准确率为 58.2%。与人类专家估计相比,GPT-5 表现出较低的敏感性(63.5%)和特异性(52.3%)。尽管 GPT-5 在筛查任务中展现出有前景的能力,但在没有针对性领域适配和优化的情况下,其性能仍不足以用于高风险的临床影像应用。然而,从 GPT-4o 到 GPT-5 的巨大性能提升表明,通用大语言模型(LLMs)辅助乳腺 X 线影像 VQA 任务的潜力呈现出令人鼓舞的趋势。

英文摘要:

Mammogram visual question answering (VQA) integrates image interpretation with clinical reasoning and has potential to support breast cancer screening. We systematically evaluated the GPT-5 family and GPT-4o model on four public mammography datasets (EMBED, InBreast, CMMD, CBIS-DDSM) for BI-RADS assessment, abnormality detection, and malignancy classification tasks. GPT-5 consistently was the best performing model but lagged behind both human experts and domain-specific fine-tuned models. On EMBED, GPT-5 achieved the highest scores among GPT variants in density (56.8%), distortion (52.5%), mass (64.5%), calcification (63.5%), and malignancy (52.8%) classification. On InBreast, it attained 36.9% BI-RADS accuracy, 45.9% abnormality detection, and 35.0% malignancy classification. On CMMD, GPT-5 reached 32.3% abnormality detection and 55.0% malignancy accuracy. On CBIS-DDSM, it achieved 69.3% BI-RADS accuracy, 66.0% abnormality detection, and 58.2% malignancy accuracy. Compared with human expert estimations, GPT-5 exhibited lower sensitivity (63.5%) and specificity (52.3%). While GPT-5 exhibits promising capabilities for screening tasks, its performance remains insufficient for high-stakes clinical imaging applications without targeted domain adaptation and optimization. However, the tremendous improvements in performance from GPT-4o to GPT-5 show a promising trend in the potential for general large language models (LLMs) to assist with mammography VQA tasks.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑