发表机构
Johns Hopkins University; Data Science and AI Institute; Department of Electrical and Computer Engineering（约翰斯·霍普金斯大学; 数据科学与人工智能学院; 电气与计算机工程系）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
研究针对脑部MRI纵向报告缺失问题，提出首个纵向视觉-语言系统BrainDiff，通过三项分析揭示图像依赖度等特性，其性能优于现有模型且在跨医院队列中表现稳定。
AI 中文摘要
神经放射科医生很少孤立地解读脑部MRI，但自动脑部MRI报告生成几乎完全是为单次研究构建的。胸部X光和胸部CT已探索了时间分析，但据我们所知，针对脑部MRI的纵向报告（其间隔变化通常是细微且空间分布的）仍未解决。我们提出BrainDiff，这是首个用于脑部MRI的纵向视觉-语言系统。在相同患者对上，BrainDiff的性能优于前沿通用模型和单次研究神经影像模型。此外，BrainDiff在外部跨医院队列上保留了91%的内部RadGraph-XL实体+关系F1（rg_er）。除系统外，我们还完成了三项分析：第一，我们确定了两个独立的基础杠杆：带有先前报告丢弃的反事实目标，其测量的图像依赖度提高了约47%，以及分阶段课程。这些干预措施共同将图像依赖度从基线提高了2.5倍；第二，我们对先前报告可用性和图像身份进行了因子分解，分离出视觉贡献为+0.0387 rg_er，且当先前报告被隐藏时该贡献会增大；第三，对候选骨干网络的廉价变化可解码性测试显示，间隔变化的可解码性远弱于单次研究病理（AUROC分别为0.60和0.77）。代码可在该https URL公开获取。
英文摘要
Neuroradiologists rarely read a brain MRI in isolation, yet automated brain-MRI report generation has been built almost entirely for single studies. Temporal analysis has been explored on chest radiography and chest CT, but to our knowledge, longitudinal reporting for brain MRI, where interval change is often subtle and spatially distributed, remains unaddressed. We present BrainDiff, the first longitudinal vision-language system for brain MRI. BrainDiff outperforms both frontier general-purpose and single-study neuroimaging models on the same patient pairs. Moreover, BrainDiff retains 91% of internal RadGraph-XL entity+relation F1 (rg_er) on an external, cross-hospital cohort. Beyond the system, we contribute three analyses. First, we identify two independent grounding levers: a counterfactual objective with prior-report dropout, which increases measured image reliance by ~47%, and a staged curriculum. Together, these interventions raise image reliance 2.5-fold from the baseline. Second, we provide a factorial over prior-report availability and image identity, isolating a visual contribution of +0.0387 rg_er, which grows when the prior report is withheld. Third, a cheap change-decodability test for candidate backbones shows that interval change is decodable far more weakly than single-study pathology (0.60 vs. 0.77 AUROC). Code is publicly available at https://github.com/jhuldr/BrainDiff.
Comments13 pages, 3 figures
