CMR-RD：面向可解释心脏磁共振诊断的长尾自适应视觉语言模型
CMR-RD: Long-Tailed Adaptive VLM for Explainable CMR Diagnosis
把在线多阶段强化学习与自适应采样（GPPO）用于视觉语言模型的后训练，让模型主动探索稀有类别，这套长尾偏置缓解思路可直接迁移到其他医学或长尾视觉-语言诊断任务。
原文依据 we propose Group Phase Policy Optimization (GPPO), which combines online multi-stage reinforcement learning (RL) with adaptive sampling
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论心脏磁共振（CMR）是评估心血管疾病的临床金标准，但其判读依赖专家经验，仍然具有挑战性，尤其是对罕见疾病的识别。现有自动化方法缺乏可解释的推理过程，限制了临床采用。尽管视觉语言模型（VLM）具备基本的视觉理解与文本生成能力，但它们在医学诊断中仍缺乏可验证的推理链，并且在长尾分布的少数类上表现不佳。为应对这些挑战，我们提出 CMR-RD，据我们所知这是首个用于 CMR 可解释诊断的 VLM，能够生成与影像证据对齐的显式诊断链。我们构建了一个反映真实临床分布的 CMR 数据集，包含五个疾病类别（其中两种为罕见病）以及正常对照。在此基础上，使用大规模医学视觉-文本数据将通用 VLM 对齐到医学与 CMR 语义，并采用冷启动训练增强其对医学概念与基础推理的理解。为提升对稀有样本的推理与性能，我们提出分组阶段策略优化（GPPO），它将在线多阶段强化学习（RL）与自适应采样相结合。GPPO 使模型能够主动探索稀有与表现不佳的类别，从而有效缓解长尾偏置。实验表明，与医学及通用 VLM 相比，CMR-RD 在准确率与推理链正确性上具有高度竞争力，并展现出更强的稀有类别识别能力。
英文原文摘要 1652 字符
Cardiac magnetic resonance (CMR) is the clinical gold standard for assessing cardiovascular diseases, but its interpretation relies on expert experience and remains challenging, particularly for identifying rare diseases. Existing automated methods lack interpretable reasoning processes, limiting clinical adoption. Although vision-language models (VLMs) possess basic visual understanding and text generation capabilities, they still lack verifiable reasoning chains in medical diagnosis and underperform on minority classes in long-tail distributions. To address these challenges, we propose CMR-RD, to our knowledge the first VLM for interpretable diagnosis in CMR, capable of generating explicit diagnostic chains aligned with imaging evidence. We construct a CMR dataset that reflects real-world clinical distributions, comprising five disease categories (including two rare conditions) plus normal controls. Building on this, the general-purpose VLM is aligned to medical and CMR semantics using large-scale medical vision-text data, and cold-start training is used to enhance its understanding of medical concepts and basic reasoning. To enhance reasoning and performance on rare samples, we propose Group Phase Policy Optimization (GPPO), which combines online multi-stage reinforcement learning (RL) with adaptive sampling. GPPO enables the model to proactively explore rare and underperforming classes, thereby effectively mitigating long-tail bias. Experiments demonstrate that CMR-RD achieves highly competitive accuracy and reasoning-chain correctness compared with medical and general VLMs, shows stronger recognition of rare categories.