{"ID":2824717,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2512.23090","arxiv_id":"2512.23090","title":"Benchmark Success, Clinical Failure: When Reinforcement Learning Optimizes for Benchmarks, Not Patients","abstract":"Recent Reinforcement Learning (RL) advances for Large Language Models (LLMs) have improved reasoning tasks, yet their resource-constrained application to medical imaging remains underexplored. We introduce ChexReason, a vision-language model trained via R1-style methodology (SFT followed by GRPO) using only 2,000 SFT samples, 1,000 RL samples, and a single A100 GPU. Evaluations on CheXpert and NIH benchmarks reveal a fundamental tension: GRPO recovers in-distribution performance (23% improvement on CheXpert, macro-F1 = 0.346) but degrades cross-dataset transferability (19% drop on NIH). This mirrors high-resource models like NV-Reason-CXR-3B, suggesting the issue stems from the RL paradigm rather than scale. We identify a generalization paradox where the SFT checkpoint uniquely improves on NIH before optimization, indicating teacher-guided reasoning captures more institution-agnostic features. Furthermore, cross-model comparisons show structured reasoning scaffolds benefit general-purpose VLMs but offer minimal gain for medically pre-trained models. Consequently, curated supervised fine-tuning may outperform aggressive RL for clinical deployment requiring robustness across diverse populations.","short_abstract":"Recent Reinforcement Learning (RL) advances for Large Language Models (LLMs) have improved reasoning tasks, yet their resource-constrained application to medical imaging remains underexplored. We introduce ChexReason, a vision-language model trained via R1-style methodology (SFT followed by GRPO) using only 2,000 SFT s...","url_abs":"https://arxiv.org/abs/2512.23090","url_pdf":"https://arxiv.org/pdf/2512.23090v2","authors":"[\"Armin Berger\",\"Manuela Bergau\",\"Helen Schneider\",\"Saad Ahmad\",\"Tom Anglim Lagones\",\"Gianluca Brugnara\",\"Martha Foltyn-Dumitru\",\"Kai Schlamp\",\"Philipp Vollmuth\",\"Rafet Sifa\"]","published":"2025-12-28T21:57:42Z","proceeding":"cs.AI","tasks":"[\"cs.AI\",\"cs.LG\"]","methods":"[\"Reinforcement Learning\",\"Large Language Model\",\"Language Model\"]","has_code":false}