Reinforcement learning improves LLM accuracy and reasoning in disease classification from radiology reports

Accurate disease classification from radiology reports is essential for many applications. While supervised fine-tuning (SFT) of lightweight LLMs improves accuracy, it can degrade reasoning. We propose a two-stage approach: SFT on disease labels followed by Group Relative Policy Optimization (GRPO) to refine predictions by optimizing accuracy and format without reasoning supervision. Across three radiologist-annotated datasets, SFT outperformed baselines and GRPO further improved classification and enhanced reasoning recall and comprehensiveness. This work was supported by the National Institute of Biomedical Imaging and Bioengineering under grant number 75N920202D00021 (to YP, GS, and AF) and NSF CAREER Award under grant number 2145640 (to YP). The funder had no role in the study design; dat... [1279 chars]