MedRepBench: A Comprehensive Benchmark for Medical Report Interpretation

2025-08-21

AI Summary

The absence of cross-departmental, multi-format standardized benchmarks hinders objective evaluation of structured medical report interpretation. Method: We introduce MedRepBench—the first comprehensive benchmark comprising 1,900 real-world Chinese clinical reports—and propose a dual-protocol evaluation framework: (i) objective assessment via field-level recall, and (ii) automated subjective scoring (factuality, explainability, reasoning quality) powered by large language models (LLMs). We further design contrastive experiments leveraging high-fidelity OCR text and propose an objective reward function for structured extraction, optimized via Group Relative Policy Optimization (GRPO) to enhance mid-scale vision-language models. Contribution/Results: Our analysis reveals inherent limitations of OCR+LLM pipelines in layout awareness and latency. GRPO yields up to 6% absolute recall gain; ablation confirms the robustness and efficiency advantages of pure-vision pathways, advancing end-to-end clinical report understanding toward real-world deployment.

Technology Category

• Natural Language Processing: Language Grounding & Multi-modal NLP
• Computer Vision: Multi-modal Vision
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work

Abstract

Medical report interpretation plays a crucial role in healthcare, enabling both patient-facing explanations and effective information flow across clinical systems. While recent vision-language models (VLMs) and large language models (LLMs) have demonstrated general document understanding capabilities, there remains a lack of standardized benchmarks to assess structured interpretation quality in medical reports. We introduce MedRepBench, a comprehensive benchmark built from 1,900 de-identified real-world Chinese medical reports spanning diverse departments, patient demographics, and acquisition formats. The benchmark is designed primarily to evaluate end-to-end VLMs for structured medical report understanding. To enable controlled comparisons, we also include a text-only evaluation setting using high-quality OCR outputs combined with LLMs, allowing us to estimate the upper-bound performance when character recognition errors are minimized. Our evaluation framework supports two complementary protocols: (1) an objective evaluation measuring field-level recall of structured clinical items, and (2) an automated subjective evaluation using a powerful LLM as a scoring agent to assess factuality, interpretability, and reasoning quality. Based on the objective metric, we further design a reward function and apply Group Relative Policy Optimization (GRPO) to improve a mid-scale VLM, achieving up to 6% recall gain. We also observe that the OCR+LLM pipeline, despite strong performance, suffers from layout-blindness and latency issues, motivating further progress toward robust, fully vision-based report understanding.

Problem

• Lack of standardized benchmarks for medical report structured interpretation
• Evaluating vision-language models on real-world medical report understanding
• Assessing structured clinical item recall and reasoning quality

Innovation

• Comprehensive benchmark using 1900 real-world medical reports
• Objective field recall and subjective LLM scoring evaluation
• GRPO optimization achieving 6% recall gain in VLM