MRScore: Evaluating Radiology Report Generation with LLM-based Reward System

TL;DR

This paper introduces MRScore, a new LLM-based evaluation metric for radiology report generation that aligns more closely with human judgment than traditional metrics like BLEU.

Contribution

The paper presents a novel LLM-based framework for radiology report evaluation, including data generation and training methods to improve assessment accuracy.

Findings

MRScore correlates better with human judgments than traditional metrics.

MRScore outperforms existing metrics in model selection tasks.

The framework effectively leverages GPT for data augmentation and training.

Abstract

In recent years, automated radiology report generation has experienced significant growth. This paper introduces MRScore, an automatic evaluation metric tailored for radiology report generation by leveraging Large Language Models (LLMs). Conventional NLG (natural language generation) metrics like BLEU are inadequate for accurately assessing the generated radiology reports, as systematically demonstrated by our observations within this paper. To address this challenge, we collaborated with radiologists to develop a framework that guides LLMs for radiology report evaluation, ensuring alignment with human analysis. Our framework includes two key components: i) utilizing GPT to generate large amounts of training data, i.e., reports with different qualities, and ii) pairing GPT-generated reports as accepted and rejected samples and training LLMs to produce MRScore as the model reward. Our experiments demonstrate MRScore's higher correlation with human judgments and superior performance in model selection compared to traditional metrics. Our code and datasets will be available on GitHub.