Saved in:
Bibliographic Details
Main Authors: Wang, Zilong, Luo, Xufang, Jiang, Xinyang, Li, Dongsheng, Qiu, Lili
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2404.00998
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910393749733376
author Wang, Zilong
Luo, Xufang
Jiang, Xinyang
Li, Dongsheng
Qiu, Lili
author_facet Wang, Zilong
Luo, Xufang
Jiang, Xinyang
Li, Dongsheng
Qiu, Lili
contents Evaluating generated radiology reports is crucial for the development of radiology AI, but existing metrics fail to reflect the task's clinical requirements. This study proposes a novel evaluation framework using large language models (LLMs) to compare radiology reports for assessment. We compare the performance of various LLMs and demonstrate that, when using GPT-4, our proposed metric achieves evaluation consistency close to that of radiologists. Furthermore, to reduce costs and improve accessibility, making this method practical, we construct a dataset using LLM evaluation results and perform knowledge distillation to train a smaller model. The distilled model achieves evaluation capabilities comparable to GPT-4. Our framework and distilled model offer an accessible and efficient evaluation method for radiology report generation, facilitating the development of more clinically relevant models. The model will be further open-sourced and accessible.
format Preprint
id arxiv_https___arxiv_org_abs_2404_00998
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle LLM-RadJudge: Achieving Radiologist-Level Evaluation for X-Ray Report Generation
Wang, Zilong
Luo, Xufang
Jiang, Xinyang
Li, Dongsheng
Qiu, Lili
Computation and Language
Artificial Intelligence
Evaluating generated radiology reports is crucial for the development of radiology AI, but existing metrics fail to reflect the task's clinical requirements. This study proposes a novel evaluation framework using large language models (LLMs) to compare radiology reports for assessment. We compare the performance of various LLMs and demonstrate that, when using GPT-4, our proposed metric achieves evaluation consistency close to that of radiologists. Furthermore, to reduce costs and improve accessibility, making this method practical, we construct a dataset using LLM evaluation results and perform knowledge distillation to train a smaller model. The distilled model achieves evaluation capabilities comparable to GPT-4. Our framework and distilled model offer an accessible and efficient evaluation method for radiology report generation, facilitating the development of more clinically relevant models. The model will be further open-sourced and accessible.
title LLM-RadJudge: Achieving Radiologist-Level Evaluation for X-Ray Report Generation
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2404.00998