RSNA Large Language Model Benchmark Dataset for Chest Radiographs of Cardiothoracic Disease: Radiologist Evaluation and Validation Enhanced by AI Labels (REVEAL-CXR)
This work addresses the problem of benchmark scarcity for domain experts in radiology, providing a validated dataset to evaluate multimodal LLMs, though it is incremental as it builds on existing AI-assisted methods for data curation.
The authors tackled the need for high-quality benchmarks in developing clinically useful multimodal large language models for chest radiographs by creating a dataset of 200 chest radiographic studies with 12 benchmark labels, verified by three radiologists, and made publicly available. They developed an AI-assisted labeling procedure to enhance efficiency, with at least two radiologists fully agreeing on labels for 381 out of 1,000 sampled radiographs.
Multimodal large language models have demonstrated comparable performance to that of radiology trainees on multiple-choice board-style exams. However, to develop clinically useful multimodal LLM tools, high-quality benchmarks curated by domain experts are essential. To curate released and holdout datasets of 100 chest radiographic studies each and propose an artificial intelligence (AI)-assisted expert labeling procedure to allow radiologists to label studies more efficiently. A total of 13,735 deidentified chest radiographs and their corresponding reports from the MIDRC were used. GPT-4o extracted abnormal findings from the reports, which were then mapped to 12 benchmark labels with a locally hosted LLM (Phi-4-Reasoning). From these studies, 1,000 were sampled on the basis of the AI-suggested benchmark labels for expert review; the sampling algorithm ensured that the selected studies were clinically relevant and captured a range of difficulty levels. Seventeen chest radiologists participated, and they marked "Agree all", "Agree mostly" or "Disagree" to indicate their assessment of the correctness of the LLM suggested labels. Each chest radiograph was evaluated by three experts. Of these, at least two radiologists selected "Agree All" for 381 radiographs. From this set, 200 were selected, prioritizing those with less common or multiple finding labels, and divided into 100 released radiographs and 100 reserved as the holdout dataset. The holdout dataset is used exclusively by RSNA to independently evaluate different models. A benchmark of 200 chest radiographic studies with 12 benchmark labels was created and made publicly available https://imaging.rsna.org, with each chest radiograph verified by three radiologists. In addition, an AI-assisted labeling procedure was developed to help radiologists label at scale, minimize unnecessary omissions, and support a semicollaborative environment.