RaTEScore: A Metric for Radiology Report Generation
This addresses the need for better evaluation metrics in radiology report generation, which is crucial for clinical applications, though it is incremental as it builds on existing metric approaches.
The paper tackles the problem of evaluating AI-generated radiology reports by introducing RaTEScore, a novel entity-aware metric that emphasizes medical entities and is robust to synonyms and negation. The result is a metric that aligns more closely with human preference than existing ones, as validated on public and new benchmarks.
This paper introduces a novel, entity-aware metric, termed as Radiological Report (Text) Evaluation (RaTEScore), to assess the quality of medical reports generated by AI models. RaTEScore emphasizes crucial medical entities such as diagnostic outcomes and anatomical details, and is robust against complex medical synonyms and sensitive to negation expressions. Technically, we developed a comprehensive medical NER dataset, RaTE-NER, and trained an NER model specifically for this purpose. This model enables the decomposition of complex radiological reports into constituent medical entities. The metric itself is derived by comparing the similarity of entity embeddings, obtained from a language model, based on their types and relevance to clinical significance. Our evaluations demonstrate that RaTEScore aligns more closely with human preference than existing metrics, validated both on established public benchmarks and our newly proposed RaTE-Eval benchmark.