Improving Factual Completeness and Consistency of Image-to-Text Radiology Report Generation

Yasuhide Miura, Yuhao Zhang, Emily Tsai, Curtis Langlotz, Dan Jurafsky
2021 Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies   unpublished
Neural image-to-text radiology report generation systems offer the potential to improve radiology reporting by reducing the repetitive process of report drafting and identifying possible medical errors. However, existing report generation systems, despite achieving high performances on natural language generation metrics such as CIDEr or BLEU, still suffer from incomplete and inconsistent generations. Here we introduce two new simple rewards to encourage the generation of factually complete and
more » ... consistent radiology reports: one that encourages the system to generate radiology domain entities consistent with the reference, and one that uses natural language inference to encourage these entities to be described in inferentially consistent ways. We combine these with the novel use of an existing semantic equivalence metric (BERTScore). We further propose a report generation system that optimizes these rewards via reinforcement learning. On two open radiology report datasets, our system substantially improved the F 1 score of a clinical information extraction performance by +22.1 (∆ + 63.9%). We further show via a human evaluation and a qualitative analysis that our system leads to generations that are more factually complete and consistent compared to the baselines. . 2020a. Asking and answering questions to evaluate the factual consistency of summaries. In . 2020b. Towards faithful neural table-to-text generation with content-matching constraints. In
doi:10.18653/v1/2021.naacl-main.416 fatcat:qmm7zzojzrezvjmkv6kumttkhm