Learning to Summarize Radiology Findings
This work addresses the time-consuming and error-prone task of summarizing radiology findings for radiologists and physicians, representing a first attempt in this direction.
The authors tackled the problem of automating the generation of radiology impressions, which summarize crucial findings, by proposing a neural sequence-to-sequence model that encodes study background information to guide decoding. Their model outperformed existing baselines on ROUGE metrics, and in a blind experiment, 67% of system summaries were rated at least as good as human-written ones by a radiologist.
The Impression section of a radiology report summarizes crucial radiology findings in natural language and plays a central role in communicating these findings to physicians. However, the process of generating impressions by summarizing findings is time-consuming for radiologists and prone to errors. We propose to automate the generation of radiology impressions with neural sequence-to-sequence learning. We further propose a customized neural model for this task which learns to encode the study background information and use this information to guide the decoding process. On a large dataset of radiology reports collected from actual hospital studies, our model outperforms existing non-neural and neural baselines under the ROUGE metrics. In a blind experiment, a board-certified radiologist indicated that 67% of sampled system summaries are at least as good as the corresponding human-written summaries, suggesting significant clinical validity. To our knowledge our work represents the first attempt in this direction.