Multilingual Natural Language Processing Model for Radiology Reports -- The Summary is all you need!
Mariana Lindo, Ana Sofia Santos, André Ferreira, Jianning Li, Gijs Luijten, Gustavo Correia, Moon Kim, Benedikt Michael Schaarschmidt, Cornelius Deuschl, Johannes Haubold, Jens Kleesiek, Jan Egger, Victor Alves
TL;DR
The paper tackles the lack of multilingual radiology report summarization by building a multilingual RR summarization model based on mT5. It first adapts English RR summarization, then extends to Portuguese and German via translation and native data, culminating in a balanced multilingual model $M^{rr-1000}_{EN, PT, GE}$. Quantitatively, it achieves higher ROUGE scores than monolingual baselines and shows strong qualitative agreement with radiologists, with the multilingual model outperforming specialized single-language models and surpassing ChatGPT in this task. The study highlights the practical potential to reduce radiologists' workload and enable cross-language, multimodal DL applications, with public availability of the model.
Abstract
The impression section of a radiology report summarizes important radiology findings and plays a critical role in communicating these findings to physicians. However, the preparation of these summaries is time-consuming and error-prone for radiologists. Recently, numerous models for radiology report summarization have been developed. Nevertheless, there is currently no model that can summarize these reports in multiple languages. Such a model could greatly improve future research and the development of Deep Learning models that incorporate data from patients with different ethnic backgrounds. In this study, the generation of radiology impressions in different languages was automated by fine-tuning a model, publicly available, based on a multilingual text-to-text Transformer to summarize findings available in English, Portuguese, and German radiology reports. In a blind test, two board-certified radiologists indicated that for at least 70% of the system-generated summaries, the quality matched or exceeded the corresponding human-written summaries, suggesting substantial clinical reliability. Furthermore, this study showed that the multilingual model outperformed other models that specialized in summarizing radiology reports in only one language, as well as models that were not specifically designed for summarizing radiology reports, such as ChatGPT.