Radiology report generation for proximal femur fractures using deep classification and language generation models

作者：

Highlights：

• State-of-the-art fracture classification performance with minimal preprocessing

• Committee voting based evaluation increases performance.

• User studies show similar language quality for generated and ground truth reports.

• Medical domain experts note equivalency in several cases.

摘要

•State-of-the-art fracture classification performance with minimal preprocessing•Committee voting based evaluation increases performance.•User studies show similar language quality for generated and ground truth reports.•Medical domain experts note equivalency in several cases.

论文关键词：Proximal femur fractures,Radiology report generation,Fracture classification,Radiology language model,User study

论文评审过程：Received 3 February 2021, Revised 25 February 2022, Accepted 15 March 2022, Available online 26 March 2022, Version of Record 5 April 2022.

论文官网地址：https://doi.org/10.1016/j.artmed.2022.102281