Attribute-aware interpretation learning for thyroid ultrasound diagnosis

作者：

Highlights：

• We investigate the thyroid nodule diagnosis interpretation from both model logic and model result perspectives.

• We propose a thyroid nodule diagnosis model with outstanding performance and a variety of doctor-acceptable interpretations.

• Human-computer collaborative experiments show that model interpretations can improve the radiologists' diagnostic ability.

摘要

•We investigate the thyroid nodule diagnosis interpretation from both model logic and model result perspectives.•We propose a thyroid nodule diagnosis model with outstanding performance and a variety of doctor-acceptable interpretations.•Human-computer collaborative experiments show that model interpretations can improve the radiologists' diagnostic ability.

论文关键词：Ultrasound image,Thyroid nodule diagnosis,Graph attention network,Computer-aided diagnosis

论文评审过程：Received 29 October 2021, Revised 20 June 2022, Accepted 27 June 2022, Available online 1 July 2022, Version of Record 5 July 2022.

论文官网地址：https://doi.org/10.1016/j.artmed.2022.102344