Predicting postoperative non-small cell lung cancer prognosis via long short-term relational regularization

作者：

Highlights：

• LSTRR is a regularization term designed to exploit the relations between short and long-term prognoses.

• The model with LSTRR achieves better AUC values for short-term prognostic prediction.

• Risk factors recognized by the model with LSTRR are more consistent for both short and long-term prognoses.

• Less unreasonable risk factors are recognized using LSTRR under the in-depth review of clinical experts.

摘要

•LSTRR is a regularization term designed to exploit the relations between short and long-term prognoses.•The model with LSTRR achieves better AUC values for short-term prognostic prediction.•Risk factors recognized by the model with LSTRR are more consistent for both short and long-term prognoses.•Less unreasonable risk factors are recognized using LSTRR under the in-depth review of clinical experts.

论文关键词：Prognostic prediction,Non-small cell lung cancer,Risk factor recognition,Long short-term relational regularization

论文评审过程：Received 21 January 2020, Revised 19 May 2020, Accepted 29 June 2020, Available online 30 June 2020, Version of Record 16 July 2020.

论文官网地址：https://doi.org/10.1016/j.artmed.2020.101921