Deep-learning model can spot patients at high risk of COPD

Thursday, November 30 | 8:20 a.m.-8:30 a.m. | R1-SSCH09-3 | Room E352

This scientific presentation will present external validation results for a deep-learning model in identifying individuals at high risk of incident chronic obstructive pulmonary disease (COPD) on routine outpatient chest x-rays (CXR).

Saman Doroodgar Jorshery, MD, and colleagues will discuss findings with CXR-Lung-Risk, a previously developed convolutional neural network designed to predict lung-related mortality. For the external validation, patients with no history of lung cancer, COPD, or emphysema, and who had a chest radiograph, were identified.

After analyzing 10,913 patients, the CXR-Lung-Risk model was significantly and independently associated with incident COPD. As determined via receiver operating characteristic (ROC) analysis, CXR-Lung-Risk added predictive value in comparison with the TargetCOPD clinical risk score, according to Jorshery and colleagues.

"Opportunistic screening of existing CXRs using deep learning based model (CXR-Lung-Risk) could help in identifying high-risk individuals and guide COPD prevention," the researchers concluded.

What else did they find? Check out this session on Thursday to get all of the results.

Page 1 of 2
Next Page