Preoperative CT-based Deep Learning Model for Predicting Disease-Free Survival in Patients with Lung Adenocarcinomas

Hyungjin Kim, Jin Mo Goo, Kyung Hee Lee, Young Tae Kim, Chang Min Park
Related Product


Date Published

This study aimed to develop and validate a preoperative CT-based deep learning model to predict disease-free survival in lung adenocarcinoma patients. Using data from patients with resected T1–4N0M0 adenocarcinoma (2009-2015) for training and patients with clinical stage I adenocarcinoma (2014) for external validation, the model was trained to extract prognostic information from preoperative CT scans. Lung nodules were manually annotated using Aview software (Coreline Soft, Seoul, Korea). The model’s performance, measured by Harrell C index, showed good discrimination and calibration, comparable to the clinical T category. Internal validation C indexes ranged from 0.74 to 0.80, and external validation ranged from 0.71 to 0.78. The model's outputs were independent prognostic factors, with hazard ratios indicating significant predictive power. The only other significant factor was smoking status. The deep learning model effectively predicted disease-free survival in patients undergoing surgery for clinical stage I lung adenocarcinoma.


제품, 인재 채용, 투자 관련 또는 기타 문의사항이 있으신 경우 편하신 방법으로 연락주시기 바랍니다