Un modelo de base autorregulado basado en tomografías computarizadas de tórax tridimensionales para el diagnóstico y...

Purpose To develop a self-supervised chest CT foundation model and evaluate its performance in lung cancer clinical tasks. Materials and Methods In this retrospective multicenter study, the authors developed the Unified CT-Based Lung Cancer Imaging Foundation (UCLIF) model using self-supervised learning on 33 901 three-dimensional chest CT scans acquired between June 1958 and February 2019. The model was pretrained with a contrastive masked image modeling task and then fine-tuned for lung cancer histologic subtype classification, cancer staging, survival, and recurrence prediction using multicenter patient datasets. Histopathology, TNM stage, and follow-up outcomes served as reference standards. UCLIF was compared with mainstream deep learning and machine learning algorithms, and model performance was assessed by accuracy, sensitivity, specificity, and area under the receiver operating characteristic curve (AUC); superiority was tested using the DeLong test. Results A total of 656 patients were included for downstream evaluation (mean age, 68.55 years ± 10.01 [SD]; 450 males). Compared with self-supervised pretraining on natural images or single tumor regions, UCLIF achieved superior performance (DeLong test, P < .001) and provided high AUCs for histologic subtype (AUC, 0.96 [95% CI: 0.88, 1.00]; AUC, 0.82 [95% CI: 0.60, 0.98]; and AUC, 0.93 [95% CI: 0.80, 0.99] for adenocarcinoma, large cell lung cancer, and squamous cell carcinoma, respectively), cancer staging (AUC, 0.95 [95% CI: 0.79, 1.00]; AUC, 0.99 [95% CI: 0.96, 1.00]; AUC, 0.92 [95% CI: 0.74, 1.00]; and AUC, 0.91 [95% CI: 0.78, 1.00] for stages I-IV, respectively), survival (AUC, 0.97 [95% CI: 0.92, 1.00]; AUC, 0.90 [95% CI: 0.72, 0.98]; and AUC, 0.90 [95% CI: 0.77, 1.00] for 1-, 3-, and 5-year survival, respectively), and recurrence (AUC, 0.95; 95% CI: 0.88, 0.99). Conclusion The UCLIF model accurately predicted lung cancer histologic subtype, stage, survival, and recurrence. Keywords: Lung Cancer, CT, Foundation Model, Diagnosis, Classification Supplemental material is available for this article. © RSNA 2026.