A Self-Supervised Foundation Model Based on Three-Dimensional Chest CT Scans for Lung Cancer Diagnosis and Prognosis Prediction

Key Points

• To develop a self-supervised model for lung cancer diagnosis and prognosis using chest CT scans.
• Trained on 33,901 chest CT scans
• Utilized a self-supervised learning approach
• Compared performance against existing algorithms
• Surpassed state-of-the-art algorithms
• Accurately predicted lung cancer subtype, stage, and survival
• Reduced the need for manual annotation