Core Viewpoint - Early cancer screening is crucial for reducing incidence and mortality rates, with a need for cost-effective, high-throughput screening methods that can address the demands of asymptomatic populations [2][5][6]. Group 1: Research Development - The research team developed OMAFound, a foundational model for multi-cancer screening using non-contrast computed tomography (CT), enabling simultaneous detection of lung and breast cancer [3][7]. - OMAFound's predictive performance matches that of specialized organ AI models and surpasses experienced radiologists in sensitivity for screening scenarios [3][11]. Group 2: Cancer Statistics and Screening Needs - In 2022, approximately 20 million new cancer cases and 9.7 million deaths were reported globally, highlighting the ongoing rise in cancer burden due to aging populations and lifestyle factors [5][6]. - Early-stage cancer patients have significantly higher five-year survival rates compared to late-stage patients, emphasizing the urgency for effective early screening strategies in high-risk, asymptomatic populations [5][6]. Group 3: Screening Methodology and Model Features - Traditional single-cancer screening methods are inadequate for high-throughput needs, necessitating the exploration of cost-effective multi-cancer early screening strategies [6][9]. - OMAFound utilizes over 200,000 CT images for pre-training and employs self-supervised learning to enhance robustness against variations in equipment and settings, improving multi-cancer screening capabilities [7][9]. Group 4: Performance Metrics - In a cohort of over 20,000 individuals, OMAFound achieved detection accuracies of 82.2% for breast cancer and 88.0% for lung cancer in women, and 86.1% for lung cancer in men [9][11]. - The model significantly improved the sensitivity of experienced radiologists by 38.9% for breast cancer and 16.0% for lung cancer, while maintaining specificity [9][11]. Group 5: Clinical Implications - OMAFound emphasizes model interpretability, aiding physicians in identifying potential lesions more effectively, thus enhancing the clinical applicability and acceptance of opportunistic breast cancer screening [11]. - The model represents a practical technological tool for AI-enabled multi-cancer screening, aiming to facilitate early detection and diagnosis, with significant clinical and societal implications [11].