Researchers at The University of Texas MD Anderson Cancer Center developed an AI model that reads routine chest CT scans to flag lung cancer patients likely to develop immunotherapy-related pneumonitis before the first dose. Pneumonitis affects roughly 10% of lung cancer patients treated with immunotherapy, creating a meaningful need for earlier risk stratification. The system’s clinical target is proactive monitoring—identifying patients at elevated risk so clinicians can adjust observation plans, management strategies, or treatment decisions. The approach also fits a broader pattern of AI moving from retrospective prediction into prospective clinical workflow support. While performance metrics and external validation will be critical for adoption, the study positions imaging-based prediction as a practical route to reduce severe immune-related lung toxicity.