AI Clinical Decision Support System Uses Digital Twins for Personalized Treatment

Clinical Decision Support AI Systems (CDSASs) must operate safely and adapt in real-time to a patient's evolving condition. Researchers have developed an online adaptive framework that addresses these requirements by integrating several advanced AI techniques. The system estimates Treatment Effects (TE) to quantify clinical benefits, utilizes a patient Digital Twin (DT) to simulate various treatment trajectories, and employs Reinforcement Learning (RL) for making sequential treatment decisions. The AI system is initially trained using historical medical records and then operates in a continuous learning loop, constantly refining its recommendations. To ensure patient safety, a rule-based module actively monitors vital signs and automatically blocks any contraindicated treatments. Furthermore, cases where the internal model exhibits significant disagreement or uncertainty are flagged for review by a human clinician, a process simulated in experiments using a pre-trained outcome model. Validation of this framework was conducted using both a synthetic clinical simulator and a real-world ovarian cancer dataset from The Cancer Genome Atlas (TCGA). In both simulated and actual clinical settings, the proposed method demonstrated superior effectiveness and stability in recommending treatments compared to standard computational baselines. The AI system also maintained low latency and required expert consultation for only a minority of cases during experimental validation, highlighting its potential as a safe, clinician-supervised tool for personalized medicine that improves with practical use.