Abstract
Background: In Nuclear Medicine, managing radioiodine (I-131) treated thyroid cancer patients remains challenging due to variations in global practices, regulatory criteria, and the need for individualized care. The lack of harmonization in patient release protocols has prompted the development of innovative, adaptive approaches.
Methods: This research introduces a novel artificial intelligence (AI)-based “Dynamic Patient Release Management System” designed to predict patient discharge timing using machine learning models. The system integrates patient-specific parameters (e.g., administered dose, biokinetics, contact scenarios, and socio-economic factors) while adhering to international safety standards. Approximately 350 patients treated with 50, 100, 150, and 200 mCi of I-131 were retrospectively analyzed to train and validate the model.
Result: The system reliably predicted safe discharge timelines: 1.5, 2.0, 2.5, and 3.0 days for 50, 100, 150, and 200 mCi doses, respectively. It also generated tailored recommendations for patient-specific occupancy factors, ensuring compliance with dose limits for caregivers and the public. Importantly, it allowed minor regional adaptability while maintaining regulatory compliance, particularly relevant for resource-limited settings.
Conclusion: The proposed AI model offers a personalized, safe, and flexible approach to managing I-131 patient release. It minimizes human judgment errors, standardizes practices across institutions, and supports policy implementation in diverse healthcare environments. This tool represents a significant advancement in balancing regulatory compliance with contextual needs.
Index Terms - AI patient release, Patient Release criteria of I-131, I-131 treated patient release via AI, one-click decision of patient
release.

Key words: Artificial Intelligence Patient Release, Iodine Radioisotopes, Radiation Protection, Patient Discharge, Radiation Dosage, Radiation Monitoring, Machine Learning, Clinical Decision Support Systems, Patient Release criteria of I-131vvia AI, one-click decision of patient release