Academic Journal
Deep reinforcement learning for Type 1 Diabetes: Dual PPO controller for personalized insulin management
| Τίτλος: | Deep reinforcement learning for Type 1 Diabetes: Dual PPO controller for personalized insulin management |
|---|---|
| Συγγραφείς: | Marchetti A., Sasso D., D'Antoni F., Morandin F., Parton M., Matarrese M. A. G., Merone M. |
| Συνεισφορές: | Marchetti, A., Sasso, D., D'Antoni, F., Morandin, F., Parton, M., Matarrese, M. A. G., Merone, M. |
| Έτος έκδοσης: | 2025 |
| Συλλογή: | ARPI - Archivio della Ricerca dell'Università di Pisa |
| Θεματικοί όροι: | Diabete, Glycemic control, Precision medicine, Proximal Policy Optimization, Reinforcement learning |
| Περιγραφή: | Background: Managing blood glucose levels in Type 1 Diabetes Mellitus (T1DM) is essential to prevent complications. Traditional insulin delivery methods often require significant patient involvement, limiting automation. Reinforcement Learning (RL)-based controllers offer a promising approach for improving automated insulin administration. Methods: We propose a Dual Proximal Policy Optimization (Dual PPO) controller for personalized insulin delivery in a hybrid closed-loop system. The controller optimizes patient-specific insulin bounds through a grid search on pre-trained models to manage both hyperglycemia and hypoglycemia. A safe-control mechanism prevents insulin administration when glucose levels drop below a predefined threshold. The system was evaluated on 10 in silico adult patients using the UVA/Padova simulator, with five-day randomized meal scenarios. Results: The Dual PPO controller significantly improved Time in Range (TIR) (69.30% ±1.61) compared to a single PPO model (61.69% ±1.54). The system effectively reduced severe hyperglycemia while maintaining a low incidence of severe hypoglycemia. Unlike conventional open-loop methods such as Basal-Bolus (BBC) and Proportional–Integral–Derivative (PIDC) controllers, our system requires minimal patient interaction, eliminating the need for carbohydrate estimation. Conclusions: The Dual PPO controller enhances personalized insulin delivery in T1DM, improving glycemic control while reducing patient burden. This approach advances precision medicine in diabetes management, with potential for future real-world applications. |
| Τύπος εγγράφου: | article in journal/newspaper |
| Γλώσσα: | English |
| Relation: | info:eu-repo/semantics/altIdentifier/pmid/40239234; volume:191; numberofpages:13; journal:COMPUTERS IN BIOLOGY AND MEDICINE; https://hdl.handle.net/11568/1343693; https://www.sciencedirect.com/science/article/pii/S0010482525004986?via=ihub |
| DOI: | 10.1016/j.compbiomed.2025.110147 |
| Διαθεσιμότητα: | https://hdl.handle.net/11568/1343693 https://doi.org/10.1016/j.compbiomed.2025.110147 https://www.sciencedirect.com/science/article/pii/S0010482525004986?via=ihub |
| Αριθμός Καταχώρησης: | edsbas.12B1DB49 |
| Βάση Δεδομένων: | BASE |
| DOI: | 10.1016/j.compbiomed.2025.110147 |
|---|