Meal detection and size estimation for artificial pancreas devices using machine learning

­Summary
The current technology can be used within an automated insulin delivery system to automate insulin delivery (AID) in response to meals. This technology has the potential to reduce the burden of a person living with type 1 diabetes using an AID and potentially improving glucose management by accurately and automatically delivering insulin in response to detected meals.

Technology Overview
Closed-loop systems, also known as artificial pancreas devices, for automated insulin delivery (AID) facilitate diabetes management, but still require patients to manually enter meals and estimate carbohydrate consumption, which is prone to inaccuracy.  The current technology is a machine-learning algorithm for closed loop systems, that allows for meal detection and size estimation to improve glucose management for patients with type-1 diabetes.  Features of the algorithm include:

• Opportunity for full automation for closed-loop insulin delivery systems to reduce patient burden in entering meals and carbohydrates.
• More accurate calculation of insulin doses to reduce incidents of post-prandial glucose dysregulation.
• Validation both in silico and in a large real-world dataset, demonstrating meal detection sensitivity of 83.3%, false discovery rate of 16.6%, and mean detection time of 25.9 minutes
• Compatible for use with smartphones, closed-loop insulin delivery systems, and on cloud-based computing systems to allow a broad range of accessibility.
• Demonstrated accuracy for patients without diabetes for utility in diet coaching settings.

Publication
Mosquera-Lopez C, et al., Enabling fully automated insulin delivery through meal detection and size estimation using Artificial Intelligence. NPJ Digit Med. 2023 Mar 13;6(1):39. Link

Licensing Opportunity
This technology is available for licensing.