Minimally-Invasive and Efficient Method to Accurately Fit the Bergman Minimal Model to Diabetes Type 2

Area of Science:

• Biomedical Engineering
• Computational Biology
• Endocrinology

Background:

• Diabetes mellitus presents a growing global health challenge, necessitating advanced prevention, diagnosis, and treatment strategies.
• Individualized computational models are crucial for understanding and managing inter- and intra-individual variations in diabetes care.
• Personalized medicine approaches are vital for optimizing treatment outcomes in diabetic patients.

Purpose of the Study:

• To develop and validate a method for accurately parameterizing the five-parameter Bergman Minimal Model.
• To utilize experimental data from diabetic rats to refine the Bergman Minimal Model.
• To advance personalized medicine for diabetes through precise physiological characterization.

Main Methods:

• Twenty experiments were conducted on Sprague-Dawley rats with streptozotocin-induced diabetes.
• Insulin-glucose response curves were recorded using an insulin pump and percutaneous glucose sensor over 60-100 minutes.
• A genetic algorithm with root-mean-squared optimization was employed to fit the Bergman Minimal Model to experimental data.

Main Results:

• Bergman Minimal Model parameters were estimated with high accuracy.
• The model demonstrated low prediction bias.
• A low average root-mean-squared error of 15.27 mg/dl glucose was achieved.

Conclusions:

• A straightforward method for accurate Bergman Minimal Model parameterization was demonstrated.
• Estimated parameters objectively characterize diabetes severity, aiding in personalized treatment planning.
• Quantified glucose effectiveness and insulin sensitivity reflect patient condition, supporting personalized medicine and research insights.