Related Experiment Video
Updated: Aug 25, 2025
07:58
Computational Reconstruction of Pancreatic Islets as a Tool for Structural and Functional Analysis
Published on: March 9, 2022
1.7K
A predictive computational platform for optimizing the design of bioartificial pancreas devices
Alexander U Ernst1, Long-Hai Wang2,3, Scott C Worland1
1Biological and Environmental Engineering, Cornell University, Ithaca, NY, USA.
Nature Communications
|October 13, 2022
Summary
A new computational platform models bioartificial pancreas devices for type 1 diabetes, revealing cell size distribution impacts efficacy. This tool optimizes device design and estimates curative cell doses for improved diabetes treatment.
Area of Science:
- Biomedical Engineering
- Computational Biology
- Endocrinology
Background:
- Bioartificial pancreas devices offer a potential cure for type 1 diabetes by encapsulating insulin-producing cells.
- Current models of these devices often oversimplify complex biological factors, limiting their predictive power.
- Mass transport limitations within devices significantly impact therapeutic efficacy.
Purpose of the Study:
- To develop a computational platform for simulating bioartificial pancreas devices, incorporating stochastic cellular properties.
- To investigate the influence of cell size distribution and localization on device performance.
- To optimize bioartificial pancreas design and estimate therapeutic cell doses.
Main Methods:
- Development of a computational platform for simulating encapsulated cell therapies.
- Inclusion of cell size distribution and random localization as key parameters.
- Application of the platform to study device potency and optimize structural parameters.
- Creation of a device-specific islet equivalence conversion table and a machine learning surrogate model.
Main Results:
- Endogenous islet size distribution variance was found to significantly impact device potency.
- Optimized device structures and estimates for curative cell doses were determined.
- A novel, device-specific islet equivalence conversion table was proposed.
- A machine learning model was developed to rapidly generate conversion coefficients for user-defined devices.
Conclusions:
- Accurate modeling of stochastic cellular properties is crucial for understanding bioartificial pancreas efficacy.
- The developed computational platform and associated tools can guide the design and optimization of next-generation diabetes therapies.
- This work provides a framework for personalized therapeutic dose estimations and device development.
