Machine Learning-Enabled Optimization of Interstitial Fluid Collection via a Sweeping Microneedle Design

Area of Science:

• Biomedical Engineering
• Materials Science
• Computational Science

Background:

• Microneedles (MNs) are crucial for minimally invasive diagnostics and drug delivery.
• Traditional MN fabrication relies on empirical data and trial-and-error optimization.
• Performance enhancement necessitates advanced data analysis techniques.

Purpose of the Study:

• To integrate finite element methods (FEMs) and machine learning (ML) for optimal MN design.
• To maximize fluid collection by identifying ideal physical and geometrical parameters.
• To advance the application of MNs in wearable point-of-care devices.

Main Methods:

• FEM simulations were used to model fluid behavior in MN patches with varying parameters.
• A comprehensive dataset from FEM simulations was used to train ML algorithms.
• ML models evaluated included multiple linear regression, random forest, support vector regression, and neural networks.

Main Results:

• Decision Tree Regression (DTR) demonstrated the highest accuracy in predicting optimal MN parameters.
• The integrated FEM and ML approach successfully identified key design parameters for enhanced fluid collection.
• ML modeling proved effective in optimizing MN geometrical designs.

Conclusions:

• Machine learning modeling offers a powerful approach to optimize microneedle design.
• This AI-driven methodology can significantly enhance fluid sampling and drug delivery via MNs.
• Optimized MNs hold promise for improved wearable diagnostics and targeted therapeutics.