AI-aided Geometric Design of Anti-infection Catheters

TL;DR

This paper introduces an AI-optimized geometric design for catheters that significantly reduces bacterial contamination by disrupting upstream swimming, demonstrated through experiments with E. coli and 3D-printed prototypes.

Contribution

It presents a novel AI-based approach using Fourier neural operators to design catheter surfaces that inhibit bacterial upstream movement, addressing a critical clinical challenge.

Findings

Achieved 10-100x reduction in bacterial contamination upstream

Validated design effectiveness with microfluidic experiments

Demonstrated improved performance in 3D-printed prototypes

Abstract

Bacteria can swim upstream due to hydrodynamic interactions with the fluid flow in a narrow tube, and pose a clinical threat of urinary tract infection to patients implanted with catheters. Coatings and structured surfaces have been proposed as a way to suppress bacterial contamination in catheters. However, there is no surface structuring or coating approach to date that thoroughly addresses the contamination problem. Here, based on the physical mechanism of upstream swimming, we propose a novel geometric design, optimized by an AI model predicting in-flow bacterial dynamics. The AI method, based on Fourier neural operator, offers significant speedups over traditional simulation methods. Using Escherichia coli, we demonstrate the anti-infection mechanism in quasi-2D micro-fluidic experiments and evaluate the effectiveness of the design in 3Dprinted prototype catheters under clinical flow rates. Our catheter design shows 1-2 orders of magnitude improved suppression of bacterial contamination at the upstream end of the catheter, potentially prolonging the in-dwelling time for catheter use and reducing the overall risk of catheter-associated urinary tract infections.