Rapid Single-Cell Phenotypic Antifungal Susceptibility Testing on a SlipChip Enabled by Deep Learning.

Area of Science:

• Medical Microbiology
• Biotechnology
• Computational Biology

Background:

• Fungal infections are a major global health threat, causing over 1.5 million deaths annually.
• Current antifungal susceptibility testing (AFST) methods are time-consuming, requiring 24-48 hours, which delays critical treatment decisions.
• There is an urgent need for rapid and accurate AFST methods to guide antifungal therapy.

Purpose of the Study:

• To develop and validate a rapid, label-free, single-cell AFST (sc-AFST) platform.
• To significantly reduce the time required for AFST from days to hours.
• To enable high-resolution single-cell growth analysis and population-level drug response profiling.

Main Methods:

• Integration of a SlipChip microfluidic device with deep learning-based image analysis (ResNet-34 classifier and U-Net segmentation model).
• Partitioning of fungal cells into picoliter droplets for exposure to antifungal gradients.
• Automated identification and quantification of cellular growth from bright-field images.

Main Results:

• Achieved >98.8% sensitivity in cell detection and >93% segmentation accuracy.
• Demonstrated 100% categorical agreement with the gold standard broth microdilution test across 10 clinical isolates.
• Successfully determined minimal inhibitory concentrations (MICs) within 4 hours.

Conclusions:

• The developed sc-AFST platform provides a scalable, rapid, and clinically actionable tool for antifungal susceptibility testing.
• This technology enables precise single-cell analysis and population-level drug response assessment.
• The rapid turnaround time of sc-AFST can significantly improve patient outcomes by enabling timely treatment decisions for fungal infections.