Point-of-care cervical cancer screening using deep learning-based microholography

Area of Science:

• Biotechnology
• Medical Diagnostics
• Nanotechnology

Background:

• Cervical cancer disproportionately affects regions with limited screening infrastructure.
• Developed nations increasingly use human papillomavirus (HPV) testing, widening the resource gap.
• There is a critical need for accessible, point-of-care HPV screening solutions.

Purpose of the Study:

• To develop a DNA-focused digital microholography platform for point-of-care HPV screening.
• To integrate deep-learning algorithms for automated and rapid analysis of HPV DNA.
• To assess the platform's performance in detecting high-risk HPV types.

Main Methods:

• Development of a microholography platform utilizing microbeads for HPV DNA target binding.
• Design of microbeads to form dimers upon binding HPV 16 or 18 DNA.
• Application of deep-learning algorithms for rapid holographic signature analysis.

Main Results:

• The HPV DNA assay demonstrated high sensitivity (single-cell level) and 100% specificity for HPV 16 and 18.
• Deep learning analysis was 120x faster than traditional methods, completing in under 2 minutes.
• The platform's performance was successfully benchmarked against an FDA-approved HPV assay in a clinical study.

Conclusions:

• The developed platform offers reliable and decentralized HPV testing capabilities.
• This technology can aid in mapping high-risk HPV prevalence in underserved populations.
• Findings can inform HPV vaccination strategies and identify coverage gaps.