BiQancer: A Novel Quantitative Mathematical Biomarker and Automated End-to-End System for Decentralized Cervical Pre-Cancer Detection

Overview

The lack of rapid, accessible, and precise quantitative biomarkers for lesion grading remains the primary barrier to preventing cervical cancer—a leading cause of cancer death among women worldwide, despite being ~100% preventable if detected at the precancerous stage. Without such biomarkers, current Pap smear screening relies heavily on a shortage of pathologists across 160+ countries and suffers from high diagnostic disagreement (=54%). This creates bottlenecks, delaying treatment and increasing the risk of progression to cancer. To address this, we developed BiQancer. In Phase-1, PreCervAI integrates the Cyto-Quantitative Marker (CQ-Marker)—a novel quantitative ruler for traceable cervical lesion grading. It maps 4 key cellular features into a standardized lesion-grade scale. A specialized algorithm is used as a measurement tool to extract micron-level cellular primitives (nucleus, cytoplasm, koilocyte), ensuring all downstream CQ-Marker computations remain traceable to physical morphology, trained on 61,720 images. PreCervAI achieved 97.54%/95.66% Sn; 99.51%/99.13% Sp (controlled/blind), outperforming ResNet50 by 10%. In phase-2, to operationalize PreCervAI in primary care via standard microscopes, we engineered SigmaEyepiece, a $196 hardware add-on that automates standard microscope scanning and image capture. It meets FDA WSI and ISO 19264 at 99.98% lower cost than WSI benchmarks while operating 7.21× faster. Field validation on 240 cases across 5 countries yielded 93.75% Acc, 93.75% Sn, and 98.75% Sp. The system reduced screening time by 72.8% and can increase screening throughput by 80%. BiQancer is an accessible assistive tool for primary care facilities, enabling end-to-end decentralized screening for timely clinical decision-making and early intervention.

Awards (1)

• Second Award of $2,400 $2,400

Competition history

• ISEF 2026 Translational Medical Science · Entry TMED069T

Resources

• ISEF abstract page ↗

Related projects

Closest projects by meaning, across every fair and year in the corpus.