Digital Pathology
Ongoing
Industry Partnered
PathoScan-AI: Gigapixel Histology Biomarker Profiling
Lead Investigator: Dr. Elena Rostova
Executive Summary
Foundation vision transformer models trained on over 50,000 whole-slide biopsies for rapid automated cancer grading and margin profiling.
Technology Stack & Domain Frameworks
Vision Transformers
Gigapixel WSI
Digital Pathology
Methodology & Clinical Validation
Project Overview:
PathoScan-AI leverages multiple instance learning (MIL) and hierarchical patch encoding to evaluate whole-slide pathology images up to 100,000 x 100,000 pixels in under 3 seconds.