Satvika Bharadwaj
BME PhD Proposal Presentation

Date: 2025-11-24
Time: 12 pm - 1pm
Location / Meeting Link: https://emory.zoom.us/meetings/96692034108/invitations?signature=NAFEA0-xSa9xsmP6cQ5GvlhW_D9imw5lGunsdJMUeX8

Committee Members:
Anant Madabhushi, PhD (Advisor); Jaydev Desai, PhD; Saurabh Sinha, PhD; Sunil Badve, MD, FRCPath; Farzad Fereidouni, PhD;


Title: Interpretable biomarker discovery using deep phenotyping of histopathology slides to predict disease progression and treatment response

Abstract:
Advances in digital pathology have made it possible to extract quantitative information from routine histopathology slides to study disease biology and therapy response. While these approaches have shown value in diagnosis, their use for prognostication and treatment prediction remains limited. There is a need for interpretable histologic features to predict outcomes and support treatment decisions. This proposal will develop a deep phenotyping pipeline that extracts quantitative tissue and cellular features to predict clinical endpoints across multiple disease states. The research will integrate computer vision and machine learning algorithms with pathology domain knowledge to develop image-based biomarkers. It will first demonstrate utility in bone biopsies, where clinically relevant measures of bone remodeling will be quantified and used to classify bone turnover. The same deep phenotyping pipeline will then be applied for biomarker discovery in HER2-positive (HER2+) and ER-positive (ER+) breast cancers to evaluate structural and cellular phenotypes linked to disease progression and therapeutic response. The proposed research will establish a computational pathology pipeline for biomarker discovery to predict outcomes and support treatment planning.