Computational Prediction of Gene Expression and Clinical Outcomes from Digital Pathology Images
Principal Investigator
Name
Mingyao Li
Degrees
Ph.D
Institution
University of Pennsylvania
Position Title
Professor
About this CDAS Project
Study
PLCO
(Learn more about this study)
Project ID
PLCOI-2022
Initial CDAS Request Approval
Feb 13, 2026
Title
Computational Prediction of Gene Expression and Clinical Outcomes from Digital Pathology Images
Summary
This project aims to develop and evaluate computational pathology methods for predicting gene expression patterns and clinically relevant molecular signatures from digital whole slide histopathology images using the PLCO dataset.
Using de-identified hematoxylin and eosin (H&E) whole slide images and linked clinical metadata, we will train and validate machine learning models to extract quantitative image-derived features and infer molecular characteristics. These predicted molecular signatures will be evaluated for their association with tumor characteristics and survival outcomes.
The study involves retrospective analysis of de-identified data only. No attempt will be made to re-identify study participants, and no clinical intervention or patient contact is involved. All data will be stored and analyzed on secure institutional servers in accordance with NIH data use policies. The results are intended to advance computational methods for integrative pathology analysis and improve understanding of image–molecular associations in cancer.
Aims
Aim 1: Develop machine learning models to extract quantitative morphological features from cancer whole slide histopathology images.
Aim 2: Train and evaluate computational models to predict gene expression patterns and molecular signatures from image-derived features using de-identified PLCO pathology and linked clinical data.
Aim 3: Validate the robustness and generalizability of the developed computational framework within the PLCO cancer cohort.
Collaborators
Mingyao Li University of Pennsylvania
WEI LI University of Pennsylvania
Sijia Huang University of Pennsylvania