Transfer Learning for Male Breast Cancer

Principal Investigator

Name
Lawrence Fulton

Degrees
Ph.D. MSStat MHA MS MMAS MSS

Institution
Texas State University

Position Title
Associate Professor

Email
lf25@txstate.edu

About this CDAS Project

Study
PLCO (Learn more about this study)

Project ID
PLCOI-742

Initial CDAS Request Approval
Mar 9, 2021

Title
Transfer Learning for Male Breast Cancer

Summary
Men account for just over 1% of breast cancer diagnoses each year. With so few cases, training deep learning models is difficult. Transfer learning, the process by which learning acquired from one domain might inform learning in another, is a technique for analyzing imagery that might be useful for analyzing male breast cancer identification and classification based on knowledge of female breast cancer.

Aims

1. Develop highly specific, highly sensitive, automated, novel, single stage deep learning methods to improve PPV and NPV of screening imagery analysis with known-labels using image pre-processing, Deep Convolutional Neural Networks, and other techniques to address the problem of overdiagnosis and overtreatment.
2. Construct Deep Learning Transfer Models for identification and classification of breast cancer in males using available imagery and data from females.

Collaborators

Dr. Yan Yan, Dr. Alex McCleod, Dr. Barbara Hewitt, Dr. Zo Ramamonjiarevelo, Dr. Diane Dolezel, and Dr. Tahir Ekin