Prostate Cancer Screening using Machine Learning

Principal Investigator

Name
Jerry John Rawlings Mensah

Institution
Virginia Commonwealth University

Position Title
Graduate Student

Email
mensahj3@vcu.edu

About this CDAS Project

Study
PLCO (Learn more about this study)

Project ID
PLCO-1732

Initial CDAS Request Approval
Nov 5, 2024

Title
Prostate Cancer Screening using Machine Learning

Summary
Early detection and risk estimation of Prostate Cancer (PCa) is especially important for effective treatment and good clinical outcomes. Early screenings using the serum Prostate-Specific Antigen (PSA) test are usually less invasive and cheaper than other PCa tests which involve biopsies, MRI, or rectal exams – that can be invasive, expensive, or both. Although PSA testing is very common, it is a controversial topic arising from the lack of clear guidelines on serum PSA threshold for diagnoses and treatment, as well as accuracy issues such as sensitivity and specificity that can lead to overdiagnosis or overtreatment. As a result, organizations such as the United States Preventive Services Task Force (USPSTF) recommend against PSA screening for other measures such as shared decision-making between providers and patients, with consideration of individual patient risk factors including age, life expectancy, family history, and race/ ethnicity. This project will use machine learners that will combine the USPSTF measures with PSA, digital rectal exam screening results, and other important baseline clinical and epidemiologic variables to train and evaluate machine learning models to predict prostate cancer risk and incidence over 10-13 years.

Aims

The goal of this study is to explore another equally cheaper and alternative method for first-line Prostate Cancer screening that builds on the advantages of PSA testing and improves upon its limitations.

Collaborators

Dr Emmanuel Taylor - VCU School of Public Health/ Massey Comprehensive Cancer Center