Accession Number:

AD1190495

Title:

Analysis of Tissue Architecture to Identify Lethal Prostate Cancer in the Veteran Population

Descriptive Note:

[Technical Report, Annual Report]

Corporate Author:

UTAH UNIV SALT LAKE CITY

Personal Author(s):

• Knudsen, Beatrice

Report Date:

2022-10-01

Pagination or Media Count:

30

Abstract:

The project in this grant is aimed to improve the precision medicine treatment of men with high-grade prostate cancer who are diagnosed within the VA healthcare system. Recent clinical trials demonstrate that multi-modal upfront treatment with combination of two-three cancer drugs improve overall survival and disease-free survival of men with aggressive prostate cancer. However, clinical parameters and genomic test results do not reliably identify men who are at greatest risk of dying from lethal prostate cancer. The Gleason grade of the cancer is a good predictor of disease course overall, but does not predict who is at risk for metastatic progression within the high-grade cancer group. We propose that in addition to the cancer grade, computer can detect salient features of the cancer that are associated with risk of metastasis. Using artificial intelligence frameworks, we propose to train models on high-grade cancer regions that predict metastatic risk. We identified approximately 12,000 men at the VA which we propose to study. Of those, we will enroll at least 600 for the initial development of algorithms that is funded by this grant. During the first year of funding, we identified the study cohorts, established the basic enrolment system using STARLIMS and generated 1337 digital slides from 279 cases. We also consolidated multiple A.I. algorithms that can be used to identify regions of high-grade cancer and are testing which of them works best with VA cases. To perform above tasks, we completed all IR Band regulatory requirements and recruited all the necessary expertise for the project. The generous funding by the Department of Defense will help us with the development of an affordable and easily deployable software tool that has a chance to improve the care of veterans diagnosed with aggressive prostate cancer.

Descriptors:

• health services
• medical personnel
• artificial intelligence
• machine learning
• computer vision
• artificial intelligence software
• information science
• students
• automata theory
• computer science
• network science
• statistical analysis
• image recognition
• neural networks
• application software
• computers
• identification

Subject Categories:

• Medicine and Medical Research

Distribution Statement:

[A, Approved For Public Release]