Deep Learning MRI-based Model for Prediction of Clinically Significant Prostate Cancer

Prostate cancer is one of the leading causes of death for men globally, despite many men being diagnosed with indolent tumors that do not warrant treatment. Increasingly, magnetic resonance imaging (MRI) is being used as a risk assessment tool, before more invasive prostate biopsies are performed for patients at suspicion of prostate cancer. We hypothesize that we can train a deep learning model that combines multi-parametric MRI images with clinical factors to accurately predict patient risk of developing clinically significant prostate cancer. We train an image model and combined image and clinical factors model on a set of 9391 MRIs from the Massachusetts General Brigham (MGB) hospital system, which achieved an area under the receiver-operator curve (AUROC) of 0.80 and 0.84, respectively, for 1-year prediction of clinically significant prostate cancer, surpassing current human baselines and existing risk models’ performance.

Name thesis:degree_name

Master

Department dc:contributor.department

Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science

Grantor dc:publisher

Massachusetts Institute of Technology

Year dc:date.issued

2023

Author dc:creator

• Yang, Janice

Advisor dc:contributor.advisor

• Barzilay, Regina

Statement dc:rights

• In Copyright - Educational Use Permitted
• Copyright retained by author(s)

Licence dc:rights.uri

• https://rightsstatements.org/page/InC-EDU/1.0/

Handle dc:identifier.uri

https://hdl.handle.net/1721.1/151663

OAI identifier oai:identifier

oai:dspace.mit.edu:1721.1/151663