L2M – Cross-Modality Translation in Medical Imaging: Bridging Diagnostic Domains

Medical imaging is vital to modern healthcare, yet high costs and limited access slow patients’ paths to
diagnosis. Radiologists often correlate imaging findings across multiple imaging modalities and sequences
to arrive at a specific diagnosis, but each extra scan or sequence increases scanner time, healthcare costs,
wait-list pressure, radiation for ionizing exams, and energy use. An AI image-to-image translator can begin
with the quickest, least-burdensome scan or sequence and digitally create the other needed images, so
patients avoid additional exams or sequences. This saves scanner hours, contrast agents, and energy,
lowers healthcare costs and radiation exposure, and still gives clinicians information required for
diagnosis. The project will survey existing image-to-image translation technologies, gather stakeholder
priorities, and deliver a roadmap that moves synthetic imaging from research into routine care, offering
multiple potential advantages for health systems and patients.

Faculty Supervisor:

Scott Adams

Student:

Partner:

DMZ Ventures Inc

Discipline:

Computer science

Sector:

Professional, scientific and technical services

University:

University of Saskatchewan

Program: