Machine Learning-based Acceleration for Planning of Magnetic Resonance Imaging (MRI)-guided transurethral ultrasound ablation – Phase I

This one-year Accelerate project is the initial phase of a larger collaboration aiming to improve outcomes for men undergoing a specific prostate cancer treatment. The treatment, called MRI-guided transurethral ultrasound ablation (TULSA), is a minimally invasive procedure that uses high-intensity ultrasound to destroy prostate tissue, with MRI technology helping to plan and monitor the process.

The TULSA-PRO device, created by Profound Medical in Canada, is already approved by major health authorities and is used by doctors. However, feedback from clinical use suggests that several software improvements could make the treatment more effective, safer, faster, and more predictable. This project aims to develop these improvements using machine learning (ML).

The specific objectives are of this first phase of our collaboration are:
1) Develop a method to predict the deformation of the prostate and surrounding structures based on screening prostate MRI images.
2) On the deformed prostate, adapt existing GPU-based numerical implementations for acoustic and thermal modelling to TULSA procedures, simulating the deformed prostate

Faculty Supervisor:

Samuel Pichardo

Student:

Partner:

Profound Medical

Discipline:

Engineering

Sector:

Manufacturing

University:

University of Calgary

Program: