Deformable Image Registration for Multimodal Radiotherapy Treatments in Gynaecological Cancers: External Beam and Brachytherapy

The Radiation Medicine Program at the Princess Margaret Cancer Centre delivers curative radiation therapy (RT) to cancer patients, including those with gynecological cancers, using a combination of external beam radiation therapy (EBRT) and brachytherapy (BT). A key clinical challenge is the accurate accumulation of radiation dose delivered across these modalities. Current methods rely on deformable image registration (DIR), which is hindered by large uncertainties due to anatomical changes caused by the BT applicator, variable bladder and rectum filling, and tumor shrinkage during treatment (Fu et al., 2023). These limitations reduce the accuracy of longitudinal dose accumulation and compromise treatment effectiveness. This project addresses that challenge by developing deep generative models to remove BT applicators from MR images, enabling accurate DIR and dose mapping between EBRT and BT sessions. The partner organization will benefit clinically by improving treatment precision and enabling better-informed re-irradiation strategies. Socially, the project supports safer, more effective cancer therapy, while economically, it may reduce planning errors and treatment complications, ultimately improving healthcare resource efficiency. By advancing AI-driven radiotherapy tools, this project also enhances the partner’s leadership in integrating machine learning into clinical cancer care.

Faculty Supervisor:

Karthik Kuber;Arvind Gupta

Student:

Partner:

Princess Margaret Cancer Centre

Discipline:

Computer science

Sector:

Health and Related Sciences & Technology

University:

University of Toronto

Program: