Predicting Breast Cancer Outcome based on tumour morphology with Explainable AI

Breast cancer (BC) diagnosis and treatment generate a significant and increasing portion of healthcare cost in Canada. Detection of a potential BC by mammogram screening triggers a confirmatory biopsy. When a BC is confirmed, additional tests are performed on the biopsy material to guide therapy. Since 2015 a new expansive (~ 5000.00 CAD per patient) molecular predictive test is also performed to predict the likelihood of chemotherapy benefit and 10-year risk of distant recurrence to inform adjuvant treatment decisions in certain women with early-stage invasive breast cancer. This new test adds financial burden on Canadian health care system. Use of artificial intelligence to predict and guide therapy for breast cancer patients can potentially reduce these costs and turn around time significantly. In this research, Alberta health Services (AHS) is collaborating with Computing Science, University of Alberta (U of A) to design such AI algorithms. While many AI-based algorithms, especially, deep learning-based algorithms lack explanations of the decision made by the AI, this research will focus on explanation capability of deep learning. The proposed MITACS internship will foster this U of A developed expertise along with the pathological and diagnostic acumen of the AHS laboratory personnel.

Faculty Supervisor:

Nilanjan Ray

Student:

Partner:

Alberta Health Services

Discipline:

Computer science

Sector:

Health and Related Sciences & Technology; Public administration; Retail trade

University:

University of Alberta

Program: