Predicting Treatment Response to 1st-Line Pembrolizumab in Highly Expressing PDL1 Advanced NSCLC Patients using Artificial Intelligence Models

Late-stage lung cancer patients with non-small cell lung cancers who demonstrate high levels of PDL1 expression have better outcomes when placed on a combined treatment program of both immunotherapy and chemotherapy. There is a subset of patients that respond well to pembrolizumab, an immunotherapy drug, alone as an initial treatment path, but there are no clinical methods or biomarkers able to identify good candidates for this approach. A preliminary study with a simple discriminating model pipeline identified 10 features from CT scans that can identify which patients would respond well to pembrolizumab. We suppose that we can boost the discriminating ability of these radiological biomarkers through more complex AI model architectures. By following strict and transparent reporting guidelines, we explore the clinical relevance of this model as a tool to assist oncologists in providing personalized cancer care.

Faculty Supervisor:

Calum MacAulay

Student:

Partner:

BC Cancer

Discipline:

Life Sciences

Sector:

Health and Related Sciences & Technology; Professional, scientific and technical services

University:

The University of British Columbia

Program: