Abstract 14- 1245-1300
Category: Clinical

At the end of the session,
participants will be able to:

  • Understand how AI tools may be used to resolve molecular heterogeneity.
  • Understanding how pharmacogenomic tools can be used to inform potential therapeutic drug sensitivities in GBM

Dr. Anglin Dent is a University of Toronto graduate student pursuing her MSc in Laboratory Medicine & Pathobiology. Her graduate work has focused on developing tools that aim to support and increase efficiency within the established pathology workflow while prioritizing transparency and generalizability. These efforts have led to the development of workflows capable of identifying clinically relevant intra- and inter-tumoral differences across a variety of tumor subclassifications. 


Anglin Dent1, Brian Lam1, Kevin Faust1, Alberto J. Leon3, Phedias Diamandis1
1Princess Margaret Cancer Centre, University Health Network, Toronto, ON, Canada

Target Audience:
Pathologists, Residents

Communicator, Collaborator, Scholar, Professional

Deep learning approaches to deciphering intra-tumoural heterogeneity in glioblastoma


BACKGROUND: Emerging evidence strongly implicates intra-tumoral heterogeneous biology in treatment resistance and
disease progression across many cancer types(1). Using glioblastoma (GBM) as a prototype, I have aimed to leverage the
computational power of Artificial Intelligence (AI) and deep learning (2) to develop an autonomous workflow for the objective
definition and quantification of biologically distinct tumour subpopulations.
OBJECTIVES: I hypothesize that AI-defined tumoral clusters predict spatially distinct molecular profiles and therapeutic
METHODS: I apply our developed image clustering workflows (3) to quantify AI-defined subregions within a clinical cohort of
10 GBM patient tumors and use laser capture microdissection and mass spectrometry-based proteomics to address if AIdefined
subregions show intra-tumoral molecular variation. Further, I leverage existing pharmacogenomic databases (4) and
carry out drug sensitivity and transcriptional clustering to define biomarkers and validate their intra-tumoral expression in my
clinical GBM cohort.
RESULTS: Preliminary data shows that region to region heterogeneity can be found in IDH wild-type GBM using our unbiased
omics approach, in addition to predicting different pharmacogenomic sensitivities.
CONCLUSIONS: This project aims to develop the first AI-driven tool to guide the routine and systematic molecular analysis of
spatial morphogenomic heterogeneity. Further, this tool may have the potential to provide novel approaches for personalized
care by selecting drug combinations that target a larger fraction of a tumor’s true biology.