Grant: Master's Award: Canada Graduate Research Scholarships - Master's program
University of Toronto
$27,000.00
Mapping Spatial Immune Heterogeneity in DIPG By AI
EF 2025-26
Résumé
Diffuse intrinsic pontine glioma (DIPG) is an incurable pediatric brain cancer with 9-11 month median survival. Immunotherapy has cured patients in other cancers but fails in DIPG, attributed to an immune cold tumor microenvironment with minimal immune activity. However, this conclusion comes from bulk sequencing, which averages immune levels across a tumor, and cannot reveal spatial immune patterns. If distinct immune-active warm pockets” exist within cold tumors, they could explain why some patients do respond to immunotherapy, and could reveal therapeutic opportunities. Spatial transcriptomics can detect these patterns but costs over $3,000 per sample, prohibiting large-scale analysis of DIPG.
I will develop an AI algorithm that predicts spatial transcriptomics from histopathology images, enabling analysis of thousands of archival DIPG histopathology images at near-zero cost. I hypothesize that DIPG's cold tumor microenvironment is not spatially uniform, but has immune warm regions with varying immune levels predictable from histopathology images.
I will apply this algorithm to thousands of archival DIPG histopathology images to generate spatial immune maps, then classify tumors as uniformly cold, predominantly warm with cold pockets, or uniformly warm. This will reveal whether DIPG shows consistent spatial organization or heterogeneous patterns explaining variable treatment responses. If heterogeneity exists, patients with immune-warm regions could benefit from localized immunotherapy, while those with uniformly cold tumors might require alternative strategies. This work establishes proof-of-concept that AI can reveal spatial biology invisible to standard sequencing, potentially transforming how we study tumor microenvironments in rare pediatric cancers.
Programme
Master's Award: Canada Graduate Research Scholarships - Master's program
Bénéficiaire
Emplacement
Ontario, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
CIHR Funding Decisions Database
Tel que publié
Id
572656
Name
Gupta Sidharth
Region
Ontario
Country
Canada
Orgtype
Institution Paid; Research Institution
Prcname2
Special Cases - Awards Programs
Projecttitle
Mapping Spatial Immune Heterogeneity in DIPG By AI
Programtype2
Studentship Programs
Orgnameinp2
University of Toronto
Cihramount2
$27,000
Cihrequipment2
$0
Programname2
Master's Award: Canada Graduate Research Scholarships - Master's program
Keyworddelim
Artificial Intelligence; Genetics; Histopathology; Machine Learning; Oncology; Pediatric Brain Cancer; Spatial Transcriptomics
Abstract
Diffuse intrinsic pontine glioma (DIPG) is an incurable pediatric brain cancer with 9-11 month median survival. Immunotherapy has cured patients in other cancers but fails in DIPG, attributed to an immune cold tumor microenvironment with minimal immune activity. However, this conclusion comes from bulk sequencing, which averages immune levels across a tumor, and cannot reveal spatial immune patterns. If distinct immune-active warm pockets” exist within cold tumors, they could explain why some patients do respond to immunotherapy, and could reveal therapeutic opportunities. Spatial transcriptomics can detect these patterns but costs over $3,000 per sample, prohibiting large-scale analysis of DIPG.
I will develop an AI algorithm that predicts spatial transcriptomics from histopathology images, enabling analysis of thousands of archival DIPG histopathology images at near-zero cost. I hypothesize that DIPG's cold tumor microenvironment is not spatially uniform, but has immune warm regions with varying immune levels predictable from histopathology images.
I will apply this algorithm to thousands of archival DIPG histopathology images to generate spatial immune maps, then classify tumors as uniformly cold, predominantly warm with cold pockets, or uniformly warm. This will reveal whether DIPG shows consistent spatial organization or heterogeneous patterns explaining variable treatment responses. If heterogeneity exists, patients with immune-warm regions could benefit from localized immunotherapy, while those with uniformly cold tumors might require alternative strategies. This work establishes proof-of-concept that AI can reveal spatial biology invisible to standard sequencing, potentially transforming how we study tumor microenvironments in rare pediatric cancers.
Competitiondate
202512
Approvedterm2
1 yr 0 mth
Pinamesdelim
Gupta, Sidharth
Approvedterm
1 an 0 mois
Cihrcontribution2
~ $27,000~ $0~