Anne Martel: Using AI to personalize cancer treatment

2026-06-11 · Source: Vector Institute for Artificial Intelligence · Field: Health & Wellbeing — Medical Devices & Health Technology, Clinical Care & Medical Practice, Health & Medical Research · Depth: Advanced, short

Summary

Anne Martel, a Professor at the University of Toronto and Vector Institute Faculty Member, develops AI systems to personalize cancer treatment by extracting actionable information from medical data. Her research primarily utilizes medical imaging, including digital pathology and radiology, combined with clinical text from patient reports, to offer oncologists predictive insights for treatment planning. Martel's work balances methodological innovation with clinical urgency, applying robust existing methods to establish benchmarks while simultaneously developing new approaches to improve accuracy. A key focus is multimodal integration, where large foundation models trained on digital pathology images are fine-tuned with genomic data and clinical text to guide learning, enabling comprehensive patient understanding. She notes medical imaging AI has shifted from adapting general computer science advances to originating novel methods, like the U-Net architecture. The Vector Institute supports this interdisciplinary research with collaboration, AI engineering, and computational resources.

Key takeaway

For oncologists planning personalized cancer treatment, integrating multimodal AI models that combine imaging, genomic, and clinical text data offers significantly enhanced predictive insights. This approach moves beyond single-data-type analyses, enabling more precise treatment recommendations and improving patient quality of life. AI scientists and Directors of AI/ML should prioritize developing and deploying systems capable of multi-task learning across these diverse data streams to build comprehensive patient understanding.

Key insights

AI integrating multimodal medical data (imaging, genomics, text) enables personalized cancer treatment and advances methodological innovation.

Principles

• Balance novel AI methods with urgent clinical problem-solving.
• Multimodal data integration yields comprehensive patient understanding.
• Domain-specific AI research can drive broader methodological innovation.

Method

Develop AI systems that combine digital pathology, radiology, and clinical text data, adapting foundation models for multimodal, multi-task learning to predict treatment outcomes and survival.

In practice

• Apply existing AI methods to establish clinical benchmarks.
• Tune foundation models with genomics and clinical text.
• Integrate imaging, genomic, and clinical data sources.

Topics

• Personalized Cancer Treatment
• Medical Imaging AI
• Digital Pathology
• Multimodal Data Integration
• Foundation Models
• U-Net Architecture

Best for: AI Scientist, Research Scientist, Director of AI/ML

Related on AIssential

• 🔬 Training Transformers to solve 95% failure rate of Cancer Trials — Ron Alfa & Daniel Bear, Noetik — Improving cancer drug success requires advanced AI models for precise patient selection based on multimodal biological data.
• Sasi Kumar Kolla Examines Multimodal Foundation Models for Precision Medicine Research — Multimodal foundation models can integrate diverse biomedical data for precision medicine, overcoming single-modality limitations.
• BreastGPT: A Multimodal Large Language Model for the Full Spectrum of Breast Cancer Clinical Routine — Workflow-aligned data and cross-scale visual modeling are critical for clinically grounded medical MLLMs.
• MedRLM: Recursive Multimodal Health Intelligence for Long-Context Clinical Reasoning, Sensor-Guided Screening, Evidence-Grounded Decision Support, and Community-to-Tertiary Referral Optimization — MedRLM enables recursive, multimodal clinical reasoning by treating patient data as an inspectable external environment.
• OMIDIENT: Multiomics Integration for Cancer by Dirichlet Auto-Encoder Networks — OMIDIENT integrates multi-omics cancer data using Dirichlet auto-encoder networks for superior unsupervised analysis and interpretability.
• Trustworthy Privacy-Preserving Multimodal Federated Learning for Personalised Breast Cancer Prediction — Federated learning can enable privacy-preserving, multimodal breast cancer prediction comparable to centralized models.

See Counsel's argued verdicts on the open AI decisions leaders are weighing →

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Vector Institute for Artificial Intelligence.