Recent Advances in Cancer Research and AI Applications
Expert Perspectives · Waqas Haque, MD, MPH · 2024-11-21
November’s edition of Dr. Waqas Haque’s blog series on artificial intelligence (AI) and technology features new uses of AI in a multidimensional approach to cancer biology, diagnostic capabilities for glioma, and biomarker discoveries for immunotherapy, emphasizing the promise of this cutting-edge field while also underscoring the knowledge that AI should complement rather than replace clinical judgment.
Multidimensional Approach to Cancer Biology
Dr. Eric Topol’s blog post last week covers a number of exciting advancements in spatial biology and tumor omics, in what is part of the push towards “high-resolution humans.” The Human Tumor Atlas Network (HTAN) is delivering groundbreaking insights through spatial omics and 4D mapping of cancer progression. Recent findings reveal that colon cancer originates from multiple cooperating cell clones that consolidate during the pre-cancerous stage. Similarly, research on ovarian cancer identified interleukin-4 as a key driver of immunotherapy resistance, suggesting new therapeutic targets. This multidimensional approach, combining spatial-temporal analysis with multiomics data, marks a significant advance in our understanding of cancer biology and treatment strategies.
Deep Learning for Diagnostic Capabilities in Glioma
Building on these molecular insights, AI is transforming diagnostic capabilities. A blog post last week demonstrates impressive progress in understanding non-invasive glioma diagnosis. Using deep learning for automatic segmentation of MRI scans, researchers achieved remarkable accuracy in distinguishing glioma subtypes, with area under the curve (AUC) scores exceeding 0.9 for key classifications. The model successfully differentiated between IDH-wildtype glioblastoma, IDH-mutant astrocytoma, and IDH-mutant oligodendroglioma with 1p/19q codeletion. This automated approach could significantly enhance preoperative diagnosis and treatment planning.
AI-Based Biomarker Discovery for Immunotherapies
In addition to assisting companion diagnostics, AI also holds promise in immunotherapy biomarker discovery. A comprehensive review of 90 studies reveals accelerating progress in AI-based biomarker discovery for immunotherapy. Non–small cell lung cancer and melanoma lead the research focus, with emerging applications in radiomics, pathomics, and multimodal data integration. While most studies show promise, the field needs prospective trial designs to validate these AI-based biomarkers. The integration of multimodal data is opening new possibilities for meta-biomarkers that could better predict immunotherapy response.
Complementing, Not Replacing, Clinical Judgment
As AI continues to evolve with different applications in oncology, remaining cognizant of its limitations is ever more vital. For example, while AI is reshaping cancer care, current systems cannot fully replicate clinical decision-making. Research shows that incorporating patient-reported outcomes with electronic health record (EHR) data can improve prediction accuracy by 4%. Human-in-the-loop models that combine AI predictions with clinical expertise are emerging as a promising approach. The key for our clinical practice is to recognize that AI should complement rather than replace clinical judgment.
About Dr. Haque
Waqas Haque, MD, MPH, is a Hematology/Oncology Fellow at the University of Chicago. He recently completed his Internal Medicine Residency at New York University (NYU) in a Clinical Investigator Track. Dr. Haque has balanced his patient care work with a variety of research projects. During his fellowship training at the University of Chicago, he plans to further his work in innovative clinical trial design, value-based care delivery to cancer patients, and clinical investigation.