PRESENTATIONS ON AUGMENTED INTELLIGENCE ( AI ) NEURO - ONCOLOGY APPLICATIONS
Virginia Hill, Todd Parrish, and Aggelos Katsaggelos were honored to give presentations on augmented intelligence (AI) neuro-oncology applications at the September 24 and January 24 Lou and Jean Malnati Brain Tumor Institute research retreats. In September, Hill, Parrish, and Katsaggelos prioritized their research presentation on AI models of recurrence and textural MRI titled “Unlocking the Potential of AI.” They reviewed different AI techniques such as deep learning, texture analysis, and random forest models and demonstrated how they could be applied clinically. The future of MR acquisitions was illuminated by quantitative techniques, including synthetic MRI that can detect treatment-induced microscopic changes before clinical imaging. Shortcomings of current AI studies in neuro- oncology and ways to overcome these limitations by integrating gene set enrichment analysis, convolutional neural networks, and radiomics were discussed, along with Northwestern’s unique large brain tumor data sets, innovative treatments, and cutting edge neuropathology techniques. Our site is one of few that performs complete next-generation sequencing and now DNA methylation profiling on all brain tumors.
In January, Hill, Parrish, and Katsaggelos, along with Robbie Cantrell, Eric Russell, Adam Sonabend, Craig Horbinski, Amy Heimberger, Rimas Lukas, Sean Sachdev, and Lee Cooper, built upon this framework by presenting two pitches for funding to the group: “Using AI to Predict Glioma Molecular Markers and Subcategories of Glioblastomas, Analyze the Tumor Microenvironment and Predict Survival” and “True Progression vs. Pseudoprogression in Treated Brain Tumors with an Emphasis on Blood-Brain Barrier Disruptive Therapies Such as the Sonocloud and Immunotherapy.” The first pitch detailed how to use AI to noninvasively gather information needed to prevent sequela of surgery, allow personalized precision treatment planning, and analyze the entirety of the tumor habitat (instead of the small portion that can be examined with biopsy or resection). The second pitch emphasized how AI tools can be used to overcome the shortcomings of MR perfusion in differentiating actual progression from pseudoprogression.
- Virginia Hill, MD
THE READ
8
Made with FlippingBook - Online magazine maker