AIBIL: Artificial Intelligence in Biomedical Imaging Laboratory: Dr. Davatzikos. University of Pennsylvania

Dr. Davatzikos and his laboratory (AIBIL: Artificial Intelligence in Biomedical Imaging Laboratory) work on various aspects of quantitative image analysis and machine learning (QIML) methods applied to the field of cancer imaging, primarily on gliomas. AIBIL has been using QIML methods to establish precision diagnostic markers for Glioblastoma Multiforme and has shown radiographic subtypes of GBM relate to patient prognosis. Moreover, Dr. Davatzikos and his team have developed methods for estimating tumor infiltration beyond standard tumor margins and have shown that these infiltration maps predict future tumor recurrence. A clinical trial at the Univ. of Pennsylvania is currently treating patients with elevated radiation based on these predictive maps. Moreover, a similar trial of extended surgical resection is being planned in collaboration with the NRG. AIBIL has also established radio genomic markers of various mutations, including EGFRvIII, IDH1, and MGMT promoter methylation. Current emphasis involves better characterization of mutational heterogeneity in GBM, using QIML methods. Dr. Davatzikos has initiated an emerging and evolving international consortium sharing over 3,000 brain MRI scans of patients with GBM and is interested in catalyzing broad-scale data sharing and application of QIML methods to glioblastoma and other cancers.