Bayesian learning algorithms for identifying and classifying heterogeneity of cell types in variety of solid tumors

Medical College Of Wisconsin, Milwaukee WI

Investigators

Linked publications, trials & patents

Abstract

Abstract Solid tumors are comprise the large majority of cancers (>90%). Standard of diagnosis and care often involved serial MR imaging for monitoring tumor response. It is not well understood how heterogeneity at the cellular and molecular levels affects the macroscopic imaging characteristics of these tumors. The long-term goal of this project is to provide computation tools for understanding the histopathologic variation of solid tumors and aid the development of image guided therapies for individualizing solid tumor treatment. The overall objective is to combine radiographic imaging with histopathological samples (i.e., radio-pathomics) to create and validate predictive tools for accurately defining tumor margins and spatial molecular profiles. Our central hypothesis is that microscopic cytological features and spatially dependent molecular profiles are reliably detectable and quantifiable with macroscopic MR imaging and that they can contribute towards the development of adaptive radiation therapy (ART). Two specific aims will objectively test this hypothesis by first determining which microscopic tissue features contribute to distinct measurements with MR imaging, and second, determining the development of fast and accurate computational tools for using MR image guided radiation therapy in clinical settings.