Tumor heterogeneity, may give insight into natural selection through detection of tumor sub-regions, referred as imaging habitats. We used statistical clustering of multiple pixels based on multiple MRI parameter maps to identify tumor habitats in pre-clinical models of sarcoma and breast cancer using T2, T2*, ADC and three model free parameter maps determined from dynamic contrast enhanced images. MRI-derived habitat maps were determined by clustering multidimensional voxels using a Gaussian mixture model. 3D-printed tumor molds were used to successfully co-register MR imaging slices with their histological habitat-counterparts. Four distinct tumor habitats were detected by MRI and biologically corroborated by histology.
This abstract and the presentation materials are available to members only; a login is required.