Loading…
US-IALE 2018 has ended
Back To Schedule
Wednesday, April 11 • 3:30pm - 3:45pm
REMOTE SENSING/IMAGE ANALYSIS: Evolutionary Principals to Find Habitats in Prostate Cancer Imaging

Sign up or log in to save this to your schedule, view media, leave feedback and see who's attending!

AUTHORS: Yoganand Balagurunathan, Ph.D., Parra Andres Nestor, Ph.D. – Cancer Imaging and Metabolism, Radiology; Hong Li, MD, Cancer Imaging and Metabolism, Radiology and Tianjin Medical University and Cancer Hospital, China; Robert Gatenby, MD, H.L. Moffitt Cancer Center; Joel Brown, Ph.D., Department of Ecology, University of Chicago; Robert J. Gillies, Ph.D., Cancer Imaging and Metabolism, Radiology and H.L. Moffitt Cancer Center

ABSTRACT: Cancer Imaging provides a non-invasive diagnostic tool for disease identification and monitoring. Recently, quantification of tumor regions and sub-regions (Radiomics) provides an objective means for relating imaging to patient outcome. We see parallels between regions within tumors and the ecological concept of Biomes. Tumor regions vary in cell density, nutrient profusion, immune infiltration and “physical” characteristics such as hypoxia and pH. Genomics assists in identifying sub clonal populations. We see the structure of tumors as resulting from Darwinian evolution and ecological feedbacks between the evolving cancer cells and their environment including normal cells. Ultimately, intra-tumoral heterogeneities and the evolution of resistance limit the effectiveness of targeted therapies. Functional imaging, such as multi-parametric magnetic resonance (mpMRI) allows us to characterize the cancer tissues diffusion (cell density) and perfusion (blood flow); features are known to differ distinctly from the normal tissues. Limitation of targeted therapies driven by intra-tumoral heterogeneities has been a leading cause for proliferative cancers. We use principles of ecology and evolution to identify habitats in prostate cancer based on imaging of the cancer regions. The Image observed habitats (tumor like) are then characterized by imaging features across a patient population. After removing redundant features in the population, discriminant functional analysis on these features was used to relate the habitats to conventional grade of the cancer. Habitat features identify aggressive forms of prostate cancer (Gleason = 7) from the indolent forms (Gleason = 6), the Area under receiver operator characteristics was 0.72. Adding descriptors from the habitats onto the mpMRI characteristics (T2-weighted) improves the ability to discriminate between high grade prostate cancers (AUC of 0.84). The use of habitats in imaging shows promise in finding cancer sub-populations with distinct morphological characteristics. Which are analogically similar to ecological principals to find terrain suitable for sustenance of species of interest (aggressive cancer).

Wednesday April 11, 2018 3:30pm - 3:45pm CDT
Water Tower Parlor

Attendees (5)