Imaging genomics is one of the new “hot topics” and “buzzwords” in diagnostic imaging alongside radiomics, computer-aided diagnosis, and deep learning and artificial intelligence. We wanted to test if a very simple, location-based approach would shed new light on the cellular origins of molecular medulloblastoma groups and validate some of the recent discoveries in developmental neurobiology. Our initial results with the wingless (WNT) medulloblastoma group suggest that such an approach actually works, especially in patients with pediatric brain tumors.
We do not expect radiologists to provide a “molecular grade” diagnosis of medulloblastoma in each and every future case, though the location of the tumor may be very characteristic in WNT and sonic hedgehog (SHH) medulloblastoma. More importantly, if radiologists wish to remain participants in the discussion at clinical management meetings and scientific symposia, they need to understand the issues pertinent to the new molecular tumor classifications, including implications for risk stratification and targeted therapies.
The neuroimaging community shows great interest in the new concepts and principles of tumorigenesis and very willingly applies the new knowledge to enhance our understanding of the phenotypic variations of tumors previously believed to represent one entity, such as medulloblastoma.
We believe that the location-based approach to analyzing pediatric brain tumors is worthy of pursuit, and we intend to continue that research and explore the value of new, innovative analytic and display tools. We expect to show the preliminary results arising from broader research based on hundreds of medulloblastomas at the ASNR 56th Annual Meeting in Vancouver.