Abnormality Detection in Brain Tissue: A deep Learning Approach

Glioblastoma, the most prevalent and aggressive form of brain cancer in adults, presents substantial challenges in both diagnosis and treatment. A crucial prognostic factor and an indicator of chemotherapy responsiveness in this context is the identification of MGMT promoter methylation within the tumour. However, the existing genetic analysis procedures for brain cancer involve invasive methods and time-intensive processes. The primary aim of this initiative is to enhance diagnosis and treatment strategies for glioblastoma patients, with a focus on reducing the reliance on invasive procedures and streamlining the genetic analysis workflow. Through the utilization of radio genomics, the goal is to develop a non-invasive approach that predicts the genetic profile of the tumour exclusively through imaging to tackle these challenges, The Radiological Society of North America (RSNA) and the Medical Image Computing and Computer Assisted Intervention Society (MICCAI Society) have collaborated to host a competition focused on glioblastoma diagnosis and treatment planning. The competition revolves around the use of MRI scans to construct a model capable of accurately forecasting the genetic subtype of glioblastoma by identifying the presence of MGMT promoter methylation. The positive outcomes derived from this competition are poised to significantly contribute to less invasive diagnostic procedures and more personalized treatment strategies for individuals battling brain cancer. This abstract provides a comprehensive overview of the project’s objectives, the contextual backdrop of the competition, and the potential transformative impact on the management and survival rates of those affected by glioblastoma.