Understanding prostate cancer progression
- Challenge: Applying computer vision to medical imaging
Prostate cancer can take considerable time to progress to lethal disease and for some typical risk nomograms are inadequate predictors. Recent innovations in imaging such as mpMRI enrich for those at highest risk but do not resolve the challenges of predicting hard clinical endpoints such as lethality from the disease. Oxford researchers are identifying lethality features in PSA detected localised prostate cancer by interrogating archival ProtecT and ProMPT cohorts, and novel 3-D microscopy platforms link unique morphological features, genomic instability, and alterations in genomic loci that researchers have found to associate with poor prognosis disease in ICGC and other large-scale genomics studies.