A set of diagnostic-focused studies and imaging tools aim to improve risk stratification and decision-making before or during treatment. A dual-center MRI radiomics-and-AI nomogram for advanced rectal cancer validated a method to predict preoperative tumor deposits, combining T2-weighted MRI features with deep learning and CA19-9. In pediatric glioma, a locked 15-microRNA diagnostic model reportedly identified diffuse intrinsic pontine glioma with 95.6% accuracy in an independent retrospective test cohort. For prostate cancer, PSMA PET/CT staging reportedly detected recurrence before the Phoenix threshold with reliable performance in a retrospective cohort. Across these reports, the theme is earlier detection and more precise selection of next steps using noninvasive markers, with multiple strategies spanning radiomics, molecular assays, and PET imaging.
Get the Daily Brief