Editor's choice: radiomics

Once considered to be of limited clinical utility, medical imaging features are becoming increasingly useful in medical research and diagnostics. Extraction of hundreds of quantitative 'radiomic' features from CT, PET or MR scans, often combined with machine learning techniques, has shed new light on disease characteristics. Radiomic profiling has increasingly been shown to be superior to conventional approaches in predicting patient outcomes. This collection shines a spotlight on some of our most recent papers on this subject.