Radiomics

Radiomics turns medical images into quantitative features—such as tumor shape, texture, and intensity patterns—that can be related to disease biology and patient outcomes. Typical pipelines segment tissues, extract and select features, then use statistical or machine-learning models for tasks including cancer classification, survival prediction, treatment-response assessment, and radiogenomics. Current approaches also combine handcrafted features with deep-learning or foundation-model representations and clinical or genomic data. A central challenge is ensuring features and predictions remain reproducible across scanners, acquisition settings, voxel spacing, segmentation choices, and external patient cohorts.