Vessel morphology radiomic phenotypes

Abstract

Biomarkers for treatment response and other outcomes in the treatment of lesions, such as cancerous tumors, are derived from routine clinical medical images, such as computed tomography (CT) and magnetic resonance imaging (MRI) scans, using radiomic machine learning techniques. The biomarkers are based, at least in part, on phenotypes derived from vessel morphology features descriptive of vasculature associated with a lesion shown in the medical images and may also include other radiomic and pathomic types of features. The biomarkers and phenotypes may be used at multiple points in the treatment process, and treatment decisions may be based on changes in radiomic features or phenotypes longitudinally through the course of treatment. The biomarkers and phenotypes may also be used in clinical trials, to establish trial parameters, patient enrollment, drug dosage, and other trial parameters.

Images