Quantitative CT Biomarkers to Predict Metastatic RCC Response to Antiangiogenic Therapy
Abstract
Targeted therapies have transformed the treatment of advanced and metastatic renal cell carcinoma (RCC), but not all patients respond favorably. The Vascular Tumor Burden (VTB), a measure of vascularized tumor on CT images, has potential to be a predictive biomarker for metastatic RCC response to targeted agents. The VTB can be easily measured using an augmented intelligence image viewer to standardize image evaluation, capture multiple CT metrics, and generate data for machine-learning algorithms. The study aimed to validate the VTB as a quantitative CT biomarker and develop a machine-learning algorithm that utilizes patient CT images. Patients classified as VTB criteria nonresponders were more likely to experience progression of disease than responders across all treatments. Intraobserver and interobserver agreement for assessing VTB were good. The machine-learning algorithm achieved a C-index of 0.78 for predicting PFS greater than 1 year using patient data and images from those treated with Sunitinib. When applied to axitinib and sorafenib cohorts, the algorithm had slightly reduced predictive accuracy (C-index values of 0.67 and 0.69, respectively). These results suggest that the VTB biomarker has the potential to aid in clinical decision-making for patients with metastatic RCC.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 01, 2023
- Accession Number
- AD1216873
Entities
People
- Andrew D. Smith
Organizations
- University of Alabama