Purpose: Tumor-associated vasculature is distinguished its convolutedness, leakiness, and chaotic architecture, which facilitates the creation of a treatment resistant tumor microenvironment. Measurable differences in these attributes might help stratify patients by potential benefit of systemic therapy. In this work, we present a new category of radiomic quantitative tumor-associated vasculature (QuanTAV) biomarkers and investigate their ability to predict outcomes across multiple cancers, imaging modalities, and treatment regimens. Experimental Design: We isolated tumor vasculature and extracted mathematical measurements of twistedness and organization using routine pre-treatment radiology (computed tomography or contrast-enhanced MRI) from 558 patients total, who received one of four therapeutic intervention strategies for breast (n=371) or non-small cell lung cancer (NSCLC, n=187). QuanTAV response scores and risk scores/groups were derived and then assessed for response and survival prediction for each therapy. Results:Classifiers of QuanTAV measurements significantly (p<.05) predicted response in held out testing cohorts alone (AUCs=0.63-0.71). Similarly, QuanTAV risk scores were prognostic of recurrence-free survival in treatment cohorts for breast cancer chemotherapy-only (p=0.0022, HR=1.25, 95% CI 1.08-1.44, C-index=.66) and NSCLC chemoradiation+surgery (p=0.039, HR=1.28, 95% CI 1.01-1.62, C-index=0.66) cohorts. QuanTAV high/low risk groups were independently prognostic among all treatments, including chemotherapy-only NSCLC patients (p=0.034, HR=2.29, 95% CI 1.07-4.94, C-index=0.62). Conclusions: Across domains, we observed an association of vascular morphology on CT and MRI – as captured by metrics of vessel curvature, torsion, and organizational heterogeneity – and treatment outcome. Our findings suggest the potential of shape and structure of the tumor-associated vasculature as biomarkers for multiple cancers and treatments.

This content is only available via PDF.

Article PDF first page preview

Article PDF first page preview
You do not currently have access to this content.