Meeting Banner
Abstract #0166

Assessing Breast Cancer Angiogenesis in vivo: Which MRI Biomarkers Are Relevant?

SUMMA25Eugene Kim1, Jana Cebulla2, B. Douglas Ward3, Kevin Rhie2, Jiangyang Zhang2, Arvind P. Pathak2, 4

1The Whitaker Biomedical Engineering Institute, Johns Hopkins University, Baltimore, MD, United States; 2The Russell H. Morgan Department of Radiology and Radiological Science, Johns Hopkins University; 3Department of Biophysics, Medical College of Wisconsin, Milwaukee, WI, United States; 4The Johns Hopkins University in vivo Cellular and Molecular Imaging Center Program

There is a crucial need for noninvasive biomarkers of breast cancer angiogenesis to evaluate the efficacy of new anti-angiogenic therapies in vivo. Here, we validated in vivo steady-state susceptibility contrast (SSC)-MRI biomarkers of angiogenesis in an orthotopic human breast cancer model against the 3D vascular morphology obtained from high-resolution micro-computed tomography (CT). Based on cross-validation analysis, the simple parameters that only require ΔR2 and ΔR2* measurements were better predictors of their CT analogs than were the complex parameters that require additional measurements. Thus, a stand-alone SSC-MRI experiment provides promising candidates for noninvasive, in vivo biomarkers of breast cancer angiogenesis.

Abstract PDF Presentation Video

Keywords

absolute accurately additionally agreement analogs analyses angiography anti apparent appropriate assessing bars basis better black blood bold breast broadly bulk cancer candidates capable cells coded collectively comparing computation computed concordance contrast correlation cross cure currently decreased density derived detection diffusion directly displayed ease efficacy equivalent error evaluate examined exception excised fractional global good graphs grid hallmark horizontal human important in vivo indicate indicated indicates injection inoculated intercept interval invasive involved johns leave linear makes mammary maps measure measured measures measuring median medical metastasis mice micro modalities model models molecular morphology none noninvasive normalized occurred overestimate parametric phenotype post predicted predictive predictor predictors proportional providing radius reflected regression relevant require resolution respective respectively response root scanner scatter science segmented sensitivity significantly since slice slope spin sprouting square steady still suggests susceptibility table temporal therapeutic therapy treat treating true tumor tumors unit validated various vascular vessel vessels volume ward week whereas whole whose wise zero