Investigation of a logistic model for T2* dynamic susceptibility contrast magnetic resonance imaging (dscMRI) perfusion studies.

There are a number of T1- and T2-based dynamic contrast-enhanced magnetic resonance imaging pharmacokinetic modeling approaches to study cancer microvasculature. Alternatively, model-free approaches offer an easy, quantitative assessment of microcirculation. In this work, we investigate a 6-parameter model-free approach applied to a T2*-weighted echo-planar imaging bolus response curve. We tested this new approach on a small cohort of patients with clinically diagnosed primary rectal carcinoma before adjuvant chemoradiotherapy and surgical excision. Comparison with healthy muscle tissue shows that logistic parameters P-1/P-2, P-4, and P-5 offer good discrimination between tumor and healthy tissue. Bolus response logistic parameters P4 and P5 have been implicated in previous T1-based works as being important in the assessment of cancer malignancy. Further comparison of T2* parameters with signal attenuation amplitude (maximum signal drop) and percentage baseline signal loss also corroborates the models' ability to quantify the microenvironment.