Development and Evaluation of Advanced Non-Contrast Perfusion MRI for Monitoring Treatment Response in Brain Metastases

Project Abstract Brain metastases (BM) are the most commonly diagnosed type of central nervous system tumor, more frequent than primary intracranial neoplasms. Progress on various therapies have accelerated over the past decade through ongoing clinical trials, and the historically poor outcomes for patients with BM have been

markedly improved. Contrast-enhanced T1-weighted MRI is routinely applied to depict BM with the size of the enhanced lesions for assessing treatment response. Lesion enhancement due to the disruption of blood-brain barrier is rather nonspecific of the functions of brain tumors. The most studied MRI methodology for perfusion

measurement is dynamic susceptibility contrast perfusion weighted imaging (DSC-PWI), which measures cerebral blood flow (CBF) and cerebral blood volume (CBV). CBV is the widely adopted perfusion measure as a sensitive marker of tumor vascularity. However, its clinical applicability in BM studies is hampered by its lack

of absolute quantification, the contrast-leakage effect, and frequent susceptibility artifacts. Arterial spin labeling (ASL) is ideal for frequent non-invasive longitudinal monitoring of tumor vascularity. The standardized spatially selective ASL technique for CBF mapping is the pseudo-continuous ASL (PCASL) method using a single post-

labeling delay, which may render underestimation of CBF due to transit time delay caused by slow arterial flow typical in elderly patients. Velocity-selective ASL (VSASL) was proposed to remove the time-delay sensitivity. Our group has implemented the first velocity-selective inversion (VSI) based VSASL with 3D segmented

GRASE acquisition and demonstrated its higher sensitivity to perfusion signal over conventional ASL methods. Additionally, our group first developed VSASL based CBV mapping by removing labeling delay, which delivered much higher SNR than ASL based CBF mapping. Furthermore, our preliminary data showed that VSASL

with 3D stack-of-spiral based FLASH acquisition delivered better perfusion image quality with less artifacts than using GRASE, and high temporal resolution potentially allowing adequate retrospective motion correction. The purpose of this study is: Aim 1, to conduct further technical developments for VSASL based CBF and CBV

mapping protocols with accelerated acquisitions; Aim 2, to evaluate the sensitivity of the two optimized VSASL protocols to CBF and CBV changes within a month after the radiation therapy, and assess their early prediction to treatment outcomes; Aim 3, to compare the specificity of VSASL derived CBF and CBV values in the

distinction of metastatic recurrence from radiation-induced effects; Aim 4, to ensure high reproducibility of the VSASL protocols between multiple scanners with different vendors and field strength. By completing the proposed aims, the advanced VSASL based CBF and CBV mapping methods are expected to demonstrate

important values for monitoring treatment response in BM, will be readily available for large-scale clinical studies, and can benefit for studies of all primary and metastatic tumors in the body.