Encrypted login | home

Program Information

Comparison of Cerebral Blood Volume Estimated by DSC- and DCE-MRI

M Aryal

M Aryal*, Y Cao , The University of Michigan, Ann Arbor, MI


SU-G-IeP1-4 (Sunday, July 31, 2016) 4:00 PM - 4:30 PM Room: ePoster Theater

Purpose: The cerebral blood volume (CBV) is commonly estimated from T2*/T2-weighted dynamic susceptibility contrast (DSC) MRI acquired by an EPI sequence. There are several challenges with DSC imaging, including geometric distortion, limited brain spatial coverage, and unreliable arterial input function (AIF) due to the partial volume effect. However, CBV can be estimated from T1-weighted dynamic contrast enhanced (DCE) MRI. This study aimed to examine whether CBV estimated from DCE-MRI is comparable to one from DSC-MRI.

Methods and Materials: Of 16 patients with brain metastases, both DCE and DSC series were acquired on a 3T scanner. 3D T1-weighted DCE images were acquired using a gradient-echo (GRE) sequence, while 2D multi-slice DSC (14-19 slices) were obtained using a GRE EPI sequence. CBV maps were generated from DCE images by the modified Toft model and from DSC images by a pharmacokinetic model with leakage correction. After image registration, CBV values estimated by the two methods were compared in two volumes of interest (VOIs) of normal appearing gray matter and white matter (NAGMWM) and in metastatic lesions at both voxel-level and the VOI-level.

Results: In the two normal tissue VOIs, means (±std) of CBV were 2.04(±0.45)% and 1.86(±0.45)% from DCE imaging and 1.95(±0.46)% and 1.82(±0.44)% from DSC imaging, which were not significantly different in either VOIs. The voxel-by-voxel correlations within the two normal tissue VOIs led to averaged R=0.53 and 0.59 across 16 patients, suggesting high correlation between the two methods.

Conclusion: The CBV estimated from DCE imaging was highly comparable to one from DSC imaging. The CBV estimated from DCE imaging is less affected by the susceptibility effect, and can cover the whole brain, which could be an alternative approach for the CBV estimation.

Funding Support, Disclosures, and Conflict of Interest: This research work was supported by NIH grant: U01 CA183848

Contact Email: