CT cerebral blood flow maps optimally correlate with admission diffusion-weighted imaging in acute stroke but thresholds vary by postprocessing platform

Abstract

Background and purpose: Admission infarct core lesion size is an important determinant of management and outcome in acute (<9 hours) stroke. Our purposes were to: (1) determine the optimal CT perfusion parameter to define infarct core using various postprocessing platforms; and (2) establish the degree of variability in threshold values between these different platforms.

Methods: We evaluated 48 consecutive cases with vessel occlusion and admission CT perfusion and diffusion-weighted imaging within 3 hours of each other. CT perfusion was acquired with a "second-generation" 66-second biphasic cine protocol and postprocessed using "standard" (from 2 vendors, "A-std" and "B-std") and "delay-corrected" (from 1 vendor, "A-dc") commercial software. Receiver operating characteristic curve analysis was performed comparing each CT perfusion parameter-both absolute and normalized to the contralateral uninvolved hemisphere-between infarcted and noninfarcted regions as defined by coregistered diffusion-weighted imaging.

Results: Cerebral blood flow had the highest accuracy (receiver operating characteristic area under the curve) for all 3 platforms (P<0.01). The maximal areas under the curve for each parameter were: absolute cerebral blood flow 0.88, cerebral blood volume 0.81, and mean transit time 0.82 and relative Cerebral blood flow 0.88, cerebral blood volume 0.83, and mean transit time 0.82. Optimal receiver operating characteristic operating point thresholds varied significantly between different platforms (Friedman test, P<0.01).

Conclusions: Admission absolute and normalized "second-generation" cine acquired CT cerebral blood flow lesion volumes correlate more closely with diffusion-weighted imaging-defined infarct core than do those of CT cerebral blood volume or mean transit time. Although limited availability of diffusion-weighted imaging for some patients creates impetus to develop alternative methods of estimating core, the marked variability in quantification among different postprocessing software limits generalizability of parameter map thresholds between platforms.

Figure 1

Sample image segmentation methodology for right hemispheric infarct. A: Admission DWI lesion (red outline). A mirrored ROI (green outline) over the contralateral uninvolved hemisphere served for normalization of the absolute voxel values. B: Temporally averaged cine CTP image served as a template for the segmentation of GM, WM, and BG (red outline).

Figure 2

A: Sample ROC curves for CTP delineation of DWI defined core using “A-dc” post-processing software (absolute parameter values only; Red: CBF; Purple: CBV*CBF; Green: CBV, and Blue: MTT). B: Bar graph of AUC (area under curve), sensitivity, and specificity (Y-axis) at the optimal ROC operating point for each CTP parameter and post-processing platform (“A-std”, “A-dc” and “B-std”; X-axis), for delineation of core (“r” = relative).

Figure 3

Sample optimal absolute CBF and CBV pixel thresholds (red overlay applied to the temporally averaged cine CTP template) for right hemispheric stroke. A: Admission DWI (left) and temporally averaged CTP template (right); B: Software “A-std”: CBV (left), CBF (right); C: Software “A-dc”: CBV (left), CBF (right); D: Software “B-std”: CBV (left), CBF (right).