Which is the best advanced MR imaging protocol for predicting recurrent metastatic brain tumor following gamma-knife radiosurgery: focused on perfusion method.

Abstract

High spatial resolution of dynamic contrast-enhanced (DCE) MR imaging allows characterization of heterogenous tumor microenvironment. Our purpose was to determine which is the best advanced MR imaging protocol, focused on additional MR perfusion method, for predicting recurrent metastatic brain tumor following gamma-knife radiosurgery (GKRS). Seventy-two consecutive patients with post-GKRS metastatic brain tumor were enrolled. Two readers independently calculated the percentile histogram cutoffs for normalized cerebral blood volume (nCBV) from dynamic susceptibility contrast (DSC) imaging and initial area under the time signal-intensity curve (IAUC) from DCE imaging, respectively. Area under the receiver operating characteristic curve (AUC) and interreader agreement were assessed. For differentiating tumor recurrence from therapy effect, adding DCE imaging to diffusion-weighted imaging (DWI) significantly improved AUC from 0.79 to 0.95 for reader 1 and from 0.80 to 0.96 for reader 2, respectively. There was no significant difference of AUC between the combination of DWI with DSC imaging and the combination of DWI with DCE imaging for both readers. With the combination of DWI and DCE imaging, the sensitivity and specificity were 86.7 and 88.1% for reader 1 and 90.0 and 85.7% for reader 2, respectively. The intraclass correlation coefficient (ICC) between readers was highest for calculation of the 90th percentile histogram cutoffs for IAUC (ICC, 0.87). Adding perfusion MR imaging to DWI significantly improves the prediction of recurrent metastatic tumor; however, the diagnostic performance is not affected by selection of either DSC or DCE MR perfusion method.