gms | German Medical Science

Objective: Moyamoya disease (MMD) and atherosclerotic cerebrovascular disease (ACVD) require exact characterization of hemodynamics for adequate therapy. As MMD affects both hemispheres, SPECT is not sufficiently reliable. This study compared perfusion SPECT and PET with O-15-water in the same patients.

Method: Static SPECT and dynamic PET were performed within 1-5d in 5 MMD (18-49y) and 5 ACVD patients (47-67y), both in resting state and after vasodilatory challenge with acetazolamide. PET was analyzed with the Watabe reference tissue method for the regional cerebral blood flow. Vasodilatation and resting images were coregistered, stereotactically normalised and smoothed. The relative cerebrovascular reserve (rCVR) was computed voxelwise according to rCVR = 100* (s*vasodilatation – resting)/resting. The individual scale factor (s) was determined to a mean rCVR of 50%. The resulting rCVR maps were assessed visually and using standard ROIs for ACA, MCA and PCA territory (left/right). Effects of modality and territory on rCVR were tested by the general linear model.

Results: Visual analysis revealed concordant findings of PET and SPECT in 4 patients. PET showed extended and more pronounced reduction of rCVR in 4 patients. Discordant results were observed in 2 MMD patients: rCVR-reduction in PET but not in SPECT in one, no overlap of the pattern of rCVR reduction in PET and SPECT in the other. ROI analysis revealed a significant effect of modality (p=0.022) and a highly significant modality*territory interaction (p=0.000). PET-rCVR was lower than SPECT-rCVR in the ACA and MCA territory, whereas it was higher in the PCA territory.

Conclusions: There was a substantial difference in the cerebrovascular reserve capacity as measured by PET versus SPECT, both visually and quantitatively. The modality*territory interaction suggests that PET is more accurate than SPECT, as MMD often spares the posterior circulation. However, this has to be confirmed in further studies.