Clinical evaluation of the statistical images for the cerebral blood flow SPECT using common normal database constructed with CT based-attenuation correction and resolution compensation: a multicenter study

Objectives: We have established the common normal database (NDB) for statistical analysis of brain perfusion SPECT with triple energy window (TEW) scatter correction, CT based non-uniform attenuation correction (CTAC), and the spatial resolution compensation as a multicenter study. However, the clinical utility has been uncertain. The aim of this study was to evaluate the statistical images using the common NDB with CTAC (CTAC-NDB) method and compare with the conventional NDB (C-NDB) with uniform attenuation correction of Chang's method clinically. Methods: We recruited total 41 patients (20 women and 21 men, mean age, 74.1 ± 5.7 years, MMSE score, 20.8 ± 6.4) with mild cognitive impairment (MCI) or dementias such as Alzheimer's disease (AD), dementia with Lewy bodies (DLB), frontotemporal dementia (FTD), and AD with vascular dementia from three institutions. The 3-dimensional stereotactic surface projections (3D-SSP) technique was used to analyze the data obtained from the 123I-IMP brain perfusion SPECT images comparing with both CTAC-NDB and C-NBD. We assessed each 3D-SSP z-score map visually by two experts how the specific findings such as AD/DLB pattern or FTD pattern changed. In addition, we applied stereotactic extraction estimation (SEE) analysis software to measure the regional z-score and extent as semiquantitative assessment. Results: In the visual assessment, 95% of cases demonstrated clearer finding in parietotemporal association cortex, inferior temporal, frontal, and lateral occipital cortex when we used CTAC-NDB comparing to C-NDB. In contrast, the findings of the medial cerebral regions including precuneus and posterior cingulate became inconspicuous in 68% of cases and 32% of cases did not show any change. In the SEE analysis, the mean z-score of frontal, temporal, parietal cortex with CTAC-NDB was significantly higher in comparison with C-NDB (frontal, parietal; P < 0.0001, temporal; P < 0.05). The extent of frontal, temporal, parietal cortex with CTAC-NDB was also significantly higher, while that of precuneus and posterior cingulate was significantly lower in comparison with C-NDB. These tendencies were similar in three institutions. Conclusion: The common CTAC-NDB with the spatial resolution compensation would be available in multicenter clinically, because the findings of the statistical images in patients with MCI and dementias showed similar tendency.