Three brain SPECT region-of-interest templates in elderly people: Normative values, hemispheric asymmetries, and a comparison of single- and multihead cameras

The purpose of this study was to generate anatomically guided region-of-int erest (ROI) brain SPECT templates based on scans of elderly healthy volunte ers. We describe normal tracer uptake and hemispheric asymmetries for each of 3 camera systems and compare these characteristics among systems. Method s: (TC)-T-99m-hexamethyl propyleneamine oxime SPECT scans were acquired fro m 28 elderly healthy volunteers (mean age [+/-SD], 70.3 +/- 6.5 y) on a sin gle-head rotating y camera (n = 15) or on dual(n = 18) or triple-head (n = 13) cameras. The average number of counts in each ROI was calculated and re ferenced to counts in a cerebellar ROI, providing semiquantitative regional cerebral blood flow (rCBF) ratios. For the templates and ROI map, base ima ges of a healthy volunteer were obtained with each camera. Data from indivi duals scanned with 2 cameras on the same day (n = 15)were used to evaluate rCBF differences across cameras. For each camera, averaged SPECT templates were made using automated image registration. The base volunteer's T1-weigh ted MR image was converted to stereotactic space with dimensions similar to those of the SPECT templates, and 79 bilateral ROIs were defined. To obtai n ROI rCBF ratios, we aligned individual images to their appropriate templa te and then to this modified MR image. Results: The ROI coefficients of var iation indicated that the fit of the ROIs was acceptable (0.07-0.35). Mean rCBF ratios ranged from 0.57 to 1.0, 0.67 to 1.01, and 0.63 to 1.00 for sin gle-, dual-, and triple-head cameras, respectively. The cuneus, occipital c ortex, occipital pole, middle temporal gyrus, and posterior middle frontal gyrus showed consistent hemispheric asymmetry (right side greater than left side in 83%-100% of individuals). Mean rCBF ratios did not differ between dual- and triple-head cameras, whereas the ratios for single- and dual-head cameras differed significantly (39 ROIs differed), even after smoothing an d filtering the dual-head images to the level of the single-head images. Co nclusion: The use of SPECT templates based on elderly healthy volunteers is an important feature of this technique because most available templates ha ve used young individuals. Another important feature is the use of MR image based ROIs. These procedures are versatile because they use more than 1 cam era. They can easily be implemented in clinical and research settings to de tect camera-specific, abnormal deviations in rCBF ROI ratios and asymmetry magnitudes in diseases associated with aging, such as stroke and dementia.