Subjects

Twenty-eight cases with DLB and 42 cases with AD were analyzed retrospectively. All cases were diagnosed by neurologists and there were no organic lesions such as cerebrovascular lesions. As for DLB, all cases were diagnosed as probable DLB based on the diagnostic criteria advocated by the 4th International Workshop [2]. The cases with AD were analyzed by the diagnostic criteria based on National Institute of Neurologic, Communicative Disorders and Stroke AD and Related Disorders Association (NINCDS-ADRDA) research group [15]. All patients were undertaken three-dimensional T1-weighted MRI using 1.5-T MRI scanners within 4 weeks before or after Tc-99m-ECD SPECT studies.

SPECT and eZIS examination

All cases were undertaken Tc-99m-ECD SPECT, and all taken data were retrospectively analyzed statistically using easy Z score Imaging System (eZIS) (PDR radiopharma Inc. Tokyo, Japan) [16].

All SPECT scans were performed using a triple-headed gamma camera (GCA9300R; Canon Medical Systems Corp., Otawara City, Japan) and high-resolution fan-beam collimator. SPECT images were obtained with a 128 × 128 matrix and continuous repetitive data acquisition mode with 120° Rotation, acquisition time was 16 min (3 detector, 30 view/detector, 4repeat 2cycle/repeat, time/cycle 120 s). For SPECT image reconstruction, a Butterworth filter (cutoff frequency, 0.76 cycles/cm; order, 4) was used. Attenuation correction was performed using Chang’s method (µ = 0.15/cm) and scatter correction was performed with a triple energy window (TEW).

After the eZIS analysis, we computed the CIScore [12, 13]. This program also allows statistical parametric mapping results to be incorporated into an automated analysis of the Z score values in the volume of interest (VOI). Using eZIS program, two VOIs were set automatically on the posterior cingulate gyrus (VOI-1) and on the occipital lobe (VOI-2). The CIScore is defined as the ratio of the VOI-2 to the VOI-1, which is automatically calculated.

MRI and VSRAD examination

In the present study, 3-dimensional T1-weighted sagittal images were taken with an Intera Achieva 1.5 Tesla (Philips, Tokyo, Japan) under the following setting: FOV 240, matrix 256 × 256, slice thickness 1.2 mm, 220 slices, TR 10 ms, TE 5 ms, flip angle 25°.

VSRAD, we assessed brain atrophy in local gray matter or white matter was a free software designed to evaluate relative local brain volume of individual patients comparing with brain MRI database for healthy individuals by means of voxel-based morphometry (VBM) [10]. Using this program, the entire brain by 3-demensional T1-weighted image taken with a 1.5-Tesla MRI device is processed with SPM8 (Statistical Parametric Mapping, 2008 Edition) to isolate grey matter based on the anatomical standardization followed by statistical analysis of grey matter density.

The results of analysis are displayed as a colored scale map on the standard brain. These Z-score maps were displayed by overlay on tomographic sections and surface rendering of the standardized brain. We registered target VOIs in the MTL and DBS as specifically atrophied areas in AD and DLB, respectively, according to previous studies [10, 17].

Description of present method and data analyses

First, we computed the CIScore. Although the threshold value is 0.281 in the default setting, 0.26 was the best cut-off value from our receiver operating characteristic (ROC) analysis in the current study of our institute. Thus, we diagnosed as DLB in case with CIScore is 0.260 or less. Then, we calculated the differential diagnostic ability (Sensitivity: Sen, Specificity: Spe, Accuracy: ACC) of CIscore.

Second, we evaluated the Z score of the volumes of interest using VSRAD. The obtained data were analyzed using software VSRAD advance® on a PC. By the result of ROC analysis in the present study, the threshold values were determined. Namely in MTL, we diagnosed as DLB in case with the Z score is 2.06 or less. As for the ratio of Z score in DBS (gray matter) to MTL, we diagnosed as DLB in case with the value is 0.37 or more. Regarding the ratio of Z score in DBS (white matter) to MTL, we diagnosed as DLB in case with the value is 0.42 or more. Then, we calculated the differential diagnostic ability (Sen, Spe, ACC) by the Z score of MTL and by DBS.

Namely, each parameter and diagnostic criteria were as follows.

(1)

CIScore as a parameter of rCBF SPECT, we diagnosed as DLB in case which is 0.259 or less.

(2)

Z score value of MTL atrophy by MRI, we diagnosed as DLB in case which is 2.05 or less.

(3)

The atrophic ratio of dorsal brainstem to medial temporal gray matter (DBS) by MRI, we diagnosed as DLB in case which is 0.38 or more.

(4)

The atrophic ratio of dorsal brainstem to medial temporal white matter (DBS) by MRI, we diagnosed as DLB in case which is 0.42 or more.

Finally, we compared the clinical parameters and differential diagnostic ability among the single parameter and the combination of these parameters, including and omitting the elderly (over 75 years old). The clinical parameters were statistically evaluated by unpaired t test or chi-square test. The diagnostic accuracy was statistically evaluated using McNemar test. All statistical analyses were performed using StatFlex version 7 (Artek Osaka Japan).