SEARCH

SEARCH BY CITATION

Keywords:

  • brain;
  • computed tomography;
  • dementia;
  • diagnosis;
  • magnetic resonance imaging;
  • X-ray

Abstract

Background:  The present study was designed to find a good alternative to the voxel-based specific region analysis of Alzheimer's disease (VSRAD) on magnetic resonance images, which has high accuracy for discriminating between very early Alzheimer's disease (AD) patients and healthy controls. Because magnetic resonance imaging is not necessarily available in ordinary psychiatric hospitals in Japan, this type of study is of clinical significance.

Methods:  In 30 psychiatric inpatients with a variety of diagnoses, the mean area of the inferior horn (Area), the mean width of the medial temporal lobe (Width), and their ratio (Area/Width) were measured on computed tomography (CT) section that was parallel to the orbitomeatus line. These three indices of medial temporal atrophy were compared separately with Z-scores from VSRAD, which indicate the degree to which the mean values deviate from control. Specifically, higher Z-scores indicate a smaller volume of the medical temporal lobe.

Results:  The mean (± SD) of the CT indices Area, Width, and Area/Width were 39.6 ± 22.1 mm2, 10.3 ± 3.01 mm, and 4.65 ± 4.03, respectively. The mean (± SD) of the Z-scores from VSRAD was 1.47 ± 0.76. The CT indices were all significantly correlated with the Z-scores while controlling for age (Area, r = 0.38, d.f. = 27, P = 0.04; Width, r = −0.45, d.f. = 27, P = 0.01; Area/Width, r = 0.57, d.f. = 27, P = 0.001). Of the three CT indices, Area/Width had the largest area under the curve (AUC) in receiver operating characteristic (ROC) curve analysis regarding Z-scores ≥1.0 as the gold standard for the detection of subtle medial temporal atrophy.

Conclusion:  Although we cannot conclude that Area/Width from CT scans is a perfect alternative to Z-scores from VSRAD ≥1.0 because of its lower sensitivity, it was found that we can expect Z-scores ≥1.0 at an Area/Width ratio ≥4.0. Further research with a larger sample size and prospective design is required to confirm the usefulness of our CT method in the evaluation of medial temporal atrophy in psychiatric patients.