New visual rating system for medial temporal lobe atrophy: a simple diagnostic tool for routine examinations

Psychogeriatrics. 2012 Jun;12(2):88-92. doi: 10.1111/j.1479-8301.2011.00390.x.


Aim: We estimated the usefulness of our new scale to rate medial temporal atrophy with short inversion time inversion recovery images.

Methods: Alzheimer's disease (AD) subjects (n= 34) and non-demented subjects (n= 19) were recruited for this study. First, coronal short inversion time inversion recovery images were scanned vertical to the long axis of hippocampus. Next, the single image in which peduncles appeared widest was adopted for estimation. The parahippocampal cerebrospinal fluid space was divided into three parts: the outer, upper and inner parts. The hippocampus was defined as a structure being of equal intensity to grey matter. Two radiologists compared each part of the parahippocampal cerebrospinal fluid space with the hippocampus and rated them on a 0-3 scale. Interrater and intrarater kappa statistics and sensitivity/specificity for the diagnosis of AD were calculated using the scores of the right, left and both sides combined.

Results: There were no significant differences between AD and ND subjects with regards to sex. AD subjects had lower Mini-Mental State Examination scores and were older than non-demented subjects. Interrater and intrarater kappa statistics were 0.52-0.68 and 0.76-0.83, respectively. Sensitivity was 88.2% using the scores of both sides.

Conclusions: Interrater and intrarater agreements were fair to good and good to excellent, respectively. Our new visual rating method detected medial temporal atrophy in AD patients at a highly sensitive rate. As such, we conclude that this visual rating scale is useful for judging medial temporal atrophy simply and objectively in clinical use, and it is helpful in establishing an AD diagnosis.

MeSH terms

  • Aged
  • Aged, 80 and over
  • Alzheimer Disease / diagnosis*
  • Atrophy
  • Dominance, Cerebral / physiology
  • Female
  • Hippocampus / pathology
  • Humans
  • Image Interpretation, Computer-Assisted / methods*
  • Japan
  • Magnetic Resonance Imaging / methods*
  • Male
  • Mental Status Schedule / statistics & numerical data
  • Observer Variation
  • Psychometrics
  • Reference Values
  • Sensitivity and Specificity
  • Temporal Lobe / pathology*