Visual rating system for assessing magnetic resonance images: a tool in the diagnosis of mild cognitive impairment and Alzheimer disease

J Comput Assist Tomogr. 2009 Jan-Feb;33(1):73-8. doi: 10.1097/RCT.0b013e31816373d8.

Abstract

Objective: Subjects with mild cognitive impairment (MCI) and mild Alzheimer disease (AD) can be distinguished from elderly subjects with no cognitive impairment (NCI) by the degree of atrophy in the entorhinal cortex (ERC) and the hippocampus (HPC), quantified by volumetric magnetic resonance image (MRI) studies. Because volumetric MRI requires rigorous standards for image acquisition and analysis and is not suitable for routine clinical use, we have used calibrated visual rating to measure atrophy in the ERC, HPC, and perirhinal cortex (PRC) and evaluated its utility in the diagnosis of very early AD.

Methods: Thus far, visual rating methods, which have been found to be reliable and sensitive only for measurement of atrophy of the HPC or for the entire medial temporal region, have been found to be relatively insensitive for discriminating mild AD from elderly NCI subjects. We have developed a computer-based visual rating system (VRS) using reference images for calibration of atrophy ratings in several discrete brain regions, including the ERC, HPC, and PRC. The VRS reference images facilitate training of raters and promote standardization of all atrophy ratings. Interrater and intrarater reliability measurements were assessed; subsequently, the ability of VRS to discriminate the diagnoses among 73 elderly subjects was studied (NCI = 27, MCI = 23, and AD = 23).

Results: Kappa values for interrater reliability of the ERC, HPC, and PRC were between 0.75 and 0.94, and for intrarater reliability, they were between 0.84 and 0.93, indicating that VRS enables highly reliable ratings to be obtained. Atrophy ratings in the ERC, HPC, and PRC distinguished AD from NCI subjects but did not distinguish AD from MCI subjects who tended to have intermediate levels of atrophy. Right and left ERC ratings and the right HPC rating distinguished MCI from NCI subjects.

Conclusions: The visual rating system is the first semiquantitative method that enables reliable measurements of ERC atrophy, and ERC measurement was found to be the best discriminator between MCI and NCI subjects. Visual rating system is a user-friendly tool that can allow a radiologist or a clinician to use structural MRI scans to be used as a biomarker in the diagnosis of prodromal AD.

Publication types

  • Research Support, N.I.H., Extramural
  • Research Support, Non-U.S. Gov't

MeSH terms

  • Aged
  • Aged, 80 and over
  • Algorithms*
  • Alzheimer Disease / complications
  • Alzheimer Disease / diagnosis*
  • Brain / pathology*
  • Cognition Disorders / complications
  • Cognition Disorders / diagnosis*
  • Female
  • Humans
  • Image Enhancement / methods
  • Image Interpretation, Computer-Assisted / methods*
  • Male
  • Reproducibility of Results
  • Sensitivity and Specificity