Sensitivity and specificity of medial temporal lobe visual ratings and multivariate regional MRI classification in Alzheimer's disease

TY - JOUR

T1 - Sensitivity and specificity of medial temporal lobe visual ratings and multivariate regional MRI classification in Alzheimer's disease

AU - Westman, Eric

AU - Cavallin, Lena

AU - Muehlboeck, J-Sebastian

AU - Zhang, Yi

AU - Mecocci, Patrizia

AU - Vellas, Bruno

AU - Tsolaki, Magda

AU - Kloszewska, Iwona

AU - Soininen, Hilkka

AU - Spenger, Christian

AU - Lovestone, Simon

AU - Simmons, Andrew

AU - Wahlund, Lars-Olof

PY - 2011

Y1 - 2011

N2 - Background: Visual assessment rating scales for medial temporal lobe (MTL) atrophy have been used by neuroradiologists in clinical practice to aid the diagnosis of Alzheimer's disease (AD). Recently multivariate classification methods for magnetic resonance imaging (MRI) data have been suggested as alternative tools. If computerized methods are to be implemented in clinical practice they need to be as good as, or better than experienced neuroradiologists and carefully validated. The aims of this study were: (1) To compare the ability of MTL atrophy visual assessment rating scales, a multivariate MRI classification method and manually measured hippocampal volumes to distinguish between subjects with AD and healthy elderly controls (CTL). (2) To assess how well the three techniques perform when predicting future conversion from mild cognitive impairment (MCI) to AD. Methods: High resolution sagittal 3D T1w MP-RAGE datasets were acquired from 75 AD patients, 101 subjects with MCI and 81 CTL from the multi-centre AddNeuroMed study. An automated analysis method was used to generate regional volume and regional cortical thickness measures, providing 57 variables for multivariate analysis (orthogonal partial least squares to latent structures using seven-fold cross-validation). Manual hippocampal measurements were also determined for each subject. Visual rating assessment of MTL atrophy was performed by an experienced neuroradiologist according to the approach of Scheltens et al. Results: We found prediction accuracies for distinguishing between AD and CTL of 83% for multivariate classification, 81% for the visual rating assessments and 89% for manual measurements of total hippocampal volume. The three different techniques showed similar accuracy in predicting conversion from MCI to AD at one year follow-up. Conclusion: Visual rating assessment of the MTL gave similar prediction accuracy to multivariate classification and manual hippocampal volumes. This suggests a potential future role for computerized methods as a complement to clinical assessment of AD.

AB - Background: Visual assessment rating scales for medial temporal lobe (MTL) atrophy have been used by neuroradiologists in clinical practice to aid the diagnosis of Alzheimer's disease (AD). Recently multivariate classification methods for magnetic resonance imaging (MRI) data have been suggested as alternative tools. If computerized methods are to be implemented in clinical practice they need to be as good as, or better than experienced neuroradiologists and carefully validated. The aims of this study were: (1) To compare the ability of MTL atrophy visual assessment rating scales, a multivariate MRI classification method and manually measured hippocampal volumes to distinguish between subjects with AD and healthy elderly controls (CTL). (2) To assess how well the three techniques perform when predicting future conversion from mild cognitive impairment (MCI) to AD. Methods: High resolution sagittal 3D T1w MP-RAGE datasets were acquired from 75 AD patients, 101 subjects with MCI and 81 CTL from the multi-centre AddNeuroMed study. An automated analysis method was used to generate regional volume and regional cortical thickness measures, providing 57 variables for multivariate analysis (orthogonal partial least squares to latent structures using seven-fold cross-validation). Manual hippocampal measurements were also determined for each subject. Visual rating assessment of MTL atrophy was performed by an experienced neuroradiologist according to the approach of Scheltens et al. Results: We found prediction accuracies for distinguishing between AD and CTL of 83% for multivariate classification, 81% for the visual rating assessments and 89% for manual measurements of total hippocampal volume. The three different techniques showed similar accuracy in predicting conversion from MCI to AD at one year follow-up. Conclusion: Visual rating assessment of the MTL gave similar prediction accuracy to multivariate classification and manual hippocampal volumes. This suggests a potential future role for computerized methods as a complement to clinical assessment of AD.

KW - Sensitivity and Specificity

KW - Magnetic Resonance Imaging

KW - Reproducibility of Results

KW - Humans

KW - Alzheimer Disease

KW - Aged

KW - Likelihood Functions

KW - Cognition Disorders

KW - Multivariate Analysis

KW - Temporal Lobe

KW - Cohort Studies

KW - Least-Squares Analysis

KW - Follow-Up Studies

KW - Female

KW - Male

U2 - 10.1371/journal.pone.0022506

DO - 10.1371/journal.pone.0022506

M3 - Article

C2 - 21811624

SN - 1932-6203

VL - 6

JO - PL o S One

JF - PL o S One

IS - 7

M1 - e22506

ER -