A comparison of classification methods for differentiating fronto-temporal dementia from Alzheimer's disease using FDG-PET imaging

Roger Higdon, Norman L. Foster, Robert A. Koeppe, Charles DeCarli, William J. Jagust, Christopher M. Clark, Nancy R. Barbas, Steven E. Arnold, R. Scott Turner, Judith L. Heidebrink, Satoshi Minoshima

Research output: Contribution to journalArticle

70 Citations (Scopus)

Abstract

Flurodeoxyglucose positron emission tomography (FDG-PET) is being explored to determine its ability to differentiate between a diagnosis of Alzheimer's disease (AD) and fronto-temporal dementia (FTD). We have examined statistical discrimination procedures to help achieve this purpose and compared the results to visual ratings of FDG-PET images. The methods are applied to a data set of 48 subjects with autopsy confirmed diagnoses of AD or FTD (these subjects come from a multi-centre collaborative study funded by the National Alzheimer's Coordinating Center). FDG-PET images are composed of thousands of voxels (volume elements) so one is left with a situation where there are vastly more variables than subjects. Therefore, it is necessary to perform a data reduction before a statistical procedure can be applied. Approaches using both the entire image and summary statistics calculated on a number of volumes of interest (VOI) are examined. We performed the data reduction techniques of principal components analysis (PCA) and partial least-squares (PLS) on the entire image and then used linear discriminant analysis (LDA), quadratic (QDA) or logistic regression (LR) to classify subjects as having AD or FTD. Some of these methods achieve diagnostic accuracy (as assessed by leave-one-out cross-validation) that is similar to visual ratings by expert raters. Methods using PLS appear to be more successful. Averaging or using VOI data may also be helpful.

Original languageEnglish (US)
Pages (from-to)315-326
Number of pages12
JournalStatistics in Medicine
Volume23
Issue number2
DOIs
StatePublished - Jan 30 2004

Fingerprint

Positron Emission Tomography
Dementia
Alzheimer's Disease
Positron-Emission Tomography
Alzheimer Disease
Imaging
Data Reduction
Partial Least Squares
Least-Squares Analysis
Entire
Diagnostic Accuracy
Voxel
Logistic Regression
Discriminant Analysis
Differentiate
Cross-validation
Aptitude
Principal Component Analysis
Discrimination
Averaging

Keywords

  • Alzheimer's disease
  • Discrimination analysis
  • FDG-PET imaging
  • PCS
  • PLS

ASJC Scopus subject areas

  • Epidemiology

Cite this

A comparison of classification methods for differentiating fronto-temporal dementia from Alzheimer's disease using FDG-PET imaging. / Higdon, Roger; Foster, Norman L.; Koeppe, Robert A.; DeCarli, Charles; Jagust, William J.; Clark, Christopher M.; Barbas, Nancy R.; Arnold, Steven E.; Turner, R. Scott; Heidebrink, Judith L.; Minoshima, Satoshi.

In: Statistics in Medicine, Vol. 23, No. 2, 30.01.2004, p. 315-326.

Research output: Contribution to journalArticle

Higdon, R, Foster, NL, Koeppe, RA, DeCarli, C, Jagust, WJ, Clark, CM, Barbas, NR, Arnold, SE, Turner, RS, Heidebrink, JL & Minoshima, S 2004, 'A comparison of classification methods for differentiating fronto-temporal dementia from Alzheimer's disease using FDG-PET imaging', Statistics in Medicine, vol. 23, no. 2, pp. 315-326. https://doi.org/10.1002/sim.1719
Higdon, Roger ; Foster, Norman L. ; Koeppe, Robert A. ; DeCarli, Charles ; Jagust, William J. ; Clark, Christopher M. ; Barbas, Nancy R. ; Arnold, Steven E. ; Turner, R. Scott ; Heidebrink, Judith L. ; Minoshima, Satoshi. / A comparison of classification methods for differentiating fronto-temporal dementia from Alzheimer's disease using FDG-PET imaging. In: Statistics in Medicine. 2004 ; Vol. 23, No. 2. pp. 315-326.
@article{5e6cd963761a4d53a7ecd30579e18c81,
title = "A comparison of classification methods for differentiating fronto-temporal dementia from Alzheimer's disease using FDG-PET imaging",
abstract = "Flurodeoxyglucose positron emission tomography (FDG-PET) is being explored to determine its ability to differentiate between a diagnosis of Alzheimer's disease (AD) and fronto-temporal dementia (FTD). We have examined statistical discrimination procedures to help achieve this purpose and compared the results to visual ratings of FDG-PET images. The methods are applied to a data set of 48 subjects with autopsy confirmed diagnoses of AD or FTD (these subjects come from a multi-centre collaborative study funded by the National Alzheimer's Coordinating Center). FDG-PET images are composed of thousands of voxels (volume elements) so one is left with a situation where there are vastly more variables than subjects. Therefore, it is necessary to perform a data reduction before a statistical procedure can be applied. Approaches using both the entire image and summary statistics calculated on a number of volumes of interest (VOI) are examined. We performed the data reduction techniques of principal components analysis (PCA) and partial least-squares (PLS) on the entire image and then used linear discriminant analysis (LDA), quadratic (QDA) or logistic regression (LR) to classify subjects as having AD or FTD. Some of these methods achieve diagnostic accuracy (as assessed by leave-one-out cross-validation) that is similar to visual ratings by expert raters. Methods using PLS appear to be more successful. Averaging or using VOI data may also be helpful.",
keywords = "Alzheimer's disease, Discrimination analysis, FDG-PET imaging, PCS, PLS",
author = "Roger Higdon and Foster, {Norman L.} and Koeppe, {Robert A.} and Charles DeCarli and Jagust, {William J.} and Clark, {Christopher M.} and Barbas, {Nancy R.} and Arnold, {Steven E.} and Turner, {R. Scott} and Heidebrink, {Judith L.} and Satoshi Minoshima",
year = "2004",
month = "1",
day = "30",
doi = "10.1002/sim.1719",
language = "English (US)",
volume = "23",
pages = "315--326",
journal = "Statistics in Medicine",
issn = "0277-6715",
publisher = "John Wiley and Sons Ltd",
number = "2",

}

TY - JOUR

T1 - A comparison of classification methods for differentiating fronto-temporal dementia from Alzheimer's disease using FDG-PET imaging

AU - Higdon, Roger

AU - Foster, Norman L.

AU - Koeppe, Robert A.

AU - DeCarli, Charles

AU - Jagust, William J.

AU - Clark, Christopher M.

AU - Barbas, Nancy R.

AU - Arnold, Steven E.

AU - Turner, R. Scott

AU - Heidebrink, Judith L.

AU - Minoshima, Satoshi

PY - 2004/1/30

Y1 - 2004/1/30

N2 - Flurodeoxyglucose positron emission tomography (FDG-PET) is being explored to determine its ability to differentiate between a diagnosis of Alzheimer's disease (AD) and fronto-temporal dementia (FTD). We have examined statistical discrimination procedures to help achieve this purpose and compared the results to visual ratings of FDG-PET images. The methods are applied to a data set of 48 subjects with autopsy confirmed diagnoses of AD or FTD (these subjects come from a multi-centre collaborative study funded by the National Alzheimer's Coordinating Center). FDG-PET images are composed of thousands of voxels (volume elements) so one is left with a situation where there are vastly more variables than subjects. Therefore, it is necessary to perform a data reduction before a statistical procedure can be applied. Approaches using both the entire image and summary statistics calculated on a number of volumes of interest (VOI) are examined. We performed the data reduction techniques of principal components analysis (PCA) and partial least-squares (PLS) on the entire image and then used linear discriminant analysis (LDA), quadratic (QDA) or logistic regression (LR) to classify subjects as having AD or FTD. Some of these methods achieve diagnostic accuracy (as assessed by leave-one-out cross-validation) that is similar to visual ratings by expert raters. Methods using PLS appear to be more successful. Averaging or using VOI data may also be helpful.

AB - Flurodeoxyglucose positron emission tomography (FDG-PET) is being explored to determine its ability to differentiate between a diagnosis of Alzheimer's disease (AD) and fronto-temporal dementia (FTD). We have examined statistical discrimination procedures to help achieve this purpose and compared the results to visual ratings of FDG-PET images. The methods are applied to a data set of 48 subjects with autopsy confirmed diagnoses of AD or FTD (these subjects come from a multi-centre collaborative study funded by the National Alzheimer's Coordinating Center). FDG-PET images are composed of thousands of voxels (volume elements) so one is left with a situation where there are vastly more variables than subjects. Therefore, it is necessary to perform a data reduction before a statistical procedure can be applied. Approaches using both the entire image and summary statistics calculated on a number of volumes of interest (VOI) are examined. We performed the data reduction techniques of principal components analysis (PCA) and partial least-squares (PLS) on the entire image and then used linear discriminant analysis (LDA), quadratic (QDA) or logistic regression (LR) to classify subjects as having AD or FTD. Some of these methods achieve diagnostic accuracy (as assessed by leave-one-out cross-validation) that is similar to visual ratings by expert raters. Methods using PLS appear to be more successful. Averaging or using VOI data may also be helpful.

KW - Alzheimer's disease

KW - Discrimination analysis

KW - FDG-PET imaging

KW - PCS

KW - PLS

UR - http://www.scopus.com/inward/record.url?scp=9144238091&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=9144238091&partnerID=8YFLogxK

U2 - 10.1002/sim.1719

DO - 10.1002/sim.1719

M3 - Article

C2 - 14716732

AN - SCOPUS:9144238091

VL - 23

SP - 315

EP - 326

JO - Statistics in Medicine

JF - Statistics in Medicine

SN - 0277-6715

IS - 2

ER -