Diagnosis of mild cognitive impairment using multiple neuroimaging modalities in addition to the Mini-Mental State Examination.

Abstract

This prospective study investigated the additional utility of combining multiple imaging modalities with the Mini-Mental State Examination (MMSE) in discriminating individuals with mild cognitive impairment (MCI) from cognitively normal individuals. A total of 103 individuals with amnestic MCI and normal cognitive function were recruited from outpatients; they underwent examination using the MMSE, magnetic resonance imaging, 18 F-fluorodeoxyglucose positron emission tomography and 11 C-Pittsburgh compound B positron emission tomography. Binary logistic regression with receiver operator characteristic analysis was used to assess the combination of the abovementioned single or multiple imaging modalities with the MMSE. Eight models were constructed, and their area under the curve values and misclassification rates in discriminating individuals with amnestic MCI from cognitively normal individuals were evaluated. The addition of each of the imaging modalities improved the discrimination accuracy over that of the MMSE alone; the accuracy obtained with the addition of 18 F-fluorodeoxyglucose positron emission tomography was the highest. The best model with the highest accuracy included a combination of all the imaging modalities in addition to the MMSE (area under the curve value 0.902). The combination of all the imaging modalities in addition to the MMSE could facilitate more accurate diagnosis of amnestic MCI. Of the individual imaging modalities, the highest accuracy resulted from the addition of 18 F-fluorodeoxyglucose positron emission tomography to the MMSE. Geriatr Gerontol Int 2019; 19: 1193-1197.