Premium
Visual rating and automated brain volumetry for medial temporal lobe atrophy in differential diagnosis and early detection of Alzheimer’s disease
Author(s) -
Zhao Lei,
Liu Wenyan,
Luo Yishan,
Kung Wing Kit,
Xiong Li,
Mok Vincent C.T.,
Shi Lin
Publication year - 2020
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1002/alz.041374
Subject(s) - atrophy , medicine , receiver operating characteristic , dementia , hippocampus , third ventricle , temporal lobe , biomarker , audiology , disease , psychiatry , biochemistry , chemistry , epilepsy
Background Visual rating of medial temporal lobe atrophy (MTA‐scale) has been well recognized as a biomarker for Alzheimer’s disease (AD), and automatic quantification of MTA (QMTA) based on the volume of hippocampus and inferior lateral ventricle (ILV) has been proposed. However, a comprehensive comparison of MTA‐scale, QMTA as well as the components of MTA (i.e. volume of hippocampus and ILV) is still lacking in differential diagnosis or early detection of AD. Also, it’s not clear about the difference of performance between the conventional MTA measures and the recently proposed AD resemblance atrophy index (AD‐RAI) that integrates multiple brain regions more than MTA. Method We included 132 AD, 132 normal control (NC) subjects, and 353 subjects with mild cognitive impairment (MCI) from ADNI. Among the MCI patients, 290 remained stable (MCIs) and 63 converted to AD (MCIc) in a two‐year follow‐up. The AD and NC subjects, and the MCIs and MCIc subjects were matched in age, gender and education, respectively (Table 1). MTA‐scale was visually rated by an experienced neuroradiologist. Baseline T1‐weighted images were processed with AccuBrain to quantify the volume of hippocampus and ILV as well as the AD‐RAI. Receiver operating characteristic (ROC) curve analyses were performed to compare these measured when differentiating AD from NC and MCIc from MCIs. Sub‐analyses were conducted by grouping the study cohort by age (>75 years or not) and gender. The optimal threshold was determined by Youden Index. Result The Spearmen correlation of QMTA and MTA‐scale was 0.805 (p<0.001). Regarding the ROC analyses, the AD‐RAI performed the best in differentiating AD from NC (AUC=0.860, Youden Index=59.1% for all individuals) and MCIc from MCIs (AUC=0.768, Youden Index=41.6% for all individuals), followed by the hippocampal fraction/volume, QMTA, MTA, and ILV fraction/volume (Table 2 and Table 3). For all the MRI‐based measures, the differentiation performance was better for those with age≤75 years in both classifications, and better for men in AD vs. NC and women for MCIc vs. MCIs. Conclusion MTA measures based on automated brain volumetry performed better than visual rating, and the superiority of AD‐RAI over MTA measures needs further validations in a larger population.