Automated cross-sectional and longitudinal hippocampal volume measurement in mild cognitive impairment and Alzheimer's disease

Abstract

Volume and change in volume of the hippocampus are both important markers of Alzheimer's disease (AD). Delineation of the structure on MRI is time-consuming and therefore reliable automated methods are required. We describe an improvement (multiple-atlas propagation and segmentation (MAPS)) to our template library-based segmentation technique. The improved technique uses non-linear registration of the best-matched templates from our manually segmented library to generate multiple segmentations and combines them using the simultaneous truth and performance level estimation (STAPLE) algorithm. Change in volume over 12months (MAPS-HBSI) was measured by applying the boundary shift integral using MAPS regions. Methods were developed and validated against manual measures using subsets from Alzheimer's Disease Neuroimaging Initiative (ADNI). The best method was applied to 682 ADNI subjects, at baseline and 12-month follow-up, enabling assessment of volumes and atrophy rates in control, mild cognitive impairment (MCI) and AD groups, and within MCI subgroups classified by subsequent clinical outcome. We compared our measures with those generated by Surgical Navigation Technologies (SNT) available from ADNI. The accuracy of our volumes was one of the highest reported (mean(SD) Jaccard Index 0.80(0.04) (N=30)). Both MAPS baseline volume and MAPS-HBSI atrophy rate distinguished between control, MCI and AD groups. Comparing MCI subgroups (reverters, stable and converters): volumes were lower and rates higher in converters compared with stable and reverter groups (p< or =0.03). MAPS-HBSI required the lowest sample sizes (78 subjects) for a hypothetical trial. In conclusion, the MAPS and MAPS-HBSI methods give accurate and reliable volumes and atrophy rates across the clinical spectrum from healthy aging to AD.

Figure 1

These examples illustrate the wide range of morphological variation in hippocampi from subjects in the Alzheimer's Disease Neuroimaging Initiative database (http://www.loni.ucla.edu/ADNI/). (A) A large hippocampal cyst and lack of temporal horn and (B) malrotation of the hippocampus (tall and narrow). Atrophy causing changes from (C) normal hippocampus to (D) MCI hippocampus (considerable atrophy) and (E) AD hippocampus (marked atrophy).

Figure 2

Flow chart showing how the best methods and parameters for the automated hippocampal segmentation are selected.

Figure 3

Average Jaccard index of the left hippocampal regions from the baseline images of 15 randomly selected subjects for vote rule, STAPLE and SBA used to assess the optimal number of templates to combine in each method. Error bars denote standard errors.

Figure 4

The largest outlier of the automated hippocampal segmentation in the subset of 15 subjects in terms of difference compared with manual measures. This difference in volume can be largely attributed to the automated region including more of the tail of the hippocampus (lower panel) and more of the medial aspect of body and tail (upper panels) than was included in the manual segmentations.

Figure 5

Automated hippocampal segmentation errors. (A) Thresholding excluding hippocampal tissue, (B) extra-hippocampal tissue included (white and grey matter of the temporal lobe) and (C) exclusion of lateral hippocampal grey matter due to large hippocampal cyst and lack of temporal horn.

Figure 6

Group comparisons using baseline diagnosis (control (n=200), MCI (n=335), AD (n=147)) in box plots*. (a) Unadjusted total (left+right) hippocampal volumes; (b) Unadjusted atrophy rates from automated MAPS-HBSI. *The horizontal line in the box represents the median value, and the box represents the interquartile range (IQR). The whiskers represent the upper and lower adjacent values, which are the highest value not greater than 75th percentile + 1.5 times IQR and the lowest value not less than 25th percentile - 1.5 times IQR. Values outside the whiskers are marked as dots.

Figure 7

MCI subgroup comparisons (reverters (n=8), stable (n=204), converters (n=123)) in box plots. (a) Unadjusted total (left+right) hippocampal volumes; (b) Unadjusted atrophy rates from automated MAPS-HBSI.