Longitudinal Cortical Atrophy in Amyotrophic Lateral Sclerosis With Frontotemporal Dementia

Frontotemporal dementia (FTD) with amyotrophic lateral sclerosis (ALS) presents with impaired language or behavior and declining motor function. Frontotemporal lobar degeneration with ubiquinated transactivating responsive sequence DNA-binding protein (TDP-43) inclusions is found postmortem in the affected brain areas of patients with ALS, FTD/ALS, and many patients with FTD.¹ Prior magnetic resonance imaging (MRI) observations revealed cross-sectional atrophy in the motor and/or premotor cortices of patients with FTD/ALS,² while a longitudinal study using diffusion tensor imaging revealed corticospinal tract changes.³ We used high-resolution diffeomorphic image normalization^4,5 and serial MRI to provide the first assessment of longitudinal cortical atrophy in patients with FTD/ALS relative to controls.

Methods

Subjects

We contrasted 4 elderly controls with 4 patients with FTD/ALS, performed by an experienced neurologist (M.G.) at the University of Pennsylvania Department of Neurology. Two trained reviewers (M.G. and L.M.) of a consensus committee confirmed the presence of specific diagnostic criteria based on an independent review of the semistructured history, detailed mental status examination, and complete neurological examination. These patients and their legal representatives participated in an informed consent procedure approved by the institutional review board at the University of Pennsylvania. The age-matched (P<.61) patients (mean [SD] age, 61.3 [6.1] years) were right-handed, high school-educated (mean [SD] education, 17.5 [1.9] years), native English speakers with a mean (SD) Mini-Mental State Examination score at the first examination of 27.0(3.2). The second evaluation was conducted a mean (SD) of 5.3 (0.5) months after the first evaluation, and the mean (SD) score at the second evaluation was 21.3(8.5).

Imaging

Baseline and follow-up image acquisitions (Trio 3.0T MRI scanner; Siemens, Munich, Germany) began with a sagittal T1-weighted localizer. A T1 structural axial image was acquired with a repetition time of 1620 milliseconds; TE echo time, 3 seconds; slice thickness, 1 mm; in-plane resolution, 0.9766 mm×0.9766 mm; and field of view, 256×256×192 voxels.

Imaging Normalization and Longitudinal Atrophy Assessment

We use a longitudinal extension⁴ of large deformation tensor-based morphometry (LDTBM)⁵ in this study. This framework first generates an unbiased intra-subject measurement of atrophy from each subject’s baseline and follow-up image.⁴ One high-resolution voxelwise map of annual atrophy for each individual is transferred to a local template space that allows statistical contrast of subject and control longitudinal change via the t test, computed within an explicit gray matter mask.

Results

We rendered cortical regions with significant annual atrophy due to FTD/ALS on the local template in the Figure. Significant effects occur in the premotor cortex, primary motor cortex, and parietal lobe bilaterally in Brodmann areas (BA) 4, 6, and 7. The average annual cortical atrophy over significant voxels in FTD/ALS on the right and left is 8.5% and 7.6%, respectively, in BA4; 8.1% and 5.9% in BA6; and 3.6% and 2.2% in BA7. For all cortices in FTD/ALS, the atrophy rate was 1.0% per year; in elderly controls, 0.25% per year. The local atrophy rate did not correlate with global brain atrophy; the age and global brain atrophy rates did not correlate.

Figure.

The significant areas of annual atrophy in our amyotrophic lateral sclerosis (ALS) population are highlighted and color-coded by neuroanatomical area on the cortex of a population-specific template. Atrophy was measured by a fine-grained quantitative structural analysis based on diffeomorphic image normalization. Such methods quantify structure in the spirit of voxel-based morphometry, but with higher anatomic fidelity and sensitivity.⁴,⁵ Significance is defined as a voxelwise false discovery rate-corrected P value of .05 with contiguous gray matter voxel clusters larger than 500 voxels. These areas indicate regions of cortical gray matter undergoing annual atrophy that are consistently greater in ALS than in the age- and education-matched elderly control population. In contrast, a cross-sectional morphometric comparison of the elderly and motor neuron degeneration cortical volumes in this cohort produced no significant effects.

Comment

We found significant differences in longitudinal cortical atrophy in motor and premotor areas in patients with clinical features of both ALS and FTD. Amyotrophic lateral sclerosis cooccurs with FTD in 5% to 10% of cases. This is owing, in part, to the presence of ubiquitinated TDP-43 underlying both FTD and ALS.⁶ Presumably, regional motor and premotor cortical atrophy reflect the motor and cognitive changes, respectively, that are characteristic of this condition. Additional study is needed to establish whether there are clinically observable changes corresponding to these parietal changes.

Cortical atrophy is thought to be difficult to detect in ALS because the relatively rapid rate of clinical progression minimizes the opportunity for noticeable cortical atrophy to emerge and motor disease limits the practical limitation of follow-up assessments. This may be the first demonstration of longitudinal cortical atrophy in FTD/ALS because normalization with LDTBM reduces residual intersubject variance in neuroanatomy while retaining sensitivity to intrasubject effects, augmenting detection power in neuromorphometry.