Whole-Brain and Regional Brain Atrophy in ORIGINAL RESEARCH Amyotrophic Lateral Sclerosis D.M. Mezzapesa BACKGROUND AND PURPOSE: Recent evidence from neuropsychologic and neuroimaging studies A. Ceccarelli suggests that central nervous system involvement in amyotrophic lateral sclerosis (ALS) extends beyond motor neurons. Our purpose was to obtain measures of global and regional atrophy in F. Dicuonzo nondemented patients with ALS to assess subtle structural brain changes. A. Carella M.F. De Caro METHODS: MR images, acquired from 16 patients and 9 healthy subjects (HS), were processed by using the Structural Imaging Evaluation of Normalized Atrophy (SIENA) software to estimate whole- M. Lopez brain atrophy measures and the voxel-based morphometry (VBM) method to highlight the selective V. Samarelli volumetric decrease of single cerebral areas. In addition, each subject underwent a neuropsychologic P. Livrea examination. I.L. Simone RESULTS: In patients with ALS, brain parenchymal fraction was slightly lower compared with HS (P ϭ .012), and seemed to be related to the presence of cognitive impairment. Patients showed a gray matter volume decrease in several frontal and temporal areas bilaterally (P Ͻ .001 uncorrected) compared with HS, with a slight prevalence in the right hemisphere. No volume reduction in primary motor cortices of patients was detected. Performances on Symbol Digit Modalities Test were signif- icantly worse in patients compared with HS (P ϭ .025). CONCLUSIONS: The presence of mild whole-brain volume loss and regional frontotemporal atrophy in patients with ALS could explain the presence of cognitive impairment and confirms the idea of ALS as a degenerative brain disease not confined to motor system. myotrophic lateral sclerosis (ALS) is a chronic progressive es.14-15 Nevertheless, MR imaging evidence of nonmotor in- Aneuromuscular disease of unknown etiology character- volvement, particularly in the frontal lobe, derives from brain ized by upper and lower motor neuron degeneration, leading atrophy evaluation rather than signal intensity abnormalities.16 to a progressive muscle weakness of bulbar, limb, thoracic, In recent years, the development of advanced automated 1 and abdominal regions and, eventually, to death. Although imaging analysis, based upon construction of statistical para- BRAIN the motor system is the cardinal target of the neuropathologic metric maps, allowed detailed anatomic studies of brain mor- insult, brain degeneration extends beyond motor neurons, as phometry. Compared with traditional volumetric methods, suggested by the clinical evidence of variable degrees of cogni- which used a preselected, manually delineated region of inter- tive impairment detected in patients with ALS, with typical est, these new tools do not require an a priori assumption of 2 involvement of frontal executive skills. In the 2 largest epide- brain district affected by disease and determine structural ORIGINAL RESEARCH miologic studies, the occurrence of cognitive impairment has anomalies directly from the data in an unbiased way. Few 3 4 been reported in 36% and 51% of patients with ALS. The studies have been performed in patients with ALS, with and clinical association between frontotemporal dementia (FTD) without cognitive changes, that have reported atrophy on a and ALS, together with the evidence of common neuropatho- regional basis.17-19 However, the extent of the extramotor in- logic findings, supports the impression of a continuum among volvement in ALS remains unclear, particularly in absence of 5-7 these entities. cognitive impairment. Neuroimaging studies, both structural and functional, sus- The aim of this study was to quantify volumetric whole- tain the hypothesis of a brain involvement beyond the motor brain and regional brain changes from MR imaging in nonde- areas and the idea of ALS as a multisystem disorder. Func- mented patients with ALS and to assess their neuropsycho- tional imaging showed hypoperfusion in frontal and temporal logic profiles. Two fully automated methods have been used: areas correlated with the impairment of neuropsychologic the Structural Imaging Evaluation of Normalized Atrophy 8-9 functions. MR spectroscopy revealed a relevant N-acetyl- (SIENA) software,20 for whole-brain volume estimation, and aspartate decrease in the primary motor cortex as well as in the the voxel-based morphometry (VBM) method, to highlight frontal lobe, the primary sensory cortex, the superior parietal the selective volumetric decrease of single cerebral areas.21 gyrus, and the anterior cingulate gyrus.10-13 Main MR imaging signal intensity abnormalities have been found on the pyrami- Methods dal tract, such as the presence of hyperintensity in subcortical white matter and the dark line in the precentral gyrus, partic- Subjects ularly on fluid-attenuated inversion recovery (FLAIR) imag- Sixteen patients (9 men; mean age Ϯ SD, 58.6 Ϯ 10.2 years) were recruited from the ALS patient population of the Department of Neu- Received March 4, 2006; accepted after revision April 24. rologic Sciences of our University. All subjects fulfilled El Escorial From the Department of Neurological and Psychiatric Sciences, University of Bari, Bari, criteria for probable or definite ALS,22 had a sporadic form of disease, Italy. and had a mean disease duration of 38.1 (range, 8–89) months. No Address correspondence to Prof. I. L. Simone, Department of Neurological and Psychiatric Sciences, University of Bari, Piazza G. Cesare 12, 70124 Bari, Italy; e-mail: patient had clinically evident dementia according to the DSM-IV cri- [email protected] teria23 and the Mini Mental State Examination (MMSE) score,24 cor- AJNR Am J Neuroradiol 28:255–59 ͉ Feb 2007 ͉ www.ajnr.org 255 rected for age and schooling for Italian population,25 was higher than tiguous sagittal sections (1.00 ϫ 0.97 ϫ 0.97) covering the whole 23 (the most widely diffuse cutoff value).26 In addition, no patient had brain. behavioral impairment or change in personality to rule out initial FTD.27 Patients with several bulbar signs were excluded from the SIENA Software analysis because they were unable to perform neuropsychologic tests. MR imaging analysis was performed by 2 coauthors blinded to clinical Nine age-matched healthy subjects (HS) (6 men; mean age Ϯ SD, and neuropsychologic data. On 3D T1-weighted images, normalized 51.8 Ϯ 12.7 years) were also studied. All subjects were right-handed. volumes of the whole of the brain were obtained by means of the Written informed consent was given by all subjects, and the study was SIENA software, a fully automated and accurate method for measur- 39 approved by the local ethics committee. ing longitudinal and cross-sectional changes in brain volume, which is freely available as part of the FMRIB Software Library (http:// Clinical and Neuropsychologic Assessment www.fmrib.ox.ac.uk/fsl). In the current study, the cross-sectional Each subject underwent clinical, neuropsychologic, and MR exami- version (SIENAX) was used. This software performs segmentation of nation on the same day. Patients completed the ALS functional rating brain from nonbrain tissue in the head, estimates the outer skull sur- scale (ALSFRS) questionnaire28 and were examined with manual face, and uses these results to drive the spatial transformation to a muscle testing (MMT).29 After a screening global cognitive evalua- standard template to normalize for skull size. Next, a probabilistic tion performed with MMSE to check the absence of dementia, both brain mask derived in standard space is applied to make sure that patients and controls underwent a battery of neuropsychologic tests certain structures such as eyes/optic nerves have not been included in exploring the following domains: 1) long-term memory (Spinnler the brain segmentation. Finally, MR imaging random field model- 40 Prose Memory Test [SPMT])30; 2) working memory (Brown-Peter- based segmentation is used to segment the brain image into differ- son Interference Test [BPIT]31-32; 3) language (Verbal Fluency Test ent tissue types, including partial volume models, giving a normalized [FAS])33-34; and 4) sustained and complex attention (Stroop Color/ volume of total brain, gray matter (GM), and white matter (WM) as Word Interference [Stroop] Test)35-36 and oral version of Symbol output. In addition, the brain parenchymal fraction (BPF), which is the ratio of brain parenchymal volume to the intracranial volume, Digit Modalities Test [SDMT].37 We chose motor-free tests to avoid may be calculated. confounding effects of motor impairment. The SPMT provides a measure of recall of prose material in condition of voluntary and VBM Analysis planned learning. The 2, 3, and 4 domains are considered elements of The theory and algorithm of VBM were well-documented by Ash- frontal executive function. The BPIT, a measure of working memory, burner and Friston.21 This is a fully automated whole-brain technique records the ability of recall trigrams after various distractor intervals for characterizing regional volume and tissue “concentration” differ- (of 10 or 30 seconds). The verbal fluency test (FAS) requires that the ences in structural MRIs. subject generate as many words as possible beginning with the letters 3D T1-weighted images were analyzed with Statistical Parametric F, A, and S, within 60 seconds for each letter. The Stroop test measures Mapping (SPM2; Wellcome Department of Imaging Neurosciences, the ability to focus attention on an attribute of a complex stimulus (eg, London, UK; http://www.fil.ion.ucl.ac.uk/spm)41 running on Matlab the color in which words are written) ignoring another competing (Math Work, Natick, Mass). The preprocessing was performed ac- attribute (eg, the semantic meaning of the word); the time to com- cording to the optimized VBM protocol described by Good et al.42 plete the requested task and the number of errors made were re- First step is generation of a customized template and prior probability corded. The oral version of the SDMT measures sustained attention, maps (specific for GM, WM, and CSF), obtained by normalizing the concentration, and processing speed.