Introduction

Amyotrophic lateral sclerosis (ALS) is progressive and fatal neurodegenerative disorder with TAR-DNA binding protein-43 (TDP-43) inclusion bodies in neurons, which affects upper and lower motor neurons [1]. The revised El Escorial criteria has been accepted as a tool for a clinical diagnosis, and required upper and lower motor neuron sign [2]. However, quite a few patients pathologically diagnosed as ALS showed only lower motor neuron sign during the whole course of disease [3]. Furthermore, there are no biomarkers to track disease progression. To directly address these issues, state-of-the-art magnetic resonance imaging (MRI) techniques have been applied for diagnosis of ALS and assessment of disease progression, such as diffusion weighted images (DWI). The fractional anisotropy (FA) is derived from DWI, and is an index of the degree of directionality of water diffusivity. A voxel-based meta-analysis revealed the significant FA reduction in corticospinal tract in ALS patients [4]. FA values in the widespread brain region including corticospinal tract and frontal lobe significantly correlated with disease severity [5–7]. We hypothesized that the region showing FA correlation with both disease severity and duration could be a more useful biomarker to track disease progression because these two factors are derived from two different sources (i.e. physical examination or interview). To test this hypothesis, we investigated where FA values correlated with both disease severity and duration in ALS patients.

Materials and methods

Subjects

Fourteen patients with ALS [9 men and 5 women, age 64.5 ± 9.2 (mean ± SD)] were diagnosed according to the revised El Escorial criteria for possible, probable-laboratory-supported, probable or definite ALS, and enrolled in this study. Clinical characteristics of these subjects are summarized in Table 1. Informed consent was obtained from all participants under protocols approved by the local ethical committee.

Table 1 Clinical characteristics of ALS patients
Full size table

MRI acquisition and analysis

The present study was performed using a 3.0 T MRI (GE, Milwaukee, WI, USA) with a standard head coil. DWI were acquired with 6 diffusion gradient directions using echo-planar single-shot spin-echo sequence. Scan parameters were as follows: b value, 1000 s/mm2; FOV, 240 mm; matrix, 128 × 128; TR, 9500 ms; TE, 82.3 ms; flip angle 90°; slice thickness, 4.0 mm. In-plane resolution of reconstructed images were 1.9 × 1.9 mm, respectively.

Images were corrected for Eddy current-induced distortions and head motion using the Eddy current correction routine in FSL. Non-brain tissue was removed from the images using the brain extraction tool in the FMRIB software library. After brain extraction, diffusion tensor components for each brain pixel were calculated, and FA maps were determined for each subject using the DTIFIT routine in FSL to fit the diffusion tensor model. B0 maps were registered to the Montreal Neurological Institute (MNI-152: 2 × 2 × 2 mm3) template using a 12-parameter affine transformation (FLIRT) and the resulting transformation was then applied to the FA maps to register them to MNI space. These images were smoothed using 8 × 8 × 8 mm3 kernel [8]. SPM5 (http://www.fil.ion.ucl.ac.uk/spm/software/spm5/) were used to identify the regions showing the significant correlation. The regions were considered significant for both a voxel-level threshold of p < 0.001 and cluster-level threshold of p < 0.05. The sphere-shaped volumes of interest (VOI) were generated with a radius of 5 mm and center at the highest t value voxel in each region detected by SPM analysis using WFU Pickatlas [9]. Finally, Spearman rank correlation coefficient was assessed between FA value in each VOI and ALSFRS-R or disease duration with IBM SPSS Statistics version 21(IBM Corp., Armonk, NY, USA).

Results

Significant negative correlation was found between FA value and disease duration in left and middle supplementary motor area (SMA; Table 2; Fig. 1a, b). There were no regions which showed significant positive correlation (Table 2). In contrast, significant positive correlation was found between FA value and ALSFRS-R in right cerebellum (IX, X), right posterior medulla, left anterior medulla including pyramid and left middle cerebellar peduncle (Table 3). No regions were found which showed significant negative correlation between FA value and ALSFRS-R (Table 3).

Table 2 The clusters which present correlation of FA value with disease duration
Full size table
Fig. 1
figure 1

Fractional anisotropy (FA) value in volume of interest (VOI) in the supplementary motor area (SMA) correlated with disease severity and duration. a, b The VOI in the left SMA in the axial (a) and sagittal view (b). The sphere-shaped VOI was identified as the region where FA value correlated with a revised amyotrophic lateral sclerosis functional rating scale (ALSFRS-R) and disease duration both. c, d Rank correlations between FA value and ALSFRS-R (c), or disease duration (d) in the left SMA VOI. Significant correlations were observed, respectively (rho = −0.59, p = 0.025; rho = −0.69, p = 0.006). The vertical axis depicted FA value, and the horizontal axis depicted ALSFRS-R (c) or disease duration (d)

Full size image
Table 3 The clusters which present correlation of FA value with ALSFRS-R
Full size table

Furthermore, we examined whether FA value correlated with disease duration and ALSFRS-R both in each VOI. In the left SMA VOI, FA value significantly correlated with disease duration and ALSFRS-R both (disease duration, rho = −0.69, p = 0.006; ALSFRS-R, rho = 0.59, p = 0.025; Fig. 1c, d).

Discussion

We identified the region in which FA value correlated with disease duration and ALSFRS-R both in the left SMA. The present finding suggested the possibility that the abnormality in motor-related region could reflect disease progression. Several lines of evidence have supported the notion that the abnormality in the SMA is involved in ALS pathology. ALS patients showed decreased activity in the SMA in the sensorimotor network identified in resting state functional MRI compared to healthy volunteers [10]. Furthermore, cortical thinning was reported in the right SMA in ALS [11]. These results were not surprising because the key protein of ALS, TDP-43, was common to frontotemporal lobar degeneration (FTLD)-TDP and ALS was considered as a same spectrum disorders of FTLD-TDP [12]. It is possible that the potential abnormality of frontal lobe was detected in our study.

Keil and colleagues reported that FA value in the right SMA correlated not with ALSFRS-R but with disease duration [13]. The other study found significant correlation between FA value and ALSFRS in the prefrontal cortex including SMA [6]. Their results, based on only a voxel-level analysis, supported our finding, however, each study found separate correlation (i.e. significant correlation between FA value and disease duration or disease severity). In contrast, less-biased VOI centered at the highest t value voxel allowed for identification of the region which reflected disease duration and severity both, even though the VOI was derived from a voxel-based analysis. Detailed condition was different among studies, such as ALS patients (i.e. we included “possible” “laboratory-supported” “probable” and “definite” patients), disease duration, severity, MRI machine and analysis software. Therefore, other studies detected different regions which presented significant association between FA value and disease duration or severity [6, 7, 13, 14].

In terms of a biomarker, cerebrospinal fluid (CSF) is another candidate [15]. As suggested previously [16], the integration of multimodal methods including MRI and CSF was expected to allow for development of more reliable biomarkers, which could be adjusted for clinical information and genetic data.

Our results raised further question on how low FA value in the SMA was caused and contributed to pathophysiology of ALS. Given glutamate-induced excitotoxicity [1], we speculated that neurons in the SMA were damaged by excitatory projection, and reduced integration of fibers originated from SMA. Further studies are required to determine whether FA value is related to glutamate concentration in the SMA.