Introduction

Motor Neuron Disease (MND) is a group of neurodegenerative disorders primarily affecting motor neurons. MND exhibits different phenotypes and can present with a widely variable involvement of upper and lower motor neurons. Moreover, regional variants restricted to the arms, legs or bulbar region as well as different patterns of clinical expression (distal or proximal, symmetric or asymmetric) are well-known [1]. Amyotrophic Lateral Sclerosis (ALS) is the most common MND and is characterized by degeneration of both upper and lower motor neurons. Precentral gyrus susceptibility changes have been investigated in ALS with T2*-weighed imaging, Susceptibility-weighted imaging (SWI) and Quantitative susceptibility mapping (QSM) based on hand-drawn ROIs and/or visual inspection [2,3,4]. Susceptibility increase is observed in both cortex and subcortical white matter probably due to iron overload and myelin content decrease respectively. However, hand-drawn ROIs-based measures and visual inspection-based scoring are strongly user-dependent and time consuming. We developed and applied a fully-automatic image processing pipeline to investigate the susceptibility properties of the precentral gyrus in MND.

Methods

51 MND (61.21 ±9.63 y) and 25 Healthy Controls (HC, 57.32 ±7.30 y) were enrolled and scanned at IRCCS Istituto Auxologico Italiano-San Luca Hospital, Milan (Italy). A 3D sagittal FSPGR BRAVO T1w (TR=8.7ms, TE=3.2ms, TI=450ms; Pixel 0.5x0.5mm, thickness=1mm, spacing=1mm, FA=12°, matrix 256x256) and a spoiled gradient-echo multiecho (TR=39ms, 7 equally spaced echoes centered at 24ms, Pixel 0.47x0.47mm, thickness=1.4mm, spacing=0.7mm, FA=20°, matrix 416x320) whole-brain sequences were acquired at 3T General Electric (GE) SIGNA unit. Images were visually assessed and processed at Neuroradiology Deparment of Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico, Milan (Italy). FSL Brain Extraction Tool provided the brain mask from magnitude image. Phase image and mask image were used to calculate the QSM by using the Matlab toolbox STI Suite [5]. Streaking artifacts reduction (STAR) QSM algorithm was adopted [6]. QSM was coregistered to T1w image with a mutual information-based rigid transformation in SPM12 and automatic segmentation of brain regions was performed in Freesurfer (Fig.1). Precentral gyrus cortex (PreGC) and subcortical white matter (PreGSubcWM) ROIs were extracted (Fig.2). Mean susceptibility and skewness of the susceptibility distribution were calculated in the PreGC and PreCSubcWM ROIs and statistically analyzed in SPSS.

Results

In PreGC, mean susceptibility was higher in MND but not statistically different from HC (p=0.139) while skewness was statistically significantly higher (p=0.002) in MND compared to HC. In PreGSubcWM, mean susceptibility (p=0.005) and skewness (p=0.039) were significantly higher in MND compared to HC. Mean susceptibility in PreGSubcWM showed a significant correlation with disease duration (Sig=0.033, r=0.26) and ALS Functional Rating Scale (ALSFRS, Sig=0.026, r=−0.42).

Discussion

The automatic ROI-based approach allows to obtain measurements that are irrespective of both pathology localization and lateralization. Automatic ROIs are not guided by the pathological changes occuring in precentral gyrus in MND patients that on the contrary may influence the ROI manual drawing. The hand drawn ROI-based approach is time consuming, user-dependent and difficult to perform due to the small size of the target structure leading to poor measure reproducibility. On the other hand, the metric of susceptibility mean calculated on this large automatic cortical ROI may lose significance: the values of the voxels with increased susceptibility are indeed averaged with all the other voxels compounding the bilateral PreGC. In addition, cortical voxels largely suffer from partial volume effects especially in standard resolution scans. On the contrary, skewness of susceptibility distribution in PreGC is sensitive to susceptibility changes in MND measured on the automatically-segmented bilateral PreGC. In PreGSubcWM, which is less affected by partial volume effects, automatically-measured mean susceptibility is able to highlight white matter anomalies likely linked to degeneration of myelinated fibers. The diagnosis of MND is a long process and there is no single definitive test, the process involves careful clinical and neurological examination during a long period of time [7]. An automatic non-invasive tool able to characterize the precentral gyrus with quantification of susceptibility properties of cortex and subcortical white matter would be beneficial in building a biomarker of pathology in MND.

Conclusion

For the first time, a fully-automatic pipeline have been applied to quantitatively study the susceptibility properties of the precentral gyrus in MND. Our study suggests that the building of a MND biomarker might rely on susceptibility skewness in PreGC and susceptibility mean in PreGSubcWM automatically measured on clinical images. The pipeline may be easily adapted to widen the measurements pool and be applied on other neurodegenerative disorders.

Acknowledgements

No acknowledgement found.

References

[1] Jeffrey M. Statland, Richard J. Barohn, April L. McVey, Jonathan Katz, Mazen M. Dimachkie. Patterns of Weakness, Classification of Motor Neuron Disease & Clinical Diagnosis of Sporadic ALS Neurol Clin. 2015 Nov; 33(4): 735–748. [2] Donatelli G, Retico A, Caldarazzo Ienco E, Cecchi P, Costagli M, Frosini D, Biagi L, Tosetti M, Siciliano G, Cosottini M. Semiautomated Evaluation of the Primary Motor Cortex in Patients with Amyotrophic Lateral Sclerosis at 3T. AJNR Am J Neuroradiol. 2018 Jan;39(1):63-69. [3] Vázquez-Costa JF, Mazón M, Carreres-Polo J, Hervás D, Pérez-Tur J, Martí-Bonmatí L, Sevilla T. Brain signal intensity changes as biomarkers in amyotrophic lateral sclerosis. Acta Neurol Scand. 2018 Feb;137(2):262-271. [4] Costagli M, Donatelli G, Biagi L, Caldarazzo Ienco E, Siciliano G, Tosetti M, Cosottini M. Magnetic susceptibility in the deep layers of the primary motor cortex in Amyotrophic Lateral Sclerosis. Neuroimage Clin. 2016 May 2;12:965-969. [5] Li W, Avram AV, Wu B, Xiao X, Liu C. Integrated Laplacian-based phase unwrapping and background phase removal for quantitative susceptibility mapping. NMR Biomed. 2014 Feb;27(2):219-27. [6] Wei H, Dibb R, Zhou Y, Sun Y, Xu J, Wang N, Liu C. Streaking artifact reduction for quantitative susceptibility mapping of sources with large dynamic range. NMR Biomed. 2015 Oct;28(10):1294-303. [7] Simon NG, Huynh W, Vucic S, Talbot K, Kiernan MC. Motor neuron disease: current management and future prospects. Intern Med J. 2015 Oct;45(10):1005-13.