University of Maryland (Maryland)
Winston Liu, Raya Massoud, Giovanna Brunetto, Daniel S. Reich, Govind Nair, and Steve Jacobson
Introduction: Human T-cell lymphotropic virus-1 associated myelopathy/tropical spastic paraparesis (HAM/TSP) and multiple sclerosis (MS) can both lead to a progressive inflammatory myelopathy. Current advances in MRI have enhanced its use as a tool to monitor such disease processes. The typical radiologic finding in HAM/TSP is thoracic cord atrophy without abnormal cord signal. Although spinal cord atrophy can sometimes be qualitatively detected on routine clinical MRI, a robust and sensitive method to quantify changes in spinal cord size might better characterize the disease severity and progression. Such a method could be used to not only develop an imaging marker for disease, but also to serve as a surrogate end point in clinical trials.
Materials and Methods: MRI was performed on 18 HAM/TSP, 9 MS (8 relapsing remitting, 1 primary progressive) and 10 healthy volunteers on a Siemens 3T Skyra system equipped with head-neck and spine array coils. Cross-sectional area of the spinal cord was measured using a novel and fast algorithm developed in-house that traces axial contours of the spinal cord, perpendicular to the cord-edge along each point from C1 to T10 vertebral body segments in mid-sagittal T1-weighted images. Total post processing time was <5 minutes with minimal user input. HAM/TSP and MS patient disability was measured through the Scripps Neurologic Rating Scale (SNRS), Extended Disability Status Scale (EDSS), Hauser Ambulation Index (AI) and peripheral blood pro-viral loads (PVL) of patients were measured using digital droplet PCR. A pairwise Tukey Honestly Significant Difference test was used to compare cross-sectional areas across disease types. A Pearson correlation test adjusted for multiple comparisons was used to assess relationships between disability scores and cross- sectional area and a p<0.05 was considered significant.
Results and Discussion: Average cross-sectional area in both the T-spine and C-spine were significantly lower in HAM/TSP patients (c-spine: 50.75 ± 10.02 mm2; T-spine: 24.76 ± 5.01 mm2) as compared to healthy controls (C- spine: 72.02 ± 5.792 mm2, p<0.0001; T-spine: 38.8 ± 6.048 mm2, p<0.0001). However, cross-sectional area was only significantly different in the T-spine when comparing HAM/TSP with MS patients (C-spine: 61.84 ± 9.54
mm2; T-spine: 34.84 ± 5.72 mm2, p<0.001). In HAM/TSP, the cross-sectional areas from multiple cord segments
correlated with disease duration (C1-C3, C6, T1-T2), EDSS (T2), and AI (T2, T6). Trends were observed between HAM/TSP average cross-sectional area and SNRS, but no correlation was observed between HAM/TSP cross-sectional area and PVL. MS cross-sectional area correlated with SNRS (C1-C5, T1-T7) and EDSS (C1-C3, T1-T6). Imaging artifacts due to swallowing (typically affecting C6-T1), and abdominal motion (typically affecting T9-T12) increased the unreliability of the area calculation.
Conclusions: These results suggest that the pattern of spinal cord tissue damage is specific to the underlying inflammatory disease, a finding that has direct implications for the use of average cross-sectional spinal cord area as a surrogate end point for clinical trials. This comparative study shows the utility of investigating the entire cord, which will improve our understanding of the patterns of cord atrophy in inflammatory diseases. Furthermore, longitudinal studies are underway to help improve our understanding of the disease process.
Acknowledgements: This study was supported by the Intramural Research Program of NINDS. Special thanks to the NINDS Summer Internship Program, and Kory Johnson for his advice relating to the statistics.
Last updated November 27, 2013