Gray vs. White Matter Segmentation of the Conus Medullaris: Reliability and Variability in Healthy Volunteers

Abstract

BACKGROUND AND PURPOSE Magnetic resonance imaging (MRI)‐derived spinal cord (SC) gray and white matter (GM/WM) volume are useful indirect measures of atrophy and neurodegeneration over time, typically obtained in the upper SC. Neuropathological evidence suggests that in certain neurological conditions, early degeneration may occur as low as the sacral SC. In this study, the feasibility of GM/WM segmentation of the conus medullaris (CM) was assessed in vivo. METHODS Twenty‐three healthy volunteers (11 female, mean age 47 years) underwent high‐resolution 3T MRI of the CM using a 3‐dimensional fast field echo sequence. Reproducibility of the volume measurements was assessed in 5 subjects (2 female, 25‐37 years) by one rater who repeated the analysis 3 times and also with 2 additional raters working independently in order to calculate the intra‐ and interrater coefficient of variation (COV), respectively. Furthermore, the influence of age, gender, spine and SC metrics on tissue‐specific measures of the CM was investigated. RESULTS Volumetric CM analyses (N = 23) for the SC, GM, and WM revealed a mean (SD) total volume of CM‐TV = 1746.9 (296.7) mm3, CM‐GM‐TV = 731.2 (106.0) mm3, and CM‐WM‐TV = 1014.6 (211.3) mm3, respectively. The intra‐rater COV for measuring the CM‐TV and CM‐GM‐TV was 3.38% and 7.42%, respectively; the interrater COV was 3.43% and 10.80%, respectively. Using age, gender, spine and SC metrics in regression models substantially reduced group variability for CM‐TV, CM‐WM‐TV, and CM‐GM‐TV by up to 39.2%, 42.7%, and 21.2%, respectively. CONCLUSIONS The results from this study demonstrate the feasibility of obtaining tissue‐specific volume measurements in the CM by means of MRI with good reproducibility and provide normative data for future applications in neurological diseases affecting the lower SC.

Intersubject Variability and Normalization Strategies for Spinal Cord Total Cross‐Sectional and Gray Matter Areas

“…5 We also show for the first time that the measurement of the spinal canal area performed on spinal cord scans can significantly reduce inter-subject variability of SC areas, particularly when accounting for sex (TCA %RSD reduction by 15.52%, 12.15% for GMA). This finding is in contrast to a recent study performed at the lumbar level of the SC that showed no improvement of TCA inter-subject variability after normalization by spinal canal area, 39 which might be explained by the higher inter-subject variability of cord and spinal canal anatomy at the lumbar compared to the cervical level. Normalizing for head size using the SienaX-derived V-scale also reduced inter-subject variability, but to a lesser extent (10.24% for TCA, 8.28% for GMA).…”

“…The variables non-significantly contributing to a model in an exploratory analysis were excluded, and the effect of the normalization by a particular selected model was evaluated by comparing the relative standard deviations (ratios of the group standard deviation and the respective means, %RSD) between the measured data and the values generated by the respective correction formulas. 38,39 Sample size calculations-To inform the design of research protocols for clinical trials using spinal cord areas as an outcome measure, we performed sample size calculations for two distinct clinical trial scenarios. Sample size calculations (with 90% power at 0.05 statistical significance) were made for the raw cord area values and for the values normalized with the best model, using a two-sided test.…”

“…Biological differences have been described as a source of inter-subject variability, including sex and, to a lesser extent, age. [36][37][38][39] Earlier studies in MS patients sought to reduce variability due to biological differences by applying normalization strategies based on correlations between TCA and brain/skull size in healthy subjects (e.g. total intracranial volume, TICV), [40][41][42][43][44] between cord volume/TCA and lumbar enlargement cord area (LECA) and volume/TCA and number of slices/length of the spinal cord.…”

Early presentation of urinary retention in multiple system atrophy: can the disease begin in the sacral spinal cord?

“…Preganglionic neurons of the IML column of the sacral spinal cord are responsible for the innervation of the detrusor [25] and involvement of this region is likely to be responsible for detrusor failure. Routine MRI sequences of the lower cord are insensitive to identify degenerative changes in the lumbosacral spinal cord; however, high-resolution MR imaging protocols have successfully imaged the lumbosacral enlargement [26] and conus medullaris [27] and changes of grey matter atrophy have been demonstrated in MSA [28].…”

MRI Measurement of Upper Cervical Spinal Cord Cross‐Sectional Area in Children

“…The effect of the normalization by a particular selected model was evaluated by comparing the coefficient of variation (%COV, ratio of the group standard deviation and the respective means expressed as percentage) between the measured data and the values generated by the respective correction formulas 10,15,16 …”