Eric Chin, Presenter: Nothing to Disclose

Asim F. Choudhri MD, Abstract Co-Author: Nothing to Disclose

We have previously developed fractional myelination (FM), a histogram-based diffusion tensor imaging (DTI) metric which attempts to quantify global myelin maturity beyond the age of 3 years, which is not possible with conventional MRI. Here we investigate whether FM can be interpreted as fully quantitative 1) across scanners of differing field strength and 2) longitudinally.

Cross-scanner validation: Six months of MRI scans (N=914) in a primarily pediatric population from a single institution were evaluated. Contiguous datasets were then identified for both 1.5T and 3T scanners (from two vendors). Longitudinal follow-up: All patients (N=40) who had multiple MRI scans at least 2 years apart since the start of routine DTI use (July 2011) were identified. Progression of FM over all DTI scans was tracked for these patients. FM calculation: Studies were excluded if there was any definable structural abnormality as determined by neuroradiologist review. All included studies had a volumetric T1 weighted sequence as well as DTI with 12 to 25 non-collinear directions of encoding, a b-value of 1000 msec and a single b-0 acquisition. Registration and segmentation were performed automatically using SPM8. FA was analyzed for intracranial white matter as a whole. FM, a ratio of mature to total white matter volume was then calculated based on the FA histograms of each patient. Nomograms of FM over age using the two scanners were then calculated and compared. Regression was based on an exponential model FM(FA,t)=FMmax – A*e-t/τ with 5th and 95th percentile bounds based on a Student’s t-distribution.

Mean FA and FM both show exponential convergence to adult values with age in all subgroups, in agreement with findings in previous studies. FM shows better contrast-to-noise and better fit to an exponential model than mean FA. Using FM, curves obtained do not differ significantly across scanners or field strengths. FM of patients with follow-up largely tracked predicted percentile curves.

Statistical analysis of histogram-based DTI metrics confirms the ability to follow myelin maturation from infancy through adolescence. FM may serve as the foundation for automated myelination age determination.

Histogram-based DTI metrics offer the ability to follow myelin maturation from birth through adolescence and may serve as the foundation for automated myelination age determination.

Chin, E, Choudhri, A, Myelination Age: Validation of a Histogram-based Fractional Anisotropy Metric across Multiple Scanners and Field Strengths with Longitudinal Follow-up.  Radiological Society of North America 2014 Scientific Assembly and Annual Meeting, - ,Chicago IL. http://archive.rsna.org/2014/14014614.html