Abstract #2: Diffusion Tensor Imaging As a Neuroprotection Outcome Metric in Multiple Sclerosis: Clinical Trial Sample Size Estimates

Abstract

Background Conventional MRI methods detect areas of brain tissue damage, but are poor measures of the degree of injury. Advanced MRI modalities offer imaging metrics of tissue integrity and may be useful in the proof-of-concept evaluation of putative neuroprotective therapies. Previous studies have shown that DTI can be successfully implemented in a multi-center study with good comparability between centers. However, advanced imaging metrics will only be useful if they provide acceptable sample size estimates, which is currently unknown. Methods A longitudinal MS patient population was used to estimate the sample size needed to measure a putative neuroprotective therapy. Patients were imaged at baseline, 1, 2, 6, and 12 months. Twenty regions of interest (ROIs) were placed on normal-appearance brain regions from baseline images in each subject and were coregistered longitudinally using FSL and AFNI. A hypothetical neuroprotective therapy was assumed to slow the worsening of DTI measures by 30% or 50% relative to those observed in this sample. F tests with multivariate analysis of variance were used to estimate sample size, with correlation over the repeated measured assumed to be 0.85 and α = 0.05. Results For 30% reduction in FA, 74 subjects per group are needed for 80% power, and 94 subjects per group for 90% power. For 50% reduction in FA, 29 subjects per group were needed for 80% power, and 36 subjects per group for 90% power. Additional analyses will include power analyses for other DTI measures and different scanning intervals, as well as additional projections using longitudinal designs models. Conclusions Using a longitudinal dataset, feasible sample size estimates are projected for using DTI in a clinical trial of a putative neuroprotective agent. DTI appears to be a feasible candidate marker of neuroprotection in MS clinical trials. Support This study was supported by grants from the National Institutes of Health ( K23 NS 47211 ) and the National Multiple Sclerosis Society ( RG 3548A2 ).