Jacobian integration method increases the statistical power to measure gray matter atrophy in multiple sclerosis

• We compared sample sizes to detect a treatment effect on gray matter atrophy.
• Jacobian integration method outperformed SIENAX, SPM, and FreeSurfer.
• Sample sizes were reduced by 4 to 5 fold with Jacobian integration method.
• Use of Jacobian integration may be advantageous for assessing neuroprotection.

Gray matter atrophy provides important insights into neurodegeneration in multiple sclerosis (MS) and can be used as a marker of neuroprotection in clinical trials. Jacobian integration is a method for measuring volume change that uses integration of the local Jacobian determinants of the nonlinear deformation field registering two images, and is a promising tool for measuring gray matter atrophy. Our main objective was to compare the statistical power of the Jacobian integration method to commonly used methods in terms of the sample size required to detect a treatment effect on gray matter atrophy. We used multi-center longitudinal data from relapsing–remitting MS patients and evaluated combinations of cross-sectional and longitudinal pre-processing with SIENAX/FSL, SPM, and FreeSurfer, as well as the Jacobian integration method. The Jacobian integration method outperformed these other commonly used methods, reducing the required sample size by a factor of 4–5. The results demonstrate the advantage of using the Jacobian integration method to assess neuroprotection in MS clinical trials.