Accurate measurement of brain changes in longitudinal MRI scans using tensor-based morphometry.

Area of Science:

• Neuroimaging
• Biomedical Engineering
• Medical Statistics

Background:

• Tensor-based morphometry (TBM) is used to estimate brain changes from serial MRI.
• Previous TBM methods exhibited an unexplained offset in atrophy rates between initial scans.
• This offset could bias power calculations for clinical trials.

Purpose of the Study:

• To address the unexplained offset in TBM-derived atrophy rates.
• To improve the accuracy of TBM for longitudinal neuroimaging analysis.
• To refine sample size estimations for drug trials targeting neurodegenerative diseases.

Main Methods:

• Implemented inverse-consistency enforcement within the TBM framework.
• Analyzed serial MRI data from 431 subjects scanned over 2 years.
• Quantified remaining offsets and attributed them to transitivity errors.

Main Results:

• Enforcing inverse-consistency reduced the initial offset from 1.4% to 0.28%.
• Revised TBM metrics yielded plausible anatomical trajectories.
• Drug trial sample size estimates using the improved TBM were competitive and lower than critique estimates.

Conclusions:

• Inverse-consistency is crucial for accurate TBM in longitudinal studies.
• The refined TBM method provides reliable metrics for clinical trial design.
• Improved atrophy rate estimation enhances the efficiency of neurodegenerative disease drug development.