Impact of truncating diffusion MRI scans on diffusional kurtosis imaging.

Area of Science:

• Neuroimaging
• Quantitative MRI

Background:

• Diffusional kurtosis imaging (DKI) offers advanced characterization of diffusion compared to diffusion tensor imaging (DTI).
• DKI requires longer acquisition times, posing challenges for scan robustness and interruptions.
• Optimizing data acquisition order is crucial for reliable DKI parameter estimation.

Purpose of the Study:

• To evaluate the impact of reduced diffusion MRI scan durations on histogram-based DKI metrics.
• To compare the effectiveness of different data acquisition ordering strategies: electrostatic repulsion model (OptEEM), spherical codes (OptSC), and random truncation (RandomTRUNC).

Main Methods:

• Diffusion multi-shell data from 14 healthy volunteers were reordered using OptEEM, OptSC, and RandomTRUNC strategies.
• Subsets of varying sizes were generated to simulate shortened scans.
• Effects on histogram-based DKI metrics were assessed using tract-based spatial statistics (TBSS) skeletonized maps.
• Simulations and repeated measures ANOVA were used to evaluate subsampling effects on data with varying signal-to-noise ratios (SNRs) and in vivo data.

Main Results:

• Subsampling effects varied by DKI parameter; fractional anisotropy was most stable (up to 5% error), while radial kurtosis was least stable (up to 26% error).
• RandomTRUNC performed worst, while OptEEM and OptSC showed comparable results.
• Histogram peak value was least affected (up to 7% error), while peak height was most affected (up to 11% error).

Conclusions:

• The impact of scan shortening on DKI metrics depends on the specific metric and histogram characteristic.
• Employing optimized acquisition order strategies enhances DKI robustness against scan interruptions or shortening.
• This optimization is advisable for clinical applications requiring efficient MRI acquisition.