Skip to main content

Medical-image Analysis and Statistical Interpretation Lab

Empirical field mapping for gradient nonlinearity correction of multi-site diffusion weighted MRI

Posted by on Wednesday, November 25, 2020 in Diffusion Weighted MRI, Harmonization.

Colin B. Hansen, Baxter P. Rogers, Kurt G. Schilling, Vishwesh Nath, Justin A. Blaber, Okan Irfanoglu, Alan Barnett, Carlo Pierpaoli, Adam W. Anderson, Bennett A. Landman. “Empirical field mapping for gradient nonlinearity correction of multi-site diffusion weighted MRI.” Magnetic Resonance Imaging 2020.

Full Text

Abstract

Background: Achieving inter-site / inter-scanner reproducibility of diffusion weighted magnetic resonance imaging (DW-MRI) metrics has been challenging given differences in acquisition protocols, analysis models, and hardware factors.

Purpose: Magnetic field gradients impart scanner-dependent spatial variations in the applied diffusion weighting that can be corrected if the gradient nonlinearities are known. However, retrieving manufacturer nonlinearity specifications is not well supported and may introduce errors in interpretation of units or coordinate systems. We propose an empirical approach to mapping the gradient nonlinearities with sequences that are supported across the major scanner vendors.

Study Type: Prospective observational study

Subjects: A spherical isotropic diffusion phantom, and a single human control volunteer

Field Strength/Sequence: 3T (two scanners). Stejskal-Tanner spin echo sequence with b-values of 1000, 2000 s/mm2 with 12, 32, and 384 diffusion gradient directions per shell.

Assessment: We compare the proposed correction with the prior approach using manufacturer specifications against typical diffusion pre-processing pipelines (i.e., ignoring spatial gradient nonlinearities). In phantom data, we evaluate metrics against the ground truth. In human and phantom data, we evaluate reproducibility across scans, sessions, and hardware.

Statistical Tests:  Wilcoxon rank-sum test between uncorrected and corrected data.

Results: In phantom data, our correction method reduces variation in mean diffusivity across sessions over uncorrected data (p<0.05). In human data, we show that this method can also reduce variation in mean diffusivity across scanners (p<0.05).

Conclusion: Our method is relatively simple, fast, and can be applied retroactively. We advocate incorporating voxel-specific b-value and b-vector maps should be incorporated in DW-MRI harmonization preprocessing pipelines to improve quantitative accuracy of measured diffusion parameters.

Figure 6. The absolute percent error (APE) in MD is shown for the human subject with one session acquired at isocenter on scanner A and another acquired 6cm superior from isocenter on scanner B. The top plot shows the sagittal and coronal view of the b0 from each session to demonstrate the shift within the scanner. The bottom plots show the APE for nine saggital slices before correction, after correction using the estimated fields, and after correction using the manufacturer specifications. The error before correction is most prominent in the superior regions of the phantom as those were the furthest from isocenter during the second acquisition.
Figure 6. The absolute percent error (APE) in MD is shown for the human subject with one session acquired at isocenter on scanner A and another acquired 6cm superior from isocenter on scanner B. The top plot shows the sagittal and coronal view of the b0 from each session to demonstrate the shift within the scanner. The bottom plots show the APE for nine saggital slices before correction, after correction using the estimated fields, and after correction using the manufacturer specifications. The error before correction is most prominent in the superior regions of the phantom as those were the furthest from isocenter during the second acquisition.