Robust constrained weighted least squares for in vivo human cardiac diffusion kurtosis imaging

Abstract

SummaryRobust estimation with convexity constraints significantly improves signal fitting for in vivo human cardiac diffusion kurtosis imaging. PurposeCardiac diffusion tensor imaging (cDTI) is an emerging technique to investigate the microstructure of heart tissue. At sufficiently high b-values, additional information on microstructure can be observed, but the data require a representation beyond DTI such as diffusion kurtosis imaging (DKI). cDTI is highly prone to image corruption, which researchers usually attempt to handle with shot-rejection. However, this can be handled more generally with robust estimation techniques. Recent work has also demonstrated the need to perform constrained fitting for DKI, as fitted parameters can otherwise violate necessary constraints on the signal behaviour, causing significant errors in estimated measures. MethodsWe developed robust constrained weighted least squares (RCWLS) by combining robust estimation with convexity constraints specifically for DKI. Using in vivo cardiac DKI data from 11 healthy volunteers collected with a Connectom scanner, we tested various combinations of fitting techniques, with/without robustness and with/without constraints. ResultsRCWLS was the only tested technique that convincingly showed radial kurtosis to be larger than axial kurtosis for all subjects, something that is expected in myocardium due to increased restrictions to diffusion in the plane perpendicular to the primary myocyte direction. RCWLS also showed the best correction of corrupted regions in diffusion parameter maps for individual subjects. ConclusionFitting techniques utilizing both robust estimation and constraints are essential to facilitate applicability of in vivo cardiac DKI.