Evaluating Variability of Commercial Liver Fibrosis Elastography Phantoms

Objective: Liver fibrosis has been found to increase the mechanical stiffness of the liver. To mimic different stages of liver fibrosis, commercially available phantoms (Model 039, CIRS, Inc.) have been produced for clinical quality assurance and research purposes. The purpose of this study was to investigate the mechanical property variability of the phantoms in two lots of CIRS Model 039 phantoms.Methods: Each lot consisted of phantoms of four stiffness types, and there were 8-10 phantoms of each type. Shear wave elastography measurements were conducted on each phantom at 10 different angles. Group velocity meas-urements and phase velocity curves were calculated for every SWE acquisition. Multilevel functional principal component analysis (MFPCA) was performed on phase velocity data, which decomposes each phase velocity curve into the sum of eigenfunctions of two levels. The variance of the component scores of levels 1 and 2 were used to represent inter-phantom and intra-phantom variability, respectively. The 95% confidence intervals of phase velocity in a phantom type were calculated to reflect curve variability.Discussion: The standard deviations of the group velocity for phantoms of any type were less than 0.04 and 0.02 m/s for lots 1 and 2, respectively. For both lots, in every type, the phase velocity curves of most individual phantoms fall within the 95% confidence interval.Conclusion: MFPCA is an effective tool for analyzing the inter-and intra-phantom variability of phase velocity curves. Given the known variability of a fully tested lot, estimation of the variability of a new lot can be performed with a reduced number of phantoms tested.

Automated summary

This study aimed to investigate the mechanical property variability of two lots of CIRS Model 039 phantoms. Shear wave elastography measurements were conducted on each phantom and multilevel functional principal component analysis was performed. The results showed that the group velocity had low standard deviations and most individual phantoms had phase velocity curves within the 95% confidence interval. Therefore, MFPCA is an effective tool for analyzing the inter- and intra-phantom variability of phase velocity curves.