Intra- and inter-observer variation and accuracy using different shear wave elastography methods to assess circumscribed objects - a phantom study.

ABSTRACT

AIMS: The elastic properties of circumscribed tissues (e.g., tendons, lymph nodes, prostates, brain tumors) are of considerable clinical interest. The purpose of this study was thus to compare the Intra-/Inter-observer variation and accuracy in vitro of point shear wave elastography (pSWE) with that of 2D-SWE and to assess 2D-SWE's precision with variable ROI (vROI) incircumscribed objects. MATERIAL AND METHODS: Round targets (Elasticity QA Phantom Model 049) were examined for varying degrees of stiffness (8, 14, 45, and 80 kPa) and diameters (20/10 mm). Three ultrasound systems and four probes were applied (pSWE Acuson/S3000 9L4/4C1 and Epiq7 C51, 2D-SWE Aplio/500 PVT375BT). Three different ROIs were used, namely fixed ROI (fROI) and variable ROI rectangular-best-fitted ROI, and round-best-fitted ROI. Each measurement was performed twice by four observers. RESULTS: A total of 3,604 measurements were conducted. The intra-observer variation of 2D-SWE measurements indicated better agreement (Intraclass Correlation Coefficient (ICC) = 0.971; 95% CI=[0.945; 0.985]), than for the pSWE measurements (ICC = 0.872; 95% CI=[0.794; 0.92]). With both methods, the shear wave elastography applied showed low inter-observer variation ICC = 0.980; 95% CI=[0.970; 0.987]. However, a significant difference was observed between fROI (pSWE) and vROI (2D-SWE) on circumscribed objects in terms of accuracy. The lowest degree of observationerror was detected in situations where the ROI was not "best fitted", but placed within the target of 3mm from the border (target diameter 20mm; mean relative error = 0.15). CONCLUSIONS: When estimating the elastic properties of circumscribed tissues, the different measurement techniques performed by commercial shear wave elastography systems reveal a strong susceptibility for observational errors, depending upon the fixed vs. variable ROI of the pSWE vs. 2D-SWE technique.