Head‐to‐Head Comparison of Four Ultrasound Elastography Techniques in Detecting Advanced Fibrosis

ABSTRACT Aims Elastography is an important noninvasive method for diagnosing advanced fibrosis (AF); however, evidence remains insufficient, particularly for acoustic radiation force impulse (ARFI) techniques. This research aims to evaluate the diagnostic performance and the optimal cutoff values of four ultrasound‐based elastography techniques for detecting AF in patients with chronic liver disease (CLD), and to assess the agreement among these methods in measuring liver stiffness (LS). Methods Adult inpatients with suspected or confirmed CLD who were scheduled for ultrasound‐guided liver biopsy were prospectively and consecutively enrolled from the hepatology departments of four institutions in Nanjing, Changzhou, and Xuzhou. Agreement in LS measurements among the four elastography techniques was assessed using Bland‐Altman analysis. Diagnostic performance for AF was evaluated using receiver‐operating characteristic (ROC) curve analysis, and comparisons were made using Delong's test. Optimal cutoff values for AF detection were determined using Youden's index with the bootstrap resampling. Results A total of 209 patients with CLD were included, 148 in the main cohort and 61 in the external validation cohort. The main cohort had a median age of 52.0 years (interquartile range [IQR], 12.0) and a median body mass index (BMI) of 24.59 kg/m 2 (IQR, 5.88). AF was diagnosed in 34 patients, corresponding to a prevalence of 23.0% (34/148). LS values measured by ARFI techniques were significantly lower than that obtained by vibration controlled transient elastography (VCTE). The mean differences were −3.82 kPa (95% confidence interval [CI], −4.59 to −3.04 kPa) for point shear wave elastography (pSWE), −3.68 kPa (95% CI, −4.52 to −2.83 kPa) for auto‐pSWE, and −4.27 kPa (95% CI, −5.17 to −3.38 kPa) for two‐dimensional shear wave elastography (2D‐SWE). For the diagnosis of AF in the main cohort, VCTE demonstrated the highest diagnostic accuracy, with an area under the ROC curve of 0.842 (95% CI, 0.773–0.896), significantly outperforming all other ultrasound‐based elastography techniques (0.713 [95% CI, 0.633–0.785] for pSWE, 0.685 [95% CI, 0.604–0.759] for auto‐pSWE, and 0.728 [95% CI, 0.649–0.798] for 2D‐SWE; all p < 0.05). The optimal cutoff values for diagnosing AF in the main cohort was ≥ 12.0 kPa for VCTE, ≥ 5.5 kPa for pSWE, ≥ 8.2 kPa for auto‐pSWE, and ≥ 8.9 kPa for 2D‐SWE. External validation and subgroup analysis confirmed the superior performance of VCTE for diagnosing AF in those with hepatic steatosis. Conclusion In this mixed‐etiology cohort, VCTE exhibited superior diagnostic performance compared with pSWE, auto‐pSWE, and 2D‐SWE for the detection of AF in patients with CLD. Choice should also consider device availability, and patient factors. Trial Registration Chinese Clinical Trial Registry at https://www.chictr.org.cn/ (registration number ChiCTR2400090256).