Real-time shear-wave elastography: applicability, reliability and accuracy for clinically significant portal hypertension

Background & aims: Real-time shear wave elastography (RT-SWE) might be useful to assess the severity of portal hypertension; reliability criteria for measurement are needed.

Methods: We prospectively included 88 consecutive patients undergoing hepatic venous pressure gradient measurement (HVPG, reference standard) for portal hypertension. Liver stiffness (LS) was measured by RT-SWE and by transient elastography (TE). Spleen stiffness (SS) was measured by RT-SWE. Reliability criteria for RT-SWE were searched, and the accuracy of these techniques to identify HVPG ⩾10mmHg (clinically significant portal hypertension, CSPH) was tested and internally validated by bootstrapping analysis.

Results: LS and SS by RT-SWE were feasible respectively in 87 (99%) and 58 (66%) patients. Both correlated with HVPG (LS: R=0.611, p<0.0001 and SS: R=0.514, p<0.0001). LS performed well for diagnosing CSPH (optimism corrected AUROC=0.858). Reliability of measurements was influenced by standard deviation (SD)/median ratio and depth. SD/median ⩽0.10 and depth of measurement <5.6cm were associated to 96.3% well classified for CSPH, while when one or none of the criteria were fulfilled the rates were 76.4% and 44.4%, respectively. Measurements fulfilling at least one criterion were considered acceptable; in these patients, RT-SWE performance to detect CSPH was excellent (AUROC=0.939; 95% CI: 0.865-1.000; p<0.0001; best cut-off: 15.4kPa). LS by RT-SWE and by TE were strongly correlated (R=0.795, p<0.0001) and performed similarly both in "per protocol" and in "intention-to-diagnose" analysis after applying reliability criteria.

Conclusions: LS by RT-SWE is an accurate method to diagnose CSPH if reliability criteria (SD/median ⩽0.10 and/or depth <5.6cm) are fulfilled.