D. Lau, Singapore, SG

Hepatic venous pressure gradient(HVPG) is the gold standard for measuring portal pressure but is invasive. Clinically significant portal hypertension(CSPH), defined as HVPG≥10mmHg is associated with varices and clinical decompensation. Spleen stiffness(SS) and spleen length(SL) are potential predictors of CSPH. We prospectively compared the SS and SL measured by magnetic resonance elastography(MRE) versus ultrasonographic shear-wave elastography(SWE) for the non-invasive prediction of CSPH.

Material and methods

Patients with chronic liver disease scheduled for HVPG measurement were prospectively recruited for this IRB-approved study. All underwent MRE and/or SWE within 8 weeks of HVPG measurement. Reliability of SS+SL to predict CSPH was assessed for each modality using AUROC.

Results

47 subjects underwent HVPG and MRE, of which 43 completed SWE. Median age was 63 years, 53% were males. 38 subjects(81%) had liver cirrhosis, 55% Child A and 45% Child B/C. Median HVPG was 14 mmHg(IQR 10-17). CSPH was present in 38(81%). Median SWE-SS was significantly higher in patients with CSPH(31.7kPa vs. 16.1kPa, p=0.041), and was a reliable predictor for CSPH(AUROC 0.749, 95% CI 0.477-1.000). A multivariable model of SWE-(SS+SL) provided an improved AUROC of 0.788(95% CI 0.545-1.000). Mean MRE-SS was not significantly different in patients with versus without CSPH(9.4kPa vs. 8.1kPa, p=0.067), and was a less reliable predictor for CSPH(AUROC 0.709, 95% CI 0.430-0.989). MRE-(SS+SL) provided an improved AUROC of 0.757(95% CI 0.504-1.000) for prediction of CSPH.

Conclusion

SWE-SS is a better predictor of CSPH compared to MRE-SS. Best results are with a combination of SWE-SS and SWE-SL, which can be measured in the same setting.

Collapse