Combined model with acoustic radiation force impulse to rule out high-risk varices in HBV-related cirrhosis with viral suppression.

AIMS: To prospectively evaluate the performance of spleen stiffness measurement (SSM) and liver stiffness measurement (LSM) via acoustic radiation force impulse (ARFI) imaging combined with platelet counts (PLT) in ruling out HRV in HBV-related cirrhotic patients with viral suppression. METHODS: Patients with cirrhosis enrolled between June 2020-March 2022 were divided into a derivation cohort and validation cohort. LSM and SSM ARFI-based, and esophagogastroduodenoscopy (EGD) were performed at enrollment. RESULTS: In the derivation cohort, overall, 236 HBV-related cirrhotic patients with maintained viral suppression were enrolled, and the prevalence of HRV was 19.5% (46/236). With the aim of identifying HRV, the most accurate LSM and SSM cut-offs were chosen of 1.46 m/s and 2.28 m/s, respectively. The combined model (LSM<1.46 m/s and PLT>150 × 109/L strategy combined with SSM ≤ 2.28 m/s) can spare 38.6% of EGDs and 4.3% of HRV cases were misclassified. In the validation cohort, we analysed 323 HBV-related cirrhotic patients with maintained viral suppression and validated the combined model can spare 33.4% (108/323) of EGD, and the HRV missed rate was 3.4%. CONCLUSIONS: A non-invasive prediction model combining LSM<1.46 m/s and PLT>150 × 109/L strategy with SSM ≤ 2.28 m/s exhibited excellent performance in ruling out HRV and avoided a significantly large number (38.6% vs 33.4%) of unnecessary EGDs in HBV-related cirrhotic patients with viral suppression.