Spleen Stiffness–Based Algorithms Are Superior to Baveno VI Criteria to Rule Out Varices Needing Treatment in Patients With Advanced Chronic Liver Disease

INTRODUCTION: The Baveno VI criteria have set the stage for noninvasive assessment of compensated advanced chronic liver disease (ACLD). The algorithm combining liver stiffness measurement (LSM, <20 kPa) and platelet count (>150,000/μL) safely avoids screening endoscopy for varices needing treatment (VNT) but identifies only a relatively low number of patients. We aimed to evaluate the value of spleen stiffness measurement (SSM) using spleen-dedicated elastography in ruling out VNT. METHODS: In this real-life multicenter retrospective derivation-validation cohort, all consecutive patients with ACLD (defined by LSM ≥10 kPa) with available upper endoscopy, laboratory results, spleen diameter, LSM, and SSM measured with spleen-dedicated transient elastography were included. VNT were defined as medium-to-large varices or small varices with red spots. RESULTS: In the derivation cohort (n = 201, 11.9% VNT), SSM demonstrated excellent capability at identifying VNT (area under the receiver operating characteristic curve [AUROC] 0.88), outperforming LSM (AUROC 0.77, P = 0.03) and platelets (AUROC 0.73, P = 0.002). In comparison with Baveno VI criteria (33.8% spared endoscopies), the sequential Baveno VI plus SSM and a novel spleen size and stiffness model were able to increase the number of patients avoiding endoscopy (66.2% and 71.1%, respectively) without missing more than 5% of VNT. These findings were confirmed in an external validation cohort of patients with more advanced liver disease (n = 176, 34.7% VNT) in which the number of spared endoscopies tripled (27.3% and 31.3% for SSM-based algorithms) compared with Baveno VI criteria (8.5%). DISCUSSION: Spleen stiffness–based algorithms are superior to Baveno VI criteria in ruling out VNT in patients with ACLD and double the number of patients avoiding screening endoscopy.