Simone Cappelli, Giovanni Petralli, et al. EASL Congress. 2024.

•The values of liver stiffness by transient elastography using different ultrasound systems could differ and jeopardize the patients’ clinical management. Our aim was to compare the results of two different technologies for liver stiffness measurement in a large cohort of chronic liver disease (CLD) patients (pts).

•In 777 consecutive CLD pts admitted at Hepatology Unit of University Hospital of Pisa from May 2022 to December 2023 liver stiffness was at the same time with two different diagnostic systems: FibroScan® (Echosens, France) equipped with M or XL probe and FT9000 (Hisky Medical, China) equipped with a single probe liver stiffness measurement (LSMFibroscan/LSMFT9000).
•Pearson correlation coefficients and Bland-Altman analyses(BAA) were conducted to study the agreement between LSMFibroscan/LSMFT9000. Delta FT9000-FibroScan® LSM (D) associations with CLD clinical parameters was observed too. AST/GGT and platelets counts/portal veins caliper/spleen long axis were used as proxies of necro-inflammatory activity and portal hypertension-vascular congestion.

•Overall LSMFibroscan was strongly correlated with LSMFT9000 (r = 0.77, p < 0.001). LSM correlation varied among BMI classes (normal weight r = 0.84, overweight r = 0.75, obese r = 0.65, p < 0.001).
•LSMs obtained with the two devices showed a bias of 0.52 kPa (p = 0.01) with-11/12 kPa 95% LoA, with a clear trend for higher values measured by LSMFT9000 for LSM < 10 kpa and vice versa for LSM > 10 kPa.
•In fact, D was equal to zero for LSMFibroscan 9.6 kPa. In LSMFibroscan > 10 kPa cohort D significantly correlated with AST, GGT, portal vein caliper, spleen long axis, and platelets (r = − 0.22, r = − 0.17, r = − 0.189, r = − 0.287, and r = 0.031, respectively). In a bivariate analysis D correlated with LSMFibroscan and AST (beta = 0.002, p = 0.001) with a p of interaction equal to 0.018 and D was also associated with spleen long axis (beta = 0.033, p = 0.027) and platelets count(beta = − 0.008, p = 0.018, respectively), with a significant interaction of both with LSMFibroscan (p < 0.001).

•LSMFibroscan and LSMFT9000 measures showed a limited agreement across different dynamic ranges with a good agreement only around 10 kPa. The discrepancies associated with anthropometry and necro-inflammatory activity and vascular congestion proxies. LSMFT9000 provides higher values than LSMFibroscan below 10 KPa whereas LSMFibroscan higher values, with wider dynamic range, over 20 KPa. Comparative studies using as gold standard reference liver histology and porto-systemic gradient will help to better understand the reasons for the discrepancies.
Whether these results will be confirmed, method specific cut-offs in clinical practice will be required.