Reproducibility of biomarkers in MASLD: A benchmark for clinically meaningful change in serially measured non-invasive tests.

Abstract

Background & aims: The use of non-invasive tests (NITs) to monitor disease severity in metabolic dysfunction-associated steatotic liver disease (MASLD) necessitates a clear understanding of clinically meaningful changes at the patient level. This study assessed the measurement variation of NITs in the pre-randomisation screening pipeline of a phase II clinical trial in patients with biopsy-confirmed at-risk MASLD.

Methods: This study was conducted across 198 centres in 11 countries. The screening analysis set (ScrAS) (n = 803) comprised patients screened for trial entry with an evaluable biopsy for eligibility assessment. The randomised analysis set (RAS) (n = 255) comprised participants with histologically confirmed at-risk steatohepatitis that fulfilled all trial entry criteria and represents a more highly characterised subset nested within the larger ScrAS dataset. Multiple NITs (including vibration controlled transient elastography [LSM-VCTE], enhanced liver fibrosis test [ELF], PRO-C3, ADAPT, PRO-C6, CK-18 M30/M65) were evaluated for measurement variation, with results reported as within-subject % coefficient of variation (wCV) and reproducibility coefficient % (RDC).

Results: The analysis of ScrAS demonstrates that LSM-VCTE exhibits a wCV of 19.9%, with a corresponding RDC of 55.2%. LSM-VCTE in RAS shows a wCV of 23.0% and RDC of 63.8%. While controlled attenuation parameter had the lowest wCV (8.8% in ScrAS; 8.7% in RAS) and RDC (24.2% in ScrAS; 24.1% in RAS) of the FibroScan-derived parameters, ELF demonstrated the lowest wCV (3.4%) and RDC (9.3%) overall.

Conclusions: These data provide a clear and robust measure of reproducibility for a range of widely adopted NITs, providing insights into the magnitude of biomarker change that may be biologically meaningful and thereby informing both longitudinal monitoring of disease progression in clinical practice and assessment of treatment response in therapeutic trials.

Clinicaltrials: gov (NCT#04321031).

Impact and implications: The lack of disease-specific biomarker reproducibility data has hampered the adoption of non-invasive tests (NITs) for longitudinal monitoring of disease natural history in clinical practice, and has hindered the regulatory adoption of NITs as tractable surrogate endpoints to support drug development, as it leads to uncertainty about what magnitude of biomarker change indicates genuine disease progression. The present study addresses both these challenges by providing a clear and robust measure of reproducibility for a range of widely adopted NITs, generating actionable data that facilitate longitudinal monitoring of disease progression and treatment response.