Challenges in Histological Endpoints for MASH Therapies: An Exercise in Statistical Modelling

Area of Science:

• Hepatology
• Clinical Trial Methodology
• Biostatistics

Background:

• Current nonalcoholic steatohepatitis (NASH) trial endpoints rely on liver biopsy histology, which has significant inter-/intra-reader variability.
• This variability complicates the interpretation of trial results, as few studies have demonstrated positive outcomes using these endpoints.

Purpose of the Study:

• To statistically evaluate the impact of imprecise histologic scoring on the results of NASH/MASH clinical trials.
• To quantify the dilution of true effect size caused by scoring variability.

Main Methods:

• Simulated kappa values based on the relationship between agreement measures and scoring sensitivity.
• Utilized published kappa values from NASH trials to determine histology parameter sensitivities.
• Conducted simulations with varying over/under-scoring probabilities to estimate effect size dilution.

Main Results:

• Simulations indicated approximately 50% dilution of the true effect size for both fibrosis improvement and MASH resolution endpoints in 2-arm trials.
• This dilution effect was observed to be constant regardless of the trial's sample size.

Conclusions:

• Imprecise histologic scoring disproportionately affects the 'superior' arm in clinical trials.
• The observed dilution necessitates consideration in the risk-benefit assessment of MASH studies.
• These findings advocate for the adoption of non-invasive biomarkers over traditional histologic endpoints.