# Collagen Proportionate Area and Histological Grading of Fibrosis as Predictors of Clinical Outcomes and Mortality in Patients with Steatotic Liver Disease

## TL;DR

This study examines how collagen proportionate area and fibrosis staging in liver biopsies predict outcomes in patients with steatotic liver disease.

## Contribution

The study evaluates the prognostic value of CPA in a population-based setting for the first time.

## Key findings

- CPA correlated significantly with fibrosis stage, especially in advanced fibrosis.
- CPA and fibrosis stage showed comparable long-term prognostic value for mortality.
- Predictive models for liver-specific outcomes had poor performance despite moderate discrimination for mortality.

## Abstract

Steatotic liver disease (SLD) including metabolic dysfunction-associated and alcoholic liver disease, is the most prevalent chronic liver condition globally. Liver biopsy remains the gold standard for fibrosis assessment, but traditional staging is difficult even for highly skilled experts. Collagen proportionate area (CPA), a quantitative digital pathology measure, may offer an alternative, yet its prognostic value in population-based settings is unclear.

Liver biopsies from a cohort of 166 adults with biopsy-confirmed SLD, diagnosed between 1995 and 2008, were retrieved. The liver biopsies underwent fibrosis staging (stages F0–F4) by a pathologist and digital CPA quantification. Clinical outcomes and mortality were tracked via national registries over a median follow-up of 14.8 years. Associations between fibrosis stage, CPA, and all-cause mortality and liver-related events were assessed using Cox regression and cause-specific hazard models, adjusting for alcohol use and age. Predictive performance was evaluated with area under the receiver operating characteristic curve (AUC), index of prediction accuracy (IPA), and calibration plots, internally validated by bootstrapping.

CPA correlated significantly with fibrosis stage (Spearman's ρ = 0.63), particularly in advanced fibrosis. Higher fibrosis stages and greater CPA were associated with increased all-cause mortality and liver-related events, but the associations for CPA were nonsignificant after adjustment for age and alcohol use. Scaled estimates of HRs from CPA were lower compared to HRs from fibrosis stages. Predictive models for mortality demonstrated comparable and moderate discrimination for CPA and fibrosis stage, improving with age adjustment. IPA values and calibration plots indicated positive predictive accuracy for mortality but poor performance for liver-specific outcomes.

CPA closely correlates with fibrosis stage and has comparable long-term prognostic value in SLD. Interpretation of study results is limited by the small sample size and low number of liver-specific events. Larger studies are needed to validate CPA's clinical utility and explore its integration into routine practice.

Image 1