# Influence of iris color on corneal densitometry

## TL;DR

This study shows that lighter-colored irises can lead to higher corneal densitometry readings due to brightness artifacts, and introduces a new metric to objectively measure iris color.

## Contribution

The study introduces a novel objective metric (IrisColor) for quantifying iris color and demonstrates its use in correcting corneal densitometry bias.

## Key findings

- Lighter-colored irises caused a 6.6% overestimation in corneal densitometry values.
- Iris brightness artifacts strongly correlate with iris color and affect densitometry measurements.
- The proposed IrisColor metric effectively classifies iris color and could improve diagnostic accuracy.

## Abstract

To determine whether iris pigmentation introduces a measurable bias in corneal densitometry (CD) values obtained via Scheimpflug imaging, and to develop an objective metric for iris color quantification.

This observational study included 91 eyes from 47 healthy adults. CD was assessed as mean pixel intensity (MPI) from 25 Scheimpflug images per eye. Brightness artefacts from the iris were quantified using automated image processing. Iris color was objectively characterized from slit-lamp photographs using an objective single-value metric (IrisColor) derived from normalized CIELAB components. Associations among CD, iris brightness, and iris pigmentation were evaluated using Pearson correlation and linear mixed-effects models (LMMs).

CD correlated positively with iris brightness artefacts (r = 0.47, β = 1.49, p < 0.001), which in turn showed a strong negative correlation with IrisColor (r = –0.83, β = –1.11, p < 0.001). Lighter-colored irises (lower IrisColor values) exhibited statistically significantly higher CD values equivalent to a 6.6% relative overestimation. Groupwise comparisons confirmed that iris pigmentation significantly influences both CD and overall image brightness.

Iris pigmentation induces a measurable bias in Scheimpflug-based CD estimates, primarily through increased brightness artefacts in light-colored eyes. The proposed IrisColor metric enables objective, continuous classification of iris color and could support future corrections for pigmentation-induced bias in CD-based diagnostics.