# Reliability and Agreement of Quantitative Pulmonary Imaging Biomarkers Between Ultra-Low-Dose and Low-Dose Chest CT: A Paired Intra-Individual Study

## TL;DR

This study compares ultra-low-dose and low-dose CT scans to see if they reliably measure lung imaging biomarkers.

## Contribution

The study provides new evidence on the reliability of ULD-CT for quantitative pulmonary imaging biomarkers.

## Key findings

- Density-based biomarkers showed high concordance and strong correlations between ULD-CT and LD-CT.
- Airway structural metrics had clinically acceptable agreement and near-perfect reproducibility.
- Voxel-based functional biomarkers were more dose-sensitive but maintained consistent directional bias.

## Abstract

Background/Objectives: Ultra-low-dose computed tomography (ULD-CT) enables substantial radiation reduction compared with routine low-dose computed tomography (LD-CT), but its quantitative reliability across lung imaging biomarkers remains insufficiently characterized. This study aimed to assess the agreement and reliability of quantitative pulmonary imaging biomarkers between paired ULD-CT and LD-CT examinations. Methods: In this prospective study, 48 patients who underwent paired LD-CT and ULD-CT on the same day were analyzed. Whole-lung quantitative biomarkers were categorized into density-derived indices, airway structural metrics, and voxel-based functional biomarkers. Agreement between LD-CT and ULD-CT was evaluated using Bland–Altman analysis and Pearson correlation. Results: Density-based biomarkers demonstrated high concordance, strong correlations, and small systematic biases, indicating robust dose stability. Airway structural metrics showed clinically acceptable agreement with near-perfect reproducibility for cluster-based indices. Voxel-based functional biomarkers exhibited greater dose sensitivity but preserved consistent directional bias. Total lung volume showed excellent reproducibility with minimal bias. Conclusions: ULD-CT enables reliable quantitative lung imaging with clinically acceptable agreement across major biomarker domains, supporting its use as a dose-efficient platform for longitudinal and screening applications.