Identifying Early ILD using Quantitative CT (QCT) Thresholds

Principal Investigator

Name
Jennifer Wang

Degrees
M.D., M.S.

Institution
University of Michigan

Position Title
Clinical Instructor

Email
wangjenn@med.umich.edu

About this CDAS Project

Study
NLST (Learn more about this study)

Project ID
NLST-1499

Initial CDAS Request Approval
Mar 23, 2026

Title
Identifying Early ILD using Quantitative CT (QCT) Thresholds

Summary
Rationale: There is increased interest in identifying early evidence of interstitial lung disease (ILD), termed interstitial lung abnormalities (ILAs), on chest computed tomography (CT) to improve clinical outcomes. Despite the widespread availability of chest CT to identify incident ILD, there remains no consensus on how quantitative CT (qCT) criteria can be used to identify early evidence of ILD and identify high-risk patients.

Objectives: To determine whether qCT-based measures of fibrosis are associated with poor outcomes in a healthcare cohort of individuals with incident ILD and validate these measures in a separate cohort, the NLST.

Methods: This was a single-center retrospective study including all adults aged 21 and older undergoing outpatient chest CT over a calendar year, excluding patients with prevalent cancer. Computer-Aided Lung Informatics for Pathology Evaluation and Ratings (CALIPER) was applied to all chest CTs to produce quantitative measures of fibrosis, which included percent reticular opacity, ground glass opacity and honeycombing as well as measures of vascular volume and percent hyperlucent as a surrogate for emphysema.

Using logistic regression, we iteratively added and removed variables to produce the best performing and most parsimonious model differentiating ILAs from chest CTs without evidence of fibrosis (using natural language processing of CT reports for fibrosis keywords). We then dichotomized this qCT measure of ILAs using cutpoint analysis at Youden's index.

The primary outcome was 3-year transplant-free survival, modeled using multivariable Cox proportional hazard regression. Following this analysis, we will validate this qCT measure of early ILD in a separate cohort, the National Lung Screening Trial (NLST), using low dose chest CTs. We will apply this qCT measure to prospectively identify cases of incident ILA/early ILD in a healthcare cohort over a 1-year period.

To extend our efforts, we will also add the percent hyperlucent in our models to identify cases of combined pulmonary fibrosis and emphysema (CPFE) as well, given the particularly poor outcomes in this patient population.

Aims

In this study, our specific aims are the following:

1. To determine whether quantitative CT measures of fibrosis using CALIPER are associated with 3-year transplant free survival (TFS) in a large, single-center historical healthcare cohort of all individuals undergoing outpatient, non-contrast chest CT in a single calendar year.

2. To derive a quantitative CT-based measure of interstitial lung disease/interstitial lung abnormalities in this single-center historical cohort and validate this novel quantitative CT measure in an independent population-based cohort, the NLST.

3. To prospectively apply this qCT-based measure to identify incident cases of interstitial lung disease/interstitial lung abnormalities in a healthcare cohort over a 1-year period.

4. To extend our models to include percent hyperlucent, validate these models similarly in the NLST, and identify cases of combined pulmonary fibrosis and emphysema (CPFE) prospectively.

Collaborators

Jennifer Wang University of Michigan
Justin Oldham University of Michigan
Charles Hatt University of Michigan
Swaraj Bose University of Michigan