Stacey-Ann Whittaker Brown

M.D

Icahn School of Medicine at Mount Sinai

Fellow Physician

NLST (Learn more about this study)

NLST-310

Jun 9, 2017

Lung Cancer Risk in Patients with Interstitial Lung Disease

Background: Interstitial lung diseases (ILDs) are a widely heterogenous group of disorders affecting the lung parenchyma. Several causes of ILD are also associated with an increased risk of lung cancer. Of these, fibrosing lung diseases have the strongest association. As attempts are made at reducing lung cancer mortality by screening smokers, other high-risk populations who may benefit from screening will need to be characterized. Using data from the NLST, we aim to determine the incidence of lung cancer and lung-cancer specific mortality in patients with ILD.

Methods:
Case Selection: We will identify subjects with ILD as those who self-reported asbestosis, silicosis, sarcoidosis or pulmonary fibrosis on baseline questionnaire, or those who on baseline CT or CXR were found to have “reticular/reticulonodular opacities, honeycombing,fibrosis, scar".

Statistical Analysis: After identifying the following cases of any ILD, we will compare baseline characteristics, tumor characteristics and spirometry with non-ILD cases, using chi-squared tests for categorical variables and Wilcoxon rank-sum tests or t-tests for continuous variables.
We will compare the relative risk of lung cancer in the two groups using Poisson regression, stratified by screening group (CT scan versus CXR) and with the following covariates: age, sex, ethnicity, marital status, smoking history. If significant differences exist between incidence of subtypes of lung cancer, we will provide analyses of relative risk for SCLC (the most aggressive form of malignancy), non-BAC NSCLC (intermediate in aggressiveness) and BAC (the most benign form of bronchogenic carcinoma).
To determine the difference in lung cancer mortality, we plan to use Kaplan-Meier curves to compare the unadjusted lung-cancer specific mortality in the group. We plan to repeat our analysis using Cox-proportional hazards model, stratified by screening group and treatment strategy (surgical, non-surgical, none) with the covariates listed above, along with tumor characteristics (TNM stage and histology).
Because ILD is associated with significant morality, we will perform a competing risk analysis using the above covariates on the subgroup of patients with mild-moderate restriction on spirometry in the CT screening arm with non-BAC NSCLC.
All tests will be conducted using SAS 9.4, at a two-sided α=0.05.