Radiologic Features of Small Pulmonary Nodules and Lung Cancer Risk in the National Lung Screening Trial: A Nested Case-Control Study

Abstract

Purpose To extract radiologic features from small pulmonary nodules (SPNs) that did not meet the original criteria for a positive screening test and identify features associated with lung cancer risk by using data and images from the National Lung Screening Trial (NLST). Materials and Methods Radiologic features in SPNs in baseline low-dose computed tomography (CT) screening studies that did not meet NLST criteria to be considered a positive screening examination were extracted. SPNs were identified for 73 incident case patients who were given a diagnosis of lung cancer at either the first or second follow-up screening study and for 157 control subjects who had undergone three consecutive negative screening studies. Multivariable logistic regression was used to assess the association between radiologic features and lung cancer risk. All statistical tests were two sided. Results Nine features were significantly different between case patients and control subjects. Backward elimination followed by bootstrap resampling identified a reduced model of highly informative radiologic features with an area under the receiver operating characteristic curve of 0.932 (95% confidence interval [CI]: 0.88, 0.96), a specificity of 92.38% (95% CI: 52.22%, 84.91%), and a sensitivity of 76.55% (95% CI: 87.50%, 95.35%) that included total emphysema score (odds ratio [OR] = 1.71; 95% CI: 1.39, 2.01), attachment to vessel (OR = 2.41; 95% CI: 0.99, 5.81), nodule location (OR = 3.25; 95% CI: 1.09, 8.55), border definition (OR = 7.56; 95% CI: 1.89, 30.8), and concavity (OR = 2.58; 95% CI: 0.89, 5.64). Conclusion A set of clinically relevant radiologic features were identified that that can be easily scored in the clinical setting and may be of use to determine lung cancer risk among participants with SPNs. ^© RSNA, 2017 Online supplemental material is available for this article.

Figure 1:

Schema of the nested case-control study. All participants in this analysis had a baseline negative screening based on the NLST protocol. Incident case patients (red boxes) were given a diagnosis of lung cancer at either the first follow-up screening examination (n = 62) or the second follow-up screening examination (n = 63) and were combined to form the case group (n = 125). Control subjects were defined as participants who had three consecutive negative screenings (green box). A 2:1 nested study design was used to identify 250 control subjects who were frequency matched to the case patients for age, sex, race, and smoking status. In the 125 patients with incident lung cancers and 250 control subjects, radiologic SPNs were detected on the baseline screen for 73 case patients and 157 control subjects, which was the final sample size for analysis.

Figure 2:

Graph shows AUCs for multivariable models containing demographic variables only (red line), imaging features only (green line), and imaging features and demographic variables together (blue line). The model containing only demographic variables included age, sex, race, smoking status, number of pack-years smoked, and family history of lung cancer. The model containing only imaging features included total emphysema, attachment to vessel, nodule location, border definition, and concavity. The model containing both demographics and imaging features included all covariates from the demographic variables–only model and the imaging features–only model.