Abstract 877: Molecular and radiographic profiling of lung cancer in never-smokers

Abstract

Abstract Gene expression changes in nasal and bronchial epithelium, and quantitative features from chest CT reflect the presence of lung cancer in current and former smokers and can serve as biomarkers. Between 10-20% of lung cancer cases are diagnosed in patients who have never smoked (<100 lifetime cigarettes) and biomarkers could assist the diagnosis of pulmonary nodules discovered incidentally in this population. We compared lung cancer-associated gene expression and radiomic features in never-smokers and ever-smokers. Nasal epithelial RNA was isolated and sequenced from brushings collected at UCLA from 73 patients presenting to interventional radiology for biopsy for suspected lung cancer. Limma was used to identify genes with lung cancer-associated expression in never-smokers by comparing 13 patients with benign nodules with 14 patients diagnosed with lung cancer. Similarities between cancer-associated genes in never-smokers and those previously identified in ever-smokers were assessed by gene set enrichment analysis (GSEA). Next, using all 73 samples, genes whose lung cancer-associated expression is modified by smoking status were identified via an interaction term. Radiomic features for 46 patients (31 ever-smokers, 15 never-smokers) were extracted using PyRadiomics, including features from the interior of the nodule and perinodular features generated 10, 15, and 20 mm away from the boundary of the nodule. A binomial model was used to identify features with a significant interaction effect between cancer and smoking status. We identified 74 genes decreased and 84 genes increased in cancer in never-smokers (p < 0.005), controlling for age and median TIN. Genes changed in ever-smoker lung cancer patients were significantly enriched among the genes most associated with lung cancer in never smokers (GSEA; p = 0.0059 and p = 4.38e-15). Next, there were 51 genes with a p-value < 0.01 for the cancer*smoking interaction term. Lastly, four radiomic features were associated with the perinodular term for cancer and smoking status (FDR< 0.2), including two nodular and two boundary features. These results suggest that lung cancer-associated gene expression differences in ever-smokers are preserved in never-smokers, suggesting the potential relevance of previously derived biomarkers for this population. There may also be never-smoker-specific gene expression patterns associated with lung cancer. Our radiomic findings suggest that both nodular and perinodular characteristics may be important predictors of malignancy. Citation Format: Minyi Lee, Anil Yadav, Gang Liu, Steven Dubinett, Jennifer Beane, William Hsu, Ashley Prosper, Denise Aberle, Marc Lenburg. Molecular and radiographic profiling of lung cancer in never-smokers [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 877.