Abstract The widespread implementation of low-dose computed tomography (LDCT) has markedly increased the detection of pulmonary nodules, yet their genetic determinants remain poorly understood. We conducted a genome-wide association study (GWAS) of 36 175 LDCT-screened individuals from Zhejiang and Jiangsu provinces in China, assessing fourteen pulmonary nodule phenotypes defined using a convolutional neural network–based computer-aided detection (CNN-CAD) system. We identified eleven independent single-nucleotide polymorphisms (SNPs) at nine loci associated with nodule phenotypes. Functional annotation prioritized six candidate genes with either missense variants or strong colocalization evidence, including TP63 at 3q28, PLA2G4F at 15q15.1, HLA-DRB6 and CYP21A2 at 6p21.32, TNFSF11 at 13q14.11, and TNFRSF11B at 8q24.12. These genes were enriched in pathways related to cell adhesion, chemotaxis, cytokine activity, and lymphocyte activation. Several of the identified variants—associated with non-solid components, nodule size, and the number of positive nodules—were also significantly associated with increased malignancy probability of pulmonary nodules. Genetic correlation analysis revealed a substantial shared genetic basis between purely ground-glass nodules (pGGNs) and lung adenocarcinoma (LUAD) (rg = 0.79; 95% CI, 0.16–1.42; P = 0.014), and Mendelian randomization (MR) further supported a potential causal relationship, with an odds ratio (OR) of 51.1 (95% CI: 1.13–2035.31). These findings offer novel insights into the genetic architecture of pulmonary nodules and highlight a substantial genetic overlap between pGGNs and LUAD, which may inform risk prediction and precision prevention in lung cancer screening.
Zhang et al. (Thu,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: