Lung cancer (LC) is the main cause of cancer-related death worldwide (Sung et al., 2021). According to official data, LC was the largest cancer in terms of incidence, accounting for about 20.4% of the 828,000 new cancer cases and the main common cause of cancer death accounting for about 27.2% of 657,000 cancer death cases in China (Zheng et al., 2022). LC prognosis is correlated with the clinical stage at first detection, with five-year survival decreasing from 85% at stage IA to 6% at stage IV (Chansky et al., 2017). However, despite using the diagnostic methods, treatments, and therapies, the five-year survival of LC is only 16% (Torre et al., 2015).

Except for the development of new treatments and therapies,early detection is necessary to decrease cancer-related death ratios. Screening of populations at high LC risk by low-dose computed tomography (LDCT) has shown a 20.0% decrease in LC mortality (Aberle et al., 2011). However, this imaging method has a dramatically high false-positive rate of 96.4% for distinguishing malignant nodules from benign nodules (de Koning et al., 2014). Recently, researchers begin to pay more attention to molecular diagnosis for the early detection of cancer-related patients. LC is caused by the interaction between nongenetic factors and genetic factors. Nongenetic factors include tobacco smoking, occupational exposure, and environmental carcinogens. Among these nongenetic factors, smoking is a high factor for LC risk accounting for 80%–90% of all LC diagnoses (Schabath and Cote, 2019). Genetic factors, such as mutations in CXCR2 (Ryan et al., 2015), DDR2 (Hammerman et al., 2011), and EGFR (Tan et al., 2016), are involved in the molecular mechanism of LC disease. A combination of the nongenetic factors and genetic factors is efficient in the detection of high-risk persons but is hardly applied for LC. Therefore, it is urgent and important to search for novel biomarkers for the early detection of LC patients.

Aberrant epigenetic changes are ubiquitous features of carcinogenesis, occurring early in the progression of cancer (Irizarry et al., 2009). DNA methylation plays a very important function in the regulation of gene expression and the architecture of cell nuclei (Robertson and Wolffe, 2000). Aberrant methylation of oncogenes and suppressor genes are early events in many cancers, suggesting that the alterations of DNA methylation levels might be one of the detectable changes in tumorigenesis (Irizarry et al., 2009; Baylin and Jones, 2011). Liquid biopsy has been reported as a noninvasive, real-time, and dynamic monitoring method for the detection of cancers. Researchers have found that circulating tumor DNA methylation markers can be applied in the diagnosis of hepatocellular carcinoma (Xu et al., 2017). A study suggested biomarkers for LC classification using plasma cell-free DNA (cfDNA) methylomes (Shen et al., 2018). Although cfDNA methylation in plasma has potential in the cancer detection, we cannot ignore its limitations in the detection of cancer patients at early stages due to the low quantity of tumor DNA in plasma, low sensitivity, and high costs by very deep sequencing (Aravanis et al., 2017; Chen et al., 2020). Alternatively, recent studies including our previous researches provided a hint that methylation biomarkers in whole blood were associated with cancer at an early stage as well as for LC (Yang et al., 2015; Zhang et al., 2015; Qiao et al., 2020), but mostly these studies contained small sample sizes and candidate approaches.

Here, Illumina 850K methylation array was used to screen for LC-related DNA methylation biomarkers in peripheral whole blood. Next, the candidate gene was validated in two independent studies via mass spectrometry. This large case-control study was based on more than 5000 subjects, including approximately 1500 stage I LC cases. The correlations between DNA methylation and clinical characteristics of LC cases were also analyzed.