Abstract
Lung adenocarcinoma (LUAD) is one of the most common cancers and lethal diseases in the world. Recognition of the undetermined lung nodules at an early stage is useful for a favorable prognosis. However, there is no good method to identify the undetermined lung nodules and predict their clinical outcome. DNA methylation alteration is frequently observed in LUAD and may play important roles in carcinogenesis, diagnosis, and prediction. This study took advantage of publicly available methylation profiling resources and a machine learning method to investigate methylation differences between LUAD and adjacent non-malignant tissue. The prediction panel was first constructed using 338 tissue samples from LUAD patients including 149 non-malignant ones. This model was then validated with data from The Cancer Genome Atlas database and clinic samples. As a result, the methylation status of four CpG loci in homeobox A9 (HOXA9), keratin-associated protein 8-1 (KRTAP8-1), cyclin D1 (CCND1), and tubby-like protein 2 (TULP2) were highlighted as informative markers. A random forest classification model with an accuracy of 94.57% and kappa of 88.96% was obtained. To evaluate this panel for LUAD, the methylation levels of four CpG loci in HOXA9, KRTAP8-1, CCND1, and TULP2 of tumor samples and matched adjacent lung samples from 25 patients with LUAD were tested. In these LUAD patients, the methylation of HOXA9 was significantly upregulated, whereas the methylation of KRTAP8-1, CCND1, and TULP2 were downregulated obviously in tumor samples compared with adjacent tissues. Our study demonstrates that the methylation of HOXA9, KRTAP8-1, CCND1, and TULP2 has great potential for the early recognition of LUAD in the undetermined lung nodules. The findings also exhibit that the application of improved mathematic algorithms can yield accurate and particularly robust and widely applicable marker panels. This approach could greatly facilitate the discovery process of biomarkers in various fields.
Highlights
Lung cancer is the leading cause of cancer-related deaths worldwide
The results further confirmed that the hypermethylation of homeobox A9 (HOXA9) and the hypomethylation of KRTAP8-1, cyclin D1 (CCND1), and tubby-like protein 2 (TULP2) were observed in Lung adenocarcinoma (LUAD) tumor samples
To develop a reliable recognition panel utilizing DNA methylation signals, we selected training samples of tumors and non-malignant tissues from the Gene Expression Omnibus (GEO) dataset to build a predication model for LUAD and tested our prediction model in two validation sets derived from the The Cancer Genome Atlas (TCGA) database
Summary
As the most frequent histological subtype of nonsmall-cell lung cancer (NSCLC), lung adenocarcinoma (LUAD) accounts for more than 40% of the incidence of lung cancer. It is usually diagnosed at an advanced stage. Effective early recognition methods and relevant biomarkers are urgently needed to reduce the mortality caused by LUAD. DNA methylation is a relatively stable biochemical modification carried out by DNA methyltransferases and can be detected in DNA molecules from tissue and the free DNAs in serum and plasma (18), making it a promising biomarker for the early recognition of the undetermined lung nodules
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
