DNA methylation biomarkers-based method for the differential diagnosis of multiple lung cancers.

BACKGROUND

Distinguishing intrapulmonary metastasis (IPM) from separate primary lung cancers (SPLC) is crucial for staging and management of multiple lung cancers (MLC). Current methods lack a gold standard, limiting precise diagnosis to a subset of patients. This study introduces DNA methylation biomarkers to differentiate IPM from SPLC.

METHODS

DNA methylation sequencing was performed on 83 specimens from patients with MLC or pulmonary brain metastasis (PBM). In this study, PBM lesions were labelled as analogues of IPM, while indolent lung lesions were designated as reference of SPLC. Unsupervised clustering and random forest algorithms classified patients as IPM or SPLC based on epigenetic similarity. Prognostic performance was assessed using Kaplan-Meier and receiver operating characteristic (ROC) curves. DNMT3B expression was evaluated by immunohistochemistry.

RESULTS

Two clusters of specimens with distinct epigenetic, clinical, pathological, and prognostic features were identified in the training set. Based on the clustering method, DNA methylation method (DM method) was developed utilizing 31 DNA methylation biomarkers to classify patients into IPM or SPLC. Kaplan-Meier curve demonstrated a significant prognostic difference between IPM and SPLC using DM method (p < 0.05). ROC curve indicated that DM method exhibited superior performance compared to other diagnostic methods in predicting progression or mortality. DNMT3B expression was significantly elevated in the IPM subgroup.

CONCLUSIONS

The DM method demonstrates strong potential for differential diagnosis of MLC and outperforms current approaches. Further efforts to develop DM method for MLC diagnosis and prediction of prognosis are warrant.