DNA methylation-assisted diagnosis of lung cancer in bronchial washings

• Main Author: Nikolaidis, Georgios
• Published: University of Liverpool 2011
• Subjects: 616.99424
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.569248

Lung cancer is the most lethal malignancy worldwide and late diagnosis is a significant factor contributing this. The Liverpool Lung Project (LLP) aims to reduce lung cancer mortality through the development of a molecular-epidemiological risk assessment model which will facilitate early detection of lung cancer and thus early intervention. LLP encompasses retrospective and prospective sub-studies. DNA methvlation is an epigenetic modification with key role in gene transcriptional control, embryonic development, imprinting and cancer. A large number of studies in lung cancer have revealed abnormal DNA methylation patterns involving a variety of genes. The aim of this study was to construct and evaluate a panel of DNA-methylation biomarkers which can be applied to bronchial washings (BWs) material and assist in diagnosis of lung cancer. The discovery and validation process followed the guidelines set by the NCI- Early Detection Research Network (EDRN) and the CR-UK diagnostic biomarker roadmap. Specific objectives included (a) the discovery of promoter targets with high frequency of hypermethylation in primary lung cancer, (b) the development of a highly sensitive and specific DNA methylation assay fitting to clinical standards and (c) the validation of these targets in BWs utilising a longitudinal retrospective case-control design. Targets from previous high throughput approaches were validated in an independent set of 48 non-small cell lung cancer samples and paired normal tissues using Pyrosequencing methylation analysis (PMA). PMA confirmed significant hypermethylation in the primary NSCLC tissue for the following promoters: RASSF1, p16, WT1, CYGB, RARB, CDH13, DAPK, p73, TMEFF2, and TERT. In addition, immunohistochemical staining for p16 and WT1 was performed in a 20 non-small cell lung cancer samples. Quantitative methylation-specific PCR (qMSP) assays were subsequently developed and tested for reliability and robustness for these ten candidates. These assays were used to screen 655 BWs from the Liverpool Lung Project (LLP) subjects divided into a training (194 cases and 214 Controls) and validation (139 cases and 109 controls) sets. Multifactor Dimensionality Reduction (MDR) was used to select the best subset of the markers with good discrimination. Analysis in the training BWs set demonstrated significant differences in the detected hypermethylation frequency in cases over controls for RASSF1 p16, WT1, CYGB, RARB and TERT. The diagnostic efficiency of this panel was evaluated in the independent validation set. A logit method was used to obtain the sensitivity and specificity of the six markers. LogicF analysis demonstrated that the top five predictors were WT1, cytology, RASSF1, TERT and p16. The overall performance the latter panel demonstrated no diagnostic bias in different groups of gender, age or smoking status. While cytology alone provides a 49.5% sensitivity, 99.5% specificity, the addition of the four methylation markers provided 76.2% sensitivity, 92.3% specificity (AUC=0.89). Although the diagnostic efficiency of this panel must be improved by incorporating additional promoters, our findings clearly demonstrate the impact of DNA methylation- based assays in the diagnosis of cytologically occult lung neoplasms.