Table 2

Methylation-, gene-expression- and miRNA-based biomarkers for risks and early detection of lung cancer
Reference Sample type Genetics/genomics platform Clinical settings Key findings
a) Epigenetic biomarkers
[59] Sputum, lung tissue, biopsies MSP Lung tissue, precursor lesions and bronchial biopsies from patients with SCC and sputum from individuals with suspicion of lung cancer CDKN21 hypermethylation more often observed in patients with cancer than with no cancer
[66] Paired serum and lung tissue MSP Lung tissue and serum from patients with NSCLC and control 73% of patients had serum DNA that reflected aberrant methylation in their tumors, specifically in CDKN2A, MGMT, DAPK, GSTP1
[60] Paired sputum and lung tissue MSP Lung tissue and sputum from smokers with SCC CDKN2A and MGMT were hypermethylated in both sputum and tumor of patients at time of diagnosis
[72] Bronchial epithelial cells, blood lymphocytes, lung tissue MSP Paired blood and bronchial epithelial samples from smokers/non-smokers with pre-neoplastic lesions and neoplastic lesions from individuals with NSCL versus controls ECAD and DAPK more likely to be methylated in smokers’ peripheral lymphocytes or bronchial epithelium and never methylated in non-smokers
[58] Peripheral blood leukocytes Illumina Beadchip and Pyrosequencing Smokers with recently diagnosed SCLC and controls Forty-three CpG sites were differentially methylated between SCLC and controls, and nine of these, validated by pyrosequencing, could discriminate SCLC with AUC of 0.86
[71] Paired serum and lung tissue MSP Paired serum and lung tissue samples from individuals with lung cancer and controls Six-gene serum panel that discriminated patients with lung cancer with 75% sensitivity and 73% specificity
b) Transcriptomics biomarkers
[76] Bronchial brushing, large airway epithelium Affymetrix array Bronchial brushings of cytologically normal large airway eptihelium obtained from smokers undergoing bronchoscopy for suspicion of lung cancer Eighty gene airway biomarker with >80% diagnostic sensitivity and specificity, and 95% sensitivity and negative predictive value when biomarker is combined with cytology collected at bronchoscopy
[78] Bronchial brushings from normal airway bronchial epithelial cells (StaRT)-PCR Normal bronchial epithelial cells of patients with lung cancer and non-lung cancer controls Fourteen gene airway biomarkers of antioxidant, DNA repair and transcription factor genes with performance in a test AUC >0.84 and an accuracy of 80%
[86] Peripheral blood mononuclear cells cDNA array Blood collection from smokers with newly diagnosed lung cancer confirmed by histopathology twenty-nine-gene blood signature with >80% sensitivity and specificity
[79] Bronchial brushing from airway epithelium Affymetrix array Bronchial airway brushings of cytologically normal epithelium from smokers with and without lung cancer or premalignancy Gene-expression signature of PI3K signaling pathway activation was differentially expressed in airways of smokers with lung cancer or dysplasia and was reversible with chemopreventive therapy
[88] Whole blood Sentrix whole genome bead chips WG6 (Illumina) PAX gene-stabilized blood samples from three independent groups consisting of patients with NSCLC and controls Genes differently expressed in whole blood of patients with NSCLC and controls were used to build a diagnostic classifier with AUC >0.82
[85] Saliva Affymetrix array Whole saliva collected from untreated patients with lung cancer with matched cancer-free controls Seven highly discriminatory transcriptomic salivary biomarker with AUC = 0.925 with >82% sensitivity and specificity
c) MicroRNA biomarkers
[93] Sputum RT-qPCR Sputum from patients with squamous lung cancer and healthy controls Three miRNA diagnosed stage I squamous cell lung cancer with AUC = 0.87
[94] Sputum RT-qPCR Sputum from patients with lung adenocarcinoma and healthy controls Four miRNA diagnosed stage I lung adenocarcinoma with AUC = 0.90
[107] Serum Genoexplorer microRNA expression system Serum from patients with lung cancer versus healthy controls Two miRNA discriminated individuals with early stages NSCLC with AUC = 0.77
[103] Serum Taqman Low Density Arrays RT-qPCR Serum from asymptomatic patients with NSCLC and healthy smokers. Patients were screened by low-dose CT and sera were collected at the time of diagnosis before the surgery Thirty-two miRNA predicted risk of developing lung cancer in asymptomatic high-risk individuals with an accuracy of 80%
[109] Plasma Taqman Low Density Arrays RT-qPCR Multiple plasma samples were collected before and at the time of disease, from two independent spiral CT-screening trials Fifteen miRNA predicted the risk of lung cancer with AUC = 0.85 and 13 miRNA diagnosed lung cancer in undetermined CT nodules with AUC = 0.88
[110] Plasma RT-qPCR Plasma from patients with lung cancer versus healthy controls Four miRNAs discriminated patients with NSCLC with AUC = 0.93
[105] Serum RT-qPCR Serum from patients with lung cancer versus healthy controls Ten miRNAs discriminated patients with NSCLC with AUC = 0.97

Brothers et al.

Brothers et al. BMC Medicine 2013 11:168   doi:10.1186/1741-7015-11-168

Open Data