The purpose of this study was to define a biomarker panel for detection of cancer cells in cytologically negative sputum and to evaluate the panel for assessment of lung cancer risk. We examined 19 genetic and epigenetic markers using a sensitive fluorescence-based method in cytologically negative sputum and in lung tumour tissues from 82 lung cancer patients. We also used these markers to test the sputum of 37 cancer-free individuals who were matched by age, sex, and smoking habit. Based on the concordance of biomarkers in lung tumours and corresponding sputum, and the low prevalence in cancer-free individuals, we selected seven markers for a nested case-control study: microsatellite instability of D9S942; loss of heterozygosity of D9S286, D9S942, GATA49D12, and D13S170; and methylation of p16INK4a and RARbeta. Based on the assumption that a lung cancer cell has alterations in two or more of the seven biomarkers, we compared the pattern of biomarker alteration in lung tumours and corresponding sputum. Our comparison yielded a sensitivity of 82%, specificity of 75%, and concordance of 79%. Three cancer-free individuals were considered to have an elevated risk based on the criterion that their sputum showed alteration in two of the seven biomarkers. One individual was indeed diagnosed as having lung cancer 18 months after sputum collection. In the nested case-control study, six biomarkers showed significantly increased odds ratios ranging from 3.14 to 11.24. Our study defines a biomarker panel for detection of cancer cells in cytologically negative sputum and verifies its use for risk assessment of lung cancer. In combination with conventional diagnostic tools, this multiple genetic and epigenetic panel should improve the detection or risk assessment of lung cancer.
(c) 2007 Pathological Society of Great Britain and Ireland