Molecular Biomarkers And Their Correlations With Clinicopathologic Parameters Of Lung Cancer

Lung cancer is a highly lethal and extremely common cancer worldwide, with a5-year survival rate of only about15%. Nearly70%of patients are diagnosed with advanced disease that is not amenable to curative therapy. Early diagnosis is of vital importance for improved prognosis. Meanwhile, even for patients diagnosed at stage Ⅰ, the5-year survival rate varies a lot. Until now, no robust biomarkers are applied into clinical practice providing diagnostic and prognostic evaluation of lung cancer.The aims of our study are:1. Screening for abnormal expressed proteins in lung cancer tissue samples and identifying a protein panel with high sensitivity and specificity in diagnosing lung cancer.2. Screening for abnormal expressed proteins and chromosomal abnormalities in bronchial brushings and/or sputum specimens of early-stage lung cancer patients; in order to identify potential protein panels or probe sets assisting conventional cytological examination.3. To find potential prognostic biomarkers in lung cancer.The study using lung cancer tissue samples was designed as two phases; one was discovery phase, in which we evaluated the expression status of186proteins immunohistochemically in nearly500cases of lung cancer. In comparing with adjacent tissues, the expression of15proteins remarkably increased in tumor tissues. Analyzing by different combinations, we found that the6-protein panel containing TP53, Ki67, MCM6, MCM7, KIAA1522and KIAA0317showed the highest sensitivity in both training set (245cases) and testing set (246cases) of82.6%(194/235) and84.4%(194/230), while in operative margin tissues without tumor involvement it showed positivity in less than5%cases. In the validation phase composed by342tissue samples, the sensitivity of the panel was81.8%(261/319), while in operative margin tissues, only2.4%(7/297) cases showed positive expression. Stratified analysis indicated that positive diagnosis rate of the panel could reach80%in detection of early-stage lung cancer in the cohorts. Furthermore, it showed a high diagnostic sensitivity and specificity of97.3%(36/37) in38cases of small cell lung cancers.In viewing that detection of tissue samples could not be used in early diagnosis or screening of lung cancer, we explored and validated biomarkers in non-invasive or mini-invasive specimens. The expression of45proteins was evaluated in880bronchial brushings using immunocytochemical technique. Nine proteins were observed with elevated expression in bronchial brushings of lung cancer patients comparing with that of benign lung lesions. The sensitivity and specificity were81.1%(309/381) and83.3%(65/78) with an optimal combination of several overexpressed proteins (TP53, Ki67, MCM6, MCM7, KIAA1522and KIAA0317) in detection of non-small-cell lung cancer, meanwhile, the positive diagnostic rate of the panel is86.8%(145/167) in detecting small-cell lung cancer, the high sensitivity also maintained in limited stage cases or peripheral lung cancer. More importantly, the panel could detected93.2%(68/73) NSCLCs and89.7%(61/68) SCLCs in outpatients. In142cases of sputum specimens, we found the combination of cytology and the panel significantly improved the sensitivity of cytological examination for lung cancer detection (P=0.002). Centromeric enumeration probes (CEPs) for chromosomes2,3,6,7,8,9,11,12and17were analyzed using FISH in74surgical resection tissues,32operative margin tissues without tumor involvement of lung cancer and174bronchial brushings. The combination of cytology and FISH using the three-probe set for chromosomes3+7+8could significantly improve the sensitivity of bronchial brushing examination for lung cancer detection (P=0.00003). Compared to cytology alone, this increased sensitivity was achieved, and high specificity was maintained. Moreover, significantly more peripheral tumors were identified using the combination of FISH and cytology compared to cytology alone (93.3%vs.58.7%; P=0.008).We also demonstrated some prognostic biomarkers of non-small-cell lung cancer. Kaplan-Meier curves indicated that patients with high expression of MCM7, KIAA1522, TIMP1, PTPN1, ICBP90or VPRBP exhibited a worse overall survival (OS) compared to those with low/no expression of the protein. Decreased P120expression was significantly associated with poor prognosis of patients. Multivariate Cox regression analysis showed that MCM7, KIAA1522, TIMP1, PTPN1and P120could act as independent prognostic indicators in tissue samples. We also revealed for the first time that KIAA1522was an independent prognostic indicator even for early-stage (stage0-Ⅱ) NSCLCs as compared with sex, stage, tumor size, N-status and tumor differentiation (P=0.00025, HR=2.317,95%CI:1.477-3.635). It is worth to mention that MCM7in bronchial brushings also showed an independent prognostic value (P=0.0318, HR=4.347,95%CI:1.136-16.627), which may be useful when biopsies are unavailable. These data suggest that MCM7overexpression could indicate a subset of patients for whom a more aggressive treatment should be considered, and the evaluation of MCM7expression may provide useful information for doctors to make optimal clinical decisions, to which further validation studies are needed.