P34

Abstract

Lung cancer (LC) is currently the world’s leading cause of cancer-related mortality with overall mortality to incidence ratio of 0.87. Only 15% of LC patients are diagnosed at early stage of disease and have 5-year survival rate of 54%, whereas 56% of patients have distant metastases and 5-year survival rate of about 4%. Current methods of lung cancer diagnostics are not efficient as screening tools due to high costs (CT), low sensitivity and health risks (radiology), invasiveness (biopsy) or low prevalence in population (genetic alterations). New strategies for preclinical lung cancer screening as well as monitoring of post therapy relapses are required. MiRNAs circulating in blood were shown to reflect the progression of disease and are thus considered as potential biomarkers for cancer diagnostics and theranostics. In this study we investigated seven-miRNA signature circulating in plasma of lung cancer patients and healthy individuals. All seven miRNAs were previously shown to be involved in either regulation of cell cycle and apoptosis and/or tumor development, invasion and vascularization. Blood samples of 50 healthy individuals were obtained from Center of New Medical Technologies (Novosibirsk, Russia) and Novosibirsk Research Institute of Circulation Pathology of E.N. Meshalkin (Novosibirsk, Russia). Samples of 75 lung cancer patients, including 52 patients with squamous cell carcinoma (SCC) and 18 patients with adenocarcinoma (AC) were obtained from Novosibirsk Research Institute of Circulation Pathology of E.N. Meshalkin (Novosibirsk, Russia) and Tomsk Cancer Research Institute RAMS (Tomsk, Russia). None of the patients have undergone surgical treatment or received chemotherapy prior to blood collection. Lung biopsy specimens and imaging techniques were applied to confirm the histopathological features and tumor stages of LC patients. Study was approved by ethical committees of all participating organizations and written informed consent was provided by all participants. Circulating miRNAs were isolated from blood plasma using a single-phase phenol-free protocol (Zaporozhchenko et al., Anal Biochem, 2015; Rus. patent application No. 2014137763, priority date 17.09.2014). Concentrations of miRNAs (miR-21, miR-19b, miR-126, miR-25, miR-205, miR-183, miR-125b) were measured by qRT-PCR and normalized to miR-16. Concentrations of four miRNAs (miR-19b, miR-21, miR-25, miR-183) were significantly different in lung cancer versus healthy individuals (p<0.05, T-test, two-sided). Two miRNAs were upregulated (miR-19b, miR-21), two were downregulated (miR-25, miR-183) in cancer patients. Four miRNAs (miR-19b, miR-126, miR-25, miR-205) were found to be differentially regulated in SCC patients when compared to healthy controls. In AC patients only two miRNAs (miR-19b, miR-183) were differently expressed. Thus, cancer subtypes have different input into overall picture. Receiver Operating Characteristic (ROC) curve analysis of differentially expressed miRNAs in total study population and subtype-based groups showed that miR-19b has highest predictive value for total population and SCC patients, while miR-183 was more effective in discriminating patients with AC. All miRNAs except miR-21 showed strong bias towards one of the subtypes. Presumably a combination of miRNAs with opposite bias should provide a more potent diagnostic tool for cancer detection in total population than individual miRNAs or panels of miRNAs that are specific to one cancer subtype. Indeed, stepwise binary logistic regression has identified the combination of miR-19b and miR-183 to be a strong prediction of disease in total population and yielded a solid increase in AUC: 0,990 (miR-19b+miR-183) versus 0,806 (miR-19b) or 0,924 (miR-183). This combination can be used to identify lung cancer with 94.7% sensitivity and 95.2% specificity. Thus, biological aspects such as tumor genetics and phenotype, stage of disease, response to therapy, and other meaningful tumor properties of such heterogeneous malignancies as LC can interfere with contents and composition of cell-free miRNA pool. Wide profiling of circulating miRNAs will specify biomarkers of LC phenotypes and improve non/less-invasive LC diagnostics.