Cytokine profile determined by data-mining analysis set into clusters of non-small-cell lung cancer patients according to prognosis

Abstract

Background Immunoregulatory cytokines may play a fundamental role in tumor growth and metastases. Their effects are mediated through complex regulatory networks. Human cytokine profiles could define patient subgroups and represent new potential biomarkers. The aim of this study was to associate a cytokine profile obtained through data mining with the clinical characteristics of patients with advanced Non-Small Cell Lung Cancer (NSCLC). Patients and Methods We conducted a prospective study of the plasma levels of 14 immunoregulatory cytokines by ELISA and a Cytometric Bead Array (CBA) assay in 110 NSCLC patients prior to chemotherapy and 25 control subjects. Cytokine levels and data mining profiles were associated with clinical, quality of life and pathological outcomes. Results NSCLC patients had higher levels of IL-6, IL-8, IL-12p70, IL-17a and IFN-g and lower levels of IL-33 and IL-29 compared to controls. The pro-inflammatory cytokines IL-1b, IL-6 and IL-8 were associated with lower hemoglobin levels, worse functional performance status (ECOG), fatigue and hyporexia. The anti-inflammatory cytokines IL-4, IL-10 and IL-33 were associated with anorexia and lower body mass index. We identified 3 clusters of patients according to data mining analysis with different overall survival (OS) (25.4, 16.8 and 5.09 months respectively, P=0.0012). Multivariate analysis showed that ECOG performance status and data mining clusters were significantly associated with OS (RR 3.59, [1.9-6.7], P<0.001 and 2.2, [1.2-3.8], P=0.005). Conclusion Our results provide evidence that complex cytokine networks may be used to identify patient subgroups with different prognoses in advanced NSCLC. These cytokines may represent potential biomarkers, particularly in the immunotherapy era in cancer research.