Differentiation of NSCLC patients and healthy controls by a urinary microRNA panel

Abstract

Background: Previous studies showed that microRNAs (miR) in tissue biopsies and blood samples are potential biomarkers for diverse types of cancer. MiRs are actively secreted to end up in the blood and pass the kidneys to be excreted through urine. Aims of this study were to investigate whether miRs in urine can be used to diagnose non-small cell lung cancer (NSCLC) and to compare diagnostic miR panels of blood and urine. Methods: We used next generation sequencing (NGS) and differential expression analyses to identify diagnostic miRs in cell-free urine and plasma of 15 stage III and IV NSCLC patients and 13 healthy individuals. An independent urine dataset derived from a previous study with 13 NSCLC patients and 15 healthy controls was used to confirm reliability. Results: Urine of NSCLC patients contained a significant increased expression of 5 miRs and this panel identified lung cancer with a sensitivity of 76%, a specificity of 92% and an AUC of 85%. In the independent urine dataset these results persisted with a sensitivity and specificity of 63% and 100% and an AUC of 73%. No differentiation could be made between plasma samples using the same miR panel, in fact a completely different miR panel was identified by differential expression analyses with only limited differentiation power. Conclusion: A 5-miR panel in urine of patients showed good diagnostic accuracy in the differentiation between NSCLC patients and healthy controls. The diagnostic potential was confirmed in an independent dataset. Remarkably there was no differentiation power of this panel in plasma samples, nor could we identify a different diagnostic miR panel.