Detection of a 5-circRNA signature to improve prognostic prediction in gastric cancer

Abstract

Gastric cancer (GC) is a commonly diagnosed malignancy with a high mortality rate worldwide. Despite advances in therapeutic approaches, the 5-year survival rate of patients with GC remains poor. A type of non-coding RNA, circular RNA (circRNA), has been discovered. Some circRNAs are dysregulated in cancer and may have important functions. Thus, circRNAs could be candidate biomarkers for disease diagnosis and prognosis. Our study used a reannotation framework to derive expression values of circRNAs from microarray data sets. We used an integrated pipeline including differential expression analysis and Cox regression analyses to detect GC-associated circRNA signatures. We validated results with 54 paired gastric tumor and adjacent normal tissues. Five circRNA signatures were highly associated with patient overall survival and disease-free survival. Real-time PCR experiments on hsa_circ_0103398 and hsa_circ_0127859 in tumors and normal tissues showed that these circRNAs were significantly upregulated in GC. High expression of hsa_circ_0103398 and hsa_circ_0127859 was closely associated with unfavorable survival time. Our findings indicated that a 5-circRNA signature might serve as a candidate prognostic biomarker of GC.