Purpose: Colorectal cancer (CRC) is one of the leading causes of cancer-related deaths worldwide. In 2016, there were estimated 95,270 newly diagnosed CRC cases, and 49,190 deaths from this disease in the United States. Survival of patients is closely associated with the tumor stage at the time of diagnosis as 5-year relative survival rates range from 65% for all stages, 90% for localized primary tumor, 71% for regional metastasis and 13% for distant metastasis. Post-surgery, adjuvant therapy is only recommended to those with high risk stage II, as well as stage III and IV tumors. However, approximately 40-50% of the patients undergo curative surgery only and 20-30% that are treated with adjuvant chemotherapy, eventually relapse and experience a metastatic disease and eventual death. The current gold standard TNM (Tumor, Node, Metastasis) staging for determining the prognosis of CRC patients remains inadequate at identification of high-risk stage II and III patients that have a high potential of developing tumor recurrence. MicroRNAs (miRNAs) play an important role in CRC development and are emerging as important disease biomarkers. Therefore, in this study we sought to determine the prognostic potential of the miRNAs using a systematic, genome-wide biomarker discovery approach, followed by validation of biomarkers in multiple patient cohorts.
Experimental design: Three independent publicly available genome-wide miRNA expression datasets were used for miRNA biomarker discovery (n=158) and validation (n=109 and n=40) of this miRNA signature for recurrence prediction in stage II and III CRC using multivariate Cox regression analysis. The eight gene miRNA signature discovered from the genome-wide analysis was analytically validated in two independent patient cohorts (n=127 and n=96) using Taqman-based RT-PCR assays.
Results: The genome-wide comprehensive analysis led to the identification of an eight gene miRNA classifier that significantly predicted recurrence free survival (RFS) in training (log-rank p=0.003) and two independent validation cohorts (log-rank p<0.0001 and p=0.002). The RT-PCR based training and validation of the eight gene classifier in two independent clinical cohorts significantly associated with poor prognosis in stage II and III CRC patients (log-rank p<0.004 and p<0.0001). Multivariate analyses performed in these two patient cohorts revealed that the eight gene miRNA classifier served as an independent predictor of poor prognosis in stage II and III CRC patients.
Conclusions: In conclusion, we have identified a novel miRNA-based classifier, which is robustly predictive of poor prognosis in patents with stage II/III CRC, and might facilitate identification and stratification of high-risk patients that are candidates for adjuvant chemotherapy and clinical surveillance.