Circular RNAs as Diagnostic and Prognostic Indicators of Colorectal Cancer: A Pooled Analysis of Individual Studies

Abstract

Background: Circular RNAs (circRNAs) have proven as a special subset of endogenous RNAs that are implicated in the tumorigenesis of various cancers. This study sought to evaluate the role of circRNAs in the diagnosis and prognosis of colorectal cancer (CRC). Methods: The online databases were searched for collecting relevant studies on circRNAs as diagnostic and prognostic biomarkers of CRC. Two researchers independently screened literature, extracted data, and evaluated the bias and risks of included studies. The diagnostic and prognostic indicators were merged and analyzed using STATA 12.0 software, and sources of heterogeneity were traced by the sensitivity analysis and the meta-regression test. Results: A total of 29 articles representing 2639 CRC patients were included. The pooled sensitivity, specificity, and area under the curve (AUC) of circRNAs in differentiating CRC from non-tumor control were 0.75 (95% CI: 0.69-0.80) and 0.74 (95% CI: 0.69-0.78) and 0.81, respectively. The survival analysis showed that up-regulations of up-regulated circRNAs were significantly related to dismal survival in CRC patients (HR = 2.38, p < 0.001). A stratified analysis showed that the comprehensive diagnostic value of up-regualted circRNAs in CRC was higher than that of down-regualted circRNAs (AUC: 0.83 vs. 0.77; Z test, p < 0.05). The efficacy of tissue-derived circRNAs in the diagnosis of CRC was equal to that of plasma/serum-derived ones (AUC: 0.81 vs. 0.82; Z test, p > 0.05). Conclusion: Abnormally expressed circRNAs as auxiliary biomarkers present underlying value in the diagnosis and prognosis prediction of CRC.

Keywords: circular RNA; colorectal cancer; diagnosis; pooled analysis; prognosis.

FIGURE 1

The flow chart of literature inclusion and exclusion process according to the PRISMA statement.

FIGURE 2

Forest plots of pooled (A) sensitivity, (B) specificity, (C) DOR, and (D) AUC for the diagnostic effect of circRNAs in CRC.

FIGURE 3

Forest plots of the prognostic effect of (A) up-regulated, (B) un-adjusted down-regulated circRNAs and (C) adjusted down-regulated circRNAs on survival in CRC patients.

FIGURE 4

The influence analysis of the homogeneity among the included studies. (A) the overall diagnostic efficacy of the studies. (B) the diagnostic efficacy of plasma/serum-based circRNA testing. (C) up-regulated and (D) down-regulated circRNAs in predicting the OS in CRC patients.

FIGURE 5

Appraising the publication bias among the included studies. (A) the overall diagnostic effect assessed by the Deek’s funnel plot. Begg’s test for the effects of (B) down-regulated and (C) up-regulated circRNAs in predicting the OS in CRC patients. (D) funnel plot of the trim-and-fill method shows no significant impacts on the combined effect of up-regulated circRNAs in CRC.