Background: Increasing evidence has identified circular RNAs (circRNAs) as ideal molecular biomarkers for cancer diagnosis, therapy, and prognosis. However, the overall diagnostic efficiency of circRNAs remains unclear. Thus, this meta-analysis aimed to comprehensively evaluate the diagnostic accuracy of circRNA expression profiles for cancer. Methods: A literature search of online databases was conducted to identify all eligible studies. The quality of the studies was assessed using the Quality Assessment of Diagnostic Accuracy Studies 2 tool. All statistical analyses were executed using STATA 14.0, Meta-DiSc 1.4, and Review Manager 5.2 software. Results: A total of 32 studies, involving 2,400 cases and 2,295 controls, were included in the diagnostic meta-analysis. The pooled sensitivity, specificity, positive likelihood ratio, negative likelihood ratio, diagnostic odds ratio, and area under the curve were 0.79 (95% CI: 0.73-0.84), 0.73 (95% CI: 0.67-0.79), 2.9 (95% CI: 2.5-3.5), 0.29 (95% CI: 0.24-0.36), 10 (95% CI: 8-13), and 0.83 (95% CI: 0.79-0.86), respectively. The overall analysis suggested that circRNAs are useful diagnostic biomarkers for cancer. Subgroup analysis indicated that plasma samples had a better diagnostic performance than cancer tissue samples for cancer detection. Studies involving >= 100 cases or gastric cancer showed higher sensitivities than those including < 100 cases or other cancers. Conclusion: This meta-analysis revealed that circRNAs were significantly correlated with cancer diagnosis. In addition, circRNAs had good diagnostic accuracy and might serve as effective diagnostic biomarkers for cancer.