Background
Colorectal cancer (CRC) is a highly aggressive, high-incidence malignancy. CRC accounted for approximately one out of every ten cancer cases and deaths. Although miRNAs are often used for medical diagnostic purposes, their diagnostic effectiveness in CRC remains uncertain.
Methods
Therefore, from January 2016 to April 2024, we conducted a comprehensive search of China National Knowledge Internet (CNKI), PubMed, Cochrane Library, Web of Science (WoS) and other resources. The pooled sensitivity, specificity, positive likelihood ratio (PLR), negative likelihood ratio (NLR), diagnostic odds ratio (DOR), area under the curve (AUC) and Fagan plot analysis were used to assess the overall test performance of machine learning approaches. Moreover, we evaluated the publication bias by the Deeks’funnel plot asymmetry test.
Results
Ultimately, a total of 23 publications were identified and incorporated into this meta-analysis. The aggregated diagnostic data were as follows: The sensitivity of the test was 0.83, with a 95% confidence interval of 0.81–0.84. The specificity was found to be 0.83 with a 95% confidence interval (CI) of 0.81–0.84. The PLR was 4.60 with a 95% CI of 3.77–5.62. The NLR was 0.22 with a 95% CI of 0.17–0.27. The DOR was 23.79 with a 95% CI of 16.26–34.81. The AUC was 0.90 with a 95% CI of 0.87–0.92. The Deek funnel plot suggests that publication bias has no statistical significance. The Fagan plot analysis that the positive probability is 50% and the nagative probability is 5%.
Conclusion
In summary, our results suggest the high accuracy of miRNAs in diagnosing CRC.