MicroRNAs as diagnostic biomarkers in primary central nervous system lymphoma: a systematic review and meta-analysis

Abstract

Background Diagnosing primary central nervous system lymphoma (PCNSL) remains a challenge. MicroRNAs (miRNAs) are promising noninvasive markers for the identification of PCNSL. The present study aims to assess the diagnostic value of miRNAs for PCNSL patients as biomarkers. Methods We systematically searched PubMed, Embase and the Cochrane library from inception to June 31, 2020. The pooled sensitivity, specificity, positive likelihood ratio (PLR), negative likelihood ratio (NLR), diagnostic odds ratio (DOR), together with the summary receiver operator characteristic (SROC) curve, and area under the SROC curve (AUC) value were used to estimate overall diagnostic performance. We used Q statistic and I2 to test heterogeneity, and used subgroup analyses to investigate the source of heterogeneity. The statistical analyses were independently performed by two investigators using Stata 14.0 and Revman 5.3. Results In total, eleven studies from six records were included in the current meta-analysis with 281 PCNSL patients and 367 controls. Our statistical analysis demonstrated that the pooled sensitivity, specificity, PLR, NLR, DOR, and AUC were 0.91 (95% CI 0.84–0.95), 0.88 (95% CI 0.84–0.91), 7.48 (95% CI 5.71–9.78), 0.11 (95% CI 0.06–0.19), 70 (95% CI 35–142), and 0.90 (95% CI 0.87–0.92), respectively. The studies had substantial heterogeneity (I2 = 54%, 95% CI 0-100). Two subgroup analyses were conducted based on the type of specimen and miRNAs profiled. Conclusions This meta-analysis indicated miRNAs were suitable as noninvasive diagnostic biomarkers for PCNSL with high accuracy. In addition, both cerebrospinal fluid -based and blood-based miRNAs assays for PCNSL detection were considered reliable for clinical application. MicroRNA-21 assays also seemed to be more accurate in the diagnosis of PCNSL. Good quality studies with large samples should be conducted to verify our results.