Clinical utility of cerebrospinal fluid interleukin-10 in central nervous system lymphoma: a molecular neuropathology-informed diagnostic meta-analysis.

Abstract

Accurate diagnosis of central nervous system lymphoma (CNSL) remains challenging, and cerebrospinal fluid (CSF) interleukin-10 (IL-10) has emerged as a promising minimally invasive biomarker. We conducted a PRISMA-DTA systematic review and diagnostic meta-analysis to evaluate the diagnostic performance of CSF IL-10 for CNSL and to examine its clinical utility and robustness using a reproducible Python-based workflow. Eleven studies comprising 1,462 participants (510 CNSL cases and 952 controls) met the inclusion criteria. Using a bivariate random-effects model, the pooled sensitivity and specificity were 79.8% (95% CI, 76.1-83.2) and 94.8% (95% CI, 93.2-96.0), respectively. The pooled positive likelihood ratio was 18.04 (95% CI, 8.35-38.98), the negative likelihood ratio was 0.185 (95% CI, 0.118-0.291), and the diagnostic odds ratio was 130.55 (95% CI, 47.85-356.14). The summary receiver operating characteristic curve demonstrated excellent overall accuracy, with an area under the curve of 0.973 and a Q* index of 0.9246. Leave-one-out sensitivity analysis showed only modest changes in the pooled estimates, supporting the stability of the results. Δ-sensitivity ranged from - 1.9% to + 3.0% and Δ-specificity from - 1.1% to + 2.0%, supporting robustness. Fagan's analysis, using a pre-test probability of 34.9%, yielded post-test probabilities of 90.6% following a positive result and 9.0% following a negative result. Publication bias was suggested by Begg's test (τ = 0.400, p = 0.0434) and Egger's test (intercept = 3.283, 95% CI, 1.337-5.229; p = 0.0021), while trim-and-fill analysis indicated potential small-study effects. Overall, CSF IL-10 demonstrates high diagnostic accuracy and strong rule-in utility for CNSL. However, prospective multicenter studies with standardized thresholds and pre-analytical protocols are needed to confirm its clinical applicability.