PATH-44. MULTIPLE BIOMARKER ALGORITHM BASED ON CXCL13, IL-10, IL-2 RECEPTOR, AND β2-MICROGLOBULIN IN CEREBROSPINAL FLUID TO DIAGNOSE CENTRAL NERVOUS SYSTEM LYMPHOMA

Abstract

Abstract BACKGROUND Brain biopsy is the gold standard for the diagnosis of primary central nervous system lymphoma (PCNSL). However, the biopsy procedure has a risk of complications, such as hemorrhage and seizure. We have reported that cerebrospinal fluid (CSF) interleukin-10 (IL-10), soluble IL-2 receptor (sIL-2R), and β2-microglobulin (β2-MG) are the useful diagnostic biomarkers for PCNSL. Recently, the C-X-C motif chemokine ligand 13 (CXCL13) has been reported to be another useful biomarker for the PCNSL. The aim of this study was to validate the diagnostic performance of the CSF CXCL13 and diagnostic algorithm based on the combination of these biomarkers. METHODS We retrospectively examined the CSF CXCL13 concentration, as well as IL-10, sIL-2R, and β2-MG, in a case-control study (n=230). We used logistic regression analyses to create diagnostic algorithms based on these CSF biomarkers. To examine the utility of the multivariate diagnostic algorithm, we conducted the prospective study and evaluated the accuracy of this algorithm. RESULTS In case-control study, we have demonstrated that CSF CXCL13 levels were significantly higher in the patients with PCNSL (area under the curve (AUC)=0.981). A total of 84 patients were enrolled in the prospective study. The multivariate diagnostic algorithm using CSF levels of CXCL13, IL-10, sIL-2R, and β2-MG demonstrated excellent diagnostic performance: positive predictive value was 89% and negative predictive value was 100%. Four markers’ combination (CXCL13 + IL-10 + β2-MG + sIL-2R) had the highest AUC (AUC=0.998). The misdiagnosis cases of the algorithm were only 3 cases; CNS Sjögren’s syndrome, histiocytic sarcoma, and glioblastoma. In addition, CSF CXCL13 levels were prognostic biomarkers in PCNSL patients. CONCLUSIONS Our study suggests that the algorithm based on 4 CSF biomarkers had excellent diagnostic performance in CNS lymphoma. However, this algorithm should be further validated in prospective cohort studies with larger numbers of patients.