Bivariate elevation of both CXCL13 and IL-10
| Biomarker . | Sensitivity . | Specificity . | AUC . | CI . |
|---|---|---|---|---|
| CXCL13 | 71% | 95% | 0.841 | 0.779-0.892 |
| IL-10 | 64% | 94.1% | 0.851 | 0.789-0.901 |
| CXCL13 or IL-10 | 84.2% | 90.5% | 0.874 | 0.815-0.919 |
| CXCL13 plus IL-10 | 50% | 99.3% | 0.746 | 0.675-0.809 |
| Cytology | 14% | Reference standard | Reference standard | Reference standard |
| Flow cytometry | 20% | Reference standard | Reference standard | Reference standard |
| Biomarker . | Sensitivity . | Specificity . | AUC . | CI . |
|---|---|---|---|---|
| CXCL13 | 71% | 95% | 0.841 | 0.779-0.892 |
| IL-10 | 64% | 94.1% | 0.851 | 0.789-0.901 |
| CXCL13 or IL-10 | 84.2% | 90.5% | 0.874 | 0.815-0.919 |
| CXCL13 plus IL-10 | 50% | 99.3% | 0.746 | 0.675-0.809 |
| Cytology | 14% | Reference standard | Reference standard | Reference standard |
| Flow cytometry | 20% | Reference standard | Reference standard | Reference standard |
Bivariate elevation of both CXCL13 and IL-10 is more than twice as sensitive as cytology (P < .001) and flow cytometry (P < .05), with equivalent specificity, in the detection of non-HIV–associated PCNSL at diagnosis (Fisher’s exact test [2-tailed]). An algorithm based on detection of elevation of either CXCL13 or IL-10 in CSF is more than 3 times as sensitive compared with the reference standard tests (P < .0001).