Table 2.
Receiver operating characteristic curve analysis of the performance of the RISK6 signature to distinguish active TB cases (ATB) from healthy donors (HD) and from latent TB infected individuals (LTBI) in cohorts from Georgia, Madagascar, Lebanon, and Bangladesh.
| TPP requirement | Cut-off | Sensitivity% | Specificity% | Cases, n | Controls, n | AUC | AUC 95%CI | |
|---|---|---|---|---|---|---|---|---|
| Screening test (ATB vs HD) | ||||||||
| Georgia | Sensitivity > 90% | > 0.2583 | 90.6 | 85.7 | 32 | 7 | 96.4% | 90.5–100% |
| Madagascar | > 0.3697 | 90.9 | 87 | 44 | 23 | 95.6% | 90.9–100% | |
| Lebanon | > 0.3171 | 90.5 | 88 | 21 | 25 | 94.7% | 88.6–100% | |
| Bangladesh | > 0.3625 | 90.9 | 68.8 | 44 | 16 | 90.1% | 80.7–99.4% | |
| All | > 0.3209 | 90.1 | 80.3 | 141 | 71 | 92.6% | 88.8–96.3% | |
| Triage test (ATB vs LTBI) | ||||||||
| Madagascar | Sensitivity > 90% | > 0.3697 | 90.9 | 88.5 | 44 | 26 | 92.8% | 85.6–100% |
| Initial TB diagnostic test to replace smear microscopy (ATB (CLT+ AFB−) vs HD) | ||||||||
| Georgia | Sensitivity ≥ 60% | > 0.3514 | 63.6 | 100 | 11 | 7 | 94.8% | 85.1–100% |
| Madagascar | > 0.4298 | 66.7 | 95.7 | 9 | 23 | 96.1% | 90.1–100% | |
| Lebanon | > 0.3217 | 75 | 88 | 4 | 25 | 90% | 78.2–100% | |
| Bangladesh | > 0.3541 | 75 | 68.8 | 4 | 16 | 79.7% | 58.8–100% | |
| All | > 0.3823 | 60.7 | 88.7 | 28 | 71 | 87.7% | 80.6–94.8% | |
| Confirmatory test (ATB (CLT+ AFB−) vs HD) | ||||||||
| Georgia | Sensitivity ≥ 65% | > 0.3131 | 72.7 | 100 | 11 | 7 | 94.8% | 85.1–100% |
| Madagascar | > 0.4298 | 66.7 | 95.7 | 9 | 23 | 96.1% | 90.1–100% | |
| Lebanon | > 0.3217 | 75 | 88 | 4 | 25 | 90% | 78.2–100% | |
| Bangladesh | > 0.3541 | 75 | 68.8 | 4 | 16 | 79.7% | 58.8–100% | |
| All | > 0.3674 | 67.9 | 87.3 | 28 | 71 | 87.8% | 80.6–94.8% | |
The performance of the signature is benchmarked against the WHO TPP for a non-sputum based screening/triage test (at a sensitivity of > 90%, the minimum specificity as set out in this TPP should be ≥ 70%), for an initial TB diagnostic test to replace sputum smear (at minimum 60% sensitivity, the minimum specificity as set out in this TPP should be > 98%) and for a confirmatory test (at minimum 65% sensitivity, the minimum specificity as set out in this TPP should be > 98%)19.
ATB active TB, LTBI latent TB infection (were only recruited from Madagascar), HD healthy donors, CLT+ positive sputum culture, AFB− negative AFB smear microscopy, AUC area under the curve, CI confidence interval, Vs versus.