Abstract
Background
Paucibacillary pulmonary tuberculosis (PTB) is diagnostically challenging because molecular assays are inherently insensitive at sub-threshold bacillary loads. The incremental value of T-SPOT.TB—a host immune-response assay (an ELISPOT-based interferon-γ release assay, IGRA)—combined with molecular testing in this setting is not well established.
Methods
In this single-center, STARD 2015-compliant prospective diagnostic-accuracy study at a regional general hospital (January 2021–December 2023), adults with suspected smear-negative PTB underwent simultaneous GeneXpert MTB/RIF, in-house IS6110 TB-DNA real-time PCR, and T-SPOT.TB. Each participant was classified as PTB or non-PTB by a blinded three-member panel using a reference standard with two PTB-confirmation routes—microbiological (positive mycobacterial culture) or clinical (compatible clinicoradiological presentation plus documented response to standard anti-TB therapy at 6 months); participants meeting neither route were non-PTB. Performance was quantified by sensitivity, specificity, predictive values, likelihood ratios, and balanced accuracy ([sensitivity + specificity]/2). Paired comparisons used the McNemar exact test with Bonferroni correction. No a priori sample-size calculation was performed; analyses are exploratory.
Results
Among 78 participants [PTB
n
= 44 (microbiological
n
= 33, clinical
n
= 11); non-PTB
n
= 34], T-SPOT.TB had the highest individual sensitivity (79.5%; 95% CI 65.5–88.8%) and balanced accuracy (0.780; 0.681–0.872), versus GeneXpert (29.5%; 0.618) and TB-DNA PCR (31.8%; specificity 100%; 0.659). Under a parallel (any-positive) rule, TB-DNA PCR + T-SPOT.TB showed the highest balanced accuracy observed in this cohort (0.826; 0.734–0.907; sensitivity 88.6%, specificity 76.5%; the sensitivity gain over T-SPOT.TB alone did not reach statistical significance, exact McNemar p = 0.125). The triple combination did not improve discrimination (0.796). T-SPOT.TB was the sole positive assay in 21/44 (47.7%) confirmed PTB cases.
Conclusion
In this exploratory single-center cohort, T-SPOT.TB contributed unique diagnostic information (47.7% exclusive detection) not captured by molecular assays. TB-DNA PCR + T-SPOT.TB showed the highest observed balanced accuracy, but, given the single-center design and absence of a priori powering, this is hypothesis-generating and requires confirmation in adequately powered, multicenter studies before any clinical recommendation. Where available, Xpert MTB/RIF Ultra is the preferred first-line molecular assay and should be prioritized, in combination with T-SPOT.TB, in future prospective evaluation.