EVALUATION OF MYCOBACTERIUM TUBERCULOSIS-SPECIFIC IFN-G, TNF-A, CXCL10, IL2, CCL2, CCL7 AND CCL4 LEVELS FOR ACTIVE TUBERCULOSIS DIAGNOSIS

BackgroundNovel diagnostic tests for active tuberculosis (ATB) are urgently needed. We aimed to efficiently and robustly assess whether seven previously identified, promising biomarkers (IFN-g, TNF-a, CXCL10, IL2, CCL2, CCL7 and CCL4) could distinguish patients with ATB within a cohort of patients presenting with the full clinical spectrum of suspected TB in routine practice.MethodsWe designed a nested case-control study (n=92) within the IDEA study.1 Uniquely, we enriched our ATB population to include ~50% patients in whom current IGRAs fail (and unmet clinical need is greatest), to assess whether any biomarker offered superior diagnostic accuracy to IFN-g. We utilised stored supernatants from QFT-GIT tests performed in the IDEA study and compared Mycobacterium tuberculosis-specific biomarker levels in patients with ATB and non-tuberculosis respiratory diseases using Meso Scale Discovery U-PLEX assays. We analysed group differences using Kruskal-Wallis tests.ResultsIn phase I, we analysed IFN-g, TNF-a, CXCL10, IL2, CCL2, CCL7 and CCL4 levels in 32 patients. MSD-measured biomarkers (except CCL4) detected higher numbers of true positives (TP) compared to QFT-GIT, however, all biomarkers lost specificity (42.9% increase in false positive (FP) results). TNF-a, IL2 and CCL7 had very low raw biomarker concentrations. In Phase II, based on above results, we analysed IFN-g, CXCL10 and CCL2 in a further 60 patients (table 1). CXCL10 achieved the highest increase in TP results for ATB diagnosis, with 43 TP compared with 26 TP results for QFT-GIT. MSD-measured IFN-g detected 41 TP results, while CCL2 only detected 26 TP. All three biomarkers demonstrated >35% loss of specificity compared to QFT-GIT. The QFT-GIT sensitivity and specificity values in our study population were 45.6% and 80% respectively. A ‘triple-test’ combining IFN-g, CXCL10 and CCL2 results achieved sensitivity 83.3% and specificity 12.1%.ConclusionOur study provides a unique and novel gating method for efficiently assessing the diagnostic performance of candidate biomarkers for ATB diagnosis. Our study population was purposely engineered to compare candidate biomarkers to QFT-GIT in a clinically-relevant manner and we hope our study design will aid future, targeted efforts for high-quality biomarker follow-up studies.ReferenceWhitworth HS, et al. Lancet Infect Dis. 2019;19(2):193–202. PMID:30655049 Please refer to page A188 for declarations of interest related to this abstract.