Abstract
Background: The availability of targeted Next Generation Sequencing (tNGS) for drug-resistant tuberculosis (DRTB) is a major advance in TB diagnostics, but its optimal implementation, yield, and cost in high TB-incidence settings remain unclear. We aimed to evaluate tNGS in relation to laboratory workflow, implementation cascade, turnaround time, and unit cost.
Methods: We prospectively evaluated tNGS implementation with Deeplex Myc-TB for sputum samples at the provincial reference laboratory in West Java, Indonesia, from 1 August 2024 to 10 October 2025. We included patients with Xpert-diagnosed rifampicin-resistant TB (RRTB) at 32 hospitals and 15 community health centres, collectively catering to 51.0% of RRTB patients in the province.
Findings: From 1210 RRTB sputum samples collected, 812 (67.1%) samples completed sequencing, with 429/812 (52.8%) yielding complete and 164/812 (20.2%) partial reads. Success of tNGS depended on bacterial load, ranging from 88.8% for Xpert high to 0% for very low samples. Majority samples were confirmed as RR/MDR TB (89.2%), with pre-XDR and XDR TB recorded in 10.6% and 0.2%, respectively. Workflow optimisation improved sequencing performance, with final workflow reaching a laboratory success of 82.4% (201/244). The median turnaround time for tNGS from sample reception until sequencing results was eight days (IQR 6-10). The total tNGS unit cost per sample was $263.6 for iSeq and $239.6 for MiSeq.
Interpretation: tNGS performed well in those with high DNA sputum load with relatively rapid turnaround time. Its optimisation for wide adoption should consider DRTB burden, sample selection, sequencing platform, and costs.