Targeted next-generation sequencing - a promising approach in the diagnosis of Mycobacterium tuberculosis and drug resistance

Abstract

Targeted next-generation sequencing (tNGS) offers a high-throughput, culture-independent approach that delivers a comprehensive resistance profile in a significantly shorter turn-around time, making it promising in enhancing tuberculosis (TB) diagnosis and informing treatment decisions. This study aims to evaluate the performance of tNGS in the TB diagnosis and drug resistance detection of Mycobacterium tuberculosis (MTB) using MTB clinical isolates and bronchoalveolar lavage fluid (BALF) samples. A total of 143 MTB clinical isolates were assessed, tNGS, phenotypic antimicrobial susceptibility testing (AST), and AST based on whole genome sequencing (WGS) exhibited high concordance rates, averaging 95.10% and 97.05%. Among 158 BALF samples, culture, Xpert MTB/RIF, and tNGS reported 29, 70 and 111 positives, respectively. In the confirmed cases with etiological evidence (smears, cultures, or molecular test), the positive rate of tNGS (73/83, 87.95%) was higher than that of Xpert MTB (67/83, 80.72%). Additionally, 45% (27/60) of clinically diagnosed cases (with imaging or immunological evidence) were positive for tNGS. Further validation on the discrepant results between tNGS and Xpert MTB/RIF with droplet digital PCR (ddPCR) yielded 35 positives, tNGS detected all, and Xpert MTB/RIF only identified 6 positives. In conclusion, tNGS demonstrates robust and rapid performance in the identification of MTB and its associated drug resistance, and can be directly applied to clinical samples, positioning it as a promising approach for laboratory testing of tuberculosis.

Keywords: Mycobacterium tuberculosis; Diagnosis; Drug resistance; Targeted next-generation sequencing.

Fig. 1

Comparative analysis and workflow schematic of MTB diagnosis and drug resistance detection. (A) Comparison of the performance of tNGS, WGS, and phenotypic AST in detecting drug resistance on MTB isolates. (B) Comparison of the performance in MTB and rifampicin-resistant gene identification between tNGS and Xpert MTB/RIF on BALF samples. (C) The workflow of the tNGS in routine clinical practice

Fig. 2

The overall performance of tNGS and WGS for each drug using phenotypic AST as the reference standard

Fig. 3

Further validation of the discrepant results Note: A) Relationship between Xpert MTB/RIF results with varying bacterial loads and the RPhK detected by tNGS. The tNGS RPhK prevalence was plotted based on different Xpert MTB/RIF semi-quantitative positive categories: RPhK_{Xpert MTB/RIF(high)} = [74323, 85882]; RPhK_{Xpert MTB/RIF(medium)} = [9105,71443]; RPhK_{Xpert MTB/RIF(low)} = [34, 37556]; RPhK_{Xpert MTB/RIF(very low)} = [0, 3036]; RPhK_{Xpert MTB/RIF(−)−ddPCR(+)} = [3, 1973]; RPhK_{Xpert MTB/RIF(−)−ddPCR(−)} = [14, 234]. Significant differences were observed in the RPhK values between the Xpert MTB/RIF(high) group and the Xpert MTB/RIF(medium) group (P = 0.02, Wilcoxon rank sum test), between the Xpert MTB/RIF(medium) group and the Xpert MTB/RIF(low) group (P = 0.035, Wilcoxon rank sum test), and between Xpert MTB/RIF(low) group and the Xpert MTB/RIF(very low) group (P = 1.3e-06, Wilcoxon rank sum test). The RPhK values for both ddPCR(+) and ddPCR(-) did not show a significant difference (P = 0.11, Wilcoxon rank sum test). B) Detection results of Non-tuberculosis. C) Validation of the inconsistent MTB identification results between Xpert MTB/RIF and tNGS using ddPCR (N = 38/51^#)