Evaluating the efficacy of whole genome sequencing in predicting susceptibility profiles for first-line antituberculosis drugs

Objectives: This study aimed to examine the efficacy of whole genome sequencing (WGS) in accurately predicting susceptibility profiles, potentially eliminating the need for conventional phenotypic drug susceptibility testing (pDST) for first-line antituberculosis drugs in routine tuberculosis diagnosis. Methods: Over the period of 2017 to 2020, 1114 Mycobacterium tuberculosis complex isolates were collected with drug susceptibility testing conducted using the MGIT960 system and WGS performed for predicting drug resistance profiles. In addition, we implemented a new algorithm with an updated WGS workflow, omitting pan-susceptible strains from pDST. Results: Results showed that out of 1075 analysed isolates, WGS-based genotypic sensitivity predictions for isoniazid, rifampicin, ethambutol, and pyrazinamide were 100% (95% CI, 99.6-100%), 100% (95% CI, 99.62-100%), 99.8% (95% CI, 99.26-99.94%), and 100% (95% CI, 99.63-100%), respectively. In contrast, the WGS-based genotypic resistance prediction, was 98.85% (95% CI, 93.77-99.79%) for isoniazid, 94.74% (95% CI, 82.71-98.54%) for rifampicin, 86.96% (95% CI, 67.87-95.46%) for ethambutol, and 75.7% (95% CI, 59.9-86.63%) for pyrazinamide. Moreover, WGS enabled the implementation of a new testing algorithm that made it unnecessary to perform pDST in 954 of all 1075 samples (88.7%) and in 890 of 901 pan-susceptible samples (98.8%). Discussion: Integrating WGS into tuberculosis management offers significant potential to replace phenotypic drug susceptibility testing, especially for problematic drugs like pyrazinamide and ethambutol, potentially improving treatment outcomes.