Tuberculosis (TB), caused by Mycobacterium tuberculosis (MTB) infection, is the ninth highest cause of death worldwide. There are an estimated 10 million cases of TB globally, with 1.6 million TB-associated deaths per year [12]. Although the incidence of TB is decreasing by 2% every year, its prevalence and mortality rate remain high, necessitating comprehensive efforts for eradication [2].

Multidrug-resistant (MDR) TB and extensively drug-resistant (XDR) TB are of particular concern as they are difficult to treat. MDR-MTB exhibits resistance to two of the most important first-line drugs, rifampin (RIF) and isoniazid (INH), while XDR-MTB demonstrates resistance to RIF and INH, as well as to at least one fluoroquinolone (FQ) and at least one second-line injectable drug (SLID; kanamycin, amikacin, and capreomycin) [3]. Treatment success rates for MDR-TB and XDR-TB are low, at 54% and 30%, respectively—and they are the main obstacles in TB eradication [1]. Therefore, rapid determination of the drug susceptibility of the TB-causing bacteria is important to ensure appropriate treatment.

Drug susceptibility testing (DST), used to select appropriate drugs, is a culture-based method and requires approximately six to eight weeks for completion [4]. To overcome the limitations of conventional assays, DST based on molecular diagnostic assays has been developed. For example, GenoType MTBDRplus (Hain Lifescience, Nehren, Germany) and GenoType MTBDRsl (Hain Lifescience) are employed for the rapid identification of gene mutations related to MDR- and XDR-MTB using a line probe assay, while GeneXpert MTB/RIF (Cepheid AB, Solna, Sweden) is used to determine RIF resistance by real-time PCR [5].

Recently, QuantaMatrix Inc. (Seoul, Korea) developed the QuantaMatrix Multiplexed Assay Platform (QMAP), which utilizes magnetic micro-particles and a reverse hybridization assay. In QMAP, a probe for a specific gene is combined with a barcoded magnetic micro-particle, which is a carboxyl-functionalized magnetic disk with a 50-µm-thick silica-coated surface and a graphical barcode that allows >100-plex coding capacity in high-throughput analysis [6]. Each probe enables the capture of PCR products with a complementary sequence and then emits fluorescence. A 100-plex capacity in a single microwell system allows the testing of 100 types of pathogens in one microwell with one sample. Previously, we developed an assay for detecting MDR-MTB based on QMAP and evaluated the utility of the assay using strains isolated from TB patients in Korea [7].

We have now developed a more sophisticated QMAP MDR/XDR assay by adding a probe to identify genetic mutations associated with resistance to ethambutol (EMB), streptomycin (SM), FQ, and SLID in addition to RIF and INH. This molecular assay can simultaneously detect MDR- and XDR-MTB within six hrs. We evaluated the utility of this assay using clinical strains isolated from TB patients in Myanmar, which is among the 22 countries with the highest TB burden and is included in the global list of 27 countries with a high incidence of MDR-TB [8].

The utility of the QMAP assay was evaluated using MTB strains isolated from TB patients. We also compared the results of QMAP and pDST; the assays showed an overall concordance rate of 87.9%. The DNA sequencing results of the 23 isolates with discordant results matched QMAP results. Because of the inherent limitation of molecular DST methods, not all strains resistant to a specific drug can be detected as many drug resistance genes remain to be identified. The purpose of molecular DST is timely detection of resistant strains as many as possible. In that context, our current assay could detect 87.9% of MDR- and XDR-MTB isolates accurately and rapidly.

Currently, several molecular DST assays, such as the AdvanSure MDR-TB GenoBlot assay kit (LG Chem, Seoul, Korea) and GenoType MTBDRplus, are clinically used. Both kits demonstrated good performance for MDR-MTB detection in clinical isolates, with concordance rates of 94.3% and 88.5–98.2% [101112], respectively. One study reported that the results of GenoType MTBDRsl, a kit for XDR-MTB detection, showed a concordance rate of 94.7% [13]. However, a pooled analysis reported lower sensitivity of 75–80%, accompanied by high specificity of 91–100%, for detection of XDR-MTB isolates [14]. Our discordant pDST and QMAP results might have occurred for the following reasons. First, not all mutations associated with drug resistance are known [15]. Currently, only approximately 95% of RIF resistance due to the rpoB gene, which contains the RIF resistance determining region (RRDR), is detectable [16]; in case of katG and inhA, which are genes related to INH resistance, the detection rates of INH resistance have been reported to be 70% and 10%, respectively [17]; and for the gyrA and gyrB regions associated with FQ resistance, the detection rates is only approximately 60% [18]. Second, the tested isolates possibly exhibited heteroresistance. Heteroresistance is detected in 20–30% of TB patients and can be caused by a mixed infection of two different isolates or by acquisition of drug resistance during treatment [192021]. Patients infected with heteroresistant isolates might convert to full resistance [22]. In such cases, ≤1% of the DNA of resistant isolates could be detected by molecular methods [23]. Third, isolates with low-level resistance and borderline resistance exhibiting susceptibility in pDST could be detected by molecular methods [24]. For these reasons, although molecular assays might not replace the pDST, they could be more widely used for detecting drug resistance, as WHO has recently reported [25].

The following study limitations should be considered. First, the QMAP MDR/XDR assay results showed low concordance (56.5%) with pDST results for detecting EMB resistance. This low detection rate could be attributed to a lack of appropriate probes. EMB resistance occurs most frequently at codons 306, 406, and 497 in the embB gene [2627]; however, we designed probes related only to codon 306. Second, this study was conducted using only selected strains, including a high proportion of drug-resistant strains, and the performance evaluation was mainly focused on the detection of resistance. However, in the real situation, there would be a larger proportion of susceptible strains, potentially leading to false-positive detection of resistance. Therefore, further studies using different and a greater number of probes and including an adequate number of susceptible strains are needed to improve the performance of QMAP MDR/XDR assay and confirm its specificity.

As not all types of mutations that cause drug resistance have been identified, using molecular assays in conjunction with pDST could enable the rapid and accurate determination of drug susceptibility of MTB isolates, thereby facilitating timely initiation of appropriate anti-TB drug therapy [28]. In particular, the 100% concordance between DNA sequencing and QMAP results indicates that QMAP could detect all known, or at least, targeted DNA mutations associated with drug resistance. Therefore, the QMAP MDR/XDR assay can be used for molecular diagnosis of MDR- and XDR-TB.