Title : Feasibility and accuracy evaluation of microfluidic chip for rapid detection of drug-resistant genes in tuberculosis clinical isolates
Abstract:
Objective: The objective of this study was to assess clinical isolates of Mycobacterium tuberculosis using a microfluidic chip, simultaneously detecting eight drug-resistant genes associated with first- and second-line anti-tuberculosis drugs. These findings were compared with phenotypic drug susceptibility results to evaluate the feasibility and accuracy of the microfluidic chip for diagnosing drug-resistant tuberculosis. Focusing specifically on clinical samples of spinal tuberculosis, the microfluidic chip was employed for species identification of Mycobacterium tuberculosis complex and detection of drug-resistant genes, aiming to offer guidance for personalized chemotherapy regimens.
Method: A total of 544 clinical isolates of Mycobacterium tuberculosis with documented drug susceptibility results were gathered from the hospital's strain repository. This cohort comprised 500 drug-resistant strains and 44 susceptible strains. The 544 clinical isolates underwent testing using the microfluidic chip, with results compared against phenotypic drug susceptibility testing to assess the chip's feasibility and accuracy in detecting resistance to eight anti-tuberculosis drugs. Additionally, 136 patients diagnosed with spinal tuberculosis via pathological and radiological assessments were included. Tissue specimens (pus, caseous materials, or granulation tissue) were procured through lesion clearance or CT-guided puncture. These specimens were evenly divided, with one portion undergoing rapid culture using the BACTEC MGIT960 liquid culture system, followed by modified Roche absolute concentration method for drug susceptibility testing on positive samples. The other portion underwent species identification of Mycobacterium tuberculosis complex and detection of mutations associated with resistance to first- and second-line anti-tuberculosis drugs using the microfluidic chip. Phenotypic drug susceptibility testing served as the gold standard to evaluate sensitivity and specificity of the microfluidic chip in detecting clinical samples of spinal tuberculosis.
Results: Among the 544 clinical isolates of Mycobacterium tuberculosis, there were 197 strains resistant to INH and 303 sensitive strains, 219 strains resistant to RIF and 281 sensitive strains, 97 strains resistant to EMB and 403 sensitive strains, 133 strains resistant to SM and 367 sensitive strains, 82 strains resistant to FQS and 418 sensitive strains, 102 strains resistant to AMK and 398 sensitive strains, 70 strains resistant to KAN and 430 sensitive strains, and 49 strains resistant to CPM and 451 sensitive strains. The sensitivity and specificity of the microfluidic chip for detecting INH resistance were 89.34% and 97.36%, respectively; for RIF resistance, 91.78% and 95.73%; for EMB resistance, 76.29% and 95.29%; for SM resistance, 87.97% and 97.00%; for FQS resistance, 95.12% and 96.65%; for AMK resistance, 88.24% and 97.74%; for KAN resistance, 84.29% and 97.21%; and for CPM resistance, 89.80% and 96.90%, respectively.
Using a microfluidic chip, identification of Mycobacterium tuberculosis species was performed on 136 clinical samples of spinal tuberculosis, with a positivity rate of 87.50%. Among culture-positive samples, the sensitivity of Mycobacterium tuberculosis species identification by the microfluidic chip was 95.45%. The sensitivity for samples positive on both smear and culture was 95.80%, while for samples negative on smear but positive on culture, the sensitivity was 31.32%.
Regarding isoniazid (INH) resistance detection, the microfluidic chip exhibited a sensitivity of 80.95% and a specificity of 91.04%. For rifampicin (RIF) resistance detection, the sensitivity of the microfluidic chip was 88.24%, with a specificity of 90.74%. The sensitivity for streptomycin (SM) resistance detection was 81.82%, with a specificity of 92.21%. For fluoroquinolones (FQS) resistance detection, the sensitivity was 88.89%, and the specificity was 97.47%. The sensitivity for ethambutol (EMB) resistance detection was 71.43%, with a specificity of 96.30%. Kanamycin (KAN) resistance detection showed a sensitivity of 75.00% and a specificity of 95.24%. Capreomycin (CPM) resistance detection had a sensitivity of 100% and a specificity of 98.84%.
The average time required for Mycobacterium tuberculosis species identification and detection of drug-resistant mutations using the microfluidic chip was 6 (range 5-9) hours.
Conclusion: The application of the chip in detecting Mycobacterium tuberculosis strains demonstrates high sensitivity and specificity in identifying common first- and second-line anti-tuberculosis drug-resistant mutation genes. The process is simple to operate and time-efficient. In the clinical testing of spinal tuberculosis samples, this microfluidic chip exhibits high sensitivity. Compared to traditional chips, it shows higher sensitivity and specificity, particularly in detecting ethambutol (EMB) and fluoroquinolones, thus providing a promising tool for tuberculosis diagnosis and drug resistance surveillance.
