A Clinical TB Detection Method Based on Molecular Typing Technique with Quality Control

Abstract

The gold standard for diagnosing pulmonary Mycobacterium tuberculosis (TB) is the detection of tubercle bacillus in patient sputum samples. However, current methods either require long waiting times to culture the bacteria or have a risk of getting false-positive results due to cross-contamination. In this study, a method to detect tubercle bacillus based on the molecular typing technique is presented. This method can detect genetic units, variable number of tandem repeat (VNTR), which are the characteristic of tuberculosis (TB), and performs quality control using a mathematical model, ensuring the reliability of the results. Compared to other methods, the proposed method was able to process and diagnose a large volume of samples in a run time of six hours, with high sensitivity and specificity. Our method is also in the pipeline for implementation in clinical testing. Reliable and confirmed results are stored into a database, and these data are used to further refine the model. As the volume of data processed from reliable samples increases, the diagnostic power of the model improves. In addition to improving the quality control scheme, the collected data can be also used to support other TB research, such as that regarding the evolution of the tubercle bacillus.