A Novel Transcriptomic Biosignature for Tuberculosis Diagnosis, Treatment Monitoring and Outcome Prediction = A Novel Transcriptomic Biosignature for Tuberculosis Diagnosis, Treatment Monitoring and Outcome Prediction

다국어 초록 (Multilingual Abstract)

Background
Transcriptomic biomarkers have been demonstrated as a promising tool for accurate diagnosis and differential diagnosis of tuberculosis (TB). However, little is known about their potential in TB treatment monitoring and outcome prediction.
Methods
We carried out a comprehensive search for available transcriptomics data. The data was used for the discovery of genes that show significant responses during the treatment of TB. Genes with consistent downtrend alterations were determined. Potential candidates were then subjected to external validation. The potential of the biomarker candidates in TB diagnosis and active TB progression risk evaluation was also evaluated.
Results
The transcriptome of the TB patients was significantly changed during the treatment. There were 371 significantly expressed genes with the monotonic downtrend alteration in the time series comparison. Among these genes, 10 genes were found to be associated with the TB treatment outcome. The 10-gene signature reflected well the downtrend kinetics of TB patients during the course of treatment. It had potential in predicting cured TB patients, when employed as covariates in a logistic regression model. In addition, biomarker candidates clearly differentiated TB patients from non-TB patients and people with latent infection.
Conclusions
The 10-gene biosignature had potential in TB diagnosis, treatment monitoring, and outcome prediction.