AUTHOR=Song Qisheng , Guo Xiaohong , Zhang Liling , Yang Lianjun , Lu Xiwei 

TITLE=New Approaches in the Classification and Prognosis of Sign Clusters on Pulmonary CT Images in Patients With Multidrug-Resistant Tuberculosis

JOURNAL=Frontiers in Microbiology

VOLUME=Volume 12 - 2021

YEAR=2021

URL=https://www.frontiersin.org/journals/microbiology/articles/10.3389/fmicb.2021.714617

DOI=10.3389/fmicb.2021.714617

ISSN=1664-302X

ABSTRACT=Abstract
Background: To date, radiographic sign clusters of multidrug-resistant pulmonary tuberculosis (MDR-TB) patients have not been reported. We conducted a study to investigate the classification and prognosis of sign clusters in pulmonary CT images from patients with MDR-TB for the first time by principal component analysis (PCA).
Methods: The clinical data and pulmonary CT findings of 108 patients with MDR-TB in the Liupanshui Third Hospital were collected(from January 2018 to December 2020). PCA was used to analyze the sign clusters on pulmonary CT in them, and receiver operating characteristic (ROC) analysis was used to analyze the predictive value of treatment outcome of MDR-TB patients. 
Results: Six cluster signs of MDR-TB were determined by PCA: nodules, infiltration, consolidation, cavities, destroyed lung and nonactive lesions. At 9 months after treatment, the area under the ROC curve (AUC) of MDR-TB patients with a cavity sign cluster was 0.818（95% CI：0.733 - 0.886）, the positive predictive value(PPV) and negative predictive value(NPV) of the treatment outcome were 79.6%（95% CI：65.7 - 89.8%）and 72.9%（95% CI：59.7 - 83.6%）, respectively. 
Conclusion: Classification of sign groups on pulmonary CT can simplify radiographic studies in patients with MDR-TB. Scores especially which of cavity sign cluster in CT images have definitive value in predicting the treatment outcome of MDR-TB patients.