Latent class analysis to define radiological subgroups in pulmonary nontuberculous mycobacterial disease

• Published in: BMC pulmonary medicine Vol. 18; no. 1; pp. 145 - 8
• Main Authors: Cowman, Steven A., Jacob, Joseph, Obaidee, Sayed, Andres Floto, R., Wilson, Robert, Haworth, Charles S., Loebinger, Michael R.
• Format: Journal Article
• Language: English
• Published: London BioMed Central 31.08.2018; BioMed Central Ltd; BMC
• Subjects: Aged; Bronchiectasis; Care and treatment; Cavitation; Computed tomography; Critical Care Medicine; Cystic fibrosis; Diagnosis; Female; High resolution computed tomography; Humans; Infections; Infectious; Intensive; Intensive care; Internal Medicine; Latent Class Analysis; Lung - diagnostic imaging; Lung - pathology; Lung diseases; Lung Diseases - diagnostic imaging; Lung Diseases - microbiology; Male; Medical imaging; Medical prognosis; Medicine; Medicine & Public Health; Middle Aged; Mortality; Mycobacterium infections; Mycobacterium Infections, Nontuberculous - diagnostic imaging; Nontuberculous Mycobacteria; Pneumology/Respiratory System; Pulmonology; Radiology; Rare and Idiopathic Pulmonary Diseases; Research Article; Risk factors; Statistical analysis; Studies; Survival Analysis; Tomography, X-Ray Computed - methods; Variance analysis; High resolution computed tomography; Latent class analysis; Nontuberculous mycobacteria
• ISSN: 1471-2466; 1471-2466
• DOI: 10.1186/s12890-018-0675-8

Background Nontuberculous mycobacterial (NTM) pulmonary disease has conventionally been classified on the basis of radiology into fibrocavitary and nodular-bronchiectatic disease. Whilst being of great clinical utility, this may not capture the full spectrum of radiological appearances present. The aim of this study was to use latent class analysis (LCA) as an unbiased method of grouping subjects with NTM-pulmonary disease based on their CT features and to compare the clinical characteristics of these groups. Methods Individuals with NTM-pulmonary disease were recruited and a contemporaneous CT scan obtained. This was scored using an NTM-specific scoring system. LCA was used to identify groups with common radiological characteristics. The analysis was then repeated in an independent cohort. Results Three classes were identified in the initial cohort of 85 subjects. Group 1 was characterised by severe bronchiectasis, cavitation and aspergillomas, Group 2 by relatively minor radiological changes, and Group 3 by predominantly bronchiectasis only. These findings were reproduced in an independent cohort of 62 subjects. Subjects in Group 1 had a lower BMI and serum albumin, higher serum CRP, and a higher mortality. Conclusions These findings suggest that NTM-pulmonary may be divided into three radiological subgroups, and that important clinical and survival differences exist between these groups.