Introduction: Non-tuberculous mycobacteria pulmonary disease (NTM-pd) commonly complicates bronchiectasis. However, clinical and radiological features of NTM-pd and bronchiectasis are very similar. We aimed to develop a radiologic prediction tool for bronchiectasis to identify NTM-pd.
Methods: We reviewed clinical, laboratory and radiological data in patients with bronchiectasis. Radiologic features on CT scans and the individual components of the Bhalla scoring system were compared between people with and without NTM-pd. Logistic regression and receiver-operating curve (ROC) analysis were performed to predict NTM-pd.
Results: People with NTM-pd had more pulmonary segments with bronchiectasis (13 ± 5 vs. 11 ± 5, p = 0.03), presence of mucus plugging (47% vs. 19%, p < 0.0001) and tree in bud infiltrates (53% vs. 28%, p = 0.004). The total modified- Bhalla score was worse among people with NTM-pd (median [IQR] 11[9,13] vs. 9[8,12], p = 0.03). Logistic regression identified the number of pulmonary segments involved, presence of bullae, consolidations, and a total score of 10 or more to be independently associated with presence of NTM-pd. ROC analysis with radiographic variables only identified an AUC of 0.706 (95% CI 0.644-0.762). When people with chronic Pseudomonas infection were excluded from the ROC analysis, prediction for NTM was improved: AUC = 0.87 (95% CI 0.796-0.945).
Discussion and conclusions: Radiological features together with advanced age and female gender may predict NTM-pd among people with bronchiectasis. Infection with Pseudomonas aeruginosa may resemble NTM radiographically, and this prediction rule may better differentiate people with and without NTM-pd when Pseudomonas infection is not present.
Keywords: Bronchiectasis; Non-tuberculous mycobacteria; Radiology.