Characteristic chest CT findings for progressive cavities in Mycobacterium avium complex pulmonary disease: a retrospective cohort study

A total of 485 outpatients with MAC-PD, diagnosed on the basis of the diagnostic criteria for NTM-PD advocated by the America Thoracic Society/Infectious Disease Society of America in 2007 [14], were identified between December 2006 and June 2016 at National Hospital Organization, Osaka Toneyama Medical Center. Among these, we extracted the data of 139 patients who had cavities, were observed for over 3 years, and could be evaluated with CT at two or more points. A cavity was defined as a radiographic opacity with an internal area of lucency. Beaded airspace enlargements that were apparently contiguous with the airways were excluded as bronchiectasis. We also excluded patients who underwent lung resection, or who were associated with lung cancer, interstitial pneumonia, or pulmonary aspergillosis because these diseases would be considered the primary condition rather than MAC-PD in these cases. Thus, 97 patients were finally enrolled into this study.

The study was approved by the National Hospital Organization, Osaka Toneyama Medical Center Review Board (approval number TNH-2019005), Osaka, Japan. The approval allowed retrospective data collection and reporting of anonymous results without acquisition of informed consent from eligible study subjects.

Initial CT was defined as the CT examination on which a cavity was first identified, and the corresponding cavity was defined as the initial cavity. The size of this cavity at the time of the final observation was compared to its initial size, and the patients were classified into the progressive or non-progressive cavity groups accordingly. If a patient had multiple cavities, the cavity exhibiting the largest internal diameter at the time of cavity confirmation was evaluated. A progressive cavity was defined by a more than two-fold increase in internal diameter in comparison with the initial cavity size or the emergence of a new cavity around the initial cavity (Fig. 1a, b). The non-progressive cavity group included all cases other than those in the progressive cavity group (Fig. 1c, d).

Baseline clinical characteristics were obtained from medical chart review, including patient observation period (the interval between the initial visit and the final observation), cavity observation period (the interval between the initial and the final CT), sex, age at initial cavity confirmation, BMI, underlying lung disease, comorbidities, steroid and/or immunosuppressive agent use, MAC species, sputum-smear positivity at initial cavity confirmation, treatment, levels of glycopeptidolipid-core antibody, the rate of all-cause death, and respiratory failure as prognosis. A multidrug regimen was defined as the use of two or more drugs containing macrolide antibiotics for at least 1 year. Respiratory failure was defined as initiation of home oxygen therapy and/or noninvasive pressure ventilation therapy.

In the initial cavity assessment, the lung area where the cavity was present, the largest internal diameter, and the shortest distance from the pleura to the outer cavity wall were evaluated. The presence or absence of multiple cavities was also evaluated. We defined the cavity that was completely tangential to the pleura as the cavity close to the chest wall. Pulmonary areas were divided into three regions: upper lobe, middle lobe or lingula, and lower lobe. Among CT findings around the initial cavity, bronchiectasis, small nodules (≤ 5 mm), nodules (> 5 mm and < 30 mm), consolidation, atelectasis, pleural thickening, pleural indentation, pulmonary emphysema, or presence of bulla were evaluated. Chest CT findings were assessed by two pulmonologists blinded to the clinical data. Discrepancies were solved through a consensus review.

Continuous data were summarized as quartiles and compared using the Wilcoxon rank-sum test for two-group comparisons. Categorical data were summarized as numbers (percentages) and compared using Fisher’s exact test. The Kaplan–Meier method was used to estimate survival curves for all-cause mortality and respiratory failure. The log-rank test was used to compare survival curves between the progressive and non-progressive cavity groups. Univariable and multivariable logistic regression analyses were used to investigate the factors associated with cavity progression. As for the chest CT findings, presence of the largest cavity in the middle lobe or the lingula, the shortest distance from the pleura to the outer cavity wall, atelectasis, and pleural thickening were excluded from the multivariable logistic regression analysis because these were strongly correlated with consolidation around the initial cavity. Receiver Operating Characteristic (ROC) analysis was used to determine the optimal cut-off value of the maximum inner diameter of cavities for the prediction of progressive cavity (Additional file 1: Figure S1). All reported p values were two-sided, and p values <0.05 were considered statistically significant. SAS software version 9.3 (SAS Institute, Inc., Cary, NC) was used for statistical analysis.

Of the 97 patients, 52 were in the progressive cavity group while 45 were classified to the non-progressive cavity group. Baseline patient characteristics in the two groups are summarized in Table 1. The median observation periods were 8.0 (range: 3.0–12.9) years in the progressive cavity group and 8.8 (3.8–12.0) years in the non-progressive cavity group (p = 0.103). The median cavity observation periods were 4.6 years (range: 1.1–11.9 years) in the progressive cavity group and 5.5 years (3.0–10.8 years) in the non-progressive cavity group (p = 0.025). None of the patients were known to be infected with the human immunodeficiency virus. The age at initial cavity confirmation was significantly higher in the progressive cavity group (p < 0.001), while BMI was significantly lower in the progressive cavity group (p = 0.043). In assessments of underlying lung diseases, both groups showed a high rate of bronchiectasis. However, in assessments of comorbidities, the proportion of patients with diabetes was significantly higher in the progressive cavity group (p = 0.005). There was no significant intergroup difference in the rates of smear positivity at initial cavity confirmation (p = 0.078). There was no significant intergroup difference in the rates of multidrug chemotherapy (p = 0.809). Glycopeptidolipid-core antibody, that is useful for diagnosing MAC-PD [15], was measured in 25 of 52 cases in the progressive group and 26 of 45 cases in the non-progressive group within 6 months of the initial CT. No significant difference in the antibody levels was observed between the two groups (5.5 U/ml [range: 2.3–11.1 U/ml] in the progressive group, 4.0 U/ml [range: 0.7–9.0 U/ml] in the non-progressive group, p = 0.271).

Table 2 shows the initial chest CT findings for cavities. The distribution of cavities was significantly different between the two groups (p = 0.024). The maximum inner diameter of the cavity was significantly longer in the progressive cavity group (p < 0.001). The shortest distance from the pleura to the cavity outer wall was significantly shorter in the progressive cavity group (p < 0.001). The rate of multiple cavities was significantly higher in the progressive cavity group (p = 0.023). Table 3 shows chest CT findings observed around the initial cavity. Consolidation (p < 0.001), atelectasis (p = 0.011), and pleural thickening (p < 0.001) around the initial cavity were more frequently observed in the progressive cavity group.

Table 4 shows the results of univariable and multivariable logistic regression analysis for the factors related to progressive cavities. Univariable logistic regression analysis revealed that age (p = 0.001), diabetes mellitus (p = 0.021), presence of the largest cavity in the middle lobe or lingula (p = 0.015), maximum internal diameter of the cavity (p = 0.002), the shortest distance from the pleura to the outer cavity wall (p = 0.010), multiple cavities (p = 0.015), and the findings around the cavity, namely, consolidation (p < 0.001), atelectasis (p < 0.001), and pleural thickening (p < 0.001), were significantly associated with progressive cavities. Multivariable logistic regression analysis revealed that consolidation around the cavity (odds ratio [OR]: 16.15, 95% confidence interval [CI]: 4.05–64.46, p < 0.001), age (OR: 1.12, 95% CI: 1.03–1.20, p = 0.005), and maximum internal diameter of the cavity (OR: 1.11, 95% CI: 1.02–1.21, p = 0.012) were factors that showed significant relationships with progressive cavities. Figure 2 shows the ratio of progressive cavities by the presence or absence of a cavity maximum inner diameter of ≥10 mm and consolidation around the initial cavity. Progression rates were 21.2% in patients with maximum inner diameter < 10 mm and without consolidation around the initial cavity, 46.4% in those with maximum inner diameter ≥ 10 mm without consolidation, 70% in patients with maximum inner diameter < 10 mm with consolidation, and 96.2%, in patients with maximum inner diameter ≥ 10 mm with consolidation.

Figure 3 shows the Kaplan–Meier survival curves for (A) all-cause mortality and (B) respiratory failure. The 10-year mortality rate was 46.7% in the progressive cavity group and 9.8% in the non-progressive cavity group (log-rank test: p < 0.001). The 10-year rate for respiratory failure was 28.1% in the progressive cavity group and 0% in the non-progressive cavity group (log-rank test: p < 0.001).