Usefullnes of chest computed tomography as a tool to predict COPD exacerbations

Abstract

Introduction: Phenotyping of COPD patients by chest computed tomography (CT) has been shown to correlate with pulmonary function parameters. Whether if it is of interest in prediction of COPD exacerbations is under investigation.

Objectives: To evaluate the association of COPD CT phenotyping with patients’ exacerbations profiles.

Methods: Retrospective analysis of COPD patients’ exacerbation history during a 1-year period. Patients were classified into normal, air dominant, emphysema dominant and mixed type based on CT scans using two cutoffs: percentage of low attenuation area (%LAA) less than a threshold of -950 Hounsfield units of 15% and the upper quartile value of the percentage of bronchial wall area (WA%/BSA). Exacerbation and its severity were defined accordingly to the GOLD 2021 report.

Results: A total of 80 patients were evaluated. Median age was 66.5 (IQR 16) years. Mean number of exacerbations was 1 (±1.2) for the normal phenotype; 0.74 (±1.5) for air-dominant; 1.9 (±2.6) for emphysema-dominant; and 1 (±1.4) for mixed-type. There were no statistical differences regarding having an exacerbation, the number of exacerbations, nor the severity of the exacerbation (respectively, p=0.937; p=0.315; and severe: p=0.579; mild: p=0.388; and light: p=0.620). COPD CT phenotyping did not statistically predict the occurrence of exacerbations of any severity (X²=0.417, p=0.937, R²=0.007), neither the number of exacerbations (F[1,78]=0.255,p=0.615, R²=0.003). Similar results were found when analyzing each level of severity individually.

Conclusions: In our sample, COPD CT phenotyping did not statistically predict COPD exacerbations. Our results lead us to recommend against its use for this purpose.