Cluster analysis of emphysema for predicting pulmonary complications after thoracoscopic lobectomy

Abstract OBJECTIVES Despite significant advances in surgical techniques, including thoracoscopic approaches and perioperative care, the morbidity rate remains high after lung resection. This study focused on a low attenuation cluster analysis, which represented the size distribution of pulmonary emphysema and assessed its utility for predicting postoperative pulmonary complications after thoracoscopic lobectomy. METHODS From April 2013 to September 2018, lung cancer patients who received spirometry and computed tomography (CT) before surgery and underwent thoracoscopic lobectomy were included. The cumulative size distribution of the low attenuation area (LAA, defined as ≤−950 Hounsfield unit on CT) clusters followed a power-law characterized by an exponent D-value, a measure of the complexity of the alveolar structure. D-value and LAA% (LAA/total lung volume) were calculated using preoperative 3-dimensional CT software. The relationship between pulmonary complications and patient characteristics, including D-value and LAA%, was investigated. RESULTS Among 471 patients, there were 61 respiratory complication cases (12.9%). Receiver operation characteristic curve analysis revealed that the best predictive cut-off value of D-value and LAA% for pulmonary complications was 2.27 and 16.5, respectively, with an area under the curve of 0.72 and 0.58, respectively. D-value was significantly correlated with % forced expiratory volume in 1 s. Per univariate analysis, gender, smoking history, forced expiratory volume in 1 s/forced vital capacity, LAA% and D-value were risk factors for predicting postoperative pulmonary complications. In the multivariate analysis, D-value remained a significant predictive factor. CONCLUSION Preoperative assessment of emphysema cluster analysis may represent the vulnerability of the operated lung and could be the novel predictor for pulmonary complications after thoracoscopic lobectomy.