Machine Learning and Prediction of All-Cause Mortality in COPD

• Published in: Chest Vol. 158; no. 3; pp. 952 - 964
• Main Authors: Moll, Matthew, Qiao, Dandi, Regan, Elizabeth A., Hunninghake, Gary M., Make, Barry J., Tal-Singer, Ruth, McGeachie, Michael.J., Castaldi, Peter J., San Jose Estepar, Raul, Washko, George R., Wells, James M., LaFon, David, Strand, Matthew, Bowler, Russell P., Han, MeiLan.K., Vestbo, Jorgen, Celli, Bartolome, Calverley, Peter, Crapo, James, Silverman, Edwin K., Hobbs, Brian D., Cho, Michael H.
• Format: Journal Article
• Language: English
• Published: United States Elsevier Inc 01.09.2020; American College of Chest Physicians
• Subjects: Cause of Death; COPD; Female; Humans; Machine Learning; Male; Middle Aged; mortality; Original Research; prediction; Predictive Value of Tests; Pulmonary Disease, Chronic Obstructive - mortality; random survival forest; Respiratory Function Tests; LAA < −950 HU; RSF; MLMP; COPDGene; machine learning; ADO; random survival forest; BODE; ECLIPSE; mortality; prediction; 6MWD; PA:A; Pi10; VIMP; COPD
• ISSN: 0012-3692; 1931-3543; 1931-3543
• DOI: 10.1016/j.chest.2020.02.079

COPD is a leading cause of mortality. We hypothesized that applying machine learning to clinical and quantitative CT imaging features would improve mortality prediction in COPD. We selected 30 clinical, spirometric, and imaging features as inputs for a random survival forest. We used top features in a Cox regression to create a machine learning mortality prediction (MLMP) in COPD model and also assessed the performance of other statistical and machine learning models. We trained the models in subjects with moderate to severe COPD from a subset of subjects in Genetic Epidemiology of COPD (COPDGene) and tested prediction performance in the remainder of individuals with moderate to severe COPD in COPDGene and Evaluation of COPD Longitudinally to Identify Predictive Surrogate Endpoints (ECLIPSE). We compared our model with the BMI, airflow obstruction, dyspnea, exercise capacity (BODE) index; BODE modifications; and the age, dyspnea, and airflow obstruction index. We included 2,632 participants from COPDGene and 1,268 participants from ECLIPSE. The top predictors of mortality were 6-min walk distance, FEV1 % predicted, and age. The top imaging predictor was pulmonary artery-to-aorta ratio. The MLMP-COPD model resulted in a C index ≥ 0.7 in both COPDGene and ECLIPSE (6.4- and 7.2-year median follow-ups, respectively), significantly better than all tested mortality indexes (P < .05). The MLMP-COPD model had fewer predictors but similar performance to that of other models. The group with the highest BODE scores (7-10) had 56% mortality, whereas the highest mortality group defined by the MLMP-COPD model had 62% mortality (P = .046). An MLMP-COPD model outperformed four existing models for predicting all-cause mortality across two COPD cohorts. Performance of machine learning was similar to that of traditional statistical methods. The model is available online at: https://cdnm.shinyapps.io/cgmortalityapp/.