Table 6: Optimal estimated parameters, best features and corresponding area under the ROC curve (AUC), and -score for each classifier (trained on micro-CT density-based descriptors, RNA cytokine expression data, and single compartment model parameters from pulmonary function tests (PFTs)) are shown.

ClassifierOptimal parameterFeaturesAUC -score

Micro-CT MLI, VBT, RMBR0.950.92
Cytokine KC, IL6, IP100.880.71
PFTs R 0.710.66