LOGICAL ANALYSIS OF COMPUTED TOMOGRAPHY DATA TO DIFFERENTIATE ENTITIES OF IDIOPATHIC INTERSTITIAL PNEUMONIAS


M.W. Brauner      N. Brauner     P.L. Hammer      I. Lozina       D. Valeyre


Abstract. The aim of this paper is to analyze computed tomography (CT) data by using the Logical Analysis of Data (LAD) methodology in order to distinguish between three types of idiopathic interstitial pneumonias (IIPs). The paper demonstrates that LAD can distinguish with high accuracy different forms (IPF, NSIP and DIP) of IIPs. It shows also that the patterns developed by LAD techniques provide additional information about outliers, redundant features, the relative significance of attributes, and makes possible the identification of promoters and blockers of various forms of IIPs.