Computer model of human lung morphology to complement SPECT analyses.

Abstract

Aerosol therapy protocols could be improved if inhaled pharmacologic drugs were selectively deposited within the human lung. The targeted delivery to specific sites, such as receptors and sensitive airway cells, would enhance the efficacies of airborne pharmaceuticals. The high spatial discrimination of deposition patterns of inhaled particles can be determined via Single Photon Emission Computer Tomography (SPECT). However, major problems continue to compromise SPECT protocols. Our work focuses on two issues: how can the spatial discrimination be improved; and how are the images to be interpreted? We present a methodology, described by a mathematical model and supercomputer code, for systematically slicing through the human lung in a prescribed manner that is conducive to implementation into SPECT analyses. The lung is divided into concentric shells, or annuli, and the airway generation-by-generation composition of the respective shells determined. This identification was accomplished by superimposing the new shell structure with the 3-D branching network presented by Martonen et al. [23]. Perhaps the key aspect of the model-code is that the supercomputer is instructed to determine the 3-D spatial coordinates of each of the airways (over 16 million) of the lung with respect to the clinician-selected contours of shells within the lung.