Quantitative Analysis of the Human Airway Tree Using High Resolution MDCT Data

Open Access
Author:
Amthor, Michael R
Area of Honors:
Electrical Engineering
Degree:
Bachelor of Science
Document Type:
Thesis
Thesis Supervisors:
  • William Evan Higgins, Thesis Supervisor
  • William Evan Higgins, Honors Advisor
  • Jeffrey Scott Mayer, Faculty Reader
Keywords:
  • image processing
  • lungs
  • medical imaging
  • quantitative
  • analysis
  • statistics
Abstract:
The Multidimensional Image Processing Lab (MIPL) at the Pennsylvania State University has developed a software suite designed to aid in navigational bronchoscopy. The MIPL suite takes a high-resolution computed tomography (CT) patient chest scan and creates a 3D reconstruction of the patient’s chest anatomy. Associated with this process is the generation of a quantitative representation of the airway tree, extracted via image processing techniques developed by the MIPL. Quantitative measurement is possible due to the known dimensions and orientation of chest CT scans. Current methods and the MIPL’s work in planning navigational bronchoscopies have led to an unprecedented availability of high-resolution 3D data. Because of the exhaustive verification of the MIPL suite’s efficacy, we take interest in using the data produced by the MIPL in a quantitative analysis of the human airway tree. The data utilized in this analysis come from 81 patient chest scans accrued over the years by the MIPL. Each case must first be individually processed by the MIPL suite and then by a C++ program we have designed to create a standardized output file. The output files of many cases are then compiled and further processed by a MATLAB program designed to calculate aggregate statistics. Finally, these statistics are organized into plots and tables to give a statistical representation of the airway tree. We observe branch diameters, lengths, and angles by generation, lobar generation and lobe for the entire population, and compare branch measurements between genders and among body mass index classifications. We note a decrease in branch size with an increase in generation, supported by Ewald Weibel’s conclusions. We also observe a larger branch size in males than in females, and an increase in branch diameters with BMI except in obese patients.