Soonho Yoon MD, Presenter: Nothing to Disclose

Jin Mo Goo MD, PhD, Abstract Co-Author: Research Consultant, INFINITT Healthcare Co, Ltd

Eun-Ah Park MD, Abstract Co-Author: Nothing to Disclose

Ju Lip Jung BEng, Abstract Co-Author: Nothing to Disclose

Helen Hong PhD, Abstract Co-Author: Nothing to Disclose

Chang Hyun Lee MD, PhD, Abstract Co-Author: Editorial Consultant, Siemens AG

Xenon ventilation CT using the dual energy technique has shown the potential to classify several phenotypes of chronic obstructive pulmonary disease (COPD). This study was performed to evaluate the efficacy of a newly proposed computer-aided classification (CAC) system for ventilation pattern analysis at xenon ventilation CT.

Thirty-eight patients with COPD (36 men, 2 women; mean age, 65.1 years; age range, 46-78) underwent two-phase xenon ventilation CT creating wash-in (WI) and wash-out (WO) xenon images in 190 lobes. Representative images showing structural abnormalities in each lobe were selected and visually classified into four patterns by comparing the xenon attenuation of structural abnormalities with those of adjacent normal-looking background on WI/WO images in consensus by two experienced radiologists: pattern A, iso- or hyperattenuation on WI and isoattenuation on WO images; pattern B, iso- or hyperatternuation on WI and hyperattenuation on WO images; pattern C, hypoattenuation on WI and WO images; pattern D, hypoattenuation on WI and iso- or hyperattenuation on WO images. Among these datasets, a total of eighty images, twenty images of each pattern, were randomly selected. The proposed CAC system matched WI and WO images via deformable registration using the combined gradient force with active cell approach and automatically assessed the ventilation patterns of selected images at two thresholds (Tlow, Thigh) using histogram analysis of the normal region on WI and WO images.

Overall agreement of the CAC system was 71% (57 of 80; 95% confidence interval, 60%-80%). The agreement was 100% for pattern A, 90% for pattern B, 25% for pattern C and 75% for pattern D. Regarding the low performance in pattern C, the CAC system had misclassified the images into patterns A or D. The agreement for pattern C was improved to 80 % by adjusting Tlow without reducing agreement in other patterns (overall agreement, 85%; 95% confidence interval, 75%-91%).

The newly proposed CAC system has the potential for ventilation pattern analysis at xenon ventilation CT using the dual-energy technique.

The CAC system can be used to classify regional ventilation abnormalities in patients with COPD, and may improve interobserver agreement.

Yoon, S, Goo, J, Park, E, Jung, J, Hong, H, Lee, C, Computer-aided Classification for Visual Ventilation Pattern Analysis at Xenon Ventilation CT Using a Dual-Energy Technique.  Radiological Society of North America 2011 Scientific Assembly and Annual Meeting, November 26 - December 2, 2011 ,Chicago IL. http://archive.rsna.org/2011/11034548.html