Ragyoung Yoon, Presenter: Nothing to Disclose

Joon Beom Seo MD, PhD, Abstract Co-Author: Speaker, Siemens AG

Dong Soon Kim MD, Abstract Co-Author: Nothing to Disclose

Jin Woo Song, Abstract Co-Author: Nothing to Disclose

Namkug Kim MS, Abstract Co-Author: Nothing to Disclose

To compare texture-based automated quantification system (CAD) and visual assessment in evaluating extent and interval change of regional disease patterns on initial and follow-up HRCTs.

Fourty one patients with clinically and/or biopsy confirmed UIP (n=32) and NSIP (n=9), whose initial and 1-yr follow-up HRCTs were available, were included. CAD to quantify five regional disease patterns (ground-glass opacity, GGO; reticular opacity, RO; honeycombing, HC; emphysema, EMPH; and consolidation, CONS) at HRCT was developed based on the texture feature analysis, Support Vector Machine classifier, automated lung segmentation and quantification of whole lung area with moving ROI function. Automated quantification of the whole lung was performed on initial and follow-up HRCTs and the interval change in disease extent was calculated by subtracting the results of initial HRCT from follow-up HRCT. The extent of total abnormal lung fraction (TA) and the fibrosis score (FS) were also calculated. For comparison, an experienced thoracic radiologist measured 5%-interval CT visual scores of five regional disease patterns at initial HRCT. Interval changes on follow-up HRCT was measured in 5% scale with side-by-side comparison. The agreement between CAD and visual assessment on initial disease extent and interval change was assessed with interclass correlation coefficient (ICC). Quantitative results on initial HRCT were correlated with DLco.

ICC test revealed acceptable agreements between the quantification results of CAD and radiologist on initial CT (R=0.60 for TA, 0.71 for FS, 0.78 for HC, 0.72 for EMPH, 0.55 for RO, 0.32 for GGO, and 0.43 for CONS). ICCs of measured interval changes in extent of each disease patterns were also acceptable, although the values were lower than those for initial CT assessment (R=0.49 for TA, 0.50 for FS, 0.64 for HC, 0.56 for RO, 0.30 for GGO, and 0.39 for CONS). TA measured on CAD and visual assessment showed significant correlation with DLco (r=-0.61 and -0.58, respectively).

CAD system is comparable to visual assessment in evaluating disease extent and interval changes of disease patterns at HRCT of fibrotic interstitial pneumonia.

CAD system may be used as an objective complement in evaluating initial extent and interval change at HRCTs of fibrotic interstitial pneumonia

Yoon, R, Seo, J, Kim, D, Song, J, Kim, N, Quantitative Assessment of Change in Regional Disease Patterns on Serial HRCT Images of Fibrotic Interstitial Pneumonia with Texture-based Automated Quantification System: Comparison with Visual Assessment.  Radiological Society of North America 2010 Scientific Assembly and Annual Meeting, November 28 - December 3, 2010 ,Chicago IL. http://archive.rsna.org/2010/9010657.html