RSNA 2004 

Abstract Archives of the RSNA, 2004


SSG05-07

Usual Interstitial Pneumonia and Other Chronic Idiopathic Interstitial Pneumonias: Univariate and Multivariate Analysis of Appearances on High-Resolution CT in 92 Patients

Scientific Papers

Presented on November 30, 2004
Presented as part of SSG05: Chest (High-Resolution CT)

Participants

Hiromitsu Sumikawa MD, Presenter: Nothing to Disclose
Takeshi Johkoh MD, Abstract Co-Author: Nothing to Disclose
Noriyuki Tomiyama MD, PhD, Abstract Co-Author: Nothing to Disclose
Seiki Hamada MD, PhD, Abstract Co-Author: Nothing to Disclose
Hironobu Nakamura MD, PhD, Abstract Co-Author: Nothing to Disclose
Atsuo Inoue, Abstract Co-Author: Nothing to Disclose
Kazuya Ichikado MD, Abstract Co-Author: Nothing to Disclose
Osamu Honda, Abstract Co-Author: Nothing to Disclose
et al, Abstract Co-Author: Nothing to Disclose

PURPOSE

The aim of this study is to compare the findings on high-resolution CT images between usual interstitial pneumonia (UIP) and other chronic idiopathic interstitial pneumonias (IIPs).

METHOD AND MATERIALS

The study included 92 patients with proved chronic IIPs both clinically and histologically. The cases were categorized histologically into the following six groups: UIP (n=20), cellular nonspecific interstitial pneumonia (cellular NSIP) (n=16), fibrotic NSIP (n=16), respiratory bronchiolitis-associated interstitial lung disease (RBILD) (n=11), desquamative interstitial pneumonia (DIP) (n=15) and lymphoid interstitial pneumonia (LIP) (n=14). Two independent observers evaluated the extent and distribution of various CT findings and assigned a likely diagnosis. CT findings of each IIP were compared with those of UIP using both univariate and multivariate technique. Agreement of observers for the extent and distribution of abnormalities were evaluated with Spearman’s rank correlation coefficient and kappa statistics respectively.

RESULTS

The rate of correct diagnosis on HRCT was 80.4% (148/184) in all observations. In univariate analysis, the areas with faint centrilobular nodules in UIP was significantly smaller than those in RBILD (Mann-Whitney U test: p<0.001), DIP (p < 0.001) and LIP (p < 0.005). The areas with ground-glass attenuation in DIP were significantly larger than those in UIP (Mann-Whitney U test: p<0.001). Multivariate logistic regression analysis indicated that the decrease of the extent of honeycombing (odds ratio: 0.34; p < 0.001) and the increase of the dimension of the most proximal bronchus with bronchiectasis (odds ratio: 2.08; p < 0.01) in cellular NSIP and the decrease of the extent of honeycombing (odds ratio: 0.57; p < 0.001) in fibrotic NSIP were cardinal features in comparison with UIP. There was moderate to good agreement between the observers for the extent (r = 0.34-0.85, p < 0.01) and distribution (κ = 048-0.74) of various abnormalities.

CONCLUSIONS

UIP has a characteristic appearance that allows differentiation from other types of chronic IIPs at HRCT. The most useful finding differentiating between UIP and NSIP was the extent of honeycombing.

Cite This Abstract

Sumikawa, H, Johkoh, T, Tomiyama, N, Hamada, S, Nakamura, H, Inoue, A, Ichikado, K, Honda, O, et al, , Usual Interstitial Pneumonia and Other Chronic Idiopathic Interstitial Pneumonias: Univariate and Multivariate Analysis of Appearances on High-Resolution CT in 92 Patients.  Radiological Society of North America 2004 Scientific Assembly and Annual Meeting, November 28 - December 3, 2004 ,Chicago IL. http://archive.rsna.org/2004/4413435.html