Japanese
Titleエロソール吸入肺スキャン画像の不均一さと肺機能
Subtitle原著
Authors手島建夫*, 井沢豊春*, 平野富男*, 蝦名昭男*, 白石晃一郎*, 今野淳*
Authors(kana)
Organization*東北大学抗酸菌病研究所内科
Journal核医学
Volume22
Number9
Page1377-1383
Year/Month1985/9
Article原著
Publisher日本核医学会
Abstract「要旨」 これまでアイソトープを用いた核医学画像の不均一さを定量的に表現する指標を定義して検討を加えてきた. 今回, 67例のエロソール吸入肺スキャン画像に対し, 変数選択を加味した重回帰分析および主成分回帰分析を行い, 実用に耐えうる肺機能の予測式を導くことができた. 例えば一秒率 (FEV1.0%) は重相関係数R=0.921, 肺活量 (%VC) はR=0.753で, それぞれ良い予測式を得ることができ, 画像診断に基づく肺機能の予測が可能となった. さらに従来の肺機能検査が全肺的であるのにたいし, 本法の結果の演繹により, 局所肺機能の情報が得られるという点で意味があると考えた.
Practice臨床医学：一般
KeywordsLung function, Aerosol inhalation lung images, Unevenness on images, Regression analysis
English
TitleUnevenness on Aerosol Inhalation Lung Images and Lung Function
SubtitleOriginal Articles
AuthorsTakeo TESHIMA, Toyoharu ISAWA, Tomio HIRANO, Akio EBINA, Koichiro SHIRAISHI, Kiyoshi KONNO
Authors(kana)
OrganizationDepartment of Medicine, The Research Institute for Chest Diseases and Cancer, Tohoku University
JournalThe Japanese Journal of nuclear medicine
Volume22
Number9
Page1377-1383
Year/Month1985/9
ArticleOriginal article
PublisherTHE JAPANESE SOCIETY OF NUCLEAR MEDICINE
Abstract[Summary] The unevenness or inhomogeneity of aerosol deposition patterns on radioaerosol inhalation lung images has been interpreted rather qualitatively in the clinical practice. We have reported our approach to quantitatively analyze the radioactive count distribution on radioaerosol inhalation lung images in relation to the actual lung function data. We have defined multiple indexes to express the shape and the unevenness of the count distribution of the lung images. To reduce as much as possible the number of indexes to be used in the regression functions, the method of selection of variables was introduced to the multiple regression analysis. Because some variables showed greater coefficients of simple correlation, while others did not, multicollinearity of variables had to be taken into consideration. For this reason, we chose a principal components regression analysis. The multiple regression function for each item of pulmonary function data thus established from analysis of 67 subjects appeared usable as a predictor of the actual lung function : for example, %VC (vital capacity) could be estimated by using four indexes out of the multiple ones with a coefficient of multiple correlation (R) of 0.753, and FEV1.0% (forced expiratory volume in one second divided by forced expiratory volume) , by 7 indexes with R=0.921. Pulmonary function data regarding lung volumes and lung mechanics were estimated more accurately with greater R's than those for lung diffusion, but even in the latter the prediction was still statistically significant at p less than 0.01. We believe the multiple regression functions thus obtained are useful for estimating not only the overall but also the regional function of the lungs.
PracticeClinical medicine
KeywordsLung function, Aerosol inhalation lung images, Unevenness on images, Regression analysis

【全文PDF】