24 March 2016 Correlation analysis between pulmonary function test parameters and CT image parameters of emphysema
Author Affiliations +
Abstract
Conventionally, diagnosis and severity classification of Chronic Obstructive Pulmonary Disease (COPD) are usually based on the pulmonary function tests (PFTs). To reduce the need of PFT for the diagnosis of COPD, this paper proposes a correlation model between the lung CT images and the crucial index of the PFT, FEV1/FVC, a severity index of COPD distinguishing a normal subject from a COPD patient. A new lung CT image index, Mirage Index (MI), has been developed to describe the severity of COPD primarily with emphysema disease. Unlike conventional Pixel Index (PI) which takes into account all voxels with HU values less than -950, the proposed approach modeled these voxels by different sizes of bullae balls and defines MI as a weighted sum of the percentages of the bullae balls of different size classes and locations in a lung. For evaluation of the efficacy of the proposed model, 45 emphysema subjects of different severity were involved in this study. In comparison with the conventional index, PI, the correlation between MI and FEV1/FVC is -0.75±0.08, which substantially outperforms the correlation between PI and FEV1/FVC, i.e., -0.63±0.11. Moreover, we have shown that the emphysematous lesion areas constituted by small bullae balls are basically irrelevant to FEV1/FVC. The statistical analysis and special case study results show that MI can offer better assessment in different analyses.
© (2016) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Cheng-Pei Liu, Cheng-Pei Liu, Chia-Chen Li, Chia-Chen Li, Chong-Jen Yu, Chong-Jen Yu, Yeun-Chung Chang, Yeun-Chung Chang, Cheng-Yi Wang, Cheng-Yi Wang, Wen-Kuang Yu, Wen-Kuang Yu, Chung-Ming Chen, Chung-Ming Chen, "Correlation analysis between pulmonary function test parameters and CT image parameters of emphysema", Proc. SPIE 9785, Medical Imaging 2016: Computer-Aided Diagnosis, 978537 (24 March 2016); doi: 10.1117/12.2214936; https://doi.org/10.1117/12.2214936
PROCEEDINGS
10 PAGES


SHARE
Back to Top