Purpose: Automated lung volume segmentation is often a preprocessing step in quantitative lung computed tomography (CT) image analysis. The objective of this study is to identify the obstacles in computerized lung volume segmentation and illustrate those explicitly using real examples. Awareness of these difficult cases may be helpful for the development of a robust and consistent lung segmentation algorithm. Methods: We collected a large diverse dataset consisting of 2768 chest CT examinations acquired on 2292 subjects from various sources. These examinations cover a wide range of diseases, including lung cancer, chronic obstructive pulmonary disease, human immunodeficiency virus, pulmonary embolism, pneumonia, asthma, and interstitial lung disease (ILD). The CT acquisition protocols, including dose, scanners, and reconstruction kernels, vary significantly. After the application of a neutral thresholding-based approach to the collected CT examinations in a batch manner, the failed cases were subjectively identified and classified into different subgroups. Results: Totally, 121 failed examinations are identified, corresponding to a failure ratio of 4.4. These failed cases are summarized as 11 different subgroups, which is further classified into 3 broad categories: (1) failure caused by diseases, (2) failure caused by anatomy variability, and (3) failure caused by external factors. The failure percentages in these categories are 62.0, 32.2, and 5.8, respectively. Conclusions: The presence of specific lung diseases (e.g., pulmonary nodules, ILD, and pneumonia) is the primary issue in computerized lung segmentation. The segmentation failures caused by external factors and anatomy variety are relatively low but unavoidable in practice. It is desirable to develop robust schemes to handle these issues in a single pass when a large number of CT examinations need to be analyzed.
Bibliographical noteFunding Information:
In the past few years, we have collected a large number of chest CT examinations from several ongoing research projects at our institute and those available in the National Biomedical Imaging Archive (NBIA) at the National Cancer Institute (NCI). Funded by the National Institutes of Health (NIH), the involved research projects include the Specialized Center of Clinically Oriented Research (SCCOR), the Severe Asthma Research Program (SARP), the Lung Tissue Research Consortium (LTRC), and the Specialized Programs of Research Excellence (SPORE) in Lung Cancer. In addition, we collected around 898 CT examinations from a public database named Lung Image Database Consortium (LIDC) at the NBIA. The LIDC dataset is the result of a multiinstitutional effort funded by NIH. Totally, there have been 2768 examinations acquired from 2292 subjects. These examinations cover a wide range of diseases, including lung cancer, COPD, HIV, pulmonary embolism (PE), pneumonia, asthma, and ILD. Because these examinations were acquired from various sources with different research/clinical backgrounds, the CT acquisition protocols, including dose, scanners, reconstruction kernels, and presence of diseases, vary significantly. For example, the section thickness of these examinations ranges from 0.625 mm to 5.0 mm, and the in-plane pixel size from 0.50 mm to 1.0 mm. Hence, our established database has a relatively large diversity in size, acquisition protocol, and disease. Testing using such a diverse dataset may lead to a relatively objective and thorough performance assessment of a lung segmentation scheme.
This work is supported in part by Grant Nos. R01 HL096613, P50 CA090440 from National Heart, Lung, and Blood Institute, National Institutes of Health (NIH), to the University of Pittsburgh, The Bonnie J. Addario Lung Cancer Foundation, and the Pittsburgh and the SPORE in Lung Cancer Career Development Program. Dr. Xin Meng and Dr. Yongqian Qiang contributed equally to this work.
- computed tomography (CT)
- computer-aided diagnosis
- lung segmentation