Computer-aided diagnosis of emphysema in COPD patients: Neural-network-based analysis of lung shape in digital chest radiographs

Abstract

Several abnormalities of the shape of lung fields (depression and flattening of the diaphragmatic contours, increased retrosternal space) are indicative of emphysema and can be accurately imaged by digital chest radiography. In this work, we aimed at developing computational descriptors of the shape of the lung silhouette able to capture the alterations associated with emphysema. We analyzed two-sided digital chest radiographs from a sample of 160 patients with chronic obstructive pulmonary disease (COPD), 60 of which were affected by emphysema, and from 160 subjects with normal lung function. Two different description schemes were considered: a first one based on lung-silhouette curvature features, and a second one based on a minimal-polyline approximation of the lung shape. Both descriptors were employed to recognize alterations of the lung shape using classifiers based on multilayer neural networks of the feed-forward type.Results indicate that pulmonary emphysema can be reliably diagnosed or excluded by using digital chest radiographs and a proper computational aid. Two-sided chest radiographs provide more accurate discrimination than single-view analysis. The minimal-polyline approximation provided significantly better results than those obtained from curvature-based features. Emphysema was detected, in the entire dataset, with an accuracy of about 90% (sensitivity 88%, specificity 90%) by using the minimal-polyline approximation.

A computer-aided diagnosis approach for emphysema recognition in chest radiography

“…It is worth noting that the availability of lung boundary in postero-anterior and lateral views may be usefully exploited to compute additional parameters such as the radiographic total lung capacity [31]. As to shape description and classification, the described system exhibits improved recognition capabilities in comparison with previous works [8,10]. In particular, the adopted shape-descriptor was able to successfully cope with alterations of lung silhouette not related to the presence of emphysema, such as skeletal distortions.…”

“…Lung shape alterations due to emphysema can be characterized by a limited set of numerical features [8]. The adopted shape descriptor is obtained by a polygonal approximation of the lung boundary.…”

“…In addition, computational approaches to the use of digital chest radiography for emphysema detection were explored [8]. Quantitative shape descriptors of hand traced pulmonary silhouettes proved quite accurate in such tasks [9].…”

“…On the other hand, lateral projection provides important diagnostic information for recognizing emphysema: expansion of the retrosternal space is appreciated only in the lateral view which also accurately describes diaphragm flattening [8]. It must be pointed out that, due to the thickness of the imaged object and the overlapping of the lung fields, lateral views usually exhibit inferior image quality, both in terms of contrast and in terms of signal-to-noise ratio.…”

Computer-aided recognition of emphysema on digital chest radiography

“…Based on a preliminary study with different net-work architectures [10], we chose a neural network featuring 18 input units, two hidden layers with 10 and 4 units respectively, and a single output unit whose activation yields the probability that the input pattern belongs to the "emphysema" category.…”