A pneumothorax is an abnormal collection of air in the pleural space between the lung and the chest wall. This air pushes on the outside of the lung, causing it to collapse. A pneumothorax can be caused by a blunt or penetrating chest injury, certain medical procedures, or from underlying lung disease, typically emphysema. Depending on its size, pneumothorax can result in complete lung collapse or collapse of only a portion of the lung. Occasionally it may occur for no obvious reason (idiopathic). Pneumothorax can potentially be life-threatening and is considered to represent a critical finding in Emergency Radiology (ER), requiring immediate reporting to the treating physician to ensure immediate medical attention. Hence, Pneumothorax detection is of critical importance in clinical care. Pneumothorax may be detected with the help of image processing and deep learning algorithms. If utilized effectively, deep learning techniques can assist radiologists with quick detection, segmentation, classification and quantification of pneumothorax. In this paper, we evaluate two deep learning architectures for the detection and segmentation of pneumothorax regions on chest radiograph images. The AI system detects regions of pneumothorax in a chest radiograph and may assist the radiologist to review on priority the cases that contain a pneumothorax and thus facilitate early management of patients.
Pneumothorax Detection and Classification on Chest Radiographs using Artificial Intelligence
