The AI-Rad solely analyzes the posterior-anterior (p.a.) view of chest X-ray images and creates secondary capture DICOM objects reporting on the results of the analysis. Each finding is marked on a copy of the analyzed X-ray image and listed in a table. Additionally, the AI-Rad provides a “confidence score” (CS) on a scale of 1 (low) to 10 (high) for each finding, which expresses the algorithm´s certainty for the presence of that particular finding. The manufacturer has preset the AI-Rad only to report findings with a CS ≥ 6, whilst findings with a CS ≤ 5 are not displayed.
The AI-Rad (version VA23A) is designed to detect five specific radiographic findings: Pulmonary lesions, consolidation, atelectasis, pneumothorax and pleural effusion. Pulmonary lesions, as defined by the AI-Rad, include lung nodules (rounded or oval opacities < 3 cm in diameter) and lung masses (pulmonary, pleural or mediastinal lesions > 3 cm in diameter). To detect pneumothoraces, the AI-Rad screens for radiographic signs suggestive of air in the pleural space. Likewise, the AI-Rad screens for radiographic signs suggestive of fluid in the pleural space for the detection of pleural effusions. Atelectasis are defined as increased opacities accompanied by volume loss, which, in turn, can be an abnormal displacement of fissures, bronchi, vessels, the diaphragm, or the mediastinum. The AI-Rad defines consolidations as increased parenchymal attenuation. This definition includes homogeneous increases of parenchymal attenuation (consolidation) that obscures pulmonary vessels and bronchi as well as hazy increases of parenchymal attenuation (ground glass opacity) that do not obscure pulmonary vessels and bronchi.
Free full text: Click here