IMAGING CRITERIA IN THE DIAGNOSIS OF ATELECTASIS

• DOI: 10.22533/at.ed.1593612317086

• Palavras-chave: Atelectasia pulmonar, clasificación de atelectasia pulmonar, tomografía contrastada con radionúcleo, tomografía computarizada, aprendizaje profundo.

• Keywords: Lung collapse, lung collapse classification, radionuclide contrasted tomography, computed tomography, deep learning.

• Abstract:

  Introduction: Atelectasis, defined as the collapse of the lung volume which is affected in its entirety or only a part of it, its classification is given by the etiology causing obstructive and non-obstructive atelectasis. The diagnosis of atelectasis includes observation and physical examination, which must be confirmed through imaging tests. Currently, the use of the different learning modalities reflects as a result a computer with the necessary data for the early recognition of atelectasis. Objective: Identify the imaging criteria present in atelectasis through a bibliographic review in the state of the art. Methodology: Narrative bibliographic review; analysis and interpretation of articles from various medical journals with a high impact on health. Conclusion: Signs of atelectasis present on plain chest radiography are direct and indirect signs; The increase in lung density and the deviation of the interlobar fissures with similar characteristics in the lower lobes but different in the upper lobes were reiterated. These 2 are considered direct signs. There are secondary signs to the loss of lung volume as compensation (indirect signs) among which the displacement of structures such as the trachea towards the affected side was observed when there is atelectasis of the upper lobe, the approximation of the ribs or elevation of the hemidiaphragm in cases of atelectasis in lower lobes. Another sign is the displacement of the hilum towards the upper part or the hyperinflation of a healthy segment or lobe that compensates for the compromise of the affected lung area. Artificial intelligence allows you to improve image quality, suppress structures and focus on a specific area through automatic segmentation.