Computation of Scores by Pattern Recognition in Lung Ultrasound Videos

MESCHINO, GUSTAVO JAVIER; PASSONI, LUCÍA ISABEL; AZPILCUETA, RODRIGO; ACOSTA, CECILIA; ECHEVERRÍA, NOELIA INÉS; TUSMAN, GERARDO

San Juan

Congreso; XXIII Congreso Argentino de Bioingeniería; 2022

Sociedad Argentina de Bioingeniería

Lung ultrasound (LUS) is a dynamic, non-invasive diagnostic tool that allows atelectasis to be observed after induction of general anesthesia. Recruitment maneuvers (RM) consist of transient increases in mean airway pressure to recruit collapsed alveolar units. This paper presents a method to recognize pulmonary states in lung ultrasound, in order to computationally estimate a score that is used by specialists. This allows comparison of lung states pre and post MR. Four pulmonary states were detected, and scores were estimated with a difference of 14% in 50 cases, with respect to the values proposed by the specialists. The discrepancy with the experts expressed as the mean of score differences was 0.22.