Flow starvation during square-flow assisted ventilation detected by supervised deep learning techniques

Abstract

Background: Flow starvation is a type of patient-ventilator asynchrony that occurs when gas delivery does not fully meet the patients' ventilatory demand due to an insufficient airflow and/or a high inspiratory effort, and it is usually identified by visual inspection of airway pressure waveform. Clinical diagnosis is cumbersome and prone to underdiagnosis, being an opportunity for artificial intelligence. Our objective is to develop a supervised artificial intelligence algorithm for identifying airway pressure deformation during square-flow assisted ventilation and patient-triggered breaths.

Methods: Multicenter, observational study. Adult critically ill patients under mechanical ventilation > 24 h on square-flow assisted ventilation were included. As the reference, 5 intensive care experts classified airway pressure deformation severity. Convolutional neural network and recurrent neural network models were trained and evaluated using accuracy, precision, recall and F1 score. In a subgroup of patients with esophageal pressure measurement (ΔP_es), we analyzed the association between the intensity of the inspiratory effort and the airway pressure deformation.

Results: 6428 breaths from 28 patients were analyzed, 42% were classified as having normal-mild, 23% moderate, and 34% severe airway pressure deformation. The accuracy of recurrent neural network algorithm and convolutional neural network were 87.9% [87.6-88.3], and 86.8% [86.6-87.4], respectively. Double triggering appeared in 8.8% of breaths, always in the presence of severe airway pressure deformation. The subgroup analysis demonstrated that 74.4% of breaths classified as severe airway pressure deformation had a ΔP_es > 10 cmH₂O and 37.2% a ΔP_es > 15 cmH₂O.

Conclusions: Recurrent neural network model appears excellent to identify airway pressure deformation due to flow starvation. It could be used as a real-time, 24-h bedside monitoring tool to minimize unrecognized periods of inappropriate patient-ventilator interaction.

Keywords: Airway pressure deformation; Artificial intelligence algorithms; Asynchronies; Flow starvation; Patient–ventilator interaction.

Fig. 1

Representative examples of the airway pressure (P_aw) deformation patterns classification on the pressure–time waveform. Red arrows show the initiation of the patient-triggered breath. The P_aw deformation in the moderate, severe and severe with double triggering tracings is represented by a solid black line on the P_aw tracings. The asterisk shows the second breath added to the first one in the severe breaths with double triggering (breath stacking)

Fig. 2

Flowchart of the breath annotation procedure from ventilator tracings

Fig. 3

Confusion matrix for the recurrent neural network (RNN) and convolutional neural network (CNN) validation processes, respectively. The implemented models provide a strong performance for normal-mild and severe patterns. The reported performance metrics are the average across the 15 repetitions

Fig. 4

Representative examples of airway pressure (P_aw), airflow and esophageal pressure (P_es) tracings during square-flow assisted control ventilation corresponding to normal-mild breath, moderate breath, severe breath and double triggering, respectively. The esophageal swing is represented by solid black lines on the P_es tracings, which increases in relation to the different patterns (the greater the swing, the greater the inspiratory effort)