Introduction:
Pseudo-Pulseless Electrical Activity (p-PEA) is a lifeless form of profound cardiac shock characterized by measurable cardiac mechanical activity without clinically detectable pulses. p-PEA may constitute up to 40% of reported cases of cardiac arrest and its management may be different from ventricular fibrillation or even true-PEA. Currently, diagnosis of p-PEA requires either echocardiography or intravascular catheterization, neither of which are ideal in the prehospital setting. Electrical impedance tomography (EIT) uses skin surface electrodes to generate images based on cross sectional resistance. We investigated the ability of EIT and machine learning (ML) to detect p-PEA.
Methods:
Female swine (N = 14) under intravenous anesthesia were instrumented with aortic and central venous micromanometer catheters. p-PEA was induced by ventilation with 6% O
2
in 94% N
2
, defined as a systolic aortic pressure less than 40 mmHg. Continuous EIT renderings were obtained from circumferential thoracic and abdominal electrode arrays. A deep learning model was utilized to detect p-PEA using EIT sequences. Twelve pigs were randomly selected as training data and 2 pigs as a test set. EIT images were saved as 30 second clips, resulting in 3033 clips generated. To increase generalizability, random epochs ranging from 30 - 100% of the total clip length were generated, resulting in a model capable of detecting this disease state with limited video fragments.
Results:
This technique yielded a receiver operator characteristic curve - area under the curve (ROC-AUC) of 0.91 for detection of p-PEA in the testing dataset (Figure 1), with 84% accuracy, 88% sensitivity, and 84% specificity.
Conclusions:
EIT combined with machine learning may be able to reliably delineate p-PEA in a hypoxic porcine model. This approach may be promising for non-invasive operator-independent p-PEA detection.
Imaging fast electrical activity in the brain during ictal epileptiform discharges with electrical impedance tomography
