Analysis during chest compressions in out-of-hospital cardiac arrest patients, a cross/sectional study: The DEFI 2022 study.

AIMS: During out-of-hospital cardiac arrest (OHCA), an automatic external defibrillator (AED) analyzes the cardiac rhythm every two minutes; however, 80% of refibrillations occur within the first minute post-shock. We have implemented an algorithm for Analyzing cardiac rhythm While performing chest Compression (AWC). When AWC detects a shockable rhythm, it shortens the time between analyses to one minute. We investigated the effect of AWC on cardiopulmonary resuscitation quality. METHOD: In this cross-sectional study, we compared patients treated in 2022 with AWC, to a historical cohort from 2017. Inclusion criteria were OHCA patients with a shockable rhythm at the first analysis. Primary endpoint was the chest compression fraction (CCF). Secondary endpoints were cardiac rhythm evolution and survival, including survival analysis of non-prespecified subgroups. RESULTS: In 2017 and 2022, 355 and 377 OHCAs met the inclusion criteria, from which we analyzed the 285 first consecutive cases in each cohort. CCF increased in 2022 compared to 2017 (77% [72-80] vs 72% [67-76]; P < 0.001) and VF recurrences were shocked more promptly (53 s [32-69] vs 117 s [90-132]). Survival did not differ between 2017 and 2022 (adjusted hazard-ratio 0.96 [95% CI, 0.78-1.18]), but was higher in 2022 within the sub-group of OHCAs that occurred in a public place and within a short time from call to AED switch-on (adjusted hazard ratio 0.85[0.76-0.96]). CONCLUSIONS: OHCA patients treated with AWC had higher CCF, shorter time spent in ventricular fibrillation, but no survival difference, except for OHCA that occurred in public places with short intervention time.

Deep Learning Strategy for Sliding ECG Analysis during Cardiopulmonary Resuscitation: Influence of the Hands-Off Time on Accuracy.

This study aims to present a novel deep learning algorithm for a sliding shock advisory decision during cardiopulmonary resuscitation (CPR) and its performance evaluation as a function of the cumulative hands-off time. We retrospectively used 13,570 CPR episodes from out-of-hospital cardiac arrest (OHCA) interventions reviewed in a period of interest from 30 s before to 10 s after regular analysis of automated external defibrillators (AEDs). Three convolutional neural networks (CNNs) with raw ECG input (duration of 5, 10, and 15 s) were applied for the shock advisory decision during CPR in 26 sequential analyses shifted by 1 s. The start and stop of chest compressions (CC) can occur at arbitrary times in sequential slides; therefore, the sliding hands-off time (sHOT) quantifies the cumulative CC-free portion of the analyzed ECG. An independent test with CPR episodes in 393 ventricular fibrillations (VF), 177 normal sinus rhythms (NSR), 1848 other non-shockable rhythms (ONR), and 3979 asystoles (ASYS) showed a substantial improvement of VF sensitivity when increasing the analysis duration from 5 s to 10 s. Specificity was not dependent on the ECG analysis duration. The 10 s CNN model presented the best performance: 92-94.4% (VF), 92.2-94% (ASYS), 96-97% (ONR), and 98.2-99.5% (NSR) for sliding decision times during CPR; 98-99% (VF), 98.2-99.8% (ASYS), 98.8-99.1 (ONR), and 100% (NSR) for sliding decision times after end of CPR. We identified the importance of sHOT as a reliable predictor of performance, accounting for the minimal sHOT interval of 2-3 s that provides a reliable rhythm detection satisfying the American Heart Association (AHA) standards for AED rhythm analysis. The presented technology for sliding shock advisory decision during CPR achieved substantial performance improvement in short hands-off periods (>2 s), such as insufflations or pre-shock pauses. The performance was competitive despite 1-2.8% point lower ASYS detection during CPR than the standard requirement (95%) for non-noisy ECG signals. The presented deep learning strategy is a basis for improved CPR practices involving both continuous CC and CC with insufflations, associated with minimal CC interruptions for reconfirmation of non-shockable rhythms (minimum hands-off time) and early treatment of VF (minimal pre-shock pauses).

Analyze Whilst Compressing algorithm for detection of ventricular fibrillation during CPR: A comparative performance evaluation for automated external defibrillators.

OBJECTIVE: The aim of this study was to present new combination of algorithms for rhythm analysis during cardiopulmonary resuscitation (CPR) in automated external defibrillators (AED), called Analyze Whilst Compressing (AWC), designed for decreasing pre-shock pause and early stopping of chest compressions (CC) for treating refibrillation. METHODS: Two stages for AED rhythm analysis were presented, namely, "Standard Analysis Stage" (conventional shock-advisory analysis run over 5 s after CC interruption every two minutes) and "AWC Stage" (two-step sequential analysis process during CPR). AWC steps were run in presence of CC (Step1), and if shockable rhythm was detected then a reconfirmation step was run in absence of CC (Step2, analysis duration 5 s). RESULTS: In total 16,057 ECG strips from 2916 out-of-hospital cardiac arrest (OHCA) patients treated with AEDs (DEFIGARD TOUCH7, Schiller Médical, France) were subjected patient-wise to AWC training (8559 strips, 1604 patients) and validation (7498 strips, 1312 patients). Considering validation results, "Standard Analysis Stage" presented ventricular fibrillation (VF) sensitivity Se = 98.3% and non-shockable rhythm specificity Sp>99%; "AWC Stage" decision after Step2 reconfirmation achieved Se = 92.1%, Sp>99%. CONCLUSION: AWC presented similar performances to other AED algorithms during CPR, fulfilling performance goals recommended by standards. AWC provided advances in the challenge for improving CPR quality by: (i) not interrupting chest compressions for prevalent part of non-shockable rhythms (66-83%); (ii) minimizing pre-shock pause for 92.1% of VF patients. AWC required hands-off reconfirmation in 34.4% of cases. Reconfirmation was also common limitation of other reported algorithms (25.7-100%) although following different protocols for triggering chest compression resumption and shock delivery.

Fully Convolutional Deep Neural Networks with Optimized Hyperparameters for Detection of Shockable and Non-Shockable Rhythms.

Deep neural networks (DNN) are state-of-the-art machine learning algorithms that can be learned to self-extract significant features of the electrocardiogram (ECG) and can generally provide high-output diagnostic accuracy if subjected to robust training and optimization on large datasets at high computational cost. So far, limited research and optimization of DNNs in shock advisory systems is found on large ECG arrhythmia databases from out-of-hospital cardiac arrests (OHCA). The objective of this study is to optimize the hyperparameters (HPs) of deep convolutional neural networks (CNN) for detection of shockable (Sh) and nonshockable (NSh) rhythms, and to validate the best HP settings for short and long analysis durations (2-10 s). Large numbers of (Sh + NSh) ECG samples were used for training (720 + 3170) and validation (739 + 5921) from Holters and defibrillators in OHCA. An end-to-end deep CNN architecture was implemented with one-lead raw ECG input layer (5 s, 125 Hz, 2.5 uV/LSB), configurable number of 5 to 23 hidden layers and output layer with diagnostic probability p ∈ [0: Sh,1: NSh]. The hidden layers contain N convolutional blocks × 3 layers (Conv1D (filters = Fi, kernel size = Ki), max-pooling (pool size = 2), dropout (rate = 0.3)), one global max-pooling and one dense layer. Random search optimization of HPs = {N, Fi, Ki}, i = 1, … N in a large grid of N = [1, 2, … 7], Fi = [5;50], Ki = [5;100] was performed. During training, the model with maximal balanced accuracy BAC = (Sensitivity + Specificity)/2 over 400 epochs was stored. The optimization principle is based on finding the common HPs space of a few top-ranked models and prediction of a robust HP setting by their median value. The optimal models for 1-7 CNN layers were trained with different learning rates LR = [10-5; 10-2] and the best model was finally validated on 2-10 s analysis durations. A number of 4216 random search models were trained. The optimal models with more than three convolutional layers did not exhibit substantial differences in performance BAC = (99.31-99.5%). Among them, the best model was found with {N = 5, Fi = {20, 15, 15, 10, 5}, Ki = {10, 10, 10, 10, 10}, 7521 trainable parameters} with maximal validation performance for 5-s analysis (BAC = 99.5%, Se = 99.6%, Sp = 99.4%) and tolerable drop in performance (<2% points) for very short 2-s analysis (BAC = 98.2%, Se = 97.6%, Sp = 98.7%). DNN application in future-generation shock advisory systems can improve the detection performance of Sh and NSh rhythms and can considerably shorten the analysis duration complying with resuscitation guidelines for minimal hands-off pauses.

Adaptive step size LMS improves ECG detection during MRI at 1.5 and 3 T.

OBJECTIVE: We describe a new real-time filter to reduce artefacts on electrocardiogram (ECG) due to magnetic field gradients during MRI. The proposed filter is a least mean square (LMS) filter able to continuously adapt its step size according to the gradient signal of the ongoing MRI acquisition. MATERIALS AND METHODS: We implemented this filter and compared it, within two databases (at 1.5 and 3 T) with over 6000 QRS complexes, to five real-time filtering strategies (no filter, low pass filter, standard LMS, and two other filters optimized within the databases: optimized LMS, and optimized Kalman filter). RESULTS: The energy of the remaining noise was significantly reduced (26 vs. 68%, p < 0.001) with the new filter vs. standard LMS. The detection error of our ventricular complex (QRS) detector was: 11% with our method vs. 25% with raw ECG, 35% with low pass filter, 17% with standard LMS, 12% with optimized Kalman filter, and 11% with optimized LMS filter. CONCLUSION: The adaptive step size LMS improves ECG denoising during MRI. QRS detection has the same F1 score with this filter than with filters optimized within the database.

Shock advisory system with minimal delay triggering after end of chest compressions: accuracy and gained hands-off time.

AIMS: Shortening hands-off intervals can improve benefits from defibrillation. This study presents the performance of a shock advisory system (SAS), which aims to decrease the pre-shock pauses by triggering fast rhythm analysis at minimal delay after end of chest compressions (CC). METHODS: The SAS is evaluated on a database of 1301 samples from 311 out-of-hospital cardiac arrests (OHCA) from automated external defibrillators (AEDs). The following rhythms are identified: 788 asystoles (ASYS), 20 normal sinus rhythms (NSR), 394 other non-shockable rythms (ONS), 81 ventricular fibrillations (VF), 18 rapid ventricular tachycardias (VThi). SAS is launched in two-stages: first stage for accurate detection of actual end of CC (ReEoCC); second stage for early "Shock"/"No-Shock" decision by using all available artifact-free ECG signals after REoCC during 3, 5, 7 s. RESULTS: Performance of the presented SAS versus AEDs is compared. The median hands-off time gained from earlier starting of ECG analysis is 5.8 s and for earlier shock advice is 12.5 s to 8.5 s when SAS rhythm analysis lasts 3 s to 7 s. The SAS accuracy at 3-7 s is: specificity 97.7-98.9% (ASYS), 100-100% (NSR), 98.5-99.2% (ONS); sensitivity 91.4-98.8% (VF), 88.9-96.7% (VThi). CONCLUSION: This study indicates that shortening the pre-shock hands-off pause by more efficient management of the SAS process in AEDs is possible. For analysis duration of 5 s (7 s), the delay between the end of chest compressions and the shock advice can be reduced by 10.5 s (8.5 s) median, while AHA requirements for rhythm detection accuracy are met. The use of this solution in AEDs could provide more reliable rhythm analysis than methods applying filtering techniques during CC.

Bench study of the accuracy of a commercial AED arrhythmia analysis algorithm in the presence of electromagnetic interferences.

This paper presents a bench study on a commercial automated external defibrillator (AED). The objective was to evaluate the performance of the defibrillation advisory system and its robustness against electromagnetic interferences (EMI) with central frequencies of 16.7, 50 and 60 Hz. The shock advisory system uses two 50 and 60 Hz band-pass filters, an adaptive filter to identify and suppress 16.7 Hz interference, and a software technique for arrhythmia analysis based on morphology and frequency ECG parameters. The testing process includes noise-free ECG strips from the internationally recognized MIT-VFDB ECG database that were superimposed with simulated EMI artifacts and supplied to the shock advisory system embedded in a real AED. Measurements under special consideration of the allowed variation of EMI frequency (15.7-17.4, 47-52, 58-62 Hz) and amplitude (1 and 8 mV) were performed to optimize external validity. The accuracy was reported using the American Heart Association (AHA) recommendations for arrhythmia analysis performance. In the case of artifact-free signals, the AHA performance goals were exceeded for both sensitivity and specificity: 99% for ventricular fibrillation (VF), 98% for rapid ventricular tachycardia (VT), 90% for slow VT, 100% for normal sinus rhythm, 100% for asystole and 99% for other non-shockable rhythms. In the presence of EMI, the specificity for some non-shockable rhythms (NSR, N) may be affected in some specific cases of a low signal-to-noise ratio and extreme frequencies, leading to a drop in the specificity with no more than 7% point. The specificity for asystole and the sensitivity for VF and rapid VT in the presence of any kind of 16.7, 50 or 60 Hz EMI simulated artifact were shown to reach the equivalence of sensitivity required for non-noisy signals. In conclusion, we proved that the shock advisory system working in a real AED operates accurately according to the AHA recommendations without artifacts and in the presence of EMI. The results may be affected for specificity in the case of a low signal-to-noise ratio or in some extreme frequency setting.