Khanolkar, U K

Abstract Background New-onset atrial fibrillation (AF) is common in patients who are hospitalized with acute illness and is associated with overall poor outcomes. Current clinical and demographic predictors of new-onset AF have limitations. Electrocardiogram (ECG) waveforms, combined with these factors as part of multimodal approach could help improve prediction of new-onset AF. Purpose The purpose of our study was to develop a multimodal, deep-learning system that combines ECG waveforms with clinical data for prediction of new-onset AF in hospitalized patients. Methods We developed and compared two separate 1-dimensional models with a fusion model for prediction of new-onset AF. (a): one-dimensional convolutional neural network (1D-CNN) that learns patterns from the raw ECG waveforms. (b): a Multi-Layer Perceptron (MLP), unsupervised machine learning model which utilizes data from structured variables such as: age, heart rate, routinely obtained laboratory data and medications. The two modules each produce latent feature vectors, which are then placed end-to-end and fed into a final Multi- MLP head that outputs the probability of New-onset AF during hospitalization. We then compared this fused MLP model with the two simpler unimodal models based solely on ECG waveforms and clinical data respectively. Results On the held-out test set data, the feature-level fused model achieved an Area Under the Curve (AUC) of 0.81. This significantly outperformed both the ECG-alone model (0.77) and the clinical-alone model (0.74). By adjusting the threshold, we can fine-tune the tool at a threshold of 0.5, it flags 83% of patients who later develop AF (207 of 248) with a precision of 10%, a setting that could be useful for wider screening. By adjusting the threshold of 0.98 for a high-precision alert configuration, it has the ability to predict 46% of future occurrence of AF cases (113 of 248) while reducing false alarms by 81% (from 1,797 to 328), increasing precision to 26% and achieving an overall accuracy of 89%, a setting useful when resources for follow-up are limited. Conclusions Combining ECG signals with clinical data significantly enhances the model’s ability to detect new-onset AF in hospitalized patients as compared to a uni-dimensional CNN model. The dynamic nature of ML model also allows for adjustment of its decision threshold which can aid in broader screening and improving adverse outcomes related to occurrence of AF.ROC Curve for Fusion model

ABSTRACT Background Ventricular tachycardia (VT) is a life‐threatening arrhythmia that requires both relatively rapid and accurate detection in intensive care units (ICUs). Continuous monitoring systems play a crucial role in detecting them. However, previous studies have reported that nearly 9 out of 10 arrhythmia alarms in ICUs tend to be false positives, which usually transpire to a well‐documented phenomenon called “alarm fatigue” that leads to desensitization, delayed responses, and increased cognitive burden on healthcare providers. Methods We developed a deep learning based, one‐dimensional convolutional neural network (1D‐CNN) to classify VT alarms using multiple raw waveform inputs, including two electrocardiogram (ECG) leads, photoplethysmogram (PPG) and arterial blood pressure (ABP) signals. The model was trained using the publicly available VTaC Arrhythmia Benchmark Dataset. We used the 10‐second waveform segments that preceded each VT alarm, pre‐processed and then used them to train the machine learning model to correctly classify the VT alarm. Results On the test set, the model achieved an area under the receiver operating characteristic curve of 0.901, overall accuracy of 83.22%, F1‐score of 73.3%, sensitivity of 77.53%, specificity of 85.63%, and positive predictive value of 69.57%. The model successfully detected over three‐quarters of them while significantly reducing false positive rates for the detection of VT. Conclusions This study demonstrates that a deep learning based 1D‐CNN model using short segments of raw waveform data can achieve robust performance in distinguishing true and false VT alarms.

Ventricular tachycardia (VT) is a life-threatening arrhythmia that requires both relatively rapid and accurate detection in intensive care units (ICUs). Continuous monitoring systems play a crucial role in detecting them. However, previous studies have reported that nearly 9 out of 10 arrhythmia alarms in ICUs tend to be false positives, which usually transpire to a well-documented phenomenon called “alarm fatigue” that leads to desensitization, delayed responses and increased cognitive burden on healthcare providers. We developed a deep learning based, one-dimensional convolutional neural network (1D-CNN) to classify VT alarms using multiple raw waveform inputs, including two electrocardiogram (ECG) leads, photoplethysmogram (PPG) and arterial blood pressure (ABP) signals. The model was trained using the publicly available VTaC Arrhythmia Benchmark Dataset. Ten-second waveform segments preceding each alarm were pre-processed and used to train the model to correctly classify the VT alarm. On the test set, the model achieved an area under the receiver operating characteristic curve (AUC-ROC) of 0.901, overall accuracy of 83.22%, F1-score of 73.3%, sensitivity of 77.53%, specificity of 85.63%, and positive predictive value (PPV) of 69.57%. The model successfully detected over three-quarters of them while significantly reducing false positive rates for the detection of VT. This study demonstrates that a deep learning based 1D-CNN model using short segments of raw waveform data can achieve robust performance in distinguishing true and false VT alarms.