Deshmukh, A D

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 Introduction Atrial fibrillation (AF) is a major contributor to stroke risk, yet many cases remain undiagnosed due to its silent and paroxysmal nature. To address this, multiple risk scores have been developed to predict incident AF. This study aimed to evaluate the predictive accuracy of five widely used clinical AF prediction tools (i.e., CHARGE-AF, CHA2DS2-VASc, mC2HEST, HATCH, and HAVOC) among patients undergoing long-term continuous monitoring with implantable loop recorders (ILRs). Methods We conducted a retrospective cohort study across three major US academic medical centers, including patients with ILRs and no prior diagnosis or ECG evidence of AF. Clinical variables used to calculate each risk score were extracted from electronic health records. Discrimination was assessed using the area under the receiver operating characteristic curve (AUC). Kaplan-Meier survival analysis and Cox proportional hazards models were used to evaluate the association between each score’s risk categories and incident AF (defined as any episode lasting ≥6 minutes). Calibration was assessed using calibration plots and Brier scores. Results A total of 3,895 patients were included (mean age 61 ± 17 yrs.; 51% male). Incident AF was detected in 879 patients (22%) over a median follow up of 17 months (IQR 8-35). Table 1 presents baseline characteristics and device detected AF event data during follow up. CHARGE-AF had the highest AUC (0.68) but also the highest Brier score (Fig. 1). Failure curves showed clear risk stratification across all tools (Fig. 2). Calibration plots (Fig. 3) demonstrated modest agreement between predicted and observed risk across all models. Using CHARGE-AF as the reference, pairwise comparisons of Brier scores via the Wilcoxon signed-rank test showed lower calibration accuracy than CHA2DS2-VASc (p = 0.03), mC2HEST (p = 0.001), and HAVOC (p = 0.001), with no significant difference compared to HATCH (p = 0.15). Conclusions In patients undergoing continuous rhythm monitoring with ILRs, commonly used AF prediction scores demonstrated modest discriminatory ability and similar calibration performance. These findings suggest that existing risk models, which were largely developed in cohorts with clinically diagnosed AF, may not optimally translate to continuously monitored populations. Future work is needed to develop and validate AF prediction models specifically in cohorts undergoing long-term continuous cardiac monitoring.Figure 1Figure 2-3