A Human-Centered AI Framework for Efficient Labelling of ECGs From Drug Safety Trials.

Drug safety trials require substantial ECG labelling like, in thorough QT studies, measurements of the QT interval, whose prolongation is a biomarker of proarrhythmic risk. The traditional method of manually measuring the QT interval is time-consuming and error-prone. Studies have demonstrated the potential of deep learning (DL)-based methods to automate this task but expert validation of these computerized measurements remains of paramount importance, particularly for abnormal ECG recordings. In this paper, we propose a highly automated framework that combines such a DL-based QT estimator with human expertise. The framework consists of 3 key components: (1) automated QT measurement with uncertainty quantification (2) expert review of a few DL-based measurements, mostly those with high model uncertainty and (3) recalibration of the unreviewed measurements based on the expert-validated data. We assess its effectiveness on 3 drug safety trials and show that it can significantly reduce effort required for ECG labelling-in our experiments only 10% of the data were reviewed per trial-while maintaining high levels of QT accuracy. Our study thus demonstrates the possibility of productive human-machine collaboration in ECG analysis without any compromise on the reliability of subsequent clinical interpretations.

AI-Assisted QT Measurements for Highly Automated Drug Safety Studies.

Rate-corrected QT interval (QTc) prolongation has been suggested as a biomarker for the risk of drug-induced torsades de pointes, and is therefore monitored during clinical trials for the assessment of drug safety. Manual QT measurements by expert ECG analysts are expensive, laborious and prone to errors. Wavelet-based delineators and other automatic methods do not generalize well to different T wave morphologies and may require laborious tuning. Our study investigates the robustness of convolutional neural networks (CNNs) for QT measurement. We trained 3 CNN-based deep learning models on a private ECG database with human expert-annotated QT intervals. Among these models, we propose a U-Net model, which is widely used for segmentation tasks, to build a novel clinically useful QT estimator that includes QT delineation for better interpretability. We tested the 3 models on four external databases, amongst which a clinical trial investigating four drugs. Our results show that the deep learning models are in stronger agreement with the experts than the state-of-the-art wavelet-based algorithm. Indeed, the deep learning models yielded up to 71% of accurate QT measurements (absolute difference between manual and automatic QT below 15 ms) whereas the wavelet-based algorithm only allowed 52% of QT accuracy. For the 2 studies of drugs with small to no QT prolonging effect, a mean absolute difference of 6 ms (std = 5 ms) was obtained between the manual and deep learning methods. For the other 2 drugs with more significant effect on the volunteers, a mean difference of up to 17 ms (std = 17 ms) was obtained. The proposed models are therefore promising for automated QT measurements during clinical trials. They can analyze various ECG morphologies from a diversity of individuals although some QT-prolonged ECGs can be challenging. The U-Net model is particularly interesting for our application as it facilitates expert review of automatic QT intervals, which is still required by regulatory bodies, by providing QRS onset and T offset positions that are consistent with the estimated QT intervals.

Deep learning analysis of electrocardiogram for risk prediction of drug-induced arrhythmias and diagnosis of long QT syndrome.

AIMS: Congenital long-QT syndromes (cLQTS) or drug-induced long-QT syndromes (diLQTS) can cause torsade de pointes (TdP), a life-threatening ventricular arrhythmia. The current strategy for the identification of drugs at the high risk of TdP relies on measuring the QT interval corrected for heart rate (QTc) on the electrocardiogram (ECG). However, QTc has a low positive predictive value. METHODS AND RESULTS: We used convolutional neural network (CNN) models to quantify ECG alterations induced by sotalol, an IKr blocker associated with TdP, aiming to provide new tools (CNN models) to enhance the prediction of drug-induced TdP (diTdP) and diagnosis of cLQTS. Tested CNN models used single or multiple 10-s recordings/patient using 8 leads or single leads in various cohorts: 1029 healthy subjects before and after sotalol intake (n = 14 135 ECGs); 487 cLQTS patients (n = 1083 ECGs: 560 type 1, 456 type 2, 67 type 3); and 48 patients with diTdP (n = 1105 ECGs, with 147 obtained within 48 h of a diTdP episode). CNN models outperformed models using QTc to identify exposure to sotalol [area under the receiver operating characteristic curve (ROC-AUC) = 0.98 vs. 0.72, P ≤ 0.001]. CNN models had higher ROC-AUC using multiple vs. single 10-s ECG (P ≤ 0.001). Performances were comparable for 8-lead vs. single-lead models. CNN models predicting sotalol exposure also accurately detected the presence and type of cLQTS vs. healthy controls, particularly for cLQT2 (AUC-ROC = 0.9) and were greatest shortly after a diTdP event and declining over time (P ≤ 0.001), after controlling for QTc and intake of culprit drugs. ECG segment analysis identified the J-Tpeak interval as the best discriminator of sotalol intake. CONCLUSION: CNN models applied to ECGs outperform QTc measurements to identify exposure to drugs altering the QT interval, congenital LQTS, and are greatest shortly after a diTdP episode.