Paired clinical 12 lead and apple watch electrocardiogram data repository from childhood cancer survivors authors

TL;DR

This paper introduces a dataset of paired clinical and Apple Watch ECGs from childhood cancer survivors to support the development of AI tools for remote cardiac monitoring.

Contribution

The dataset provides paired same-day 12-lead and wearable ECGs for AI model validation in cardiovascular surveillance.

Findings

The dataset supports the development of AI-based cardiac screening tools using paired ECG data.

Researchers can evaluate if AI models trained on clinical ECGs work with Apple Watch ECGs.

The dataset may improve long-term cardiovascular monitoring for childhood cancer survivors.

Abstract

Childhood cancer survivors (CCS), exposed to prior cardiotoxic treatments such as anthracyclines and chest radiation, are at lifelong risk of cardiovascular complications. Current guidelines recommend periodic echocardiographic surveillance, but adherence rates are as low as 41%. This dataset provides paired same-day 12-lead clinical electrocardiograms (ECG) and single-lead wearable ECG recordings from the Apple Watch, collected from adult CCS participating in the St. Jude Lifetime Cohort Study (SJLIFE). The availability of paired wearable and clinical ECGs enables the development and validation of remote AI-based cardiac screening tools, potentially leading to more precise long-term cardiovascular surveillance in this population. Using this dataset, researchers can assess whether an AI model developed using clinical ECG can be repeat when using ECG from an Apple Watch.