CODE-II: A large-scale dataset for artificial intelligence in ECG analysis

TL;DR

This paper introduces CODE-II, a large-scale, high-quality ECG dataset with detailed annotations, enabling improved AI-based ECG analysis and demonstrating its effectiveness through transfer learning on external benchmarks.

Contribution

The paper presents CODE-II, a comprehensive ECG dataset with standardized annotations and multiple subsets, facilitating advancements in AI ECG interpretation and outperforming existing datasets in transfer learning.

Findings

Pre-trained neural network on CODE-II outperforms models trained on larger datasets.

CODE-II achieves superior transfer performance on external ECG benchmarks.

High-quality, annotated dataset enhances AI ECG analysis capabilities.

Abstract

Data-driven methods for electrocardiogram (ECG) interpretation are rapidly progressing. Large datasets have enabled advances in artificial intelligence (AI) based ECG analysis, yet limitations in annotation quality, size, and scope remain major challenges. Here we present CODE-II, a large-scale real-world dataset of 2,735,269 12-lead ECGs from 2,093,807 adult patients collected by the Telehealth Network of Minas Gerais (TNMG), Brazil. Each exam was annotated using standardized diagnostic criteria and reviewed by cardiologists. A defining feature of CODE-II is a set of 66 clinically meaningful diagnostic classes, developed with cardiologist input and routinely used in telehealth practice. We additionally provide an open available subset: CODE-II-open, a public subset of 15,000 patients, and the CODE-II-test, a non-overlapping set of 8,475 exams reviewed by multiple cardiologists for blinded evaluation. A neural network pre-trained on CODE-II achieved superior transfer performance on external benchmarks (PTB-XL and CPSC 2018) and outperformed alternatives trained on larger datasets.