What is the clinical evidence from this study?

Study design: Other. Population: ST-segment elevation myocardial infarction (STEMI) (n=79). Intervention: Modified ResNet-50 and Custom CNN Ensembles vs. MobileNetV2 and LeNet5 ensembles. Primary outcome: Diagnostic accuracy for STEMI detection.

What question did this study set out to answer?

The central aim is to develop a hybrid model for accurate automated classification of STEMI using ECG data.

June 12, 2026Discover Artificial Intelligence0 citationsOpen Access

Modified ResNet-50 and custom CNN ensembles achieve near perfect STEMI detection on European ST-T database

SNSonam NagarNational Institute of Technology Delhi KVKaran VermaNational Institute of Technology Delhi SSSachin SinghManipal Academy of Higher Education

Key Result

Modified ResNet-50 and Custom CNN ensembles achieved 99.78% accuracy, 98.67% AUC, and 96.19% MCC in detecting STEMI from ambulatory ECG recordings.

Key Points

The central aim is to develop a hybrid model for accurate automated classification of STEMI using ECG data.
Hybrid model utilizes Modified ResNet-50 and ensemble techniques for ECG classification,
Data comprises 90 annotated ECG excerpts from 79 subjects, partitioned into 80% training and 20% testing,
Performance metrics compared against MobileNetV2, LeNet5, and Custom CNN in ensemble classification.
Modified ResNet-50 with Ensemble achieved 99.78% accuracy and 98.67% AUC,
Custom CNN with Ensemble also reached 99.78% accuracy with 96.30% MCC,
MobileNetV2 and LeNet5 had lower performance metrics compared to the primary models.

Structured PICO

Does a Modified ResNet-50 and Custom CNN ensemble model improve automated detection of STEMI and ischemia in ambulatory ECG excerpts?

Population

79 subjects from the European ST-T Database, yielding 13,500 ECG segments, used to train and evaluate deep learning models for automated STEMI detection.

Intervention

Modified ResNet-50 with a sparse autoencoder optimized via an ensemble approach, and Custom CNN with Ensemble

Comparator

M-MobileNetV2 and M-LeNet5 with ensemble classification

Outcome

Automated ECG classification performance metrics including accuracy, precision, recall, AUC, and Matthews Correlation Coefficient (MCC) for ST-segment abnormalities (STEMI and ischemia)

Modified ResNet-50 and custom CNN ensemble models demonstrate near-perfect accuracy for automated STEMI detection on ambulatory ECGs, offering a robust framework for AI-assisted cardiac diagnostics.

Limitations

Single-dataset training without external validation
Retrospective in silico validation without real-time clinical deployment testing

Abstract

Abstract Background ST-segment elevation myocardial infarction (STEMI) affects over 750,000 Americans annually, with 30-day mortality rate of 2.5–10%. While traditional risk factors include hypertension, hyperlipidemia, smoking, and diabetes, though SMuRF-less cases suggest genetic or non-atherosclerotic causes. Rapid diagnosis via electrocardiogram (ECG) ST- elevation and troponin levels is vital, requiring primary percutaneous coronary intervention (PCI) within 120 min, supported by dual antiplatelet and anticoagulant therapy. Recent advancements (2023–2025) in AI-enhanced ECG interpretation, PCSK9 inhibitors, and revascularization highlight the need for rapid detection. Methods Our study introduces a hybrid model for automated ECG classification of ST-segment abnormalities, including STEMI and ischemia, using the European ST-T Database, comprising 90 annotated ambulatory ECG excerpts from 79 subjects, converted into CSV format and partitioned into 80% training and 20% testing, yielding 10,800 training and 2700 testing samples despite class imbalance. The base model, Modified ResNet-50 with a sparse autoencoder, is optimized via an ensemble approach and compared with M-MobileNetV2, M-LeNet5, and Custom CNN, all employing ensemble classification. Results Modified ResNet-50 with Ensemble achieved 99.78% accuracy, 95.12–99.92% precision, 97.50–99.85% recall, 98.67% AUC, and 96.19% MCC. Custom CNN with Ensemble excelled with 99.78% accuracy, 98.76–99.81% precision, 94.12–99.96% recall, 97.04% AUC, and 96.30% MCC. MobileNetV2 and LeNet5 with Ensemble scored lower across metrics. Conclusion The proposed models enable rapid STEMI detection, enhancing PCI timeliness, telehealth monitoring, and AI assisted ECG training. Custom CNN with Ensemble offers superior precision and MCC, ideal for edge devices, while Modified ResNet-50 ensures high-sensitivity diagnostics, forming a robust framework for improved cardiac care.

Ask AI

Helpful

Bookmark

View Full Paper