44. DEVELOPMENT OF AN ARTIFICIAL INTELLIGENCE MODEL FOR PREMATURE VENTRICULAR COMPLEX DETECTION ON SINUS RHYTHM ELECTROCARDIOGRAM
Main Article Content
Abstract
Objectives: To develop an AI model capable of accurately detecting PVCs and to evaluate its screening performance on reference standardized ECG datasets.
Methods: This retrospective study utilized 24-hour Holter ECG data collected from Nguyen Trai Hospital and Nguyen Tri Phuong Hospital between 2021 and 2024. Data labeling and analysis were performed from October 2024 to April 2025. The AI model was constructed using a deep learning-based ResNet architecture.
Results: From a total of 453 Holter ECG datasets, 643675 PVCs were identified, with the rate of patients exhibiting frequent PVCs recorded as 4.0%. The prevalence of PVC couplet, bigeminy, and trigeminy was 17.0%, 31.8%, and 29.1%, respectively. The developed AI model demonstrated a sensitivity of over 80%, a specificity exceeding 90%, and an F1-score above 85% when validated against MIT-BIH, AHA, and ESC reference datasets.
Conclusion: Our AI model has strong potential for real-world application in large-scale ECG-based PVC screening, offering an efficient and scalable solution for PVC detection
Article Details
Keywords
Electrocardiogram, premature ventricular complex, artificial intelligence.
References
[2]. Zeppenfeld K, Tfelt-Hansen J, de Riva M, Winkel BG, Behr ER, Blom NA, Charron P, Corrado D, Dagres N, de Chillou C, Eckardt L, Friede T, Haugaa KH, Hocini M, Lambiase PD, Marijon E, Merino JL, Peichl P, Priori SG, Reichlin T, Schulz-Menger J, Sticherling C, Tzeis S, Verstrael A, Volterrani M; ESC Scientific Document Group. 2022 ESC Guidelines for the management of patients with ventricular arrhythmias and the prevention of sudden cardiac death. Eur Heart J. 2022 Oct 21;43(40):3997-4126. doi: 10.1093/eurheartj/ehac262.
[3]. Amoni M, Dries E, Ingelaere S, Vermoortele D, Roderick HL, Claus P, Willems R, Sipido KR. Ventricular Arrhythmias in Ischemic Cardiomyopathy-New Avenues for Mechanism-Guided Treatment. Cells. 2021 Oct 1;10(10):2629. doi: 10.3390/cells10102629.
[4]. Babayiğit E, Ulus T, Görenek B. State-of-the-art look at premature ventricular complex diagnosis and management: Key messages for practitioners from the American College of Cardiology Electrophysiology Council. Turk Kardiyol Dern Ars. 2020 Oct;48(7):707-713. English. doi: 10.5543/tkda.2020.69786.
[5]. Cai Z., Li J., Johnson A.E., Zhang X., Shen Q., Zhang J., Liu C. Rule-based rough-refined two-step-procedure for real-time premature beat detection in single-lead ECG. Physiol. Meas. 2020;41:54001–54004. doi: 10.1088/1361-6579/ab87b4.
[6]. ECG Databases. https://www.physionet.org/physiotools/wag/evnode3.htm. (Accessed on Mar 30, 2025)
[7]. Jung Y., Kim H. Detection of PVC by using a wavelet-based statistical ECG monitoring procedure. Biomed. Signal Process. Control. 2017;36:176–182. doi: 10.1016/j.bspc.2017.03.023.
[8]. Li Q., Liu C., Li Q., Shashikumar S.P., Nemati S., Shen Z., Clifford G.D. Ventricular ectopic beat detection using a wavelet transform and a convolutional neural network. Physiol. Meas. 2019;40:55001–55002. doi: 10.1088/1361-6579/ab17f0.
[9]. Sun X, Yin Y, Yang Q, Huo T. Artificial intelligence in cardiovascular diseases: diagnostic and therapeutic perspectives. Eur J Med Res. 2023 Jul 21;28(1):242. doi: 10.1186/s40001-023-01065-y.
[10]. Koulaouzidis G, Jadczyk T, Iakovidis DK, Koulaouzidis A, Bisnaire M, Charisopoulou D. Artificial Intelligence in Cardiology-A Narrative Review of Current Status. J Clin Med. 2022 Jul 5;11(13):3910. doi: 10.3390/jcm11133910.