Fast Detection of Ventricular Fibrillation and Ventricular Tachycardia in 1-Lead ECG from Three-Second Blocks

0509014 - ÚPT 2020 RIV US eng C - Conference Paper (international conference)
Plešinger, Filip - Andrla, Petr - Viščor, Ivo - Halámek, Josef - Jurák, Pavel
Fast Detection of Ventricular Fibrillation and Ventricular Tachycardia in 1-Lead ECG from Three-Second Blocks.
Computing in Cardiology 2018. Vol. 45. New York: IEEE, 2018, č. článku 8743881. E-ISSN 2325-887X.
[Computing in Cardiology Conference (CinC) /45./. Maastricht (NL), 23.09.2018-26.09.2018]
R&D Projects: GA MŠMT(CZ) LO1212
Grant - others:AV ČR(CZ) MSM100651602
Program: Program na podporu mezinárodní spolupráce začínajících výzkumných pracovníků
Institutional support: RVO:68081731
Keywords : cardiology * automated detection * ventricular fibrillation * ventricular tachycardia
OECD category: Medical engineering

Background: Ventricular tachycardia (VT) is dangerous irregularity of heart rhythm. VT may evolve into ventricular fibrillation (VF) which often leads to cardiac death. Therefore, fast automated detection of VF/VT events is of the utmost importance. Here, we present a method detecting VT and ventricular fibrillation (VF) events suitable for real-time application on continuously incoming ECG data.

Method: We designed a method for detection of VF/VT events in short-time (3 s), 1-lead ECG blocks. Five features are extracted from this block using analysis of ECG spectra, derivatives, amplitude measures and auto-correlation. The extracted features are fed into a logistic regression model showing the probability of a VF/VT event. The model was trained on the public PhysioNet CUDB dataset consisting of 393 automatically selected blocks.

Results: The model (AUC 0.99) showed a sensitivity and specificity of 95 % and 97 %, respectively (5-fold cross-validation). The model was tested on the public PhysioNet VFDB dataset, showing specificity and sensitivity of 95 % and 83 %, respectively. Both the feature extraction code (Matlab format) and the model are publicly accessible and easy implementation of the logistic regression model predetermines it for real-time applications.
Permanent Link: http://hdl.handle.net/11104/0299828