Automated Differentiation of Wide QRS Complex Tachycardia
Author Information
Author(s): May Adam M., Katbamna Bhavesh B., Shaikh Preet A., LoCoco Sarah, Deych Elena, Zhou Ruiwen, Liu Lei, Mikhova Krasimira M., Ghadban Rugheed, Cuculich Phillip S., Cooper Daniel H., Maddox Thomas M., Noseworthy Peter A., Kashou Anthony
Primary Institution: Washington University School of Medicine in St. Louis
Hypothesis
Can machine learning algorithms effectively differentiate between ventricular tachycardia and supraventricular wide complex tachycardia using QRS polarity and shifts?
Conclusion
Automated algorithms using QRS polarity and shifts can accurately differentiate wide QRS complex tachycardias, improving diagnostic accuracy.
Supporting Evidence
- Machine learning models showed AUCs ranging from 0.86 to 0.93 for differentiating WCT types.
- Models using paired ECG data improved diagnostic accuracy compared to those using WCT data alone.
- Presence of a polarity shift strongly favored a diagnosis of ventricular tachycardia.
Takeaway
Doctors can use special computer programs to tell the difference between two types of fast heartbeats, which helps them treat patients better.
Methodology
The study used machine learning models trained on ECG data from patients with wide QRS complex tachycardia, comparing features from WCT and baseline ECGs.
Potential Biases
Potential bias due to reliance on physician diagnoses without corroborating electrophysiology procedures for all patients.
Limitations
The study included WCTs diagnosed by physicians without a robust reference standard for all cases.
Participant Demographics
Patients included 235 individuals with wide QRS complex tachycardia, with varying underlying conditions.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website