If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
The current clinical paradigm to diagnose supraventricular tachycardias (SVTs) results in potential overlap between various ECG expressions. Machine learning may identify visually imperceptible ECG changes and augment the predictive accuracy of determining SVT mechanisms.
To compare a Convolutional Neural Network (CNN) with manual SVT identification among atrioventricular nodal re-entrant tachycardia (AVNRT), atrioventricular reciprocating tachycardia (AVRT), and atrial tachycardia (AT).
All patients with a 12-lead ECG of a diagnosed and successfully ablated SVT during an electrophysiology study from 2013-2020 were included. Digital ECG data ≥10 seconds were extracted from the recording system and split into training, validation, and test datasets in a ratio of approximately 7:1:2. The results were reported as the average across 10 random data splits and model initializations for robustness. We then compared the CNN performance with an independent adjudication by an experienced cardiac electrophysiologist.
From 763 patients, 1524 ECGs (371 AVNRT, 312 AVRT, 95 AT, and 746 sinus rhythm) were used to develop the CNN. CNN identified 1) AVNRT with a higher sensitivity and similar specificity; 2) AVRT with a lower sensitivity but higher specificity; and 3) AT with a lower sensitivity and similar specificity compared to the adjudicator (Table). The CNN area under the receiver operating characteristic curve for AVNRT, AVRT, and AT was 0.855, 0.880, and 0.774 respectively.