AI gaining popularity in electrophysiology

His primary area of AI research has been focused on its ability to detect atrial fibrillation (AFib) and ventricular tachycardia (VT). He has collaborated with physicians in France on a couple studies using the Cardiologs Holter ECG analysis algorithm to analyze data from the first 24 hours of using a Holter to predict diagnoses over the next 14 days. Singh's team published a study on this in December 2021 and presented an additional late-breaking study at the 2023 Heart Rhythm Society (HRS) meeting. In the late-breaking study, a novel AI model correctly identified patients at near-term risk of life-threatening sustained VT who could potentially benefit from pre-emptive interventions to prevent sudden cardiac death (SCD). The AI model utilizes a single-lead electrocardiogram (ECG) screening tool that could offer physicians a new approach to SCD risk management. 

"Early prediction is going to be key, because it allows us to put into place preventive strategies," Singh explained. 

In the EP space, there have been several algorithms developed to read ECGs to help flag events or simply read standard diagnostic ECGs. These algorithms are trained to interpret single lead ECGs from wearable device, including consumer-grade watches, and they are able to help physicians predict when patients may go on to develop AFib or other similar conditions. Combining these advanced AI models with traditional risk score assessments can determine a patient's risk of future events in the coming days or even up to five years into the future. 

"The deep-learning algorithms are able to pick up signals within the surface ECGs that the human eye cannot see," Singh explained. 

Several late-breaking studies in previous years at HRS have included AI algorithms for consumer-grade heart monitors built into devices such as the Apple Watch. Some EP experts see the algorithms as a way to increase awareness about conditions such as AFib and enable wider population level screening. There are also concerns about a tidal wave of concerned patients coming in with false positive readings. 

"The watches are able to show which patients are developing AFib. Unfortunately, not all of those algorithms are always correct, so we do have a lot of false positives. This unnecessarily increases patient anxiety and physician and clinician burden. I think it is a great technology, it just requires a lot of refining."

Mass General is also working with the vendor SentiAR to develop AI-based augmented reality (AR) to improve the the procedural workflow in the EP lab.

"We wear virtual reality glasses while we are doing procedures, so that is another form of AI we are actively engaged in," Singh said.