‘Heart attack’ is something that most people are familiar with, but not many know of the other big problem — heart rhythm abnormalities or ‘arrhythmias’, where the heart goes either too slow or too fast.
We read about young athletes suddenly collapsing on the field during a game, and studies have implied that arrhythmias and other electrical disorders may be the most common cause. The risk of heart rhythm abnormality increases after a heart attack. Where the condition runs in families, early diagnosis enables timely treatment. In others, cardiac rhythm abnormalities can arise from heart muscle abnormalities or a dilated heart. Heart rhythm abnormalities such as ‘atrial fibrillation’ (where the upper chambers of the heart beat chaotically, irregularly, and out of sync with the lower chambers) need prompt detection and treatment to prevent clot formation in the heart, and stroke.
AI in diagnosis
Artificial intelligence-based research in cardiology has made tremendous progress over the last few years. Universal use of smartphones, advancements in telecommunications, easy availability of Wi-Fi, Bluetooth and wireless technologies, internet-based storage and analysis using AI algorithms have enabled the development and rapid growth of hand-held and wearable cardiac monitors, including smartwatches and smartphones.
These devices have been tested in various studies across the globe. In the Apple Heart Study in the US, involving a population of over 400,000, a smartwatch-based irregular pulse notification AI algorithm was successfully used to detect atrial fibrillation; these algorithms are currently being used in clinical practice.
Earlier, if anyone had palpitation, medical experts would generally monitor the heart rate for at least 72 hours. Patients were required to undergo the analysis once a week or once in two weeks or sometimes once a month, so it was difficult to catch an abnormal heart rhythm at a particular time. However, now, with the advent of gadgets that use AI, it has become easier to keep track.
Computerised AI interpretation of electrocardiogram (ECG) has shown accuracy in line with opinions from expert cardiologists. Researchers at Mayo Clinic, Rochester, in the US, have developed an AI algorithm using 6.5 lakh ECG readings that can detect previous episodes or impending episodes of atrial fibrillation. This AI-enabled ‘12-lead ECG’, as well as several other AI algorithms developed in recent years can be used to identify people at risk for a stroke, heart valve problems, heart failure, or heart muscle abnormalities.
The AI algorithms are soon likely to be available for clinical use and would have a significant impact in India, where a simple, universally available test such as the ‘12-lead ECG’ can enable rapid diagnosis even in remote parts of the country. The 12-lead ECG was approved for clinical use in the US in the middle of the Covid-19 pandemic to identify patients at risk of cardiac arrest.
These contemporary technological solutions are likely to revolutionise heart rhythm management in the near future, resulting in timely diagnosis and interventions to save lives.
In sum, AI today has the capability to predict arrhythmias, heart attacks, and so on. If a heart patient is in hospital today, the technology makes it possible to say with a high probability that he will need a particular type of treatment. This is already happening in India, where patients in remote locations are being counselled and treated by a doctor sitting in a faraway city.
The writer is cardiac electrophysiologist and lead for heart rhythm and cardiac device services, Kokilaben Dhirubhai Ambani Hospital, Mumbai