Doctors must begin to prepare to put more trust in machines after hearing that a machine-learning model is able to identify heart arrhythmias from an electrocardiogram (ECG) better than an expert.

This revelation comes from a team of researchers at Stanford University, led by Andrew Ng, a prominent AI researcher and an adjunct professor there, reports MIT. The automated approach could prove vital to everyday medical treatment by making the diagnosis of potentially deadly heartbeat irregularities more reliable. Additionally, it has the potential to make quality care more readily available in areas where resources are scarce.

Advances such as this reinforce how machine learning seems to be revolutionising medicine. In recent years, researchers have shown that machine learning techniques can be used to locate all sorts of ailments, including, breast cancer, skin cancer and eye disease from medical images.

Ng explained that he has been encouraged by how many people are accepting the idea that deep learning can diagnose at an accuracy superior to doctors in select verticals. He also added that he is exhilarated to see researchers looking beyond imaging to other forms of data such as ECG.

During the process, the team trained a deep learning algorithm to identify different types of irregular heartbeats in ECG data. Some irregularities can lead to serious health complications including sudden cardiac death, but the signal can be difficult to detect, so patients are often asked to wear ECG sensor for several weeks. However, even then it can be difficult for a doctor to distinguish between irregularities that may be benign and ones that could require treatment.

The researchers partnered with iRhythm, a company that makes portable ECG devices. They collected 30,000 30 second clips from patients with different forms of arrhythmia.. To assess the accuracy of their algorithm, the team compared its performance to that of five different cardiologists on 300 undiagnosed clips. They had a panel of three expert cardiologists provide a ground truth judgement.

Deep learning focuses on feeding large quantities of data into a large simulated neural network, and fine tuning its parameters until it accurately recognised problematic ECG signals. Looking ahead to the future, it is evident that there is a chance that machine learning will be able to find traces of disease by combing through large quantities of disparate data.

However, the main hurdle to overcome is to get doctors and patients to fully entrust in algorithm’s that are often so complex that their reasoning cannot be understood. Finding ways to make it more explainable will be vital in building trust and refining treatment.

Despite this, NG believes that a revolution with this technology is on the horizon. He commented, “We still have work ahead to get these algorithms into the healthcare system’s workflow. But I think health care 10 years from now will use a lot more AI and will look very different than it does today.”