iRhythm Technologies, Inc. is now collaborating with the Stanford Machine Learning Group resulting in the development of a deep learning algorithm capable of expert level detection of 14 cardiac output classes, including 12 arrhythmias as well as Sinus Rhythm and noise from artifact. The collaboration leveraged the iRhythm data science and clinical teams’ expertise in electrocardiogram (ECG) analysis, as well as iRhythm’s expansive labeled ECG data set to produce an arrhythmia detection algorithm delivering expert level classification performance. The classifications include Atrial Fibrillation, Atrial Flutter, Complete Heart Block, 2nd Degree AV Block, and Ventricular Tachycardia, among others.
Consistent with iRhythm’s ongoing investment in artificial intelligence (A.I.) for arrhythmia detection and analysis, and because deep learning models are dependent upon vast amounts of reliable data, the company provided an annotated data set of about 30,000 unique patients, 500 times larger than standards-based databases utilized in previous studies. This enabled the Stanford researchers, in collaboration with iRhythm Machine Learning Specialist Masoumeh Haghpanahi, to develop the groundbreaking 34-layer convolutional neural network, comparable to A.I. models used in computer vision and speech recognition.
“As a digital health company, we are very pleased to leverage our large and unique cardiac data repository of over 200 million hours of labeled ECG data and our leadership in ECG analysis to enable this cutting edge deep learning research,” said Kevin King, Chief Executive Officer of iRhythm Technologies, Inc. “This project is a continuation of iRhythm’s exploration of state-of-the-art machine learning techniques as key to the future of health care delivery.”
Details about the algorithm have been published in a paper co-authored by Stanford and iRhythm researchers on ArXiv.org, “Cardiologist-Level Arrhythmia Detection with Convolutional Neural Networks”.