Heart Failure Detection in a Single Heartbeat using AI

Artificial Intelligence

has multiple applications across various fields. Whether you want intelligent robots, self-driving cars, smart homes, etc. Artificial Intelligence can handle it all!!! And among these applications, AI is also very helpful in the medical field for diagnosing various problems. Specifically, we will understand how

Artificial Intelligence can detect Heart Failure in a single heartbeat

in this article. But to understand that, let’s see what is Artificial Intelligence first.

What is Artificial Intelligence?

Artificial Intelligence (A.I.) is a multidisciplinary field with the main aim of automating activities that presently require human intelligence. The primary purpose of artificial intelligence is to improve computer behavior so that it can be called intelligent. It is a field of study based on the premise that rational thought can be regarded as a form of computation one that can be formalized and ultimately mechanized. The major problem areas addressed in A.I. can be summarized as

Perception, Manipulation, Reasoning, Communication,

and

Learning.

The success of AI in the Medical Field

Artificial intelligence has wholly revolutionized the diagnosis of cancer. The supercomputer

Watson of IBM

can see deviations in the health of the individual. Around 30% of cases, Watson puts patients with additional diagnosis missed by experts. Even more impressive results are achieved by AI at the Houston Methodist Research Institute, Texas. Artificial intelligence explores the millions of mammograms and gives on solutions with 99% accuracy. Microsoft has already demonstrated that AI for the first time, caught up with the man in the efficiency of automatic speech recognition. To get this result, the company used the so-called high-precision and recurrent neural networks. To prepare for the test, 2000 hours of recorded data were required.

What is Congestive Heart Failure?

Nearly 10 percent of adults above 65 years of age suffer from some congestive heart failure (CHF). There are many causes for CHF, but the fundamental chronic condition generally emerges from the heart being incapable to pump blood effectively through the body. X-rays, blood tests, and ultrasounds all offer useful ways to diagnose CHF. Still, one of the more common methods involves using

electrocardiogram (ECG) signals

to determine heart rate variability over several minutes or even multiple measurements over days. An impressive new approach has now been demonstrated, using a

Convolutional Neural Network (CNN)

that can identify CHF nearly instantly by checking ECG data from just one heartbeat.

How Congestive Heart Failure can be detected using AI?

Applying Artificial Intelligence to the electrocardiogram (ECG) enables

early detection of left ventricular dysfunction

and can also identify individuals at increased risk for its development in the future. The research, which is published in

Nature Medicine

, found that the accuracy of the AI/ECG compares favorably to other common screening tests such as prostate-specific antigen for prostate cancer and mammography for breast cancer. Asymptomatic left ventricular dysfunction (ALVD) is characterized by the presence of a weak heart pump with a big risk of heart failure. It is present in

3 – 6 percent

of the general population and is associated closely with reduced quality of life and longevity. However, it is treatable when found. Currently, there is no inexpensive, noninvasive, painless screening tool for ALVD available for diagnostic use. To address this, Paul Friedman and colleagues tested whether ALVD could be reliably detected in the ECG by a properly trained neural network. The team used paired 12-lead ECG and echocardiogram data, including the left ventricular ejection fraction, from 44,959 patients at the Mayo Clinic, and trained a convolutional neural network to identify patients with ventricular dysfunction, defined as ejection fraction less than 35 percent, using the ECG data. When tested on a set of 52,870 patients, the network model yielded values for the area under the curve, sensitivity, specificity, and accuracy of 0.93, 86.3 percent, 85.7 percent, and 85.7 percent, respectively. In patients not having ventricular dysfunction, those with a positive AI screen were at four folds the risk of incurring future ventricular dysfunction as compared to those with a negative filter.

“This suggests the network detected early, subclinical, metabolic or structural abnormalities that manifest in the ECG”

says Friedman.

100% Accuracy in Detection!

“We trained & tested the CNN model on publicly available large ECG datasets featuring subjects with CHF and healthy as well, non-arrhythmic hearts. Our model delivered a hundred percent accuracy by checking only one heartbeat; we can detect whether a person has heart failure or not. Our model is also the first to be able to identify the ECG’ s morphological features associated with the severity of the condition accurately.” says Sebastian Massaro, from the University of Surrey As Massaro suggests, the team’s system is reporting an unbelievable hundred percent accuracy rate, but the research is not without any limitations. The data used in the study consisted of ECG readings from severe CHF patients or healthy subjects only. The researchers do note results might not be as correct for patients with milder CHF. Therefore much work certainly needs to be done to verify a broader spectrum of CHF diagnoses before the technology is used out in clinical practice.