This Article is From Jun 28, 2023

How Google's New Algorithm Can Detect Heart Condition Through Retinal Scan

Googles new AI technique could eliminate the need to have a long series of tests and scans usually required to detect cardiovascular risks currently.

How Google's New Algorithm Can Detect Heart Condition Through Retinal Scan

The algorithm has proved to be correct in 70% of the cases. (Representational)

The use of Artificial Intelligence in healthcare is looking at a breakthrough with Google's latest announcement. In an event last week, Google CEO Sundar Pichai announced a new algorithm that could predict and help prevent cardiovascular events using a single retina scan. The technology could reveal the heart's health condition with just a look in the eyes after matching the scans with a matrix for cardiovascular risks. The algorithm has proved to be correct in 70 per cent of the cases where it has been tested so far. Google's new AI technique could eliminate the need to have a long series of tests and scans usually required to detect cardiovascular risks currently.

How Does It Work?

Explaining the concepts of Google's algorithm, Dr Mahipal S Sachdev, Chairman, Centre of Sight, said it is a two-fold process which first involves an imaging technique where a picture of the retina is taken. The picture reveals the blood vessels which are then analyzed by the AI to throw up a diagnosis. "So, if you have a blood vessel that's narrowing or even in cancer cases, leukaemia, or diabetic retinopathy can be thrown up," the doctor said.

Tushar Vashisht, CEO, HealthifyMe, said that Google's new AI technique is a new paradigm shift. "Until now we have been playing with rudimentary AI, expert systems, where you had to train manually the AI model with a specific set of data...But now with transformers this open vocabulary system has been trained on billions of data points and that can be far better than a previous system to classify results. Think of it as if an artificial intelligent brain that has billion of neurons that is trained with pretty much the entire human intelligence so far and therefore its ability to fine grain, detect, diagnose, and understand information is just so much better and so much error-free than it has ever been before."

He added that the generative AI will help get human-like competence but for medical diagnosis, it is always a better idea to have a doctor in the loop. The use of AI would aid doctors in being able to handle more patients and drag down the price point and allow for much deeper access for consumers.

Will It Take Over Diagnostic Completely?

Zianul Carbiwala CTO of Tricog Health called Google's model a partnership of AI and medical. "What we are seeing is that we are getting better diagnostic information, we are getting more inexpensive tests, [can] detect diseases that could have otherwise been detected only more expensive tests." However, he believed that the algorithm is unlikely to disrupt diagnostics completely and eliminate the need for human intelligence.

Dr Sandeep Mishra, Director Institute of Cardiology, NIMS, said all these advances add up and help in refining the diagnostics. "But ultimately these tests are tests; so, they have to be put in place by human intelligence, figure into the algorithm of management which includes diagnostics. They can be used to improve the therapy but stand-alone they can't replace physician or human intelligence."

He emphasised the need to verify the tool and set a context for its applicability.

.