AI a game changer in medical diagnostics space

Artificial Intelligence, if used effectively, will free up a doctor's time for more interactions with patients

There is newfound excitement about using artificial intelligence in healthcare. It will free up a doctor's time for more interactions with the patients.

One area is the use of AI in analysing medical images. This involves deep learning, a sophisticated type of machine learning, where a series of labelled images are fed into the AI system that picks out features in them and learns how to classify similar images. The AI system, as it learns more and more, gets ready to help doctors in the diagnosis of all kinds of diseases from colon cancer to dental work to eye problems.

Certain diseases like influenza, chickenpox, roseola and glandular fever have been predicted with a 90 to 97 per cent accuracy. In a recent study, 42,000 patient CT scans were sourced from Google 's AI system in making the diagnosis of lung cancer. They found that the detection rate improved by five per cent by using an AI system.

FDA approved one AI-based tool used to detect diabetic retinopathy. They claim this tool discovered the disease almost 90 per cent of the time. The solution was built to work with minimal doctor intervention. They primarily used images of the back of the eye to train the AI system.

AI systems are also being used to automate workflow in the medical imaging business. Radiologists review images from X-ray, CT or MRI scan or ultrasound. They check to make a diagnosis of the patient condition. Sometimes due to the quality of images, radiologists can potentially make errors. It can happen due to lack of enough experience or just bad judgement.

AI is being deployed in the process to reduce the error rate. In the ER, when time is critical, a review of the images in a timely fashion can lead to life-saving results.

We do need to take it with the proverbial pinch of salt as they say. The field is inevitably littered with some low-grade research. The challenge that the applications are facing is related to not enough AI training data readily available.

This data is generally located in hospitals, insurance companies and certain government agencies. Putting this data together to be uploaded into a data cloud for the use of AI systems is a gigantic task. Companies such as Google, Amazon and Microsoft are working on building the backbone to help medical tech companies and hospitals to make efficient use of this data in training their AI systems.

A lot of investment is going in to navigate through the regulatory and the privacy advocacy groups. Challenges aside, AI has come to stay in the medical diagnosis arena.

The main benefit of using AI here is to improve the ability of the doctors in assessing the patient condition in time and accurately. Hopefully, this will lead to a healthier population globally and increase the life spans of the individuals.

(The author is Chairman and CEO of Hyderabad-based Brightcom Group)