Augmenting health through AI

Among all the noise about artificial intelligence (AI), deep learning and machine learning, I’ve been focusing on news about AI in medical imaging. AI has truly captured my attention (and my imagination) and I think it is going to be a game changer in the pharma industry.

The first examples of how AI could help doctors are already being commercialized and will change the way they work. This change will result in improved workflow efficiency, while improving care.

Overload of patient information

We’re collecting patient data at a rate faster than ever before and it’s becoming more and more impractical for a doctor to sift through these data when diagnosing and treating a patient. These data could take many different forms including imaging, procedure/lab reports, admission and discharge information – the list goes on. The amount of patient data received is going to increase rapidly as digital tech embeds itself in healthcare. For example, patients will eventually be able to upload their own data for use by their doctors. These  amounts of data can be overwhelming. So what is the solution?

AI will play a key role in the next few years

With so much data flying around, it’s easy to miss things and for information to slip through the cracks. AI will be able to review findings and report the most important information in a short amount of time. Collecting full patient information would simply take too long if done solely by a healthcare professional and while AI as a replacement for doctors is a long way into the future (and may never happen) AI will give doctors more time to focus on patient care, help them review the relevant data they need to care for their patients presented in a simple, easy to understand way. 

What to watch out for

Many vendors have already started incorporating AI into their software systems – including IBM/Merge, Philips, Agfa and Siemens. This data science revolution we’re witnessing was started by IBM Watson, which has been at the forefront of medical AI for a few years now. There have been a few different ways that AI in imaging has been shown to work. I’ve seen examples of predictive analytics software, or software that uses AI to sift through big data to offer immediate support with a clinical decision. Oncology imaging is particularly interesting – with a few clicks on the tumor in the image, the AI can automatically quantify the tumor and then present a side-by-side comparison of previous tumor assessments, showing disease progression.

So what’s next?

The more experience the AI gets through sifting through patient data, the better it will be. It will continue to develop its clinical intelligence with millions of patient records, so access to huge amounts of patient data and images is needed. To start, AI will likely be introduced slowly and incrementally. It offers a huge opportunity to enhance and augment radiology reading, not to replace radiologists. And with time, I think AI software is ultimately going to act as a very experienced clinical assistant, helping the doctor and making workflows much more efficient.

This is a fast moving field and I believe AI could ultimately speed drug discovery and lead to better, faster medicines – not only improving patient care, but also making lives better. But let’s not get ahead of ourselves… AI is not perfect and it may never be, but it’s an ongoing process using software programmers refining algorithms until the AI gets it right, at least in the majority of cases. Are you as excited as I am for this huge shift in the imaging field?

Kris