The bonding link for better care

What is the best way to treat a stroke? It’s no secret: you send for an emergency doctor immediately. Step two: the patient is sent for a CT or MRI scan, and if he or she is eligible for catheter treatment, it starts immediately after the diagnosis has been made. Then, everything possible should be done to prevent a further stroke. If we know this, why is it so difficult to implement? Could new technologies, such as artificial intelligence (AI), help us out?

I am positive they can, but first things first. In medicine, AI is sometimes talked about as if it were a new medical device. It isn't. At Siemens Healthineers, we understand AI as an enabling technology that allows medical care to be redesigned and improved throughout the care continuum, from prevention to aftercare. For us, algorithms are a digital link, the thread that connects a wide spectrum of ever more-precise procedures, accelerates cognitive processes and helps optimize treatment through more personalized and precise care. This is the key for achieving outcomes that matter.

Expanding precision medicine

When we see AI in this way, it stops being about some elusive “neural network”, or a virtual doctor alternative. Instead, it’s about actual improvements to care processes at different levels of complexity – culminating in individual risk prediction and a personalized simulation of therapeutic measures using multidimensional datasets. In the future, these datasets could provide a virtual representation of the patient – what we refer to as the patient’s “digital twin”.*

AI projects need to be pursued in partnership – not in competition – with medical experts and institutions. Algorithms only become established when they enhance the quality of care, reduce the workload for medical staff or improve access to care. Take the example of imaging: In both CT and MRI diagnostics we are equipping the latest generation of devices with features and algorithms that support adapting the technology and scan procedure to the needs of individual patients. This improves the quality of the images. Processes can be replicated, and ultimately standardized, helping to reduce unwarranted variations.

Faster processes, fewer errors

The next step is to use algorithms in image evaluation. This is not a thing of the future, it’s happening now. The Siemens Healthineers AI-Rad Companion is a platform for radiological image analysis that is being constantly developed. Likewise, in laboratory diagnostics, self-learning algorithms enable highly automated sample handling on our Atellica Solution platform, speeding up processes and reducing errors. As of today, we have more than 45 products on the market that use AI algorithms. In addition, we hold over 500 patents in the area of machine learning.

But let’s return to stroke patient care: What role can AI-optimized processes play in their treatment? If AI algorithms are intelligently integrated into procedures, strokes can be detected faster and treated with higher precision. For example, we are working with the Medical University of South Carolina (MUSC) using these innovative technologies. Our aim is to significantly reduce the time between a patient’s arrival at the hospital and the start of catheter treatment, from 90 minutes at present to 20 minutes in the future.

AI as a link between technologies

For the patient, each second counts and can represent the difference between recovery and independence, or a lifelong disability. It also frees up time for new therapeutic procedures. I am thinking of the catheter therapies I mentioned earlier, which can be used to remove blood clots in the brain. In the future, catheters will not be inserted by a doctor – instead robotic microcatheters will be remotely controlled from a screen using a joystick. Through the acquisition of Corindus, a vascular robotics company, we want to strengthen our advanced therapy business to become a global leader in this field.

Robotic therapeutic systems, which can also be used on the heart or other organs, will only work if imaging, sensor systems and therapeutic equipment are closely interlinked via digital platforms. Once again, AI algorithms are an indispensable link between innovative technologies that, when combined, will result in higher-precision, more patient-friendly medicine, and ultimately lead to high-value care.

------

This article has been slightly modified from my contribution published by Handelsblatt Health Journal, September 24, 2019.

* This is a future concept (WIP). The concept and features are not for sale and future availability cannot be guaranteed.