News center

Jun 18, 2018

Where artificial intelligence in healthcare can make a dramatic difference for patients with cancer

Estimated reading time: 4-6 minutes

Adaptive Intelligence shortens the time it takes for image-to-plan radiation treatment

Potential clinical uses of artificial intelligence (AI) receive much attention but actual real-world applications are far less common. With clinical AI, we often get stymied by issues of regulatory approval or insurance reimbursement. However, if we shift our thinking, and focus on a particular clinical problem, we can readily see how clinical AI can go from research to implementation and make a dramatic difference in patient care today. This is how artificial intelligence becomes Adaptive Intelligence because it is applied in the context of patient specifications and specific disease conditions.

Radiation treatment planning is one clinical problem that Adaptive Intelligence can help us solve. When a patient has been diagnosed with cancer, and radiation treatment is indicated, the first step for treatment is to plan a radiation therapy session. This is not an easy task. The physician starts with a set of images, marks where the tumor boundaries are, and then must plan what the dosage would be from various directions. Then, the plan must be approved by a physicist, and then by the radiation oncologist, to make sure the dose is optimized. If the plan is not approved, as commonly occurs, then they have to re-plan and re-start this process over and over again until they get it right.
How AI in healthcare can make a difference for cancer patients

Typically, the time to “get it right” from the initial image to first therapeutic dose is about 12 days. Why? Every single tumor is a different shape, size and in a different location. Every patient’s age and severity is different and their health status is different. All of those variables impact the plan. Considering the many variables involved, and the quality assurance involved, twelve days may seem reasonable, it may even seem inconsequential, but not for patients with cancer. If you are a patient that has an aggressive form of lung cancer and you’re waiting for 12 days before therapy even begins, that’s an incredibly long period of time to wait. 

 

This is where Adaptive Intelligence can make a critical difference and enable us to go from 12 days of waiting to an image-to-plan treatment that can start in 1-2 days. With adaptive intelligence, we can eliminate going through multiple iterations manually and do it much, much faster. This is a huge advancement in the treatment of cancer.

 

Radiation therapy is an example of precision medicine - highly individualized. There are numerous parameters and constraints that you must be able to satisfy while at the same time adhere to the contour of the tumor from all the different perspectives of the radiation beam. In a radiation therapy system, the radiation doesn’t just come in from one direction – the entire contraption spins around the patient repeatedly, and as its spinning around the patient, it changes the contour of the radiation beam so that it doesn’t hit vital organs. Tumors are not perfectly spherical, there’s often some odd shape to it, so as you change your angle, you have to adapt the radiation beam to the shape of the tumor from that particular perspective. These shifting variables make radiation therapy not only highly individualized but very complicated as well.

Radiation therapy is an example of precision medicine - highly individualized. There are numerous parameters and constraints that you must be able to satisfy while at the same time adhere to the contour of the tumor from all the different perspectives of the radiation beam. In a radiation therapy system, the radiation doesn’t just come in from one direction – the entire contraption spins around the patient repeatedly, and as its spinning around the patient, it changes the contour of the radiation beam so that it doesn’t hit vital organs. Tumors are not perfectly spherical, there’s often some odd shape to it, so as you change your angle, you have to adapt the radiation beam to the shape of the tumor from that particular perspective. These shifting variables make radiation therapy not only highly individualized but very complicated as well.
How AI in healthcare can make a difference for cancer patients

AI is used to optimize that very, very complex equation and it is not trivial. Frequently, it’s not possible to satisfy all those constraints, so then you must calculate what is the minimum violation you can incur while delivering the proper dose to the patient. Adaptive Intelligence combines artificial intelligence and other methods with knowledge of the clinical, operational, or personal context in which they are used, so it is able to do a better job of satisfying all those constraints more quickly and with fewer iterations. If you have a system in which every single time it figures out a plan and that plan is optimal, then you don’t need to reapprove a plan and you don’t need lots of iterations on it. This is a radical improvement in patient care.  

 

There’s also a real benefit for hospitals overall as well because it’s meeting the Quadruple Aim. The patient outcome is improved because you’re treating patients faster. Patient satisfaction is certainly higher because it’s less time to treatment for them. Costs are lowered because hospitals are spending less time per patient because they’re spending fewer iterations on plan development. Likely, staff burn-out will improve because the shortening of image-to-plan from twelve to two days decreases the back-log or waitlist. 

 

To the best of my knowledge, Philips is unique in its ability to solve this very challenging problem and make a significant difference in patient care. To me, that’s very exciting because it’s no longer just AI hype, it’s reality – and this provides more hope to patients in the treatment of cancer.

About Innovation Matters

Innovation Matters delivers news, opinions and features about healthcare, and is focused on the professionals who work within the industry, as well as Philips as a cutting-edge health technology organization. From interviews with industry giants to how-to guides and features powered by Philips data, our goal is to deliver interesting, educational and entertaining content to empower and inspire all those who work in healthcare or related industries.

 

Read MoreRead Less

Share on social media

Topics

Author

Homer Pien

Homer Pien

Ph.D., Chief Scientific Officer, Diagnosis and Treatment, Philips
Follow me on