Researchers will use robotics and AI in cancer treatment
Researchers from the University of Southern Denmark (SDU), in collaboration with European partners, are working on using robotics and AI to diagnose and treat various cancers, with a solution expected ready in just a few years.
Cancers of the upper aero-digestive tract (UADT) are some of the most common types of cancer worldwide. And in many cases, the disease is detected late which complicates the treatment options.
Now, a group of European researchers are planning to do something about this.
Five universities across Europe have partnered with hospital clinics in Spain, Belgium, and Italy as well as several private companies to use AI and robotics to improve methods for diagnosing and treating UADT cancers.
- AI and robotics have reached a level of maturity that allows us to use them in the healthcare sector and take the field to new heights. By diagnosing and treating patients faster and better, we can address capacity issues, reduce waiting lists, and ultimately save lives, says Thiusius Rajeeth Savarimuthu, a professor of robotics at SDU and one of the researchers involved.
The project, named AirCare, is funded by the EU with a total budget of €7,9 million. It will run for the next four years, after which the project partners expect to have a whole range of products that have been tested and are ready for CE certification.
Real-time diagnosis
The researchers at SDU are specifically responsible for developing a so-called SmartProbe, which can be used to diagnose cancer without the need for tissue samples, such as biopsies.
The challenge with biopsies today is that not only are they invasive for the patient, as tissue needs to be removed. They also create delays because the samples require analysis in a laboratory.
The SDU researchers' sensor overcomes this issue. The SmartProbe will be able to pass through the working channel of a laryngoscope to measure the electrical characteristics of suspicious regions from the surface and, using an AI algorithm, provide a diagnosis straight away.
- Since it's done in real-time, many more tissue samples can be tested with immediate results. This also means – unlike biopsies, which analyze only one specific point – that we can map the margin of the cancerous tissue, says Zhuoqi Cheng, assistant professor of robotics at SDU and another project researcher.
- The margin of the cancerous tissue can help determine how advanced the cancer is and will be valuable information for subsequent surgery.
In time, the sensor may also be used as an instrument during the removal of cancerous tissue, ensuring that all affected areas are removed. Today, for patients with cancers in the upper respiratory and digestive tracts, recurrences are not uncommon.
Additionally, the researchers plan to use robotic control mechanisms to enable one person to operate the SmartProbe. Currently, biopsies often require multiple healthcare professionals.
Combining with an endoscope
The new sensor from SDU is intended to be combined with an endoscope being developed by Italian researchers. The endoscope will also use AI to analyze tissue, but instead of electrical impulses, it will rely on visual images of the tissue.
Nowadays, a surgeon and a nurse often inspect the upper respiratory and digestive tracts for cancer using an endoscope. If something looks suspicious, they can send a small tool down the working channel of the scope to take a tissue sample.
With the new solution, AI will assist the surgeon in the initial visual inspection and provide tissue assessments based on training data. The staff can then choose to send the SmartProbe through the endoscope to confirm the diagnosis.
Other groups in the research project will develop new methods for the cancer surgery itself, including the use of lasers and robotic manipulation of soft tissue.
Ethics and patient safety
Introducing robots and AI into the treatment of serious diseases like cancer naturally raises many ethical considerations. Therefore, patient safety is a top priority for the project, say the researchers.
- It’s quite unique that a single research project moves from technology development over clinical tests and almost all the way to CE certification of the product. This means that in four years, we'll have a functioning technology, and all the hard work of documentation will be done. But it also means that we have to consider ethics, law, and approvals from the very start, says Thiusius Rajeeth Savarimuthu.
Therefore, an ethics expert and a legal advisor are partners in the project, and many ethical guidelines have already been formulated.
For example, the researchers will embed safety mechanisms into the AI so the algorithms can only be used for their intended purpose and specific patient groups.
The researchers also emphasize the importance of explainable AI, says Zhuoqi Cheng:
- It’s important that there is transparency in the algorithms so you can always see how they arrived at their results. At the same time, they should also assess how confident they are in their result. So, if you're inspecting an area with the endoscope, it might say there is an 80% chance of cancer here.
In other words, doctors will still make the final diagnosis, but the new technology will help them do so faster and more accurately.
About the project
AirCare is a research project supported by the EU's Horizon Europe program. The project has a total budget of 7.9 million euros, or just under 59 million Danish kroner.
The goal is to use robotics and artificial intelligence to improve methods for diagnosing and treating cancers of the upper respiratory and digestive tracts, such as oral cancer and esophageal cancer, which are among the most common types of cancer worldwide. Oral cancer, for instance, is the sixth most common cancer.
The project partners are: Italian Institute of Technology (lead), University of Southern Denmark (SDU), University of Leuven, Sant'Anna School of Advanced Studies, University of Genoa, IDIBAPS, EL.EN. SPA, BARCO NV, Austrian Center for Medical Innovation and Technology, TIMELEX, and Inventya Ventures Limited.