Karim Lekadir «AI may help detect cancer earlier and provide more effective treatments»
Karim Lekadir is the coordinator of EuCanImage, a EUR 10 million project funded by the European Commission and involving nineteen European institutions and one from the United States. The project aims to develop a cancer imaging platform that will help improve the potential of artificial intelligence and precision medicine in oncology.
What is the project about?
The project has two objectives. It aims to build a platform that facilitates the collection of images and clinical information on cancer. This part, which involves big data, will support the subsequent development of the platform’s artificial intelligence. Once the big data part is completed, we will develop different AI techniques to provide solutions that can help in the detection of cancer, in the choice of treatments, and in the monitoring of cancer patients.
What will you do with the collected data?
Once the project is completed, the platform will remain active so that professionals can search it and consult all the information on the collected data. It will also be possible for people to continue adding data. In this way, the platform will keep growing after the project is over so that more data becomes available and can be shared.
Every day, patients enter hospitals and generate data. Once patients are cured, this data is stored away and remains ‘dormant.’ What we would like is for this data to be ‘awake’ so that it can still be used, as in our case, to develop artificial intelligence. Artificial intelligence improves the more data it gets, so it is important to create a culture where people want to share data and are motivated to do so.
We will collect images and data through six or seven institutions that are involved in the project. In Catalonia, for example, these will be the Hospital Clínic in Barcelona and the Parc Taulí in Sabadell, but we will also have images from institutions in Italy, Sweden, Poland, Lithuania, the United States and the Netherlands. Once we have gathered all this information we will upload it to the platform, and start developing the artificial intelligence.
How does artificial intelligence fit into the project?
We want to achieve three things. First, artificial intelligence has to respond to clinical needs. To this end, we will work with doctors and specialists, who will tell us what kind of artificial intelligence they need. This part is very important, because artificial intelligence developed exclusively with computer scientists and mathematicians runs the risk of not being useful to health professionals.
In a second phase, we will develop the methods of artificial intelligence depending on the requirements specified by the clinical staff. The third phase is also very important, as we will have to define how to validate and evaluate these methods. It is not enough to determine the accuracy of the tool; we must also examine it in greater depth so that, in addition to being accurate, it has other attributes: for instance, it needs to be equitable, to be equally effective for men and women, for different ethnic groups, ages, etc. We know that this is not always the case.
How could this technology improve the lives of cancer patients?
Cancer is a complex disease that often cannot be detected in time. Our hypothesis is that by using the power of machines, mathematics, statistics, and so on, a very large amount of data can be analysed that people alone would be unable to process. Human beings can look at a single image, and sometimes they cannot even see everything that is in there. A machine, in contrast, can analyse thousands of cases and produce an answer.
The same thing happens with treatments. There are currently different treatments for cancer, and doctors choose each treatment based on their experience and a number of studies. If we can have a large database of past patients and see which treatments have been successful and which have not, in the future, when a new patient comes along the machine will be able to compare them with past data and identify common traits to apply the most suitable treatment depending on its chances of success.
How do you coordinate a project with so many partners?
It is very important to have a good team and, especially, a good project manager. We are very fortunate in this respect. It is also important to design the project with qualified and experienced people who are motivated and love what they do. For me, personally, this is part of my life and I do it with passion. Finally, you have to know how to organise yourself and communicate well. We have different work groups that communicate weekly and are organised according to different themes. This is a very interesting and far-reaching project, but a good team ensures that it can be carried out without any problems. Besides, I’m lucky enough to have coordinated other projects, and that helps too. Within the University we have a lot of help from the FBG team, which has a lot of experience with European projects.
In your view, what is the significance of knowledge transfer?
In this particular case, not transferring our project results would be a failure. Knowledge transfer is important for patients, but in order to reach them it is vital to comply with different requirements, such as clinical, legal, and ethical standards, among others. We work on these aspects from the beginning to ensure that project results will reach patients. It would not make sense to develop such a platform without taking these aspects into account. We believe that companies and institutions are important in this process, which is why there are several companies involved in the project. It is also important that the data we collect remains active after the project ends, so that other companies and researchers can assess or develop new technologies and solutions based on this data.
More about Karim Lekadir
The best invention or advancement in history
Agriculture, over 11,000 years ago.
A future advancement or invention that scares you
Artificial intelligence being more present in society than human intelligence.
Something you would like to see in the future
Less inequality, because many of the things we do benefit only part of the world’s population. Research should aim to create more equality, and that’s what I’d like to see.
The FBG is…
Some people sometimes complain about the administration of universities, but for us as a research group the FBG is very professional and makes our lives much easier. We have several ongoing projects, and the FBG makes it very easy to keep everything going. I am very impressed and a fan of the FBG!