26 April 2024

AI makes it possible to quickly and accurately mark areas to be eliminated with radiation when treating cancerous tumors. But there is a lack of medical data to train AI models with. Attempts are therefore underway to train the models with so-called synthetic medical images.

The image consists of two parts: four sections views of a brain on the one side and fous sections of another brain on the other side
One of these two images is real and one is synthetic. Which one do you think is the synthetic one?*

To use radiation therapy to treat cancerous tumors, it is important to know exactly where the tumor is located. Oncologists, i.e. medical doctors with special training in diagnosing and treating cancer, also need to take organs at risk into account. An organ at risk is an important organ close to the tumor. By pinpointing the exact area, the radiation can eliminate the tumor tissue without damaging important adjacent tissue. To locate tumors and organs at risk, an MR scanner is used to collect different types of images. Before radiation therapy, a radiation oncologist must often manually draw the tumor and organs at risk in these images, which can take several hours for a single patient. By instead using artificial intelligence, AI, all areas can be drawn automatically and thus save many hours in healthcare.

Important with large training sets

AI models for radiotherapy are trained on a large number of medical images, where the radiation oncologist's drawings are used as the correct interpretation. It is important that the training set represents as large a percentage of the population as possible, so that the models can be used on all patients. However, there is a lack of medical images and medical information to train AI models with, as there are several legal aspects to consider. Among other things, ethical approval is required when such sensitive information is to be used for research.

Generating images to create large training sets

In an article published in the journal Scientific Data, LiU researchers describe how they evaluate the approach of instead generating images to train AI models with. Such images are called synthetic images. 

“Synthetic images may be easier to share with other researchers. Since synthetic images do not belong to a specific person, GDPR is not applicable,” says senior associate professor Anders Eklund.

Portraits of three researchers
Anders Eklund, senior associate professor at the Department of Biomedical Engineering (IMT) and the Department of Computer and Information Science (IDA), is together with postdoc Muhammad Usman Akbar (IMT) investigating the possibility of generating so-called synthetic images to train AI models with. Neuroradiologist Ida Blystad has visually assessed the synthetic images. She is also Adjunct Associate Professor at the Department of Health, Medicine and Caring Sciences (HMV).

The researchers created the synthetic images using an open collection of data called BraTS. BraTS includes MR images and marked tumors for brain tumor patients from a variety of hospitals. Since the images are based on open data, the research team has been able to share the synthetic images on the AIDA (Analytic Imaging Diagnostics Arena) data hub, so that other researchers can use these images.

Berzelius helped them to save years of waiting

The supercomputer Berzelius.
The supercomputer Berzelius, one of the world's fastest computers.Thor Balkhed
Anders Eklund's research group has great help from the Berzelius supercomputer at the National Supercomputer Center (NSC) at Linköping University. To create and evaluate synthetic MR images of brain tumors, the researchers have used Berzelius effectively. If they instead had used an ordinary powerful computer, the same job would have taken more than five years.

To evaluate the synthetic images, the research group compares the results from AI models trained with synthetic images with the results from AI models trained with real MR images of brain tumors.

“A challenge that remains is when we use too few real medical images to train the generative models. In these cases, we have seen that the synthetic images can be too similar to the real images, and we don’t want to share the real images”, says Anders Eklund.

Positive results

The results show that synthetic medical images can be an alternative to be able to share data more easily and thus develop AI models for medical applications. The exact legal conditions for sharing synthetic medical images remain to be investigated.


* The right answer to the image question can be found in Figure 5 in the article at nature.com

The dataset with the synthetic images shared by the research team

More examples of synthetic images.

About the project

The research project ASSIST

Anders Eklund, senior associate professor at the Department of Biomedical Engineering (IMT) and the Department of Computer and Information Science (IDA), together with senior associate professor Evren Özarslan at IMT, leads part of the European collaborative project ASSIST. Together with postdoc Muhammad Usman Akbar (IMT), Anders Eklund is investigating the possibility of generating so-called synthetic images to train AI models with.

The evaluation of the synthetic images has been done within the research project ASSIST and in collaboration with the company Eigenvision and Region Östergötland. The  collaboration is facilitated by CMIV, Center for Medical Image Science and Visualization, where medical doctors and engineers, among others, conduct interdisciplinary research. Neuroradiologist Ida Blystad has visually assessed the synthetic images.


Research activity

More about artificial intelligence at LiU

Latest news from LiU

Sheet of glass with droplet.

Next-generation sustainable electronics are doped with air

Researchers at LiU have developed a new method where organic semiconductors can become more conductive with the help of air as a dopant. The study is a significant step towards future sustainable organic semiconductors.

physicians in a clinica setting.

Healthcare interpreters important for heart attack aftercare

After a heart attack, foreign-born people are less likely to attend a relapse-preventing Heart School than native-born patients. But with access to a professional interpreter, participation increases, according to a new study.

Battery om fingertip.

Eco-friendly and affordable battery for low-income countries

A battery made from zinc and lignin that can be used over 8000 times. This has been developed by researchers at LiU with a vision to provide a cheap and sustainable battery solution for countries where access to electricity is limited.