DNA changes in a tumor determine the behavior of the tumor and the course of the disease
Using microscopy images of a tumor to see which DNA changes are present in the tumor, and thus determine which type of uterine cancer the patient has. That is impossible for a human eye. That is why pathologists at the Leiden University Medical Center (LUMC) called in the help of artificial intelligence (AI). And with success. Read here how the results, published in The Lancet Digital Health, can improve the diagnosis and treatment of uterine cancer.
DNA changes in a tumor determine the behavior of the tumor and the course of the disease, as much has become clear in recent years, partly due to work at the LUMC. There are four types of uterine cancer, each with a different disease course. It is important for patient and doctor to know what type of tumor it is, but this currently requires costly, additional DNA tests. Pathologists wondered whether these ‘molecular’ types of uterine cancer could also be distinguished under the microscope.
Artificial intelligence predicts DNA changes
To do this, they used microscopy images of uterine cancer from more than 2000 women who participated in the clinical PORTEC studies, coordinated from the LUMC by Professor Carien Creutzberg. All these patients underwent surgery and have given permission to use the residual material for scientific research. With this unique collection of images, researchers from the Pathology department created an AI model that predicts DNA abnormalities, and thus various types of uterine cancer. It is important here that the model shows the researchers where the visual information for the predictions is hidden in the tissue. So it is not a black box, like other AI models.
Improving uterine cancer diagnosis
“The application of AI to microscopy images is still in its infancy. Through this study, we wanted to learn more about the relationship between the appearance of the tumor and the underlying DNA changes. With this work, we have learned which areas in the tumors contain the most important visual information for diagnosis, and so what pathologists should focus on,” says Sarah Fremond, PhD student in Pathology.
Uterine cancer is the most common tumor of the female genital tract. A lot of research is being done at the LUMC into this type of tumour. According to the researchers, this study contributes to the further improvement of the diagnosis and treatment of uterine cancer. “As a next step, our team will now develop an AI model that can predict the risk of metastases,” adds Tjalling Bosse, pathologist.
Read the full article in The Lancet Digital Health.
This work was funded by the Hanarth Fund and was carried out in close collaboration with the University of Zurich.
Consult the source and/or provider for more information about this message. News may change, include errors or inaccuracies. Also read our disclaimer and please report messages, comments and/or images that go against our terms and conditions.
Click on the tags below for relevant posts, if any…