Avoid unnecessary surgery for metastatic testicular cancer after chemotherapy. That is the goal of a new study by UMC Utrecht. Using an AI model, the researchers would like to predict when additional surgery will be really necessary to remove remaining tumor tissue.
Testicular cancer is rare. In 2023, about 900 (young) men in the Netherlands were diagnosed with it. Treatment of testicular cancer (almost) always begins with surgery to remove the testicle. If the disease is detected in time, patients will have a high chance of being cured.
With about one-third of patients with testicular cancer, metastases are found at diagnosis (or shortly thereafter). These metastases are often located in the lymph nodes in the back of the abdomen (around the aorta) but may also occur in the lungs. Possible treatments for the metastases include radiation, surgery or, in most cases, chemotherapy. After chemotherapy has been completed, a CT scan of the body will be taken to assess whether the chemotherapy has indeed worked.
Tumor or scar tissue?
Britt Suelmann, internist-oncologist at UMC Utrecht: “Most favorably, the metastases in the abdomen have disappeared or become significantly smaller. Patients will then be checked according to a fixed follow-up schedule. Unfortunately, the metastases might also have grown, or new abnormalities may have emerged. Then, we know that we will have to give new treatments, often in the form of chemotherapy.”
However, a third scenario is also possible: the metastases have diminished, but not enough to tell whether the chemotherapy has indeed worked well. In this case, based on the scans, it is impossible to tell whether the ‘residual tissue’ consists of still active tumor tissue or, for instance, scar tissue or dead cancer cells. To prevent the growth of vital tumor cells, surgery will then be performed to remove the residual lesion.
“This is a risky surgical procedure that we hope to avoid in the near future.”
Peter-Paul Willemse, urologist at UMC Utrecht: “This is a complex procedure while people are still sick from chemotherapy. They have a lower resistance, limited condition and their wound healing has deteriorated. As a result, recovery after this operation will take a long time. Vascular structures and organs may also be damaged during the operation.”
Short- and long-term complications might thus occur with this major surgery. And in a number of cases, this turns out to have been unnecessary when only scar tissue or dead cancer cells are found during pathological examination after surgery.
“People are still sick from chemotherapy, so recovery after surgery will take a long time.”
Predicting with AI
In order to prevent such unnecessary surgery in patients with metastatic testicular cancer in the future, the researchers are going to build an AI model. This should make it possible to predict, with the help of machine learning, whether the residual tissue actually consists of active tumor cells. For this, Britt and Peter-Paul are collaborating with radiologist Bart de Keizer and associate professor Kenneth Gilhuijs, amongst others.
“In recent years, we have collected patient data to start building a model: from our own center of expertise, our partner hospitals within ONCOMID (Oncologie Midden-Nederland – ed.) and the Antoni van Leeuwenhoek Hospital,” Britt says. “From January 2025, we will start working with the data of 500 patients who have been diagnosed with metastatic testicular cancer. Our model will include not only radiological images, but also clinical (such as blood values – ed.) and pathological data.”
In January 2025, Britt and her colleagues will start this research project, which is called AI-PREDICT GCT. Once the AI model has been built, the model will be ‘validated’: newly diagnosed patients will be examined to see whether the model does indeed has given a good prediction.
Britt: “During this validation phase, we still have to operate on patients to check whether the model does indeed work properly. The project will take a total of four years but after three years, we expect to have the first conclusions, and we hope to be able to prevent this radical surgery in half of the cases from then on.”
Support research?
For the AI-PREDICT GCT research, Britt and her colleagues have been granted nearly half a million euros from the Hanarth Fund through the Vrienden UMC Utrecht & Wilhelmina Kinderziekenhuis, the hospital’s charity.
Read more on how you can support research into new treatments at UMC Utrecht.
