The Leiden University Medical Centre (LUMC) is investigating whether the POLARIX AI model is safe and reliable enough to be used as a tool for pre-selection in the diagnosis of uterine cancer. The aim is to determine which patients actually require further DNA testing, thereby reducing the pressure on diagnostic capacity. The KWF Cancer Relief Foundation is providing a grant of over 1 million euros for the research.
The characteristics of the tumour play a crucial role in the treatment of uterine cancer. A key biomarker in this context is a mutation in the so-called POLE gene. Patients in whom this mutation is detected generally have an excellent prognosis and usually do not need to undergo additional treatment, such as radiotherapy or chemotherapy, following surgery. This can be determined through DNA testing, enabling treatment to be tailored as closely as possible to individual patients.
Not carried out as standard for every patient
At present, only a DNA test can determine with certainty whether a tumour contains a POLE mutation. Whilst this test is reliable, it requires capacity and specialist expertise that are not available everywhere. As a result, the test is not carried out on every patient, either in the Netherlands or internationally. Consequently, there is a risk that relevant treatment information may not be available for some patients.
The POLARIX AI model, developed by the LUMC in collaboration with the AIRMEC AI research team, is set to provide a solution. The model analyses digital images of tissue and predicts which patients have an increased risk of a POLE mutation. A DNA test is then carried out only on patients with an increased risk of a POLE mutation.
Using available capacity more efficiently
The approach helps to significantly reduce the number of DNA tests required, without compromising diagnostic quality. This not only saves money and time, but also ensures that available capacity can be used for patients for whom it actually offers added value.
The next phase of the research will focus on determining whether POLARIX also functions safely and reliably in everyday practice. To investigate this, the model will be tested in several hospitals in the Netherlands and abroad. The focus will not only be on the accuracy of the predictions made by the AI model, but also on its performance across diverse patient groups and healthcare settings.
Could render three-quarters of DNA tests unnecessary
The LUMC and AIRMEC estimate that, if the model proves successful, around three-quarters of current DNA tests will no longer be necessary. The model could therefore have a significant impact, not only in the Netherlands but also beyond. For example, DNA diagnostics are only available to a limited extent in some countries. The model could help to allocate available capacity more effectively.
The researchers intend to use the grant from the KWF to further develop POLARIX and make it suitable for use in everyday clinical practice. They also aim to use POLARIX to enable more personalised treatments for women with uterine cancer.
