UZA/UAntwerp researchers can predict who will develop disease years in advance
Liver cancer is often diagnosed too late, as a cure is only possible in the early stages of the disease. Researchers at UZA and UAntwerp have now discovered a set of genes that predict who is likely to develop liver cancer. "This allows us to intervene earlier on, which can save lives."
Liver cancer is the thirteenth most common form of cancer. In terms of mortality, it even makes the top three. "It's rare for a patient to make a full recovery," says Prof. Thomas Vanwolleghem (UZA/UAntwerp). "The tumour has to be discovered at a very early stage, only then a transplant or removal of part of the liver can provide a cure. But all too often, this form of cancer is only detected at an advanced stage. Chemotherapy may then add several months to the patient's life, but not years."
More than half of all people with liver cancer are also infected with the hepatitis B or C virus: the risk of developing liver cancer is up to a hundred times higher in these patients. Vanwolleghem: "Hepatitis B and C patients have a viral infection in their livers. They don't feel anything, and often there are no symptoms. But this infection is harmful to the liver and often leads to fibrosis, which in turn greatly increases the risk of liver cancer.
People with a viral hepatitis infection need regular check-ups, up to twice a year. Imaging and blood tests are then used to determine whether the patient is developing liver cancer. These tests are expensive and seldom detect tumours in very early stages.
But that is about to change, thanks to intensive research by scientists at UAntwerp and UZA, backed by the Belgian Foundation Against Cancer and in collaboration with physicians from Antwerp, Limburg and Rotterdam.
"We analysed liver biopsies from patients with hepatitis B and C infections," explains researcher Stijn Van Hees (UAntwerp/UZA). "At the time of the biopsies, those people were cancer-free. We kept track of which patients went on to develop liver cancer. We then studied differences in gene expression patterns in the liver biopsies of patients that went on to develop liver cancer at a median of 8.3 years later and those that did not."
The researchers used various artificial intelligence techniques on the data and found more than 500 genes that were expressed differently in people who would later develop liver cancer. Van Hees: "These included oncogenes, which we know play a role in the development of cancer. We also found signs that various biological cancer-related processes were already ongoing in patients who would later develop liver cancer."
Shutting down genes
Together with Prof. Kris Laukens and Dr. Bart Cuypers, both experts in biomedical data mining at UAntwerp, a AI computer model was developed. This model aggregates the degree of liver damage, the biopsy data and the genetic information of the patient. "This enables us to recognise patterns that will lead to the development of liver cancer more than eight years down the line," says Vanwolleghem.
"These results are very promising. The aim is to use these new insights to develop a clinical test that will identify patients at risk of developing liver cancer, so that we can monitor them more closely and act earlier on. Our research not only opens up opportunities for predicting liver cancer, but also reveals genes that are important in the disease’s development. Further research might lead to technologies that can 'shut down' these genes and prevent the progression of the disease," says Van Hees.