Microscopic changes in the liver can be used to predict if, and where, pancreatic cancer may spread in the body. The discovery has the potential to provide new ways of predicting the course of the disease and preventing pancreatic cancer from spreading to other organs. The study, led by Weill Cornell Medicine, USA, was conducted in collaboration with researchers at Linköping University, Sweden, and published in Nature Medicine.

Cancer cells have several properties that distinguish them from healthy cells in the same tissue. One such property that many tumours develop is the ability to give rise to secondary tumours in other organs. This process is known as metastasis. But there are many questions about how it occurs and what causes some tumours of the same type to spread, whereas others do not.

The current study, published in the journal Nature Medicine, focuses on pancreatic cancer metastasis after initial surgical treatment. Pancreatic cancer spreads mainly to the liver. There is a great need to be able to carry out risk assessment at an early stage to find patients who are most at risk of developing liver metastases.

“Once the tumour has spread to the liver, most patients unfortunately end up dying from the metastatic disease. One of the goals of this study was to try to capture early steps in the initiation of liver metastasis,” says Dr. Constantinos Zambirinis, docent at Linköping University and surgeon at Linköping University Hospital, Sweden.

More than a hundred years ago, researchers observed that the spread is not random. There are patterns to which organs a particular type of cancer usually metastasises to. According to one theory, a good “fit” between the tumour and the receiving organ is required for the new secondary tumours to survive there. This theory was named the “seed and soil” hypothesis. Much research has focused on properties of the “seed”, i.e. the original tumour, that promote spread. The researchers responsible for the current study, however, wanted to investigate what role the “soil” – that is, the organ to which cancer spreads – can play in the process.

Previous research from the Weill Cornell Medicine research team, led by Dr. David Lyden, demonstrated that the original tumour can send signals to distant organs and induce changes that enable the growth of metastases later on. Or, put in another way: according to this theory, metastasis begins long before cancer cells arrive in other organs. However, these studies have not previously been performed in humans, but mainly in mice. The question was whether it works the same way in people.

In collaboration with Memorial Sloan Kettering Cancer Center, the researchers collected liver tissue samples from 49 patients who underwent surgery for pancreatic cancer and had no signs of metastasis at the time. They also took liver samples from 19 patients who did not have cancer but underwent similar procedures for benign lesions. The researchers then followed what happened to the study participants over the next three years.

It was found that there were notable differences in the livers of patients with pancreatic cancer that metastasised compared to the control group without cancer. These changes in the livers of cancer patients were related to, among other things, the immune system and increased inflammation. However, the group of patients whose pancreatic cancer did not spread had a liver profile relatively similar to the control group’s livers. These findings strengthen the hypothesis that the original tumour sends out signals that lead to changes in other organs, creating a so-called pre-metastatic niche where cancer cells are more likely to grow into secondary tumours.

Among the patients who later developed secondary tumours in the liver, the cancer seemed to spread at different rates. In their analyses, the researchers therefore divided the cancer patients into groups: early spread to the liver within six months of surgery and late spread to the liver after more than six months. There was also a group of patients whose cancer spread to the lungs or other organs without any liver spread, and a group whose cancer did not spread at all. The researchers found small but important differences in metabolism, gene expression, and immune cells in the liver in those who rapidly developed liver metastases compared to livers where the process took longer, or did not happen at all.

Based on the findings from the various analyses, the team developed a machine learning model to predict whether and where a pancreatic cancer would spread based on characteristics of a liver biopsy at the time of diagnosis of the initial cancer.

“This work changes the clinical setting for collecting pre-metastatic liver samples not only in pancreatic cancer but in other cancers, such as colorectal cancer and gastric cancer, with the hope of establishing the timing and likelihood of metastasis in these patients,” says Dr. David Lyden, the Stavros S. Niarchos Professor in Pediatric Cardiology and Professor of Pediatrics and of Cell and Developmental Biology at Weill Cornell Medicine.

Such predictions could help doctors tailor cancer treatment to the individual patient’s risk of relapse.

“The most interesting thing is that there are differences in the liver depending on how the patients are progressing. We see markers for various changes that seem to have an impact on the spread of cancer, and that will be very exciting to follow up in expanded studies,” says Linda Bojmar, assistant professor at Linköping University and researcher at Weill Cornell Medicine in the USA.

The researchers are now planning a larger study on pancreatic and other gastrointestinal cancers to see if their original findings can be confirmed and to further examine the cellular mechanisms.

The research has been funded with support from, among others, the National Cancer Institute, the National Institutes of Health, the Swedish Cancer Society, the Swedish Research Council and the Swedish Society for Medical Research, SSMF.

Article: Multi-parametric atlas of the pre-metastatic liver for prediction of metastatic outcome in early-stage pancreatic cancer, Linda Bojmar, Constantinos P. Zambirinis, Jonathan M. Hernandez, et al. Nature Medicine, published online 28 June 2024, doi: 10.1038/s41591-024-03075-7