Pancreatic carcinoma (pancreatic cancer) is a tumor disease with a very poor prognosis, as it is usually discovered too late to be cured by surgical or conservative treatment. Because the results of treatment have hardly improved in the last 20 years, it will soon be the second most common cause of cancer deaths. People with chronic pancreatitis, first-degree relatives of patients with pancreatic cancer, but also people of normal weight who develop diabetes over the age of 50 have a significantly increased risk of developing pancreatic cancer. In contrast to breast cancer or colon cancer, there is currently no effective screening or early detection method for these risk groups that would allow the tumor to be diagnosed at a curable stage. The only established tumor marker, CA19-9, is only really meaningful if the pancreatic carcinoma is already advanced or metastasized, i.e. no longer curable.

Over the past 12 years, Prof. Julia Mayerle's working group has investigated the possibility of developing a diagnostic test for the early detection of pancreatic cancer based on metabolites circulating in the blood, small metabolic products such as amino acids, ceramides or sphingolipids, which can be determined by mass spectrometry. In three consecutive studies using the blood of patients with pancreatic cancer and various control cohorts with other diseases or healthy individuals, patterns could be identified from the more than 1600 measurable metabolites, so-called metabolome signatures, which can with high probability rule out pancreatic cancer in a patient with a risk of or suspected pancreatic cancer. Such a test result would spare patients further, unnecessary and stressful examinations or even an operation.

The latest study, which has just been published in the journal Lancet Gastroenterology and Hepatology and was funded by the German Federal Ministry of Education and Research, involved 23 German hospitals in which either the surgical or gastroenterology departments specialize in the diagnosis and treatment of pancreatic diseases. The study included 1370 patients in whom imaging had revealed a pancreatic mass and in whom it was therefore necessary to rule out or treat pancreatic cancer. The patients' blood plasma was examined by mass spectrometry for the presence of two biomarker signatures, one with 12 and one with only 4 metabolites, as well as for an increase in the tumor marker CA19-9. The patients were followed up for two years to ensure whether pancreatic cancer was present or developing and whether another diagnosis was the cause of the CT findings. Both biomarker signatures were significantly more reliable than CA19-9 in ruling out the diagnosis of pancreatic cancer. Interestingly, the smaller panel, which quantifies only four metabolites and can therefore be measured cost-effectively on a single laboratory platform, performed practically as well as the more complex test (area under the curve AUC 0-846 (95% CI 0-842-0-849; p<0-0001 against CA 19-9, specificity of 93-6% (93-1-94-0) and accuracy of 79-0% (78-8-79-2). Even more important for clinical application is that the negative predictive value, i.e. the determination that pancreatic carcinoma can be ruled out in the patient, is over 90 percent.

This is the first time that a laboratory test has been developed that can be used in routine clinical practice to monitor patients with an increased risk of pancreatic carcinoma or patients with suspected pancreatic carcinoma based on imaging. "This metabolome test can spare patients more invasive diagnostic procedures and can be used to diagnose pancreatic cancer at a still curable stage," says last author of the study Prof. Julia Mayerle.