Pancreatic cancer: diagnosis via signature analysis

Interview with Prof. Julia Mayerle, Deputy Director, Clinic and Polyclinic for Internal Medicine A, Greifswald University Medicine

Pancreatic cancer is one of the deadliest forms of cancer because it is difficult to diagnose and only presents with symptoms in the later stages. In the future, a laboratory test developed at the Greifswald University Medicine could make an early detection of this type of cancer and consequently a faster and better treatment possible.


Photo: Smiling woman with short grey-black hair - Prof. Julia Mayerle; Copyright: private

Prof. Julia Mayerle; ©private

In this interview with, Prof. Julia Mayerle talks about the human metabolome and the development of mass spectrometric measurement to detect tumors.

Prof. Mayerle, what does early detection of pancreatic cancer presently look like?

Prof. Julia Mayerle: In 2014, 48,000 patients received in-patient treatment for pancreatic cancer in Germany. This type of cancer is very malignant and responds so poorly to medical treatment or radiation therapy to where only between four and eight percent of patients are still alive five years after receiving the initial diagnosis. By 2030, it is estimated pancreatic cancer will be the third leading cause of cancer-related deaths.

At this point, there is no early detection method for pancreatic cancer; its incidence rate is too low for this. A screening test would need to exhibit very high sensitivity and specificity. This is not technically attainable at the present time. With the currently known methods, we would have too many false-positive test results to where the mortality due to treatments would outweigh the benefits of early detection.

This is why we can only offer prevention screening for high-risk groups. These are patients suffering from chronic pancreatitis, patients with a family history of cancer or patients who are older than 50 and were newly diagnosed with type 2 diabetes mellitus.

Graphic: The pancreas and the surrounding organs; Copyright:

The pancreas is surrounded by different organs in the abdomen. Cancer can hardly be detected here using imaging; © Alexilus

You are developing a laboratory test for pancreatic cancer based on "metabolomics". What does this mean?

Mayerle: "Metabolomics" is the mass spectrometric measurement of all metabolites present in a body fluid - in our case in blood plasma. Mass spectrometry per se is not a new technique.

A metabolite is a substance in the body produced by the body’s metabolism and is smaller than 1.5 kilodalton. With its help, we can obtain a signature associated with a specific state of the body. A tumor changes the body’s metabolism dramatically to where the metabolite pattern also changes. We are able to detect this and determine whether a tumor is present.

Is this type of change specific to cancer?

Mayerle: Generally, metabolites change with every disease but there are also many external factors that modify this signature - for instance, stressful situations, sleep deprivation or extreme sports. We act on the assumption that every physical stress is detectable.

How many of these types of substances make up the metabolome and how many do you need to detect cancer?

Mayerle: Initially, we identified just over 1,000 metabolites. The number of metabolites could definitely still be increased if we would work with higher sensitivity. However, it is unclear whether we would increase diagnostic specificity this way. Changes are very difficult to detect in very low metabolite concentrations because they would get lost in the "technical noise" of the measurement.

Out of more than 1,000 metabolites, we identified those with the highest predictive values. Initially, there were nine metabolites plus CA 19-9, a tumor marker for pancreatic cancer. We are currently working on further reducing the number of metabolites required to reach a conclusive early diagnosis.

Is just the tumor marker not suited to detect this type of cancer?

Mayerle: No. CA19-9 is a so-called Lewis blood group antigen. This means the probability of developing this tumor marker depends on the blood type. Eight percent of the population are not able to express this type of protein. For these eight percent, this tumor marker test alone would not be conclusive. That is why we have to look for additional markers.

Photo: Blood samples on a data sheet; Copyright:

The researchers isolated a signature for the diagnoses of pancreatic cancer from more than 1,000 metabolites and a tumor marker in patient blood; © belchonok

You have already tested this method on approximately 900 patients. What did this test look like?

Mayerle: During a pilot phase, we first tested how we are able to differentiate between patients with chronic pancreatitis, cirrhosis of the liver, pancreatic cancer and healthy patients based on a signature. We subsequently identified the markers with which we are able to distinguish pancreatic cancer from other cohorts with a high level of diagnostic accuracy. We also examined the robustness of these markers in serum and plasma. In another prospective study, we once again identified and validated the markers in a training and test set of patient cohorts.

During this four-step process, we ultimately developed a signature that has a very high negative predictive value. That is to say, there is a high level of certainty a patient whose plasma we examine and where we obtain a negative result is actually not affected by pancreatic cancer.

The Federal Ministry of Education and Research (BMBF) supports the further development of the test method over the next four years. What are your plans for this follow-up study?

Mayerle: Starting mid-2016, we will collaborate in the study with about 20 university hospitals. We want to conduct a multicentric and prospective validation study for the early detection test with 1,500 patients who enter the hospital with an unclear pancreatic mass or a suspected cancer lesion. We suspect a pancreatic cancer diagnosis in circa 20 percent of these patients. They subsequently receive surgery and are cared for afterward within the scope of the study. In doing so, we want to find out whether we are able to also detect and confirm recurrences at an early stage via the same metabolite signature. These occur in 50 percent of patients within two years.

We keep on monitoring those patients where we were not able to detect cancer during the initial exam for two more years to ensure they are not developing cancer later on.

In the third arm of the study, we monitor patients with newly diagnosed diabetes who have no morphological indicator for a change of the pancreas. We also examine them initially and monitor them over two years for cancer.

The overall goal of the study is to obtain a reliable test procedure that we develop together with our industry partner Metanomics Health GmbH and offer to the healthcare market.

Photo: Timo Roth; Copyright: B. Frommann

© B. Frommann

The interview was conducted by Timo Roth and translated from German by Elena O'Meara.