In a recent report in eClinicalMedicine, researchers from Cancer Center
Amsterdam report that novel blood-based protein biomarkers are less accurate
than the standard tumor biomarker CA19-9 for detecting pancreatic cancer, and that
their added value over CA19-9 is limited, negative, or unclear.

Pancreatic cancer poses a major diagnostic and therapeutic challenge as symptoms often develop only when the disease is very advanced. At this stage, available treatments are not effective for most patients. It is therefore critical to develop methods that can accurately detect pancreatic cancer earlier. Recently, blood-based diagnostic tests have been developed that aim to improve the prognosis of pancreatic cancer by facilitating earlier detection and treatment.

Carbohydrate antigen 19-9 (CA19-9) is a tumor marker found in the serum that is commonly checked to monitor how well a patient responds to treatment. As the diagnostic performance of CA19-9 is low, researchers have continued to search for new and better biomarkers that could outperform CA19-9. The purpose of this study was to check the accuracy of recently discovered protein biomarkers and determine if they have added value over CA19-9 alone for detecting pancreatic cancer.

Contrary to expectations

In this meta-analysis, data from 28 studies (6127 participating individuals in total) were analyzed, and patient-level data from all studies were requested.

“Contrary to our expectations, we found that the new protein biomarkers were substantially worse than CA19-9 in distinguishing pancreatic cancer from clinically similar benign pancreatic diseases,” say first authors Lenka Boyd and Mahsoem Ali. In addition, the added value of protein biomarkers to CA19-9 was limited, negative, or unclear.

“We also found that the methodology of initial biomarker discovery studies was frequently flawed, resulting in biased and overly optimistic estimates of the proteins’ diagnostic accuracy.”

Better study design

The authors formulated a number of recommendations to improve future studies that aim to identify biomarkers that could be used for early detection of pancreatic cancer. Most importantly, studies should use blood samples from persons with suspected pancreatic cancer, rather than a ‘diagnostic case-control’ design in which cases (pancreatic cancer patients) and controls (healthy controls or patients diagnosed with benign diseases) are recruited separately. In addition, studies should always assess the added value of newly discovered biomarkers over commonly available, routinely measured markers.

For more information contact Lenka Boyd, or read the publication here:

Boyd, L.N.C., Ali, M. et al. (2023) Diagnostic accuracy and added value of blood-based protein biomarkers for pancreatic cancer: A meta-analysis of aggregate and individual participant data. eClinicalMedicine 55, 101747.

People involved at Amsterdam UMC:

Lenka Boyd

Mahsoem Ali

Mariska Leeflang

Ralph de Vries

Tessa Le Large

Marc Besselink

Elisa Giovannetti

Hanneke van Laarhoven

Geert Kazemier

Funders involved

Grants: Bennink Foundation, KWF Kankerbestrijding, and AIRC.

Visual abstract