Continuing the Search for Pancreatic Cancer Screening Tests

A Background on our Pancreatic Cancer Analysis

By Carol Smyth, MD


At Prognos, our mission is to improve health by tracking and predicting disease the earliest. We strive to create solutions that could help to one day eradicate disease, and we keep the same values in mind when it comes to creating our content. As we will explain in greater detail below, pancreatic cancer is a devastating condition that can be just as hard to detect as it is to treat.  Our hope is that our findings may shed some light on a patient’s journey towards diagnosis.


Key Points

  • A blood-based panel or test that reliably identifies early stages of pancreatic cancer is urgently needed.
  • Our previous research identified alkaline phosphatase and CA 19-9 as possible screening tests.
  • Our new research will look at a broader range of potential serum markers, using the Prognos database to identify anonymized patients with pancreatic cancer and working retrospectively to study their bloodwork over the previous 5 years.
  • Our goal is to identify likely, common blood tests that may be useful to screen for pancreatic cancer, to provide a basis for further study with a goal of finding effective screening tests.



Pancreatic adenocarcinoma grows slowly in the ducts of the pancreas, often taking over 10 years to go from localized to metastatic tumors. However, 80% of pancreatic cancers are found late, due to the lack of symptoms associated with the tumor until its late stages. Finding the tumor late results in poor survival rates because surgery to remove the tumor is the only known way to cure it, but most late pancreatic tumors are inoperable. This means that less than 10% of patients diagnosed with pancreatic cancer are still alive 5 years after diagnosis. Imaging studies are commonly done in high-risk individuals in an attempt to find pancreatic tumors when they are small and surgery is curative, but imaging is too expensive to conduct on everyone in the population. Therefore, the search for blood-based screening tests continues, and researchers around the world are working on identifying tests that would help to identify early pancreatic tumors, with a view to finding them in the early stages.

The most commonly used blood-based screening test for pancreatic cancer is carbohydrate antigen 19-9 (CA 19-9) – however, this test is not specific to the tumor, with levels rising in other conditions as well, including diseases affecting the pancreas, liver and colon.  In addition, about 10% of people lack the enzyme to make CA 19-9, and therefore have undetectable levels even with advanced pancreatic cancer. Used alone, CA 19-9 appears to have limited use as a screening tool for pancreatic cancer, but studies using it as part of a panel continue. One recent study suggested that using CA19-9 along with another serum test, levels of carcinoembryonic antigen (CEA) helps to predict prognosis in patients with pancreatic cancer. (see

We published an abstract at the 2015 American Society of Clinical Oncology meeting, in which we reported on analysis of over 9,000 patients in our nationwide database of lab tests who had been tested for CA 19-9 between 2010 and 2014.  Of this group, about 1,600 had an ICD-9 code for pancreatic cancer – we used the code as a proxy for diagnosis of the cancer.  We then compared the group with pancreatic cancer to patients who had also been tested for CA 19-9 but did not have the code, looking at lab tests including alkaline phosphatase (ALP) and bilirubin. We found that patients with high levels of CA 19-9 and ALP had a 46% risk of also having pancreatic cancer; patients with a high level of CA 19-9 alone had a 30% risk; patients with a high ALP alone had a 13% risk, and those with normal levels of both CA 19-9 and ALP had a 7% risk. Our study did not find bilirubin levels associated with risk of pancreatic cancer.  (see

 Two years later, our database has grown to include millions more patients and tests. In this blog series, we will present the results of a new analysis, this time finding patients with pancreatic cancer by ICD-9 and ICD-10 codes as before, but also looking retrospectively at the bloodwork conducted in these patients in the 5 years prior to diagnosis. Our goal is to see if we can identify a pattern of abnormal test results that may provide a clue, a signal in the “noise” of routine testing, that may point to a useful screening panel that can help to identify patients who are at increased risk for pancreatic cancer – and who may benefit from regular monitoring using imaging studies to try to find a developing tumor in the early stages when it can be cured by surgery.

We invite you to follow us through the next few weeks, to see how our team at Prognos conducts this groundbreaking analysis, and to see the results of this study.  For more information, please comment below or contact us at



Study funded by Prognos, New York, NY.


Reading List:

Up Next

Terms of Use | Privacy | © 2018 Prognos. All rights reserved