Novel blood test could improve outcomes in pancreatic cancer
A novel combination blood test detected nearly 70% of pancreatic cancers with a false-positive rate less than 5%, according to study findings published in Clinical Cancer Research.
CA19-9 — an antigen released by pancreatic cancer cells — is the best biomarker for pancreatic cancer. However, it is not reliable for detection or diagnosis of the disease, as a test developed for this marker nearly 40 years ago only detects about 40% of pancreatic cancers.
Haab and colleagues hypothesized that the glycan sTRA could be a biomarker for pancreatic ductal adenocarcinoma that improves upon CA19-9. They assessed sTRA and CA19-9 expression and secretion in panels of cell lines, patient-derived xenografts and primary tumors.
Results showed the sTRA glycan was produced and secreted by pancreatic tumors that did not produce and secrete CA19-9.
Two biomarker panels — one optimized for specificity and another for sensitivity — achieved statistical improvement (P < .001) compared with CA19-9 alone in a validation set. Moreover, the specificity-optimized panel achieved statistical improvement (P < .001) in a blinded set.
“Pancreatic cancer is an aggressive disease made even more devastating by its tendency to spread before detection, which is a serious roadblock to successful medical treatment,” Brian B. Haab, PhD, associate professor at Van Andel Research Institute, said in a press release. “We hope that our new test, when used in conjunction with the currently available [CA 19-9] test, will help clinicians catch and treat pancreatic cancer in high-risk individuals before the disease has spread.”
HemOnc Today spoke with Haab about the need for more effective early detection methods for pancreatic cancer, how the sTRA test was developed and what still must be confirmed before it is widely used.
Question: Can you describe the need for more effective strategies for early detection of pancreatic cancer?
Answer: Pancreatic cancer has the tendency to disseminate before it becomes clinically evident. The longer it is left untreated, the harder it is to treat. Signs of early pancreatic cancer are not specific — they can be easily confused with common maladies such as abdominal pain or intestinal problems — and a lot of times, patients present with metastatic disease. For this reason, a blood test would be valuable because it could be applied routinely at a relatively low cost.
Q: How was this test developed and how does it work?
A: The serological blood biomarker CA19-9 is used for pancreatic diagnostics and also for monitoring treatment responses. Although it is one of the most widely ordered tests in oncology, it is not ideal for diagnosis because it creates false positives. However, we thought this was a good start, so we started with the patients who were negative for CA19-9. We looked at the test in more detail and thought maybe there were other carbohydrate structures that were related and found in other cancers. The antigen sTRA is a nice marker for a distinct subset of pancreatic cancers that are not in the same group as those detected by CA19-9. By combining the two blood tests, we can pick up a greater number of patients than we can with either alone, with very few additional false positives.
Q: How do the results observed with this test compare with others that have been investigated?
A: For pancreatic cancer, these really are the best results we have seen in detection of a distinct subgroup of patients with high specificity that can be combined with CA19-9 for a significantly additive value. The fact that we can see it clearly in the tissue and cell culture models is further confirmation. We are getting closer to where we need to be able to use the test in high-risk populations, such as those with family history, genetic predisposition, chronic pancreatitis and possibly new onset diabetes. There are various clinical conditions that would place someone at high risk for pancreatic cancer, and it is among those people for whom we think our performance is at the level that could justify surveillance for an incipient cancer.
Q: What are the next steps are for research?
A: Further confirmation of the performance and additional studies are needed. This will include a prospective study in which we analyze the samples as they are collected so that we are replicating actual clinical use. We will build up the data and look at the value for physicians in their practice. Larger, prospective studies and a clinical trial are needed. The next steps include getting the test established in a diagnostics laboratory instead of in our academic lab, which is what we are working on now.
Q: How long before this becomes widely used in practice?
A: We have to see how this performs in the next round of studies. If things hold up — I am quite confident that they will —we could offer the test as a ‘research use only’ test for a little while to our clinical partners and continue to build up data for another publication. At this point, we would move toward offering the test as a lab-developed test for another year or so. After this, as long as everything looks good, we would move the test to a major platform. – by Jennifer Southall
For more information:
Brian B. Haab, PhD, can be reached at Van Andel Research Institute, 333 Bostwick NE, Grand Rapids, MI 49503; email: firstname.lastname@example.org.
Disclosure: Haab reports no relevant financial disclosures.