Blood test may improve detection of early pancreatic cancer
Click Here to Manage Email Alerts
Key takeaways:
- Biomarker combinations improved detection of pancreatic cancer in a double-blind trial.
- A phase 3 investigation is underway.
A novel evaluation of biomarkers known to be associated with pancreatic cancer significantly improved detection of early-stage disease compared with previous tests, according to results of a double-blind phase 2 trial.
“We can detect early-stage cancer relative to lots of noncancer conditions,” Brian Haab, PhD, professor in the department of cell biology at Van Andel Institute, told Healio.
Researchers have started a phase 3 investigation to assess how the biomarker combinations perform in prediagnostic specimens.
Background
Individuals diagnosed with pancreatic ductal adenocarcinoma (PDAC) have a 5-year survival rate of approximately 13%, according to study background.
Patients whose cancer is localized to the pancreas and node negative have a 5-year OS of 44%; however, those individuals account for only 15% of cases.
“Pancreas cancer is a systemic illness. Our only chance to give more patients the opportunity for cure is to detect it early,” Randall E. Brand, MD, professor of medicine at University of Pittsburgh and academic director of the gastrointestinal division at University of Pittsburgh Medical Center, Shadyside, told Healio.
Yet, screening the general population is not recommended due to rarity of the disease.
Multiple research groups have evaluated biomarkers that could be used to detect pancreatic cancer, but the best — CA19-9 — had sensitivity and specificity between 0.7-0.75 identifying the disease from other noncancerous conditions.
CA19-9 is not FDA approved for early detection of pancreatic cancer but it is approved for monitoring response to therapy.
Brand cited an example of a patient with metastatic disease who has elevated CA19-9 and receives chemotherapy or another type of treatment.
“If you see the level fall, the oncologists thinks, ‘That’s a good thing. Let’s continue this treatment,’ Brand said. “But if it goes up — it may go up before you see changes on imaging — they can factor it in with how the patient’s feeling clinically, and it may force them to look [at other options] sooner.”
Methods and results
Investigators also hypothesized that other glycans besides CA19-9 “could provide additive value,” Haab said.
Haab, Brand and colleagues investigated 13 biomarkers previously identified as candidates for PDAC detection in combination with CA19-9.
Researchers from four labs in NCI’s Early Detection Research Network conducted a double-blind study in which they assessed PDAC specimens along with control samples.
The control samples included healthy specimens as well as those with confounding factors, such as patients with jaundice or chronic pancreatitis.
“I don’t think [pancreatic cancer biomarker researchers] have ever been as strictly blinded for a study as these investigators were,” Brand said. “Often, people are given samples and they’re told you have 100 cases and 100 controls. They were completely blinded on the number of cases and controls.”
The study consisted of two sets of samples, a training set (66 PDAC samples, 66 control) and a validation set (170 PDAC, 125 control).
Researchers analyzed them for “specificity while controlling for high (~0.95) sensitivity and ... for sensitivity while controlling for high (~0.95) specificity,” they wrote.
Controlling for high specificity, CA19-9 had a sensitivity of 0.44 and a specificity of 0.98.
Adding CA199.STRA to CA19-9 increased sensitivity to 0.71, but only decreased specificity to 0.94. Other combinations had similar results, but with more biomarkers.
“CA199.STRA and CA19-9 probably give you the most bang for the buck without lowering your specificity because the more markers you put in, the more chance for false positives and optimizing sensitivity,” Brand said.
Controlling for high sensitivity, CA19-9 had a specificity of 0.16 and a sensitivity of 0.94.
Adding CA199.STRA and LRG1 increased specificity to 0.65 with a slight decrease of sensitivity to 0.89. Other biomarker groupings showed promise, as well.
“These biomarkers are showing that they’re doing better than CA19-9 alone,” Brand said. “Together, they’re enhancing the performance characteristics.”
Next steps
Researchers have begun a phase 3 trial that will evaluate the ability of biomarkers to distinguish PDAC from prediagnostic samples.
“We should hopefully have that done within a year,” Brand said. “All the assays have been run. We’re just starting to analyze this new data. It will be critical to get this to a phase 4 trial. Right now, we don’t have anything for the patient and know there are unmet needs.”
Researchers are evaluating multiple combinations of biomarkers beyond CA199.STRA and CA19-9, Brand said.
“Clearly, that’s the most promising panel,” Haab said. “It’s the smallest number of biomarkers, and it does most of the work in terms of the additive benefit. The other ones should be continued to be studied because we don’t have all the information we need about subsets of patients and value in patients prior to clinical onset.”
Subsets could include those with new-onset diabetes or those with chronic pancreatitis, among others.
Brand said the role blood tests have in detecting pancreatic cancer could mirror what has happened with colon cancer, with colonoscopies and at-home blood tests.
“Colonoscopy is still better than any of these other tests, but we also know many people don’t want invasive testing and don’t want to take the colon prep,” Brand said. “We must be realistic. There’s a lot of cost to our health care system in using colonoscopy for screening the general population. I believe in the future that the use of colonoscopies in average-risk patients will be defined by biomarkers, including stool-based and blood tests that can determine who’s at high risk or low risk for having a colorectal cancer or advanced adenoma. ... I think the same thing will hold true with pancreas cancer.
“I believe blood markers will help refine our screening tests in individuals at high risk for developing PDAC so that we can use our costly and invasive modalities more on those who are at ‘high risk’ in the high-risk group based on the blood marker result, and use the modalities less frequently in those considered ‘lower risk’ by the blood marker result,” he added.
For more information:
Randall E. Brand, MD, can be reached at brandre@upmc.edu.
Brian Haab, PhD, can be reached at brian.haab@vai.org.
Collapse
Click Here to Manage Email Alerts