An unexpected rise in the nation's pancreatic death rate, researchers say, has accelerated efforts to find molecular clues to the early development of these highly lethal malignancies.
Although relatively rare, pancreatic cancer is expected to overtake colorectal cancer as the second-leading cause of cancer death in the United States by 2024, if current trends persist, according to data from the National Cancer Institute's Surveillance, Epidemiology and End Results (SEER) program. Fewer than 11 percent of these cancers are now detected early when surgical treatments work best and 5-year relative survival remains grim-at 3 percent, or less, after these cancers have spread.
Investigators, however, are turning to collaborative studies, known as "biomarker bakeoffs," to alter this bleak scenario and improve survival. The bakeoff term, apparently inspired by cooking, refers to the fact that all of the scientists share the same "recipe" at the outset, but differ in the biomarker ingredients they use to distinguish between healthy cells and a developing cancer.
At six major centers across the country, bakeoff teams already have begun testing dozens of potential biomarkers both in the blood and the cystic fluid of similar patient groups for signs of early disease. Biomarkers, by definition, encompass a diverse range of molecules from RNA and DNA to proteins and metabolites, any of which may signal biological changes that occur as cancer takes hold.
Ultimately, finding one of these markers for pancreatic cancer that allows for early detection and "shifting patients to where their cancers are still operable," will lead to better survival, said Randall Brand, MD, the lead investigator in these studies and Director of the University of Pittsburgh Medical Center's GI Malignancy Early Detection, Diagnosis and Prevention Program.
When coupled with better treatments, such as targeted therapies and adjuvant therapies in the future, he said the ability to shift these cancers from late- to early-stage disease "should give us our best shot for a cure."
For now, Brand and others stressed, no such biomarker exists for prevention in pancreatic cancer. Although CA 19-9 has been extensively studied and works well for prognosis, they said it lacks the specificity needed to detect these difficult cancers early enough to improve clinical outcomes. And, some 5 percent of those who develop pancreatic cancer make none of this protein at all, while others have certain health conditions, like pancreatitis, which can compromise results.
"There are some biomarkers out there right now that are similar to CA 19-9, but nothing beats it," despite its lack of a prevention benefit, Brand said. So investigators are testing combination strategies with CA 19-9 and other biomarkers in double-blinded studies to see what works best.
In the blood-based bakeoffs, investigators share "a portion of the same blood draw at the same time," unaware whether samples come from individuals with pancreatic cancer or from a control group without it, Brand explained. A similar approach is being taken in the cystic fluid bakeoff, he said, where investigators hope to find biomarkers that can predict which cysts need close monitoring and which pose an immediate cancer threat. An estimated 15 percent of pancreatic cancers arise from malignant cysts.
"In all of our studies, we're looking for consistency of results," Brand said. "We're not just saying 'test it-let's just combine these markers ad hoc.'"
The process of benchmarking and validating individual biomarkers, investigators say, requires that promising candidates move through five phases of discovery before reaching the final point where their presence can trigger a clinical decision or apply to broader use in the general population.
Just how rigorous that review process is can be seen in the few biomarkers the FDA has approved so far, mostly as companion diagnostics to other tests, which assess cancer risk or confirm a biopsy result. To date, only eight cancer biomarkers have won the federal agency's approval out of the 1,000 or more biomarkers scientists have generated for study in various cancers, including pancreatic cancer.
"There's a lot of activity in the biomarker space and literally thousands of papers on biomarkers in pancreatic cancer," said Anirban Maitra, MBBS, a pancreatic cancer expert at the University of Texas MD Anderson Cancer Center, where he's Professor of Pathology and Translational Molecular Pathology, and one of the bakeoff investigators. "But, we're not there yet in terms of a better biomarker," with the vast majority still stuck in early phases of development, he said.
These include the initial phase, where researchers simply identify a promising biomarker, followed by testing samples in patients who've already been diagnosed with cancer. Then comes testing pre-diagnostic samples, which Maitra describes as crucial to identifying molecular signals before a known cancer diagnosis occurs in high-risk individuals.
"That's where we are now," he said. "What we want to know is: how far back does my cancer signal show up and how good is that signal?"
Researchers at MD Anderson and elsewhere are also analyzing 200 pancreatic cancer samples taken from patients who developed these malignancies during the large-scale Prostate, Lung, Colorectal and Ovarian Screening Trial, which ended in 2006. Maitra noted these "retrospective/prospective" studies are essential to finding specific biomarkers that may be visible long before a cancer appears. Obviously, the further out they're found, the greater the potential benefit.
The difficulties still surrounding the early detection of pancreatic cancer today lie in part to the pancreas' location deep within the belly, making it harder to find tumors during a physical exam and, in part, to symptoms that often mimic other health conditions.
"Most patients come to us with jaundice and pain and unexplained weight loss," Maitra said, although sometimes diabetes and worsening blood disease may flag one of these cancers 6 months prior to diagnosis. The median age at diagnosis is 68, with most pancreatic cancers found incidentally in abdominal scans for other health issues, because effective screening tests, comparable to mammography or colonoscopy, are not available.
But, recent advances in technology, like single-cell analysis, radiomics and artificial intelligence, hold promise for progress against this cancer and others, researchers believe.
Artificial intelligence tools, for example, may be able to detect atypical patterns in the pancreas that are currently radiographically silent and invisible to the human eye, yet reveal an early pancreatic cancer, said Sharmistha Ghosh-Janjigian, PhD, Program Director for the Cancer Biomarkers Research Group in NCI's Division of Cancer Prevention.
"We absolutely need early detection for pancreatic cancer," she said, given its lethality and the clear survival edge in surgically resectable disease. Even for Stage II pancreatic cancer, which is still regional, survival 'is not so good,' as indicated in a recent scientific article, she noted. "So, we must continue to aim for earlier diagnosis of this deadly cancer."
Susan Jenks is a contributing writer.