Could a Stool Test Help Spot Pancreatic Cancer?
The key to detecting pancreatic cancer early enough to save lives might be found in patients' poop, a new study suggests.
A couple of dozen types of microbes found in stool samples are closely linked to pancreatic cancer, and potentially predict whether a person is at high risk for the hard-to-detect malignancy, a team of European researchers reports.
This panel of 27 microbes, mostly bacteria, identified patients with pancreatic cancer with 84% accuracy, study results show.
The accuracy went up to 94% when researchers combined the microbe panel with another marker that's used to monitor the progression of pancreatic cancer - carbohydrate antigen (CA) 19-9, a type of biochemical released by pancreas tumors.
One expert was cautiously optimistic at the news.
"Currently, there is no early detection method for this disease, so patients are often only diagnosed once they've already reached a late stage when treatment options are limited, leading to low survival rates," said Lynn Matrisian, chief science officer for the Pancreatic Cancer Action Network (PanCAN).
"Pancreatic cancer patients can't afford to wait, so we're encouraged by any innovative research looking for new and better ways to detect [the disease]," she said.
Study author and lead researcher Ece Kartal cautioned that the new findings provide "a first step towards stool-based [pancreatic cancer] screening, but more steps and validation are required to develop this into a robust screening or diagnostic method."
Still, "a major advantage of this would be that the method is non-invasive, fast and, in principle, relatively inexpensive," said Kartal, who is a postdoctoral fellow with Heidelberg University Hospital in Germany.
Detection often comes too late
Kartal's team is confident enough in their research that they've applied for a patent to develop a pancreatic cancer diagnostic kit that detects these microbes in stool samples, according to a news release from Worldwide Cancer Research, a U.K. charity that partly funded the work.
Pancreatic cancer is relatively rare, representing about 3% of all cancers in the United States and 7% of all cancer deaths, according to the American Cancer Society. About 62,000 people will be diagnosed with pancreatic cancer this year, and nearly 50,000 will die of it in the United States. Pancreatic ductal adenocarcinoma (PDAC), the most common form, accounts for about 95% of cases.
Typically, pancreatic cancer isn't detected until later stages, when it's grown large and spread to other organs, the cancer society says. There are no good ways to screen for it.
"By virtue of where the pancreas sits in the abdomen, it doesn't show symptoms until usually late in the disease," said Dr. William Cance, chief medical and scientific officer for the American Cancer Society.
Cance called the new study "provocative" and "very significant."
"It's a very innovative finding at a critical time, and it provides some hope for earlier diagnosis of pancreatic cancer," Cance said.
For this study, the researchers studied saliva and stool samples from a group of 136 Spaniards, including 57 known to have pancreatic cancer, 50 healthy people and 29 with chronic pancreatitis.
Sophisticated computer analysis identified "characteristic gut microbial signatures that were distinct in PDAC patients when compared to healthy clinically matched controls," Kartal said.
The research team then tested their microbial panel on a separate group of 76 Germans, including 44 with pancreatic cancer and 32 who did not have the disease. They also tested it against publicly available data from 25 studies involving nearly 5,800 samples.
In both instances, the results validated the microbial panel as a potentially accurate way to detect pancreatic cancer.
"Our models can identify PDAC patients with good accuracy, but they do not erroneously predict PDAC among other disease groups, including PDAC risk factor groups [chronic pancreatitis patients or type 2 diabetics] and diseases with related symptomatology [such as liver diseases or other cancer types]," Kartal said.
More understanding needed
After developing their stool test, the researchers will conduct a study to see whether it can detect pancreatic cancer in people going forward, Kartal said.
It's likely the microbe panel will need to be combined with other measures to create a powerful enough tool to detect pancreatic cancer, said Christian Jobin, co-author of an editorial accompanying the study published March 8 in the journal Gut.
"The microbes by themselves are not enough to build a confident biomarker," said Jobin, a professor of medicine and program leader of cancer microbiota & host response at the University of Florida Health Cancer Center. "That's the hope going forward - multiple markers that will get you closer to a clinical test."
It's not clear why certain microbes in the gut are associated with pancreatic cancer, experts said. The microbes could be the result of cancer processes, or they might even contribute to the cancer.
"It provides an area to research - not only how does the microbiome potentially cause pancreatic cancer, but also can we adjust it, can we change the microbiome?" Cance said. "We have only seen very small incremental gains in pancreatic cancer that are minimal. We need broader gains in understanding what causes it and how to treat it."
Cance added that "this really points to the growing importance of the microbiome - those bacteria that have lived symbiotically with us over the ages - in being associated with cancer, causing cancer and potentially providing ways to diagnose and treat it."
The American Cancer Society has more about pancreatic cancer.
SOURCES: Ece Kartal, PhD, postdoctoral fellow, Heidelberg University Hospital, Germany; William Cance, MD, chief medical and scientific officer, American Cancer Society; Lynn Matrisian, chief science officer, Pancreatic Cancer Action Network (PanCAN); Christian Jobin, PhD, professor of medicine and program leader, cancer microbiota & host response, University of Florida Health Cancer Center; Gut, March 8, 2022, online