A biomarker panel distinguished individuals with pancreatic cancer from those who were healthy, had benign pancreatic cysts, or had chronic pancreatitis with about 90% accuracy, based on analyses of two cohorts.
The panel, which tests for the known pancreatic cancer biomarker CA 19-9 plus three new serum biomarkers, had a sensitivity of approximately 85%-92% and a specificity of 90%-94%, reported Dr. Ayumu Taguchi of the department of translational molecular pathology at the University of Texas M.D. Anderson Cancer Center in Houston.
After more testing, the panel might be useful for screening individuals who are known to be at high risk for pancreatic cancer on the basis of genetic studies, Dr. Taguchi said at a media briefing highlighting research to be presented at an American Association for Cancer Research meeting on pancreatic cancer research and treatment, held in New Orleans on May 19.
Dr. Ayumu Taguchi
Imaging modalities such as endoscopic ultrasound and magnetic resonance cholangiopancreatography can detect early-stage pancreatic cancer or small pancreatic cysts, but "are not suitable for screening in terms of throughput, cost effectiveness, or invasiveness," Dr. Taguchi noted.
There is currently no approved screening panel for pancreatic cancer, which annually kills more than 39,000 individuals in the United States, according to the American Cancer Society. In a 2008 study (PLos Med. 2008;5:e123), an investigational biomarker panel differentiated healthy controls from pancreatic cancer cases when samples were obtained 7-13 months before patients developed cancer symptoms, Dr. Taguchi noted.
To further build on that research, investigators in the current study tested two independent cohorts – a training set of 138 pancreatic cancer patients, 52 healthy individuals, and 29 patients with chronic pancreatitis, and a test set consisting of 42 patients with early-stage pancreatic cancer, 50 healthy individuals, 29 patients with chronic pancreatitis, and 14 individuals with benign pancreatic cysts.
For the training set, the areas under the curve for the biomarker combination and for CA 19-9 alone were 0.80 (95% confidence interval, 0.71-0.89) and 0.76 (95% CI, 0.67-0.86), respectively (P = .014), Dr. Taguchi and his associates reported. For the test set, the AUCs were 0.91, 0.85, and 0.94 when patients with early-stage pancreatic cancer were compared with individuals who were healthy, had chronic pancreatitis, or had benign pancreatic cysts.
In addition, the biomarker panel substantially improved negative predictive values at 98% sensitivity when patients with early-stage pancreatic cancer were compared with healthy controls and patients with chronic pancreatitis, the investigators reported. The negative predictive values of the panel were 96%, 95%, and 88%, respectively, while those for CA19-9 alone ranged from 83% to 88%, they added.
"Our biomarker panel was much better at distinguishing patients with pancreatic cancer from those who were healthy, had chronic pancreatitis, or had pancreatic cysts, compared with CA 19-9 alone," Dr. Taguchi concluded. "This means that our panel has the potential to substantially reduce the number of patients who would have to undergo extremely invasive screening procedures."
The next step is to further validate the panel in larger numbers of samples collected before patients were diagnosed with early-stage pancreatic cancer, Dr. Taguchi said.
The National Cancer Institute Early Detection Network and the Lustgarten Foundation funded the research. Dr. Taguchi reported no conflicts of interest.