Study examines screening for pancreatic cancer in high-risk populations
Researchers from New England report in a new study that using a tumor marker, serum CA 19-9, combined with an endoscopic ultrasound if the tumor marker is elevated, is more likely to detect stage 1 pancreatic cancer in a high-risk population than by using the standard means of detection. The study appears in the July issue of GIE: Gastrointestinal Endoscopy, the monthly peer-reviewed scientific journal of the American Society for Gastrointestinal Endoscopy (ASGE).
Pancreatic cancer is the fourth leading cause of cancer death in the United States. Advanced disease at diagnosis correlates directly with worse overall survival. Symptoms of abdominal pain, jaundice, and/or weight loss often do not appear until the tumor is locally advanced or metastatic, at which point effective treatment options are very limited. By contrast, detection and resection of pancreatic cancer, when it is confined to the pancreas (stage 1 disease), improves overall survival. An effective screening protocol is urgently needed to detect earlier stage tumors. Imaging methods that have been used for pancreatic cancer screening include endoscopic ultrasound (EUS), CT, endoscopic retrograde cholangiopancreatography (ERCP) and magnetic resonance imaging/MRCP.
There has been limited success in screening younger populations using the tumor marker CA19-9, so more recent pancreatic cancer screening protocols have focused on high-risk populations. It is estimated that 10 percent of patients in whom pancreatic cancer develops have at least one first-degree relative with the disease. Multiple cohort and case-control studies have demonstrated that a family history of a first-degree relative with pancreatic cancer significantly increases a patient's risk of the development of pancreatic cancer, approximately two to five-fold. The risk of the development of pancreatic cancer increases significantly as the number of affected family members increases. Advanced age is also a significant risk factor, and 93 percent of patients with pancreatic cancer present after the age of 50.
"Our hypothesis was that a high-risk population identified by age and at least one first-degree relative with pancreatic cancer can be successfully screened. Our objective was to determine whether early pancreatic neoplasia can be detected in a high-risk population by using tumor marker CA 19-9 followed by targeted endoscopic ultrasound. We also sought to determine whether this protocol was more likely to detect early stage pancreatic cancer than standard means of detection," said study lead author Richard Zubarik, MD, Fletcher Allen Health Care, University of Vermont. "Our results showed that potentially curative pancreatic adenocarcinoma can be identified with this screening protocol. Stage 1 pancreatic cancer is more likely to be detected by using this screening protocol than by using standard means of detection."
This prospective cohort study was conducted at the University of Vermont (UVM) and the Dartmouth-Hitchcock Medical Center (DHMC). Patients were enrolled from September 2006 to July 2009. Patients included were between the ages of 50 and 80 and had at least one first-degree relative (parent, sibling, or child) with pancreatic cancer. Enrollment was initiated at age 45 if a patient had two first-degree relatives with pancreatic cancer and at age 40 if the person had a BRCA2 mutation or Peutz-Jeghers syndrome.
Serum CA 19-9 testing was performed on all patients. It was chosen as the initial screening method because it is acceptable to patients, easily obtainable, widely available, inexpensive, and relatively sensitive for the disease. Endoscopic ultrasound was performed only in patients with an elevated CA 19-9 level (a CA 19-9 value greater than 37 U/mL was considered elevated) regardless of whether only one or more than one family member was affected with pancreatic cancer.
Patients who were diagnosed with pancreatic cancer at UVM (but were not enrolled in the CA 19-9/EUS study) during the same period were prospectively identified and used as the comparison group. These patients were identified by the Cancer Data Registry at the University of Vermont. Charts were then reviewed to verify tumor type, staging data and survival.
A total of 546 patients were enrolled in the study. CA 19-9 was elevated in 27 patients (4.9 percent). Neoplastic or malignant findings were detected in five patients (0.9 percent), and pancreatic cancer in one patient (0.2 percent). The patient with pancreatic cancer detected as part of this study was one of two patients presenting to the University of Vermont with stage 1 cancer. One-year follow-up contact was performed by telephone in 519 patients (95 percent), by chart review in 24 patients (four percent), and by review of the social security death index in three patients (less than one percent). Pancreatic cancer was not detected at the one-year follow-up in any additional patients.
In the comparison group, a total of 124 patients received a diagnosis of pancreatic cancer between September 2006 and July 2009. Staging of the comparison group at the time of presentation was as follows: stage 1, one patient (0.9 percent); stage 2, 52 patients (45.6 percent); stage 3, 20 patients (17.5 percent); stage 4, 41 patients (36 percent). The patient detected in the CA 19-9/EUS study had stage 1 disease, whereas only 0.9 percent of patients in the comparison group presented with stage 1 disease. This difference was statistically significant despite only having one patient with pancreatic cancer detected in the study group because the detection of stage 1 cancer in the comparison group was so rare. Median survival for the 122 subjects in the comparison group was seven months, with a 2-year survival rate of 10 percent.
The results conclude that potentially curative pancreatic cancer can be identified with CA 19-9 and targeted EUS. Stage 1 pancreatic cancer is more likely to be detected by using this screening protocol than by using standard means of detection. Potential advantages include acceptable rates of disease diagnosis and exclusion as well as acceptable costs (cost to detect 1 pancreatic neoplasia was $8,431, while the cost to detect 1 pancreatic cancer was $41,133). In particular, the patient with pancreatic cancer detected with this screening protocol is alive without evidence of recurrence three years after surgical resection and is the longest survivor of pancreatic cancer detected in a published screening protocol. Also, evidence of pancreatic cancer did not develop in subjects with negative screening studies, at least in short-term follow-up.
The researchers note that the sample size is adequate only to demonstrate the feasibility of this approach, but summarized that this trial successfully screened a high-risk patient population for pancreatic cancer based on age and genetic predisposition. Early pancreatic cancer, associated with prolonged disease-free survival, can be detected as part of this pancreatic screening protocol. Stage 1 pancreatic cancer was more likely to be detected with CA 19-9 and targeted EUS, and it appears to be superior to the standard means of pancreatic cancer detection.