Background: Diagnosing pancreatic cancer by EUS-FNA is a potentially appealing alternative to percutaneous biopsy.
Aim: To compare EUS-FNA with CT or US-guided FNA for diagnosing pancreatic cancer.
Design: Single center, prospective, randomized, cross-over.
Setting: Duke University Medical Center.
Population: Eighty-four patients referred with suspicious solid pancreatic mass lesions randomized to CT/US-FNA (n = 43) or EUS-FNA (n = 41).
Intervention: Patients underwent an imaging procedure/FNA. If cytology was nondiagnostic, cross over to the other modality was offered. Final outcome was determined by clinical follow-up every 6 months for 2 years and/or surgical pathology for patients with negative FNA.
Main outcome measurements: Sensitivity and accuracy of EUS-FNA versus CT/US-FNA for pancreatic cancer.
Results: There were 16 true positive (TP) by CT/US-FNA and 21 TP by EUS-FNA. Sixteen of the 20 CT/US-FNA negative patients crossed over to EUS-FNA; 12 underwent FNA, 4 had no mass at EUS. Seven of the 12 had positive EUS-FNA. Eight EUS-FNA negative crossed over to CT/US; 4 had no mass at CT/US, 3 remained true negative throughout follow-up, 1 had chronic pancreatitis at surgery. The sensitivity of CT/US-FNA and EUS-FNA for detecting malignancy was 62% and 84%, respectively. A comparison of the accuracy for CT/US-FNA and EUS-FNA was not statistically significant (P = .074, chi(2)).
Limitations: Failure to meet target enrollment resulted in an inability to demonstrate a statistically significant difference between the 2 modalities.
Conclusions: EUS-FNA is numerically (though not quite statistically) superior to CT/US-FNA for the diagnosis of pancreatic malignancy.