Background: In many institutions, computed tomography (CT) guided percutaneous fine-needle aspiration (FNA) has become the procedure of choice for biopsies of pancreatic mass lesions. This method of biopsy and others, such as endoscopic retrograde cholangiopancreatography (ERCP) cytology, are problematic because of a substantial false-negative rate. Objective: To investigate the yield of endoscopic ultrasonography-guided FNA biopsies in patients who had negative results on CT-guided biopsy or negative cytologic findings on ERCP sampling. Design: Prospective cohort study. Setting: Tertiary care university medical center. Patients: 102 patients (median age, 65 years; 58 men and 44 women) with suspected pancreatic cancer who fulfilled the above criteria were prospectively identified and underwent endoscopic ultrasonography-guided FNA biopsy. Measurements: The operating characteristics of endoscopic ultra; sonography-guided FNA for diagnosing pancreatic masses were determined. Surgical pathology or long-term follow-up (median, 24 months) was used to identify false-positive or false-negative results. Results: Median mass size was 3.5 cm x 2.7 cm. A median of 3.4 passes were performed. Cytologic results on endoscopic ultrasonography-guided FNA biopsy were positive in 57 patients, negative in 37, and inconclusive or nondiagnostic in 8. No false-positive results were observed. A diagnosis of pancreatic cancer was subsequently confirmed in 3 patients who had tested negative (false-negative results) and 1 of the 8 patients with nondiagnostic results. Of these 4 patients, 3 had cytologic evidence of chronic pancreatitis on endoscopic ultrasonography-guided FNA biopsy. The 95% CI for the likelihood ratio for a positive test result contained all values greater than or equal to 9.7. The likelihood ratio for a negative test result was 0.05 (CI, 0.02 to 0.15). The posterior probability of pancreatic cancer after a definitely positive result was at least 93.5% by a conservative lower 95% confidence limit; after a definitely negative test result, it was 6.9%. The prevalence of pancreatic cancer was 59.8% (61 of 102 patients). Self-limited complications occurred in 3 of the 102 patients (2.9% [CI, 0.6% to 8.4%]). Conclusion: Endoscopic ultrasonography-guided FNA biopsy may play a valuable role in the evaluation of a pancreatic mass when results on other biopsy methods are negative but pancreatic cancer is suspected.
