Objective: To evaluate the efficacy and the learning curve of the endoscopic ultrasound-guided fine needle aspiration biopsy (EUS-FNAB) diagnostics of pancreatic lesions over 8 years (2007-2014).
Material and methods: EUS-FNAB using a Linear Echoendoscope Olympus was performed in 531 patients, mostly without rapid on-site evaluation (ROSE). Smears plus cytoblock sections were used routinely. Immunocytochemistry was utilised as indicated.
Results: The average, satisfactory sampling rate increased from the 61.2% in the first 3-year period to 72.9% in the last 3-year period (P = 0.008). The availability of the material for cytoblocks increased from 36.4% in the first period to 75.3% in the last period (P = 0.017). The efficacy of cytoblocks increased from 39.6% to 46.2% (P = 0.086). Comparing the first and last 3-year periods, the indication for immunocytochemistry did not rise substantially (5.2% and 8.5% respectively), but the predictive value of immunocytochemistry rose from 56.3% to 100.0% (P = 0.001). The most frequent diagnostic result was malignancy confirmation - both primary (41.2%) and metastatic (2.1%). In cases with representative samples and follow-up information, the specificity, sensitivity, positive predictive value and accuracy were high from the beginning. The negative predictive value decreased slightly in the last 3-year period.
Conclusion: The pancreatic EUS-FNAB without ROSE represents a suboptimal arrangement conditioned with our staff/time/location reality. Nevertheless, within the last 3 years of our activity, nearly three-quarters of patients profit from the diagnostic contribution of this procedure.
Keywords: endoscopic ultrasound guided aspiration biopsy; fine needle aspiration; learning curve; pancreatic neoplasms; sampling efficacy.
© 2016 John Wiley & Sons Ltd.