Objectives: Endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA) is a routine technique to assess solid pancreatic lesions. The aim of this study was to analyse the effect of optimizing laboratory procedures for specimen preparation on the rate and accuracy of the procedure.
Methods: All EUS-FNAs of solid pancreatic lesions performed during the year 2000 (Period 1) and from May 2003 to May 2004 (Period 2) were analysed. During Period 1, one experienced gastroenterologist performed all EUS-FNAs, making direct smears and retrieving small fragments if present on the smear for histology. In Period 2, two endoscopists performed the EUS-FNAs and all the material was emptied into a vial containing a fixative. Slide preparation was carried out in the pathology laboratory: one slide was processed using cytocentrifugation and cell blocks were made from left-over material. Neither period utilized rapid on-site evaluation.
Results: During the two periods, 67 and 102 FNAs were analysed and showed significantly different (P < 0.001) non-diagnostic rates of 22.8% and 4.2%, respectively. The increased diagnostic yield can be explained by the modified laboratory procedures and to a lesser extent by the increased experience of the gastroenterologists. Sensitivity, specificity, PPV, NPV and accuracy in the second time period were, respectively, 90.6%, 100%, 100%, 81.8% and 93.4%, not significantly different from the first time period.
Conclusion: This study shows that accurate EUS-FNA results may be obtained with a low non-diagnostic rate comparable to those reported for rapid on-site evaluation by optimizing laboratory specimen processing in a setting of solid pancreatic lesions.