Aims To compare the performance of EUS-FNB using the SharkCore needle (SC) or the ProCore needle (PC) for the diagnosis of solid pancreatic lesions (SPLs).
Methods Patients with SPLs were randomized to be punctured with the SC or the PC needle. Two independent randomization lists were used for the 22G and the 25G. Three needle passes were performed, and specimens placed separately in formalin vials after each pass. ROSE was not available, and pathologists were blinded to the needle used. The primary endpoint was the histologic procurement yield. Secondary endpoints were: sample adequacy and quality, diagnostic accuracy, safety, number of passes needed to reach a definitive diagnosis. Atypia was considered as false negative when a final diagnosis of malignancy was ascertained. Diagnostic accuracy was measured against surgical pathology or after a clinical course of at least six months.
Results 192 patients were enrolled. The histologic yield after 1, 2 and 3 passes were 60.4%, 65.6% and 79.1% vs 92.7%, 96.9% and 100% with the SC and the PC, respectively (p=0.0001). Sample adequacy was higher after one pass in the SC group (86.5% vs. 97.9%, p=0.005) but comparable after 3 passes (95.8% vs. 100%, p=0.121). Similar diagnostic accuracy was observed in the PC and SC groups (91.67% vs. 92.71%, respectively). However, in the 22G subgroup, the accuracy after one pass was higher with the SC (93.4 vs 78.7%, p=0.034). Both needles were safe. The sample quality was better with the SC (p=0.0001). The second pass improved the diagnostic accuracy only in the 22G PC subgroup (78.7% vs. 93.4% with one or two passes, p=0.034).
Conclusions The SC provides higher histologic yield and sample quality compared with the PC. In the 22G subgroup the adequacy and accuracy were higher with the SC after one pass. At least two passes should be performed with the 22G PC.
