To determine the minimal number of EUS-guided needle passes that maintains high diagnostic adequacy in solid pancreatic masses when rapid on-site evaluation (ROSE) is not available, and to compare the marginal diagnostic yield and safety of FNA versus FNB.

In a single-centre retrospective cohort of pancreatic EUS-FNA/FNB procedures, cases were grouped according to the number of needle passes. For each stratum, diagnostic adequacy (conclusive vs non-conclusive) was calculated separately for FNA and FNB. For FNA, complication rates were compared between procedures with ≤2 versus ≥3 passes (Fisher’s exact test). ROSE was not used in any procedure.

For FNA, adequacy rates were 69.0% (20/29) with 1 pass, 73.4% (80/109) with 2 passes, 73.9% (34/46) with 3 passes and 100% (2/2, small n) with 4 passes, showing only a limited marginal gain beyond 2–3 passes. For FNB, adequacy was 81.8% (18/22) with 1 pass and 83.3% (10/12) with 2 passes, with an early plateau from 1–2 passes. Only one complication (a pseudoaneurysm) was observed in the FNA group, with no clear association with the number of passes; no complications occurred in the FNB group.

In the absence of ROSE, FNB achieves a high diagnostic yield with only 1–2 passes. For FNA, 2–3 passes appear sufficient, and the benefit of additional passes beyond this threshold seems minimal, with no signal of increased complications in this cohort.