To compare the diagnostic performance (adequacy), number of needle passes and safety of endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA) versus fine-needle biopsy (EUS-FNB) for solid pancreatic masses, and to identify predictors of non-diagnostic sampling.

Single-centre retrospective study including all pancreatic EUS-FNA/FNB procedures from a local database (FNA n = 187, FNB n = 34).The primary endpoint was diagnostic adequacy (conclusive vs non-conclusive). Secondary endpoints were number of passes and adverse events within 7 days. Proportions were compared using χ²/Fisher’s exact tests and non-parametric variables using the Mann–Whitney test. A multivariable logistic regression model (outcome: non-conclusive result) included sampling technique (FNB vs FNA), lesion size (>3 cm), location (head vs body/tail), puncture route (trans-duodenal vs trans-gastric) and number of passes. Because a reference standard (surgery/biopsy/follow-up) was not available, diagnostic performance was assessed through adequacy rather than sensitivity/specificity.

Adequacy was 73.3% (137/187) for FNA vs 82.4% (28/34) for FNB (p = 0.365; OR 1.70, 95% CI 0.66–4.36).Median passes (IQR) were 2 (2–3) for FNA vs 1 (1–2) for FNB (p < 10⁻⁸).Adverse events were rare: one pseudoaneurysm in the FNA group (0.5%, 1/187) and none in the FNB group (0/34; p = 1.00).Adequacy did not differ by size (>3 cm 75.3% vs <3 cm 73.6%; p = 0.95) or access route. In FNA, adequacy was 69.9% for head vs 78.4% for body/tail lesions; a similar, non-significant trend was seen with FNB. In multivariable analysis, no factor was significantly associated with failure, although head location showed a trend towards higher non-conclusive rates (OR 2.61, 95% CI 0.96–7.07). FNB achieved high adequacy with 1–2 passes, whereas FNA plateaued around 2–3 passes.

In this cohort, EUS-FNB required significantly fewer needle passes than EUS-FNA without an excess of complications and showed a numerically higher, though not statistically significant, adequacy. Lesion size and puncture route did not impact adequacy, while head lesions tended to be more frequently non-diagnostic. Robust sensitivity/specificity estimates will require linkage to a reference standard.