Document resource
Introduction Whilst advances in endoscopic ultrasound guided fine needle aspiration and biopsy (EUS-FNA/FNB) needle technology have improved diagnostic yields, performance has plateaued in many centres. We hypothesised that further gains require optimisation of post-sampling tissue handling. Prompted by our 2022 data demonstrating increased autolysis in specimens stored >24 hours in CytoLyt, we implemented a comprehensive quality improvement programme targeting the entire cytology pathway to replace a fragmented outsourced model.Methods Retrospective service evaluation comparing two periods: Period 1 (Jan 2021-Dec 2022, outsourced Cytopathology model) vs Period 2 (Jan 2024-Oct 2025, optimised in-house model). Identical EUS operators (n=3) used standard 22G FNAB needles (Boston/Cook). Rapid On-Site Evaluation (ROSE) was not used in either era. Interventions included: (1) switch from CytoLyt to PreservCyt to reduce pancreatic autolysis; (2) standardised specimen extrusion protocol (stylet use, <1ml saline rinse, dual 5ml air flush to minimise dilution); (3) same-day specimen processing (>90% <24 hours vs previous end-of-list batching); (4) multi-level cell block sectioning; (5) dedicated HPB cytopathologist; (6) intensive endoscopy staff training. Biliary brushing and EUS-FNA/FNB samples were analysed using National Minimum Cytology Dataset (N1-N5).Results Workload increased 11.3% .Outsourcing reduced dramatically from 77.4% to 2.1%. RCPath turnaround time standards are now consistently met (86.5% reported <7 days vs previous 8.6%).Most strikingly, definitive diagnoses increased substantially:Biliary brushings: N5 (malignant) 18.0%→57.2%; N1 (inadequate) 5.4%→0.6%; N3/N4 (indeterminate) collapsed from 53.1%→4.6%EUS-FNA/FNB solid lesions: N5 37.4%→57.8%; N3/N4 reduced from 16.1%→3.4%; N2 (no malignant cells) 40.3%→24.3%The apparent increase in EUS N1 rates (6.2%→14.5%) was attributable to cystic lesions, where 79.4% had diagnostic biochemistry (CEA, glucose, CA19-9) performed. Excluding cystic lesions, solid lesion N1 rates were stable while definitive diagnoses improvedAverage passes increased modestly (2.5→2.9), suggesting tissue quality,not quantity drove improvements.Conclusions This study demonstrates that systematic optimisation of post-sampling tissue handling can substantially improve diagnostic yields beyond what needle technology alone achieves. Rapid processing, optimised preservative, standardised handling, and dedicated cytopathology expertise transformed an underperforming outsourced service into a high-performance unit. Crucially, these results were achieved without ROSE. Our data suggest that the next frontier in diagnostic improvement lies in the specimen journey from needle to microscope. This reproducible model could be implemented across other centres to enhance diagnostic accuracy whilst reducing costs.