Fine-needle aspiration of pancreatic cystic lesions: A randomized study comparing standard and flexible needles with long-term follow-up.

Background and study aims: Pancreatic cystic lesions (PCLs) are increasingly found on cross-sectional imaging, and a majority have low risk for malignancy. The added value of fine-needle aspiration (FNA) in risk stratification remains unclear. We evaluated the impact of 3 FNA needles on diagnostic accuracy, clinical management, and ability to accrue fluid for tumor markers. Patients and methods: A multicenter prospective trial randomized 250 patients with PCLs≥13mm to 19G Flex(2):19G(1):22G(1) needles with cross-over as repeated FNA procedures. Diagnostic accuracy established at 2-year follow-up with final diagnosis from surgical histopathology or consensus diagnosis by experts based sequentially on clinical presentation, imaging, and aspirate analysis in blinded review. Results: Patients enrolled (36% symptomatic) with PCLs in head (44%), body (28%), and tail (26%). Percentage of cyst volume aspiration is 78% [72%-84%] for 19G Flex, 74% [64%-84%] for 22G, and 73% [63%-83%] for19G (p=.84). Successful FNA was significantly higher for 19G Flex (89% [82%-94%]) and 22G (82% [70%-90%]) compared to 19G (75% [63%-85%]) (p=0.02). Repeated FNA was required more frequently in head/uncinate lesions than body and tail (p<.01). Diagnostic accuracy of cyst aspirate was 84% [73%-91%] against histopathology at 2-year follow-up (n=79), and 77% [70%-83%] against consensus diagnosis among non-resective cases (n=171). Related serious adverse events occurred in 1.2% [0.2%-3.5%] of patients. Conclusions: Our study results demonstrate a statistically significant difference among the three needles in overall success rate of aspiration, but not in percentage of cyst volume aspirated. Flexible needles may be particularly valuable in sampling cystic PCLs in head/uncinate of pancreas.