Role of serial EUS-guided FNA on pancreatic cystic neoplasms: a retrospective analysis of repeat carcinoembryonic antigen measurements.

BACKGROUND AND AIMS: Pancreatic cystic neoplasms (PCNs) often need interval surveillance, including repeat EUS, but the role of repeat FNA with fluid analysis is poorly defined. The aim of this analysis is to evaluate the potential clinical significance of serial carcinoembryonic antigen (CEA) measurements by EUS-guided FNA (EUS-FNA) in the surveillance for PCNs.

PATIENTS: Patients who underwent EUS-FNA for PCNs were studied retrospectively. EUS-FNA findings were compared between index and prior procedures among patients who underwent repeat EUS-FNA.

RESULTS

A total of 400 patients with PCNs underwent EUS-FNA. Eighty-seven of those patients had prior EUS-FNA with cyst fluid analysis. Patients with repeat FNA were significantly more likely to have multiple cysts (57% vs 41%; P = .008), multilocular cysts (75% vs 62%; P = .042), connection to pancreatic duct (33% vs 18%; P = .005), and higher initial CEA levels (94.8 vs 25.6 ng/mL; P = .003) compared with patients who had only a single FNA. A comparison of prior and index FNAs did not show significant differences in EUS or cyst fluid analysis findings. After log transformation, the association between CEA level at prior and index FNA was moderate (R2 = 0.626; P < .001), but cystic fluid CEA classification with a cutoff value of 192 ng/mL changed in 17 patients (20%), without significant changes in EUS findings.

CONCLUSIONS

Repeat surveillance EUS-FNA resulted in stable CEA levels in the majority of patients, with spurious fluctuations of CEA in approximately 20% of patients. These data call into question any clinical significance attributed to an isolated interval rise in CEA level, especially in light of a stable EUS examination.