Pancreatic cyst epidemiology defined by endoscopic ultrasound: A single-center longitudinal study.

BACKGROUND

Pancreatic cystic lesions (PCLs) represent diverse cyst phenotypes that vary in etiology, frequency, and malignant potential. We describe the modern epidemiology of PCLs with extended follow-up surveillance by endoscopic ultrasound (EUS).

METHODS

We analyzed data on consecutive patients with PCLs followed by EUS from 1/2006-12/2024. EUS was used to establish phenotypic diagnoses and perform surveillance. A diagnostic algorithm assigned a cyst phenotype to each patient. Cysts that could not be assigned were labeled as unclassifiable cysts (UCs) and tracked for phenotypic maturation. Phenotypes were defined in the context of age, gender, prevalence, and malignant predisposition.

RESULTS

We analyzed 3808 PCLs from 3712 patients with median follow-up of 751 days (range, 5-6597 days). Three PCL cohorts were defined: mucinous (n = 2035, 53%), non-mucinous (n = 572, 15%), and UC (n = 1201, 32%). Most mucinous PCLs were intraductal papillary mucinous neoplasm (IPMN)-branch ducts (n = 1862, 91%) with IPMN-main ducts and mucinous cystic neoplasms (MCN) at 5% and 4%, respectively. Non-mucinous PCLs were comprised of pseudocysts (n = 300, 52%), serous cystadenomas (n = 187, 33%), and others. Females represented 60% of all patients. Seventy-seven percent of patients were over 60 years old at diagnosis. Multiple phenotypes were present in 2% of patients. Malignancy developed in 8% of mucinous and 3% of non-mucinous PCLs. UCs transformed in 2% of the cohort, primarily into IPMN-BD.

CONCLUSION

PCLs are a population health problem. Mucinous phenotypes and UCs represent most PCLs and are found in older patients. EUS has a low threshold for cyst recognition (< 2 mm) and highlights UCs as a potential precursor to IPMN-BD.