Trends and impact of endoscopic ultrasound utilization for suspected intraductal papillary mucinous neoplasms

Abstract

Background

Multiple management guidelines for intraductal papillary mucinous neoplasms (IPMNs) have been published to improve risk stratification and resource utilization. This study aims to evaluate trends in endoscopic ultrasound (EUS) use and agreement between cross-sectional imaging and EUS for specific pancreas cystic lesion (PCL) features.

Methods

This retrospective cohort study included consecutive adults undergoing EUS for suspected IPMN detected with cross-sectional imaging (CT/MRCP) between 2013 and 2015 (Cohort 1) and 2018–2020 (Cohort 2). Clinical, radiographic, EUS, cytologic, and outcomes data were collected. IPMN were defined as high-risk/worrisome (having any Fukuoka high risk (HR) or worrisome feature (WF)), low risk (cyst size 10–29.9 mm, without Fukuoka HR or WF), or very low risk (cyst size <10 mm, without Fukuoka HR or WF). Chi-square, Fisher's exact test, Wilcoxon rank-sum test and Kappa statistics were used. Impact of the COVID-19 pandemic was assessed via post-hoc sensitivity analysis.

Results

Of 711 patients, 292 (41.1 %) were in Cohort 1. More patients in Cohort 1 underwent EUS for non-high risk/non-worrisome PCL on pre-EUS imaging (65.8 % vs. 54.9 %, P < 0.01). Good agreement between pre-EUS imaging and EUS for WF (84.1 %) and HR (96.6 %) was seen. EUS-FNA impacted surveillance and surgical referral in 29.5 % (P = 0.51). Median surveillance duration was 917 days (IQR 367, 1439 days), during which pancreatic ductal adenocarcinoma (PDAC) was diagnosed in 7.5 % (P = 0.11).

Conclusions

In this series EUS over-utilization for low risk IPMN improved over time reflecting more conservative guidelines, without changing PDAC diagnosis rates. There was good agreement between cross-sectional imaging and EUS for specific PCL features.