Pancreatic cyst nodules in clinical practice: ethereal or real?

Abstract

Presenter: Oswaldo Aguirre MD | Medstar Georgetown University Hospital/Georgetown University School of Medicine Background: Risk stratification in patients with pancreatic cysts continues to rely heavily on structural characteristics, as imaged both by endoscopic ultrasound and axial imaging with CT and MRI. The most recent international guidelines of the IAP use the presence of a cyst wall nodule, the nodule’s size in mm, and the presence of flow within the nodule to help risk stratify patients under investigation for pancreatic cysts. At present, the guidelines do not differentiate between nodules or solid components as seen by EUS and those imaged by CT/MRI. Further, the rate of discordance between modalities is not well known. The aim of this study was to examine the structural characteristics as seen by axial imaging in patients with endoscopically identified nodules and to determine the rate of concordance. Methods: All patients who had endoscopic ultrasound and pancreatic cyst fluid molecular testing at a single institution over a 7 year period (2012-2019) were studied. Cyst fluid molecular alterations, including DNA quantity and quality, KRAS/GNAS mutations, and tumor suppressor gene mutations were included, in addition to routine endoscopic variables. The presence of a nodule and its associated features were recorded on the endoscopic report. Clinical variables, including imaging characteristics, clinical features, and pathologic/cytologic data, were collected retrospectively using primary source data. Continuous and categorical variables were summarized and compared using standard statistical testing. Results: A total of 791 endoscopic ultrasound procedures with associated cyst fluid samples with molecular testing were available from the selected time period. Of those, 93 patients were found to have a pancreatic cyst nodule by high resolution endoscopic ultrasound. The median age of that subset was 68 (SD 14.6), and median body mass index was 26 (SD 6.4). The median cyst size was 20 mm (SD 19 mm) in maximum diameter. 60% of patients with EUS identified nodules were categorized by cyst fluid molecular testing as low risk, and 2% were predicted to have “aggressive” phenotypes. Only 42% of cysts with nodules by EUS had CEA levels >192. Importantly, only 12.8% of patients were noted to have nodules visualized by axial imaging, and just 30% of those demonstrated enhancement. Despite that rate of discordance, 75% of patients had at least one “worrisome” structural imaging feature (e.g. cyst size, solid component, cyst nodules, main duct diameter, growth rate). Conclusion: : The presence of a pancreatic nodule has a prominent place in risk stratification tools for pancreatic cysts, and both high quality endoscopic ultrasound imaging (with cyst fluid analysis) and high resolution axial imaging are central to this determination. However, in practice, there remains a significant rate of discordance between these modalities in the determination of cyst wall nodules, likely due to the inherent differences in resolution and limitations of each. Future study into other variables which modulate the prognostic importance of pancreatic cystic nodules as imaged by both EUS and CT/MRI is needed.