Back to 2026 Abstracts
Defining the Incremental Value of Endoscopic Ultrasound in Assessing Pancreatic Cystic Neoplasms
*Noah X. Tocci, *Divya Deverakonda, *Chase J. Wehrle, *Breanna Perlmutter, *John McMichael, *Samer Naffouje, *Toms Augustin, *Robert Naples, *Daniel Joyce, *Robert Simon, R Matthew Walsh
Digestive Disease and Surgery Institute, Cleveland Clinic Foundation, Cleveland, Ohio
Objectives:Contrast-enhanced magnetic resonance imaging (MRI) is an important modality in distinguishing the type of pancreatic cystic neoplasm, while Endoscopic Ultrasound (EUS) is an invasive procedure that could confirm morphology and assess malignant risk through cyst fine needle aspiration (FNA). We aim to characterize discordance between these studies and the additive value of EUS when evaluating pancreatic cystic neoplasms.
Methods: A prospectively maintained pancreatic cyst registry was queried for patients who underwent both an MRI and EUS. Worrisome features (WF) (excluding cyst size), FNA cytology, surgical pathology, and follow-up were captured via the medical record. Cytology was grouped into atypical or high-risk (high-grade dysplasia or cancer). Test sensitivity and specificity based on surgical pathology were evaluated by univariate analysis.
Results: Among 3702 prospective registry patients, 1758(47%) met inclusion criteria. Cytology was available in 1142 patients: 97(3%) had atypia or high-risk cytology. Only 11(0.97%) had high-risk surgical pathology (HGD or cancer) with a normal MRI (without WF). In patients with abnormal cytology (97), 53(55%) had atypia, and 44(45%) had high-risk cytology. In patients with atypia, 16(30%) underwent resection vs 23(52%) patients with high-risk cytology (p=0.03). Eleven (69%) with atypia had high-risk surgical pathology, compared to 20(87%) with high-risk cytology (p=0.23). Median follow-up for non-resected patients with atypia and high-risk cytology was 20.8 and 25.2 months respectively, with 11% and 18% eventually developing cancer (p<0.01). In the 97 patients with abnormal cytology, MRI noted WF in 58(60%), most commonly wall enhancement (45%). In patients with abnormal MRIs, FNA demonstrated 30(52%) atypia and 28(48%) high-risk cytology. In this group, 27(47%) underwent resection, 20(74%) of whom had high-risk surgical pathology. In 39 normal MRIs, FNA noted 23 (59%) atypia and 16 (41%) high-risk cytology. Both MRI and EUS were normal in 24 patients: 6(25%) went to the OR and 5(21%) had high-risk surgical pathology. WF were noted on EUS in 51(53%) patients: mural nodule (63%) was most common. FNA in this group showed 20(39%) atypia, and 31(61%) high-risk cytology; 26(51%) of these patients underwent resection and 22(85%) had high-risk surgical pathology. EUS and MRI were concordant 62% of the time when FNA was abnormal. When using surgical pathology, MRI and EUS sensitivities were 65% and 71% (p=0.75) respectively, and specificities were 13% and 50% (p=0.25) respectively. The addition of EUS to MRI, increases sensitivity by 19% (p=0.04), particularly in the context of an abnormal MRI.
Conclusion: In patients with cystic pancreatic neoplasms and an MRI without WF, less than 1% had high risk surgical pathology. In patients with an abnormal MRI, the addition of an EUS provides a significant increase in sensitivity for high-risk surgical pathology and thus should be considered for this cohort.
Back to 2026 Abstracts
