Defining the Incremental Value of Endoscopic Ultrasound in Assessing Pancreatic Cystic Neoplasms

Characterize the additive value of endoscopic ultrasound (EUS) and fine-needle aspiration (FNA) when applying specific surgical thresholds for pancreatic cystic neoplasms (PCNs).

When characterizing PCNs, MRI assesses lesion morphology, while EUS can confirm morphology and assess malignant risk through cyst fluid aspiration. Currently there is limited understanding of the additive benefit of sequential testing.

An institutional registry was queried for patients with an MRI and EUS for a PCN. Morphologic high-risk features (MHRF): mural nodule, pancreatic duct dilation, and thickened/enhanced wall were assessed on both modalities. Clinical courses were assessed for development of high-grade dysplasia/cancer on surgical pathology or surveillance. Diagnostic accuracy was established on surgical pathology. Receiver operating curves and decision-curve analysis were conducted to qualify the benefit afforded by each modality.

Of 3702 registry patients, 1674 met inclusion criteria. MRI detected MHRF in 462 (28%) and EUS in 400 (24%), with discordance in 436 (26%). Morphologically negative MRI and EUS yielded a +FNA in 5%. MHRF on EUS but not MRI occurred in 187, with a confirmed upgrade rate of 66%. MHRF on MRI but not EUS occurred in 249, with 92% remaining cancer-free on surveillance [median: 49 (23-78) months]. In 215 surgical patients, EUS following normal MRI improved specificity by 14%, with MRI+EUS+FNA improving specificity to 38%. MRI+EUS+FNA demonstrated superior AUC versus MRI ( P =0.022). Decision-curve analysis demonstrated EUS provides greatest clinical benefit in MRIs without MHRF when combined with FNA.

EUS and FNA provide a considerable contribution to clinical decision-making in MRIs without MHRF.

Retrospective registry cohort study.