Pancreatic mucinous cystic neoplasms in adults: An update on the role of CT and MRI for risk stratification.

Fiche publication

Date publication

octobre 2026

Journal

European journal of radiology

Auteurs

Membres identifiés du Cancéropôle Est :
Pr HOEFFEL Christine

Tous les auteurs :
Nguyen TTL, Barat M, Marchese U, Dohan A, Hoeffel C, Rouquette A, Pellat A, Soyer P

Lien Pubmed

https://www.ncbi.nlm.nih.gov/pubmed/42933251

Résumé

Pancreatic mucinous cystic neoplasms (MCNs) are characterized histologically by an inner mucin-producing columnar epithelium supported by a dense ovarian-type stroma. MCNs account for approximately 10% of pancreatic cystic neoplasms. Surgical resection is the treatment of choice, as MCNs may harbor high-grade dysplasia or progress to invasive carcinoma. Establishing reliable imaging criteria for risk stratification is critical to identify patients who are candidates for surgery. Computed tomography and magnetic resonance imaging play pivotal roles in lesion characterization and assessment of features associated with high-grade dysplasia or invasive carcinoma. Current guidelines recognize several imaging features as high-risk characteristics for MCNs, including large tumor size, mural nodules, and main pancreatic duct dilatation. However, other potential predictors of malignancy remain poorly defined and are not sufficiently emphasized. The strongest MCN-specific evidence, as demonstrated in multivariable analyses, supports larger cyst size, mural nodules, main pancreatic duct dilatation, and, in selected cohorts, elevated serum cancer antigen 19-9. Male sex has also been associated with malignancy in retrospective surgical cohorts, whereas pancreatic head/neck location, calcification, wall/septal thickening, age, and signal heterogeneity on T2-weighted images remain less consistently validated. Importantly, most malignancy-prediction literature is based on resected, pathologically confirmed MCNs and is therefore vulnerable to selection and verification bias and cannot be directly extrapolated to all incidentally detected presumed MCNs. This article presents a transparent literature search methodology and an evidence-based synthesis of predictors of malignancy that clearly distinguishes MCN-specific multivariable evidence from lower-level or extrapolated data. It also provides a comparison of key guidelines. Unlike previous reviews, this article directly addresses the evidence gap between resection-based cohorts and the broader population of incidentally detected presumed MCNs. It also clarifies when imaging findings should prompt further evaluation rather than independently determine management.