Patient Selection

Inclusion criteria: patients with EUS or pathological diagnosis of IPMN who underwent an MRI examination between July 2015 and December 2017 at Department of Radiology of University Hospital of Ancona, Italy. Exclusion criteria: lack of MRI examination, contraindication to contrast media administration and lack of EUS, pathological and clinical data. 

MRI technique

MRI studies were performed with a 1.5-T scanner (GE Signa HdXt, Milwaukee, WI) and an eight-channel phased-array body coil. 

The examination protocol included axial and coronal fast spin-echo (FSE) T2-weighted sequences, axial T1-weighted gradient echo (GRE) in- and opposite-phase, diffusion-weighted imaging (DWI, b=0, 800 s/mm²), and 3D MR cholangiopancreatography (MRCP) with heavily T2-weighted fast recovery FSE. A multiphasic post-contrast study was performed with a 3D fast spoiled GRE sequence (LAVA) and administration of Gd-DOTA, 0.1 mmol/Kg (Dotarem, Guerbet, France) at 2 ml/s followed by a saline flush. Pre-contrast, arterial, venous, and delayed phase were obtained. 

Image Analysis

Three radiologists with different experience, reader A (30 years), reader B (10 years), and reader C (1 year), retrospectively reviewed in blind all the MRI.

MRI studies were evaluated as per Fukuoka guidelines v2017, the presence of Worrisome Features (WF) and High-Risk Stigmata (HRS) was assessed [3].

 WF included:  pancreatitis (pancreatic edema); cyst ≥3cm; cystic mural thickening or enhancement; not-enhancing mural nodule <5 mm; main pancreatic duct (MPD) diameter between 5 and 9 mm; abrupt change in the MPD diameter with pancreatic atrophy; lymphadenopathy (short axis >10 mm).

HRS included: obstructive jaundice (choledochal diameter ≥8 mm); enhancing mural nodules or solid component; MPD ≥10 mm.

Statistical Analysis

Quantitative variables were expressed as dichotomous following the thresholds in the previous paragraph. The IRA was calculated with Intraclass Correlation Coefficient (ICC) and 95% confidence interval (95%CI).