At the 35th Annual Meeting of the Japanese Society of Abdominal Radiology held in June 2022, patients with pathologically proven IOPN-P from seven institutions were presented and included in this study. This retrospective multicenter study was approved by all the institutional review boards, and the requirement for written informed consent was waived. The mean age of the 21 (15 man, six women) patients at diagnosis was 66.3 (range, 36–85) years. All the patients underwent surgery. Preoperative CT, MRI, and 18F-fluorodeoxyglucose-positron emission tomography (FDG-PET) were performed in 21, 21, and seven patients, respectively. In addition, from the pathological database of the University of Tokyo Hospital, 20 consecutive intraductal papillary mucinous adenomas (IPMAs) and 20 consecutive intraductal papillary mucinous carcinomas (IPMCs) from January 2013 to December 2016 were selected for comparison. Patients with IPMA/IPMC who did not undergo contrast-enhanced CT or MRI were excluded. All patients were anonymized.

The medical records of the patients were reviewed for clinical findings. We extracted data on patient demographics, preoperative laboratory data (serum white blood cell count, C-reactive protein, and amylase levels), and tumor markers such as carcinoembryonic antigen (CEA) and carbohydrate antigen 19-9 (CA19-9). Tumor recurrence was determined based on the imaging findings.

Abdominal CT was performed using a multi-detector CT unit (LightSpeed QX/i or Discovery CT750 HD; GE Healthcare, Milwaukee, WI, USA; SOMATOM Definition: Siemens, Erlangen, Germany; Aquilion ONE, Aquilion PRIME, Aquilion 64, Aquilion 16, Aquilion 320, or Aquilion Precision: Canon/Toshiba, Otawara, Japan; or IQon Spectral CT: Philips Healthcare, Best, the Netherlands) with a tube voltage of 120 kVp. Contrast-enhanced CT was not performed in one patient with IOPN-P because of a history of adverse reactions to iodine contrast media. After intravenous contrast medium administration, contrast-enhanced CT acquisition timing included the arterial (approximately 40 s) and portal (60–90 s) phases. Arterial phase images were not obtained in three patients. An automated power injector was used to inject 100–135 mL or 600–650 mg of iodine/kg body weight at a rate of 2.0–4.0 mL/s. CT images with a thickness 5 mm or less were used for evaluation.

MRI examinations were performed using 1.5-T (SIGNA HDxt: GE, Milwaukee, WI, USA; MAGNETOM Avanto, MAGNETOM VISION: Siemens, Erlangen, Germany; or Ingenia, Achieva: Philips Medical Systems, Best, the Netherlands) or 3.0-T (MAGNETOM Skyra, MAGNETOM Trio A Tim, or MAGNETOM Prisma: Siemens, Erlangen, Germany; Achieva D-Stream: Philips Medical Systems, Best, the Netherlands; Titan: Canon, Otawara, Japan; or Discovery MR750w or Signa HDXt: GE, Milwaukee, WI, USA) MRI scanners. Axial dual-echo T1-weighted images (repetition time [TR], 4.1–256 ms; echo time [TE], 1.1–5.3 ms; slice thickness, 3.5–8.0 mm), axial and coronal T2-weighted images (TR, 543–8000 ms; TE, 54–320 ms; slice thickness, 4.0–6.0 mm), and magnetic resonance cholangiopancreatography images (TR, 800–6000 ms; TE, 105–900 ms; slice thickness, 1.4–5.0 mm) were obtained.

FDG-PET/CT or FDG-PET/MRI was performed in seven patients (Discovery 690, Discovery MI, Discovery MI DR, or SIGNA PET/MR: GE, Milwaukee, WI, USA; or Aquiduo:Toshiba/Siemens Co., Ltd.) 1 h after the administration of 2.96–4.38 MBq/kg FDG at with noncontrast CT of 2.0–3.75 mm thickness.

All images were evaluated by a radiologist with 5 years of experience in abdominal radiology under the supervision of a board-certified radiologist with 10 years of experience.

The following features were evaluated using CT and/or MRI: the localization of the lesion; main pancreatic duct (MPD), branch duct or mixed type; diameter of the entire cyst, including the intracystic solid part (measurement of the MPD type was not performed); diameter of the solid portion; contrast enhancement effect (CT values); bile duct obstruction; maximum diameter of the MPD; and dilatation of the MPD downstream of the tumor. This was a multi-center study in which the time of the arterial phase was different; hence, we did not evaluate the CT values of solid portion in the arterial phase. We also evaluated lymph node enlargement, peripancreatic soft tissue invasion, and dilatation of the MPD due to obstruction. Peripancreatic invasion was evaluated with reference to a previous study [8]. The maximum standardized uptake (SUVmax) value of the solid part of the tumor was measured.

Initial pathological diagnoses were established by board-certified pathologists at each institution. Information on malignant changes and stromal invasion was extracted from the pathological records.

Statistical analyses were performed using EZR software (EZR version 1.55; Jichi Medical University Saitama Medical Center, Saitama, Japan) [9]. Fisher’s exact test was used to evaluate the association between two and three categorical variables. To adjust for multiple comparisons, the Holm method was used to control for the family wise error rate. In addition, we performed a one-way analysis of variance or the Kruskal–Wallis test to compare the means or medians of the three groups, depending on whether they were normally distributed. The Shapiro–Wilk test was used to assess whether a continuous variable followed a normal distribution. To determine the group(s) differed significantly from each other, we performed a post-hoc analysis using Tukey’s test (after one-way analysis of variance) or the Steel–Dwass test (after the Kruskal–Wallis test). Overall survival and recurrence-free survival were calculated from the pancreatic resection date and compared using the log-rank test. All P values were two-sided, and P <.05 was considered significant.