[Skip to Navigation]
Sign In
Table 1.  Characteristics of Patients Who Underwent Surgical Resection for IPMN
Characteristics of Patients Who Underwent Surgical Resection for IPMN
Table 2.  Factors Associated With Invasive IPMC for Patients With Branch Duct IPMN
Factors Associated With Invasive IPMC for Patients With Branch Duct IPMN
Table 3.  Factors Associated With Invasive IPMC for Patients With Mixed IPMN
Factors Associated With Invasive IPMC for Patients With Mixed IPMN
Table 4.  Factors Associated With Invasive IPMC for Patients With Main Duct IPMN
Factors Associated With Invasive IPMC for Patients With Main Duct IPMN
Table 5.  Diagnostic Values of Invasive IPMC Using the Independent Predictors
Diagnostic Values of Invasive IPMC Using the Independent Predictors
1.
Adsay  V, Fukuhima  N, Furukawa  T. Intraductal neoplasm of the pancreas. In: Bosman  FT, Carnerio  F, Hruban  RH, eds.  WHO Classification of Tumors of the Digestive System. Lyon, France: IARC Press; 2010:304-313.
2.
Tanaka  M, Chari  S, Adsay  V,  et al; International Association of Pancreatology.  International consensus guidelines for management of intraductal papillary mucinous neoplasms and mucinous cystic neoplasms of the pancreas.  Pancreatology. 2006;6(1-2):17-32.PubMedGoogle ScholarCrossref
3.
Tanaka  M, Fernández-del Castillo  C, Adsay  V,  et al; International Association of Pancreatology.  International consensus guidelines 2012 for the management of IPMN and MCN of the pancreas.  Pancreatology. 2012;12(3):183-197.PubMedGoogle ScholarCrossref
4.
Vege  SS, Ziring  B, Jain  R, Moayyedi  P; Clinical Guidelines Committee; American Gastroenterology Association.  American gastroenterological association institute guideline on the diagnosis and management of asymptomatic neoplastic pancreatic cysts.  Gastroenterology. 2015;148(4):819-822, e12-e13.PubMedGoogle ScholarCrossref
5.
Crippa  S, Bassi  C, Salvia  R,  et al.  Low progression of intraductal papillary mucinous neoplasms with worrisome features and high-risk stigmata undergoing non-operative management: a mid-term follow-up analysis  [published online January 7, 2016].  Gut. doi:10.1136/gutjnl_2015_310162PubMedGoogle Scholar
6.
Roch  AM, Ceppa  EP, Al-Haddad  MA,  et al.  The natural history of main duct-involved, mixed-type intraductal papillary mucinous neoplasm: parameters predictive of progression.  Ann Surg. 2014;260(4):680-688.PubMedGoogle ScholarCrossref
7.
Kwong  WT, Lawson  RD, Hunt  G,  et al.  Rapid growth rates of suspected pancreatic cyst branch duct intraductal papillary mucinous neoplasms predict malignancy.  Dig Dis Sci. 2015;60(9):2800-2806.PubMedGoogle ScholarCrossref
8.
Salvia  R, Crippa  S, Falconi  M,  et al.  Branch-duct intraductal papillary mucinous neoplasms of the pancreas: to operate or not to operate?  Gut. 2007;56(8):1086-1090.PubMedGoogle ScholarCrossref
9.
McMillan  MT, Lewis  RS, Drebin  JA,  et al.  The efficacy of adjuvant therapy for pancreatic invasive intraductal papillary mucinous neoplasm (IPMN).  Cancer. 2016;122(4):521-533.PubMedGoogle ScholarCrossref
10.
Hirono  S, Kawai  M, Okada  K,  et al.  Long-term surveillance is necessary after operative resection for intraductal papillary mucinous neoplasm of the pancreas.  Surgery. 2016;160(2):306-317.PubMedGoogle ScholarCrossref
11.
Kang  MJ, Jang  JY, Lee  KB, Chang  YR, Kwon  W, Kim  SW.  Long-term prospective cohort study of patients undergoing pancreatectomy for intraductal papillary mucinous neoplasm of the pancreas: implications for postoperative surveillance.  Ann Surg. 2014;260(2):356-363.PubMedGoogle ScholarCrossref
12.
Hirono  S, Tani  M, Kawai  M,  et al.  Treatment strategy for intraductal papillary mucinous neoplasm of the pancreas based on malignant predictive factors.  Arch Surg. 2009;144(4):345-349.PubMedGoogle ScholarCrossref
13.
Winter  JM, Jiang  W, Basturk  O,  et al.  Recurrence and survival after resection of small intraductal papillary mucinous neoplasm-associated carcinomas (≤20-mm invasive component): a multi-institutional analysis.  Ann Surg. 2016;263(4):793-801.PubMedGoogle ScholarCrossref
14.
Waters  JA, Schnelldorfer  T, Aguilar-Saavedra  JR,  et al.  Survival after resection for invasive intraductal papillary mucinous neoplasm and for pancreatic adenocarcinoma: a multi-institutional comparison according to American Joint Committee on Cancer Stage.  J Am Coll Surg. 2011;213(2):275-283.PubMedGoogle ScholarCrossref
15.
Wasif  N, Bentrem  DJ, Farrell  JJ,  et al.  Invasive intraductal papillary mucinous neoplasm versus sporadic pancreatic adenocarcinoma: a stage-matched comparison of outcomes.  Cancer. 2010;116(14):3369-3377.PubMedGoogle ScholarCrossref
16.
Yopp  AC, Katabi  N, Janakos  M,  et al.  Invasive carcinoma arising in intraductal papillary mucinous neoplasms of the pancreas: a matched control study with conventional pancreatic ductal adenocarcinoma.  Ann Surg. 2011;253(5):968-974.PubMedGoogle ScholarCrossref
17.
Poultsides  GA, Reddy  S, Cameron  JL,  et al.  Histopathologic basis for the favorable survival after resection of intraductal papillary mucinous neoplasm-associated invasive adenocarcinoma of the pancreas.  Ann Surg. 2010;251(3):470-476.PubMedGoogle ScholarCrossref
18.
Hackert  T, Fritz  S, Klauss  M,  et al.  Main-duct intraductal papillary mucinous neoplasm: high cancer risk in duct diameter of 5 to 9 mm.  Ann Surg. 2015;262(5):875-880.PubMedGoogle ScholarCrossref
19.
Hirono  S, Tani  M, Kawai  M,  et al.  The carcinoembryonic antigen level in pancreatic juice and mural nodule size are predictors of malignancy for branch duct type intraductal papillary mucinous neoplasms of the pancreas.  Ann Surg. 2012;255(3):517-522.PubMedGoogle ScholarCrossref
20.
Seo  N, Byun  JH, Kim  JH,  et al.  Validation of the 2012 international consensus guidelines using computed tomography and magnetic resonance imaging: branch duct and main duct intraductal papillary mucinous neoplasms of the pancreas.  Ann Surg. 2016;263(3):557-564.PubMedGoogle ScholarCrossref
21.
Rezaee  N, Khalifian  S, Cameron  JL,  et al.  Smoking is not associated with severe dysplasia or invasive carcinoma in resected intraductal papillary mucinous neoplasms.  J Gastrointest Surg. 2015;19(4):656-665.PubMedGoogle ScholarCrossref
22.
Jang  JY, Park  T, Lee  S,  et al.  Validation of international consensus guidelines for the resection of branch duct-type intraductal papillary mucinous neoplasms.  Br J Surg. 2014;101(6):686-692.PubMedGoogle ScholarCrossref
23.
Marchegiani  G, Mino-Kenudson  M, Sahora  K,  et al.  IPMN involving the main pancreatic duct: biology, epidemiology, and long-term outcomes following resection.  Ann Surg. 2015;261(5):976-983.PubMedGoogle ScholarCrossref
24.
Roch  AM, DeWitt  JM, Al-Haddad  MA,  et al.  Nonoperative management of main pancreatic duct-involved intraductal papillary mucinous neoplasm might be indicated in select patients.  J Am Coll Surg. 2014;219(1):122-129.PubMedGoogle ScholarCrossref
25.
Fong  ZV, Ferrone  CR, Lillemoe  KD, Fernández-Del Castillo  C.  Intraductal papillary mucinous neoplasm of the pancreas: current state of the art and ongoing controversies.  Ann Surg. 2016;263(5):908-917.PubMedGoogle ScholarCrossref
26.
Sahora  K, Fernández-del Castillo  C, Dong  F,  et al.  Not all mixed-type intraductal papillary mucinous neoplasms behave like main-duct lesions: implications of minimal involvement of the main pancreatic duct.  Surgery. 2014;156(3):611-621.PubMedGoogle ScholarCrossref
27.
Shimizu  Y, Yamaue  H, Maguchi  H,  et al.  Predictors of malignancy in intraductal papillary mucinous neoplasm of the pancreas: analysis of 310 pancreatic resection patients at multiple high-volume centers.  Pancreas. 2013;42(5):883-888.PubMedGoogle ScholarCrossref
28.
Kawai  M, Uchiyama  K, Tani  M,  et al.  Clinicopathological features of malignant intraductal papillary mucinous tumors of the pancreas: the differential diagnosis from benign entities.  Arch Surg. 2004;139(2):188-192.PubMedGoogle ScholarCrossref
29.
Hirooka  Y, Goto  H, Itoh  A,  et al.  Case of intraductal papillary mucinous tumor in which endosonography-guided fine-needle aspiration biopsy caused dissemination.  J Gastroenterol Hepatol. 2003;18(11):1323-1324.PubMedGoogle ScholarCrossref
30.
Katanuma  A, Maguchi  H, Hashigo  S,  et al.  Tumor seeding after endoscopic ultrasound-guided fine-needle aspiration of cancer in the body of the pancreas.  Endoscopy. 2012;44(suppl 2, UCTN):E160-E161.PubMedGoogle ScholarCrossref
31.
Paquin  SC, Gariépy  G, Lepanto  L, Bourdages  R, Raymond  G, Sahai  AV.  A first report of tumor seeding because of EUS-guided FNA of a pancreatic adenocarcinoma.  Gastrointest Endosc. 2005;61(4):610-611.PubMedGoogle ScholarCrossref
32.
Yamaguchi  K, Nakamura  M, Shirahane  K,  et al.  Pancreatic juice cytology in IPMN of the pancreas.  Pancreatology. 2005;5(4-5):416-421.PubMedGoogle ScholarCrossref
33.
Kang  MJ, Jang  JY, Lee  S, Park  T, Lee  SY, Kim  SW.  Clinicopathological meaning of size of main-duct dilatation in intraductal papillary mucinous neoplasm of pancreas: proposal of a simplified morphological classification based on the investigation on the size of main pancreatic duct.  World J Surg. 2015;39(8):2006-2013.PubMedGoogle ScholarCrossref
34.
Sobin  LH, Gospodarowicz  MK, Wittekind  C, eds.  International Union Against Cancer (UICC): TNM Classification of Malignant Tumors. 7th ed. New York, NY: Wiley Blackwell; 2010.
35.
Moris  M, Raimondo  M, Woodward  TA,  et al.  Risk factors for malignant progression of intraductal papillary mucinous neoplasms.  Dig Liver Dis. 2015;47(6):495-501.PubMedGoogle ScholarCrossref
36.
Matsumoto  K, Takeda  Y, Harada  K, Horie  Y, Yashima  K, Murawaki  Y.  Effect of pancreatic juice cytology and/or endoscopic ultrasound-guided fine-needle aspiration biopsy for pancreatic tumor.  J Gastroenterol Hepatol. 2014;29(1):223-227.PubMedGoogle ScholarCrossref
37.
Ohtsuka  T, Matsunaga  T, Kimura  H,  et al.  Role of pancreatic juice cytology in the preoperative management of intraductal papillary mucinous neoplasm of the pancreas in the era of international consensus guidelines 2012.  World J Surg. 2014;38(11):2994-3001.PubMedGoogle ScholarCrossref
38.
Fritz  S, Hackert  T, Hinz  U, Hartwig  W, Büchler  MW, Werner  J.  Role of serum carbohydrate antigen 19-9 and carcinoembryonic antigen in distinguishing between benign and invasive intraductal papillary mucinous neoplasm of the pancreas.  Br J Surg. 2011;98(1):104-110.PubMedGoogle ScholarCrossref
39.
Kim  JR, Jang  JY, Kang  MJ,  et al.  Clinical implication of serum carcinoembryonic antigen and carbohydrate antigen 19-9 for the prediction of malignancy in intraductal papillary mucinous neoplasm of pancreas.  J Hepatobiliary Pancreat Sci. 2015;22(9):699-707.PubMedGoogle ScholarCrossref
40.
Shimizu  Y, Kanemitsu  Y, Sano  T, Senda  Y, Mizuno  N, Yamao  K.  A nomogram for predicting the probability of carcinoma in patients with intraductal papillary-mucinous neoplasm.  World J Surg. 2010;34(12):2932-2938.PubMedGoogle ScholarCrossref
Original Investigation
March 15, 2017

Factors Associated With Invasive Intraductal Papillary Mucinous Carcinoma of the Pancreas

Author Affiliations
  • 1Second Department of Surgery, School of Medicine, Wakayama Medical University, Wakayama, Japan
  • 2Department of Surgical Pathology, Kyoto Prefectural University of Medicine, Kyoto, Japan
  • 3Department of Pathology, Kyoto First Red Cross Hospital, Kyoto, Japan
JAMA Surg. 2017;152(3):e165054. doi:10.1001/jamasurg.2016.5054
Key Points

Questions  What clinical factors are associated with invasive intraductal papillary mucinous carcinoma?

Findings  In this cohort study with 286 patients, a high mural nodule size for all types of carcinoma and high carcinoembryonic antigen levels in the pancreatic juice for main duct and mixed type carcinoma were found to be independently associated with invasive intraductal papillary carcinoma. These factors had an accuracy of 86% for differentiating between invasive and noninvasive intraductal papillary mucinous neoplasms.

Meaning  The measurement of mural nodule size in all types of carcinoma and carcinoembryonic antigen levels in the pancreatic juice in main duct and mixed type carcinoma might play important roles in predicting invasive intraductal papillary mucinous carcinoma.

Abstract

Importance  Invasive intraductal papillary mucinous carcinoma has aggressive malignant behavior, including lymph node metastasis. It is important to identify factors associated with invasive intraductal papillary mucinous carcinoma to determine operative procedures, but they remain unclear.

Objective  To identify the specific factors associated with invasive intraductal papillary mucinous neoplasms for branch duct, main duct, and mixed type carcinomas.

Design, Setting, and Participants  Retrospective cohort study of 286 consecutive patients who underwent surgical resection for intraductal papillary mucinous neoplasm from July 1999 to December 2015 at a tertiary institute. We compared clinical features between 96 patients (33.6%) with invasive intraductal papillary mucinous carcinoma and 190 patients (66.4%) with noninvasive intraductal papillary mucinous neoplasm for each morphological type.

Main Outcomes and Measures  Factors associated with invasive intraductal papillary mucinous carcinoma. Mural nodule size was measured by endoscopic ultrasonography.

Results  Of the 286 patients included in the cohort, the median (range) age was 71 (28-86) years, and 162 (56.6%) were male. High mural nodule size was independently associated with invasive intraductal papillary mucinous carcinoma in all types (branch duct: odds ratio [OR], 1.992; 95% CI, 1.177-3.367; P =.01; main duct: OR, 1.443; 95% CI, 1.094-1.905; P =.01; and mixed: OR, 1.178; 95% CI, 1.057-1.312; P =.04). The cutoff values for intraductal papillary mucinous neoplasms, determined by a receiver operating characteristic, were 9 mm for branch duct and 6 mm for mixed and main duct carcinoma. A high carcinoembryonic antigen level in the pancreatic juice was independently associated with mixed (OR, 1.002; 95% CI, 1.000-1.003; P =.01) and main duct (OR, 1.002; 95% CI, 1.000-1.003; P =.048) carcinomas, and the cutoff values were determined to be 150 and 300 ng/mL, respectively (to convert to micrograms per liter, multiply by 1). In addition, both being female and having an elevated serum carbohydrate antigen 19-9 level were also found to be independently associated with mixed type invasive intraductal papillary mucinous carcinoma, and using any 2 among 4 identified factors yielded the highest accuracy (79.0%) for mixed type carcinomas. For all types, the accuracy for these factors was 86.0% for differentiating between invasive and noninvasive intraductal papillary mucinous neoplasms, which was superior to the accuracies using the “high-risk stigmata” factors or “worrisome features” suggested by the international consensus guideline in 2012 (66.1% and 39.9%, respectively).

Conclusions and Relevance  The measurement of mural nodule size in all types of carcinomas and carcinoembryonic antigen level in the pancreatic juice in mixed and main duct carcinomas might play important roles in predicting invasive intraductal papillary mucinous carcinoma, but further large studies are needed to confirm these results.

Introduction

Intraductal papillary mucinous neoplasm (IPMN) of the pancreas has a broad histologic spectrum, ranging from low-grade dysplasia (LGD) to invasive intraductal papillary mucinous carcinoma (IPMC).1-3 According to radiographic morphological types, IPMNs are classified as branch duct (BD) type, main duct (MD) type, and mixed type.2,3 The revised international consensus guideline in 20123 classified clinical and radiological findings of IPMNs into “high-risk stigmata,” “worrisome features,” and “low-risk stigmata” categories, based on the potential for malignancy, and it recommended that IPMNs with low-risk stigmata or worrisome features should be treated without immediate surgical resection and that IPMNs with high-risk stigmata should be resected. The American Gastroenterological Association guideline on the diagnosis and management of asymptomatic neoplastic pancreatic cysts suggests that only tumors with at least 2 high-risk factors of malignancy should be considered for resection.4 The number of patients with IPMNs who are observed has increased recently, according to these criteria.5-8

Several reports have shown that invasive IPMC has characteristics similar to those of pancreatic ductal adenocarcinoma (PDAC), including potential lymph node or distant metastasis, postoperative recurrence, and poor survival.9-17 The 2012 international guideline defined only invasive IPMC as malignant IPMN because of its aggressive clinical behavior.3 Therefore, the invasive process of IPMN is a critical issue for understanding this entity, and it is important to be able to predict the risk of invasive IPMC when developing an IPMN treatment strategy.

Prior to the concept of malignancy being changed in the 2012 international guideline,3 most previous reports categorized high-grade dysplasia (HGD) as well as invasive IPMC as malignant IPMN.2,18-21 Additionally, mixed IPMN was often included in BD IPMN or MD IPMN categories during analyses for factors associated with malignancy,18-20,22-24 which frequently caused disagreement and confusion over whether mixed IPMN should be classified as BD IPMN or MD IPMN.5,25-27 The aim of the current study was to identify the factors associated with invasive IPMC by analyzing clinical and imaging findings separately for BD, MD, and mixed IPMNs and to evaluate the accuracy of diagnoses by using identified predictors.

Methods
Patients

Between July 1999 and December 2015, 286 consecutive patients underwent surgical resection for pathologically confirmed IPMN of the pancreas at Wakayama Medical University Hospital. The patients’ demographic information, clinicopathological features, and imaging findings were retrieved from the prospective maintained database of Wakayama Medical University Hospital. Patients with PDAC concomitant with IPMN were excluded from the present study. This study was approved by the Institutional Ethics Review Board of Wakayama Medical University Hospital, and the requirement for informed consent was waived.

Preoperative Examinations

All patients underwent clinical evaluation, laboratory testing, including serum tumor marker levels (carcinoembryonic antigen [CEA] and carbohydrate antigen 19-9 [CA19-9]), ultrasonography, computed tomography (CT), and endoscopic ultrasonography before surgery. The maximum branch duct cyst size and main pancreatic duct (MPD) diameter were determined by CT, and the mural nodule size was determined by endoscopic ultrasonography. Pancreatic juice obtained from the MPD by endoscopic retrograde pancreatography (ERP) was used for pancreatic juice cytology in 237 patients and the measurement of CEA levels in 223 patients, as reported previously.12,19,28 The cystic fluid was not collected by endoscopic ultrasonography–guided fine needle aspiration for cytology or measurement of biomarkers because of concerns about the possibility of seeding tumor cells along the needle track.29-31 Pancreatic juice cytology was classified into classes I, II, III, IV, and V in accordance with the grade of structural and cytological dysplasia,32 and classes IV and V were defined as positive cytology in this study.

Based on preoperative imaging analyses, we classified tumors as BD, MD, or mixed IPMN. Branch duct IPMN was defined as the presence of a dilated branch duct (≥10 mm) communicating with a nondilated MPD (<5 mm), MD IPMN was defined as segmental or diffuse dilation of MPD (≥5 mm) with no dilation of branch ducts (<10 mm), and mixed IPMN was defined as the presence of a dilated MPD (≥5 mm) communicating with a dilated branch duct (≥10 mm).3,5

Surgical Indications

Surgical indications of Wakayama Medical University Hospital were as follows: (1) presence of symptom; (2) MD IPMN; (3) class IV or V from the pancreatic juice cytology; (4) MPD size greater than 7 mm33; (5) gradual increase in the BD cyst size (total growth ≥10 mm) during follow-up7; (6) BD or mixed IPMN with a mural nodule size greater than 5 mm19; and (7) BD or mixed IPMN with a CEA level in pancreatic juice greater than 30 ng/mL (to convert to micrograms per liter, multiply by 1).19

Pathological Diagnosis

Pancreas specimens were serially sectioned at 5-mm intervals, and whole slides were reviewed and graded by a specialized pathologist (A.Y.) blinded to clinical outcomes. The grades were LGD, intermediate-grade dysplasia (IGD), HGD, or invasive IPMC, based on the degree of cytoarchitectural dysplasia in accordance with the World Health Organization classification system published in 2010.1 When more than 1 pathologic type was present in the same patient, the highest degree of dysplasia was recorded. Invasive IPMC was defined as the presence of a continuous invasive component from HGD in pathological findings to distinguish it from PDAC concomitant with IPMN. The type of invasive IPMC was classified as tubular or colloid based on differentiations in the invasive components.3 The invasive IPMC tumors were staged according to the Tumor, Node, and Metastasis Classification of Malignant Tumors published by the International Union against Cancer,34 and thereafter, the substage of the T1 category of invasive IPMC was classified as T1a (invasive component ≤5 mm), T1b (invasive component >5mm and ≤10 mm), and T1c (invasive component >10 mm and ≤20 mm), according to the revised international consensus guidelines in 2012.3

Statistical Analysis

We compared the demographic information, tumor markers, and radiographic and endoscopic findings between patients with noninvasive IPMN (LGD, IGD, and HGD) and those with invasive IPMC for each morphological type of carcinoma to identify factors associated with invasive IPMC. Estimates of central tendency were reported as median values. Categorical variables were compared by the χ2 test or Fisher exact test. Continuous variables were compared using the Mann-Whitney U test or paired Student t test. A multivariate analysis was performed by using a multiple logistic regression model, which included factors with P < .1 in a univariate analysis. A receiver operating characteristic was used to determine cutoff values for the mural nodule size and the CEA level in the pancreatic juice, and the cutoff values were determined to maximize the difference between noninvasive IMPN and invasive IPMC by receiver operating characteristic curves. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy were calculated to evaluate the diagnostic value. All statistical analyses were performed with SPSS version 20 (IBM Corp). Statistical significance was set at P < .05.

Results
Patient Characteristics

The demographic information, clinicopathological features, and imaging findings of the 286 patients who underwent surgical resection for IPMN are shown in Table 1. Their median (range) age was 71 (28-86) years, and 162 patients (56.6%) were male. Median (range) postoperative follow-up was 51.4 (0.5-203.7) months. With regard to radiographic morphological types, 109 patients (38.1%) had BD IPMN, 63 (22.0%) had MD IPMN, and 114 (39.9%) had mixed IPMN. Total pancreatectomy was performed in 26 patients (9.1%), pancreaticoduodenectomy in 185 (64.7%), distal pancreatectomy in 54 (18.9%), and central pancreatectomy in 21 (7.3%). Final pathological diagnosis showed that 94 patients (32.9%) had LGD or IGD, 96 (33.6%) had HGD, and 96 (33.6%) had invasive IPMC.

Regarding the type of invasive IPMC, 71 patients (74.0%) had tubular type and 25 patients (26.0%) had colloid type. Among 96 patients with invasive IPMC, 35 patients (36.5%) had lymph node metastasis, including 1 T1a patient.

Factors Associated With Invasive IPMC in Patients in Univariate and Multivariate Analyses

We separately analyzed factors associated with invasive IPMC for patients with BD IPMN, MD IPMN, and mixed IPMN. For patients with BD IPMN, 6 predictors of invasive IPMC were found on univariate analysis (Table 2): presence of symptoms, elevated serum CA19-9 levels, large branch cyst size, presence of mural nodules, high mural nodule size, and high CEA levels in the pancreatic juice. However, high mural nodule size was found to be the only factor independently associated with invasive IPMC for patients with BD IPMN on multivariate analysis (odds ratio [OR], 1.992; 95% CI, 1.177-3.367; P = .01) (Table 2).

For patients with mixed IPMN, 8 factors associated with invasive IPMC were found on univariate analysis (Table 3): female sex, presence of symptoms, elevated serum CEA levels, elevated serum CA19-9 levels, presence of mural nodule, high mural nodule size, high CEA levels in the pancreatic juice, and positive pancreatic juice cytology. Furthermore, 4 factors independently associated with invasive IPMC were found on multivariate analysis (Table 3): female sex (OR, 6.135; 95% CI, 1.454-25.641; P = .01), elevated serum CA19-9 levels (OR, 29.412; 95% CI, 2.817-333.333; P = .009), high mural nodule size (OR, 1.178; 95% CI, 1.057-1.312; P = .04), and high CEA levels in the pancreatic juice (OR, 1.002; 95% CI, 1.000-1.003; P = .011).

For patients with MD IPMN, 7 factors associated with invasive IPMC were found on univariate analysis (Table 4): history of pancreatitis, diffuse MPD dilatation in the entire pancreas, large MPD diameter, presence of mural nodules, high mural nodule size, high CEA levels in the pancreatic juice, and positive pancreatic juice cytology. Furthermore, high mural nodule size and high CEA levels in the pancreatic juice were found to be independently associated with invasive IPMC on multivariate analysis (mural nodule size: OR, 1.443; 95% CI, 1.094-1.905; P = .010; CEA in the pancreatic juice: OR, 1.002; 95% CI, 1.000-1.003; P = .048) (Table 4).

Diagnostic Cutoff Values for Mural Nodule Size and CEA Levels in the Pancreatic Juice

The cutoff levels for mural nodule size in BD, MD, and mixed IPMN and for CEA in the pancreatic juice in MD and mixed IPMN, which were identified as independently associated with invasive IPMC, were determined by using receiver operating characteristic curves to differentiate between invasive IPMC and noninvasive IPMN. Mathematically, the cutoff values were defined as those corresponding to points on the receiver operating characteristic curves situated furthest from the reference line.

The area under the curve for mural nodule size in patients with BD IPMN was 0.926 (P < .001), and 9 mm was determined to be the cutoff level for differentiating between invasive IPMC and noninvasive IPMN. The area under the curve value for mural nodule size and CEA level in the pancreatic juice in patients with mixed IPMN were 0.760 (P < .001) and 0.796 (P < .001), respectively, and the cutoff levels were determined to be 6 mm and 150 ng/mL, respectively. The area under the curve values for mural nodule size and CEA level in the pancreatic juice in patients with MD IPMN were 0.788 (P < .001) and 0.877 (P < .001), respectively, and the cutoff levels were 6 mm and 300 ng/mL, respectively.

Diagnostic Values for Factors Associated With Invasive IPMC

To evaluate the diagnostic values of the factors associated with invasive IPMC identified in this study, sensitivity, specificity, PPV, NPV, and accuracy were investigated (Table 5). Branch duct IPMN with a mural nodule size greater than 9 mm was associated with invasive IPMC, with a sensitivity of 86.4%, specificity of 96.6%, PPV of 86.4%, NPV of 96.6%, and accuracy of 94.5%. In combination analysis of 4 independent factors associated with mixed invasive IPMC, including female sex, elevated serum CA19-9 levels, mural nodule size greater than 6 mm, and a CEA levels greater than 150 ng/mL in the pancreatic juice, an increase in the number of factors increased the specificity and PPV while decreasing the sensitivity, and the accuracy of predicting mixed invasive IPMC was highest with any 2 factors (79.0%) (Table 5). Main duct IPMN with a mural nodule size greater than 6 mm or CEA level greater than 300 ng/mL in the pancreatic juice was associated with invasive IPMC, with a sensitivity of 95.7%, specificity of 77.5%, PPV of 71.0%, NPV of 96.9%, and accuracy of 84.1%. When taken together, BD IPMN with a mural nodule size greater than 9 mm, mixed IPMN with any 2 of 4 predictive factors, and MD IPMN with a mural nodule size greater than 6 mm or a CEA level greater than 300 ng/mL in the pancreatic juice were associated with invasive IPMC, with a sensitivity of 89.6%, specificity of 84.2%, PPV of 74.1%, NPV of 94.1%, and accuracy of 86.0% (Table 5).

When the ability to predict invasive IPMC was analyzed by using high-risk stigmata or worrisome features, as suggested by the 2012 international guideline,3 a sensitivity of 85.4%, specificity of 56.3%, PPV of 49.7%, NPV of 88.4%, and accuracy of 66.1% were found for high-risk stigmata factors, including mural nodule with enhancement, MPD diameter of 10 mm or greater, and obstructive jaundice. A sensitivity of 99.0%, specificity of 10.0%, PPV of 35.7%, NPV of 95.0%, and accuracy of 39.9% were found for worrisome features, including a history of pancreatitis, branch duct cyst size of 30 mm or greater, thickened and enhancing cyst wall, an MPD diameter of 5 to 9 mm, nonenhancing mural nodule, and abrupt change in caliber of the MPD with distal pancreatic atrophy.

For 156 patients with asymptomatic IPMNs, a sensitivity of 86.1%, specificity of 84.2%, PPV of 62.0%, NPV of 95.3%, and accuracy of 84.6% were found by using of the factors associated with invasive IPMC identified in this study. When the ability to predict asymptomatic invasive IPMC was analyzed by using at least 2 high-risk features, including a branch duct cyst size of 30 mm or greater, dilated MPD (>5 mm), and solid component, as suggested by the American Gastroenterological Association guideline,4 a sensitivity of 83.3%, specificity of 63.3%, PPV of 40.5%, NPV of 92.7%, and accuracy of 68.0% were found.

Discussion

In general, the malignant behavior of IPMN may be considered to be lower than that of PDAC.6,12 Indeed, survival after surgical resection for patients with noninvasive IPMN is excellent.10,12 In fact, HGD is the optimal situation for operation at a point when it has the optimal prognosis/outcome. Analysis including patients with HGD are necessary to provide information that is practically useful. However, 36.5% of patients with invasive IPMC had lymph node metastasis, including 1 patient with T1a (invasive component ≤5 mm) in this series. Therefore, this study has focused only on the factors associated with invasive IPMC; similar treatment should be mandatory for patients with PDAC if the tumor is diagnosed with invasive IPMC.9,10 Thus, it is important to predict invasive IPMC to determine the treatment strategy for patients with IPMN.

High mural nodule size measured by endoscopic ultrasonography was found to be significantly associated with invasive IPMC for all morphological types, which is consistent with a malignant predictor shown in several reports.19,20,27,35 High accuracies of invasive IPMC were found in all types of carcinomas by using the determined cutoff levels of 9 mm in patients with BD IPMN and 6 mm in patients with MD IPMN and mixed IPMN for mural nodule size in this study. Therefore, immediate surgical resection, with the same lymph node dissection as for PDAC, might be necessary in these patients. However, some invasive IPMC did not have a higher mural nodule than these cutoff levels, indicating that it might not be enough to determine whether the tumor should be resected or observed only by measurement of mural nodule size. Therefore, more useful combinations of other factors, such as new biomarkers and mural nodule size, should be developed by further studies to have a higher accuracy of invasive IPMC.

A high CEA level in the pancreatic juice obtained from the MPD by ERP was independently associated with MD and mixed invasive IPMCs in this study. The international guideline from 20123 did not recommend routine ERP for collecting pancreatic juice during evaluation of the malignant potential of IPMN because the procedure is invasive and could cause post-ERP pancreatitis. In the present study, 2 patients (0.8%) had post-ERP pancreatitis, who presented symptoms and required conservative treatment. The measurement of the CEA level in the pancreatic juice might have important roles in determining the surgical indications and operative procedures for patients with IPMN, as we reported previously.12,19,28 It might be implied that the rate of HGD or invasive IPMC was high in this series compared with other reports, and this factor was found to be independently associated with MD and mixed invasive IPMCs in this study. However, regarding BD IPMN, although the CEA level in the pancreatic juice was found to be associated with invasive IPMC in the univariate analysis, it was dropped after the multiple regression analysis in this study. Further, ERP is indeed an invasive procedure, as many studies have indicated.4,36,37 Therefore, ERP should be performed only as a diagnostic adjunct, not as a routine examination.

Serum CA19-9 level is a well-known predictor of PDAC and poor survival. Elevated serum CA19-9 level was found to be associated with mixed invasive IPMC, which is consistent with several previous reports,38,39 and this finding might imply that invasive IPMC has characteristics similar to PDAC. Female sex was also associated with mixed invasive IPMC. Several reports showed that female sex was a malignancy predictor for IPMN,27,40 although further large studies are necessary to confirm this result.

When all types of IPMN were taken together, using factors associated with invasive IPMC for each, a high diagnostic accuracy of 86.0% was found in this study. When the diagnostic ability of invasive IPMC was analyzed using high-risk stigmata or worrisome features, as suggested by the 2012 international guideline,3 lower accuracies were found than the identified predictors in this study. Furthermore, for patients with asymptomatic IPMN, a higher accuracy was found by using the identified predictors in this study than by using high-risk factors of malignancy suggested by the American Gastroenterological Association guidelines.4 Thus, the predictors indicated in this study might be superior to those of the 2012 international guideline and the American Gastroenterological Association guideline.

Limitations

This study has several limitations. The study included a retrospective cohort at a single institution and had a small sample size. Therefore, selection bias may have occurred, particularly with respect to operative indications and variations in preoperative examinations and operations between physicians, surgeons, and institutional characteristics. The measurement of the CEA level in the pancreatic juice by ERP is not universally used, and therefore, it is difficult to confirm and use these data worldwide. Furthermore, a cohort of patients with continuously observed IPMN was not used, so the natural history of IPMN could not be assessed.

Conclusions

In this study, the specific factors associated with each morphological type of invasive IPMC were analyzed. High mural nodule size was independently associated with all types of carcinomas, and a high CEA level in the pancreatic juice was independently associated with MD and mixed invasive IPMCs. A high diagnostic accuracy of invasive IPMC was found using identified factors, although further large studies are necessary to confirm these results.

Back to top
Article Information

Corresponding Author: Hiroki Yamaue, MD, Second Department of Surgery, School of Medicine, Wakayama Medical University, 811-1 Kimiidera, Wakayama 641-8510, Japan (yamaue-h@wakayama-med.ac.jp).

Accepted for Publication: November 1, 2016.

Published Online: January 25, 2017. doi:10.1001/jamasurg.2016.5054

Author Contributions: Drs Yamaue and Hirono had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Hirono, Kitahata, Yamaue.

Acquisition, analysis, or interpretation of data: Hirono, Kawai, Okada, Miyazawa, Shimizu, Ueno, Yanagisawa, Yamaue.

Drafting of the manuscript: Hirono, Shimizu, Yanagisawa.

Critical revision of the manuscript for important intellectual content: Hirono, Kawai, Okada, Miyazawa, Kitahata, Ueno, Yamaue.

Statistical analysis: Hirono, Ueno.

Administrative, technical, or material support: Kawai, Okada, Miyazawa, Shimizu, Kitahata, Ueno, Yanagisawa.

Study Supervision: Yamaue.

Conflict of Interest Disclosures: None reported.

References
1.
Adsay  V, Fukuhima  N, Furukawa  T. Intraductal neoplasm of the pancreas. In: Bosman  FT, Carnerio  F, Hruban  RH, eds.  WHO Classification of Tumors of the Digestive System. Lyon, France: IARC Press; 2010:304-313.
2.
Tanaka  M, Chari  S, Adsay  V,  et al; International Association of Pancreatology.  International consensus guidelines for management of intraductal papillary mucinous neoplasms and mucinous cystic neoplasms of the pancreas.  Pancreatology. 2006;6(1-2):17-32.PubMedGoogle ScholarCrossref
3.
Tanaka  M, Fernández-del Castillo  C, Adsay  V,  et al; International Association of Pancreatology.  International consensus guidelines 2012 for the management of IPMN and MCN of the pancreas.  Pancreatology. 2012;12(3):183-197.PubMedGoogle ScholarCrossref
4.
Vege  SS, Ziring  B, Jain  R, Moayyedi  P; Clinical Guidelines Committee; American Gastroenterology Association.  American gastroenterological association institute guideline on the diagnosis and management of asymptomatic neoplastic pancreatic cysts.  Gastroenterology. 2015;148(4):819-822, e12-e13.PubMedGoogle ScholarCrossref
5.
Crippa  S, Bassi  C, Salvia  R,  et al.  Low progression of intraductal papillary mucinous neoplasms with worrisome features and high-risk stigmata undergoing non-operative management: a mid-term follow-up analysis  [published online January 7, 2016].  Gut. doi:10.1136/gutjnl_2015_310162PubMedGoogle Scholar
6.
Roch  AM, Ceppa  EP, Al-Haddad  MA,  et al.  The natural history of main duct-involved, mixed-type intraductal papillary mucinous neoplasm: parameters predictive of progression.  Ann Surg. 2014;260(4):680-688.PubMedGoogle ScholarCrossref
7.
Kwong  WT, Lawson  RD, Hunt  G,  et al.  Rapid growth rates of suspected pancreatic cyst branch duct intraductal papillary mucinous neoplasms predict malignancy.  Dig Dis Sci. 2015;60(9):2800-2806.PubMedGoogle ScholarCrossref
8.
Salvia  R, Crippa  S, Falconi  M,  et al.  Branch-duct intraductal papillary mucinous neoplasms of the pancreas: to operate or not to operate?  Gut. 2007;56(8):1086-1090.PubMedGoogle ScholarCrossref
9.
McMillan  MT, Lewis  RS, Drebin  JA,  et al.  The efficacy of adjuvant therapy for pancreatic invasive intraductal papillary mucinous neoplasm (IPMN).  Cancer. 2016;122(4):521-533.PubMedGoogle ScholarCrossref
10.
Hirono  S, Kawai  M, Okada  K,  et al.  Long-term surveillance is necessary after operative resection for intraductal papillary mucinous neoplasm of the pancreas.  Surgery. 2016;160(2):306-317.PubMedGoogle ScholarCrossref
11.
Kang  MJ, Jang  JY, Lee  KB, Chang  YR, Kwon  W, Kim  SW.  Long-term prospective cohort study of patients undergoing pancreatectomy for intraductal papillary mucinous neoplasm of the pancreas: implications for postoperative surveillance.  Ann Surg. 2014;260(2):356-363.PubMedGoogle ScholarCrossref
12.
Hirono  S, Tani  M, Kawai  M,  et al.  Treatment strategy for intraductal papillary mucinous neoplasm of the pancreas based on malignant predictive factors.  Arch Surg. 2009;144(4):345-349.PubMedGoogle ScholarCrossref
13.
Winter  JM, Jiang  W, Basturk  O,  et al.  Recurrence and survival after resection of small intraductal papillary mucinous neoplasm-associated carcinomas (≤20-mm invasive component): a multi-institutional analysis.  Ann Surg. 2016;263(4):793-801.PubMedGoogle ScholarCrossref
14.
Waters  JA, Schnelldorfer  T, Aguilar-Saavedra  JR,  et al.  Survival after resection for invasive intraductal papillary mucinous neoplasm and for pancreatic adenocarcinoma: a multi-institutional comparison according to American Joint Committee on Cancer Stage.  J Am Coll Surg. 2011;213(2):275-283.PubMedGoogle ScholarCrossref
15.
Wasif  N, Bentrem  DJ, Farrell  JJ,  et al.  Invasive intraductal papillary mucinous neoplasm versus sporadic pancreatic adenocarcinoma: a stage-matched comparison of outcomes.  Cancer. 2010;116(14):3369-3377.PubMedGoogle ScholarCrossref
16.
Yopp  AC, Katabi  N, Janakos  M,  et al.  Invasive carcinoma arising in intraductal papillary mucinous neoplasms of the pancreas: a matched control study with conventional pancreatic ductal adenocarcinoma.  Ann Surg. 2011;253(5):968-974.PubMedGoogle ScholarCrossref
17.
Poultsides  GA, Reddy  S, Cameron  JL,  et al.  Histopathologic basis for the favorable survival after resection of intraductal papillary mucinous neoplasm-associated invasive adenocarcinoma of the pancreas.  Ann Surg. 2010;251(3):470-476.PubMedGoogle ScholarCrossref
18.
Hackert  T, Fritz  S, Klauss  M,  et al.  Main-duct intraductal papillary mucinous neoplasm: high cancer risk in duct diameter of 5 to 9 mm.  Ann Surg. 2015;262(5):875-880.PubMedGoogle ScholarCrossref
19.
Hirono  S, Tani  M, Kawai  M,  et al.  The carcinoembryonic antigen level in pancreatic juice and mural nodule size are predictors of malignancy for branch duct type intraductal papillary mucinous neoplasms of the pancreas.  Ann Surg. 2012;255(3):517-522.PubMedGoogle ScholarCrossref
20.
Seo  N, Byun  JH, Kim  JH,  et al.  Validation of the 2012 international consensus guidelines using computed tomography and magnetic resonance imaging: branch duct and main duct intraductal papillary mucinous neoplasms of the pancreas.  Ann Surg. 2016;263(3):557-564.PubMedGoogle ScholarCrossref
21.
Rezaee  N, Khalifian  S, Cameron  JL,  et al.  Smoking is not associated with severe dysplasia or invasive carcinoma in resected intraductal papillary mucinous neoplasms.  J Gastrointest Surg. 2015;19(4):656-665.PubMedGoogle ScholarCrossref
22.
Jang  JY, Park  T, Lee  S,  et al.  Validation of international consensus guidelines for the resection of branch duct-type intraductal papillary mucinous neoplasms.  Br J Surg. 2014;101(6):686-692.PubMedGoogle ScholarCrossref
23.
Marchegiani  G, Mino-Kenudson  M, Sahora  K,  et al.  IPMN involving the main pancreatic duct: biology, epidemiology, and long-term outcomes following resection.  Ann Surg. 2015;261(5):976-983.PubMedGoogle ScholarCrossref
24.
Roch  AM, DeWitt  JM, Al-Haddad  MA,  et al.  Nonoperative management of main pancreatic duct-involved intraductal papillary mucinous neoplasm might be indicated in select patients.  J Am Coll Surg. 2014;219(1):122-129.PubMedGoogle ScholarCrossref
25.
Fong  ZV, Ferrone  CR, Lillemoe  KD, Fernández-Del Castillo  C.  Intraductal papillary mucinous neoplasm of the pancreas: current state of the art and ongoing controversies.  Ann Surg. 2016;263(5):908-917.PubMedGoogle ScholarCrossref
26.
Sahora  K, Fernández-del Castillo  C, Dong  F,  et al.  Not all mixed-type intraductal papillary mucinous neoplasms behave like main-duct lesions: implications of minimal involvement of the main pancreatic duct.  Surgery. 2014;156(3):611-621.PubMedGoogle ScholarCrossref
27.
Shimizu  Y, Yamaue  H, Maguchi  H,  et al.  Predictors of malignancy in intraductal papillary mucinous neoplasm of the pancreas: analysis of 310 pancreatic resection patients at multiple high-volume centers.  Pancreas. 2013;42(5):883-888.PubMedGoogle ScholarCrossref
28.
Kawai  M, Uchiyama  K, Tani  M,  et al.  Clinicopathological features of malignant intraductal papillary mucinous tumors of the pancreas: the differential diagnosis from benign entities.  Arch Surg. 2004;139(2):188-192.PubMedGoogle ScholarCrossref
29.
Hirooka  Y, Goto  H, Itoh  A,  et al.  Case of intraductal papillary mucinous tumor in which endosonography-guided fine-needle aspiration biopsy caused dissemination.  J Gastroenterol Hepatol. 2003;18(11):1323-1324.PubMedGoogle ScholarCrossref
30.
Katanuma  A, Maguchi  H, Hashigo  S,  et al.  Tumor seeding after endoscopic ultrasound-guided fine-needle aspiration of cancer in the body of the pancreas.  Endoscopy. 2012;44(suppl 2, UCTN):E160-E161.PubMedGoogle ScholarCrossref
31.
Paquin  SC, Gariépy  G, Lepanto  L, Bourdages  R, Raymond  G, Sahai  AV.  A first report of tumor seeding because of EUS-guided FNA of a pancreatic adenocarcinoma.  Gastrointest Endosc. 2005;61(4):610-611.PubMedGoogle ScholarCrossref
32.
Yamaguchi  K, Nakamura  M, Shirahane  K,  et al.  Pancreatic juice cytology in IPMN of the pancreas.  Pancreatology. 2005;5(4-5):416-421.PubMedGoogle ScholarCrossref
33.
Kang  MJ, Jang  JY, Lee  S, Park  T, Lee  SY, Kim  SW.  Clinicopathological meaning of size of main-duct dilatation in intraductal papillary mucinous neoplasm of pancreas: proposal of a simplified morphological classification based on the investigation on the size of main pancreatic duct.  World J Surg. 2015;39(8):2006-2013.PubMedGoogle ScholarCrossref
34.
Sobin  LH, Gospodarowicz  MK, Wittekind  C, eds.  International Union Against Cancer (UICC): TNM Classification of Malignant Tumors. 7th ed. New York, NY: Wiley Blackwell; 2010.
35.
Moris  M, Raimondo  M, Woodward  TA,  et al.  Risk factors for malignant progression of intraductal papillary mucinous neoplasms.  Dig Liver Dis. 2015;47(6):495-501.PubMedGoogle ScholarCrossref
36.
Matsumoto  K, Takeda  Y, Harada  K, Horie  Y, Yashima  K, Murawaki  Y.  Effect of pancreatic juice cytology and/or endoscopic ultrasound-guided fine-needle aspiration biopsy for pancreatic tumor.  J Gastroenterol Hepatol. 2014;29(1):223-227.PubMedGoogle ScholarCrossref
37.
Ohtsuka  T, Matsunaga  T, Kimura  H,  et al.  Role of pancreatic juice cytology in the preoperative management of intraductal papillary mucinous neoplasm of the pancreas in the era of international consensus guidelines 2012.  World J Surg. 2014;38(11):2994-3001.PubMedGoogle ScholarCrossref
38.
Fritz  S, Hackert  T, Hinz  U, Hartwig  W, Büchler  MW, Werner  J.  Role of serum carbohydrate antigen 19-9 and carcinoembryonic antigen in distinguishing between benign and invasive intraductal papillary mucinous neoplasm of the pancreas.  Br J Surg. 2011;98(1):104-110.PubMedGoogle ScholarCrossref
39.
Kim  JR, Jang  JY, Kang  MJ,  et al.  Clinical implication of serum carcinoembryonic antigen and carbohydrate antigen 19-9 for the prediction of malignancy in intraductal papillary mucinous neoplasm of pancreas.  J Hepatobiliary Pancreat Sci. 2015;22(9):699-707.PubMedGoogle ScholarCrossref
40.
Shimizu  Y, Kanemitsu  Y, Sano  T, Senda  Y, Mizuno  N, Yamao  K.  A nomogram for predicting the probability of carcinoma in patients with intraductal papillary-mucinous neoplasm.  World J Surg. 2010;34(12):2932-2938.PubMedGoogle ScholarCrossref
×