572 Intraductal Papillary Mucinous Neoplasms of the Pancreas
Intraductal Papillary
Mucinous Neoplasms of
the Pancreas
Francesca M. Dimou, MD, Jennifer A. Perone, MD,
and Taylor S. Riall, MD, PhD
I nitially described by Ohashi and colleagues in 1982, intraductal
papillary mucinous neoplasms (IPMNs) were not recognized for-
mally as distinct lesions from mucinous cystic neoplasms (MCNs)
until 1999. Over the past few decades the incidence of IPMNs has
increased, likely secondary to both increased frequency of abdominal
imaging and clearly defined diagnostic criteria. IPMNs are a group
of intraductal pancreatic epithelial neoplasms characterized by (1)
mucin production, and (2) diffuse or segmental involvement of the
main pancreatic duct or major side branches. IPMNs lack the ovarian
stroma characteristic of MCNs, differentiating them pathologically
from these entities.
IPMNs represent a spectrum of epithelial changes, similar to the
evolution of pancreatic intraepithelial neoplasias (PanINs) to inva-
sive pancreatic adenocarcinoma. However, unlike PanINs, these
lesions can be identified grossly and radiographically before the
development of invasive cancer, providing an opportunity for early
intervention analogous to the removal of colonic adenomatous
polyps. The management of IPMNs is complicated by the fact that
recurrences have been documented in the remnant pancreas after
removal of both benign and malignant lesions. This biologic behav-
ior suggests a field defect, with increased risk of neoplastic transfor-
mation in all pancreatic ductal epithelium. Further complicating
factors include uncertainty surrounding the time for progression to
malignancy, the differing malignant potential based on the anatomic
and histologic characteristics of the IPMN, and the significant mor-
bidity and mortality associated with pancreatic resections. The great-
est controversy involves the appropriate timing and indications for
resection and the criteria for safe observation of IPMNs with lower
malignant potential. When resection is indicated, controversy
remains regarding the extent of resection in the setting of residual
IPMN at the pancreatic margin and surveillance of the pancreatic
remnant.
CLASSIFICATION
Anatomic Classification
IPMNs are classified as main duct (MD-IPMN), branch duct (BD-
IPMN), or mixed type IPMNs with both main and branch duct
components. MD-IPMNs are characterized by diffuse or segmental
involvement of the main pancreatic duct, with radiographic findings
of main pancreatic duct dilation greater than 5 mm without any
other causes of obstruction (Figure 1, A). Previously, criteria for
MD-IPMN had a cutoff of main duct dilation greater than 10 mm.
However, in developing the most current International Consensus
Guidelines (2012) for the management of IPMN and MCN of the
pancreas, expert review of available data concluded that using greater
than 5 mm as the diagnostic criteria increased the sensitivity for
radiologic diagnosis without losing specificity. The greatest concern
with MD-IPMN is risk of high-grade dysplasia (carcinoma in situ)
or invasive carcinoma; in retrospective case series of resected IPMNs,
the rate of high-grade dysplasia or invasive cancer ranged between
35% and 100%, regardless of symptomatology (Table 1).
BD-IPMNs (Figure 1, B) involve the pancreatic duct side branches
but not the main pancreatic duct, and typically occur in younger
patients. They are more common than MD-IPMNs, at a rate of 10:1.
BD-IPMNs can occur anywhere within the pancreas and 40% to 60%
Downloaded for Paul Yoffe (payoffe@montefiore.org) at Montefiore Medical Center - JCon from ClinicalKey.com by Elsevier on May 26, 2018.
For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.
are multifocal. Despite the high incidence of multifocal disease, there
is no evidence that an increased number of lesions amplifies the risk
of invasive disease. BD-IPMNs are thought to have a lower risk of
malignant transformation. Because of this lower risk of malignancy,
BD-IPMNs are more commonly observed, but in resected series the
rate of malignancy ranged from 6% to 56%. The overall rate of
malignancy in BD-IPMNs is likely much lower, as those with con-
cerning features are more likely to be removed.
Mixed type IPMNs are characterized by the involvement of both
the main and branch pancreatic ducts and behave clinically and
pathologically like MD-IPMNs (Figure 1, C).
Histologic Classification
IPMNs are composed of mucin-producing columnar cells, which
show papillary proliferation, cyst formation, and varying degrees of
cellular atypia. Tumors are graded according to the most atypical area
in the lesion, following an orderly progression from low-grade dys-
plasia to moderate dysplasia to high-grade dysplasia and finally to
invasive carcinoma. The 2012 International Consensus Guidelines
now favor using the terms low-grade, moderate, and high-grade dys-
plasia in lieu of the previously used terms IPMN adenoma, borderline
IPMN, and carcinoma in situ.
Four histologic subtypes of IPMNs have been identified: gastric,
intestinal, pancreatobiliary, and oncocytic (Table 2; Figure 2). All can
have increasing degrees of dysplasia. Within the four subtypes there
is increasing understanding of the proclivity of each subtype to
develop into invasive disease, the type of invasive malignancy that
develops, and the varying disease prognosis.
The gastric subtype occurs primarily in BD-IPMN and is the most
common subtype overall (in keeping with the greater incidence of
BD-IPMN than MD-IPMN). These neoplasms are found in the
periphery of the pancreatic parenchyma and in the uncinate process.
Histologically, they resemble gastric foveolar cells and form pyloric
gland-like structures at the base of the papillae (see Figure 2, A).
Based on immunohistochemistry, the gastric subtype expresses
mucin 5AC (MUC5AC) and mucin 6 (MUC6) proteins, with scat-
tered expression of mucin 2 (MUC2) and no expression of mucin 1
(MUC1). When compared with oncocytic and pancreatobiliary sub-
types, the gastric subtype was found to have a superior 5-year
survival rate and a lower likelihood of recurrent disease. They
are typically low grade, with only 10% to 30% developing into
invasive disease. However, once invasive disease develops, the overall
prognosis is poor.
The intestinal subtype is the second most common type of IPMN
and the most common type of MD-IPMN. As is characteristic of
MD-IPMNs, these neoplasms are typically found in the pancreatic
head but can involve the entire duct. They have a characteristic
villous growth pattern and express MUC2, MUC5AC, and caudal-
type homeobox 2 (CDX2). Thirty to fifty percent of patients with the
intestinal subtype will develop invasive malignancy, which is a muci-
nous (colloid) carcinoma that behaves in a relatively indolent fashion
when compared with tubular carcinoma.
The pancreatobiliary subtype also typically involves the main duct
and is found in the pancreatic head. It is regarded by some as a high-
grade version of the gastric type of IPMN, histologically character-
ized by complex papillae and cells that resemble pancreatic and
biliary duct cells. This subtype expresses MUC1 and MUC5AC.
Among all of the histologic subtypes, it is associated with the greatest
likelihood of malignancy, with more than 50% of patients developing
invasive disease. The invasive form of this subtype is histologically
indistinguishable from a conventional ductal (tubular) adenocarci-
noma and is differentiated by the presence of noninvasive IPMN in
the specimen. The pancreaticobiliary type is associated with a high
recurrence rate and an overall poor prognosis.
The oncocytic subtype, commonly misdiagnosed as cystadeno-
carcinoma, is a rare subtype found in the main pancreatic duct. This
 T he Pa n c r e a s 573

A B C

FIGURE 1  A, Sagittal multidetector computed tomographic (MDCT) image showing main duct intraductal papillary mucinous neoplasm (MD-IPMN)
with notable dilatation (arrow) of the main pancreatic duct. B, Axial MDCT image showing a small cystic lesion in the pancreatic tail (arrow) consistent
with branch duct IPMN (BD-IPMN). No dilation is observed in the main pancreatic duct. C, Mixed-type IPMN on CT scan. There is notable diffuse
dilatation of the pancreatic duct in the head and body of the gland (arrowheads) and multiple small branch duct lesions throughout the body and
tail (arrows).

TABLE 1:  Studies Investigating Malignancy in Main Duct and Branch Duct Intraductal Papillary Mucinous
Neoplasms (IPMNs) Before and After Establishment of the Sendai Criteria
Main Duct Branch Duct
First Author Year Total N Malignant* N (%) N Malignant* N (%) Additional Findings
Sugiyama 2003 62 30 21 (70.0) 32 13 (40.6) Mural nodules and MPD ≥7 mm associated with
malignancy
Salvia 2004 140 140 83 (59.3) Jaundice and diabetes significantly associated with
malignancy
Sohn 2004 136 36 18 (50.0) 60 18 (30.0) Lymph node status predictive of survival for invasive
disease in univariate analysis
Suzuki 2004 1024 201 120 (59.7) 509 150 (29.4) Malignant disease more common in older patients, main
duct disease, enlarged papilla orifice, and symptomatic
older patients who were symptomatic, had main duct
disease, or have an enlarged papilla orifice
Lee 2005 67 27 12 (44.4) 35 10 (28.6)
Serikawa 2006 103 47 30 (63.8) 56 11 (19.6)
Schmidt 2007 156 53 30 (56.6) 103 20 (19.4) Mural nodularity and atypical cytopathologic features
predictive of malignancy and/or invasive disease
Rodriguez 2007 145 145 32 (22.1) Findings associated with malignancy included presence
of mural nodules, cyst size ≥30 mm, and presence of a
thick wall
Schnelldorfer 2008 208 76 49 (64.4) 84 15 (17.8) In patients with invasive disease, 58% had recurrent disease
compared with 10% in those without invasive disease
Kim 2008 118 70 25 (35.7) 48 3 (6.3) Relative risk of malignancy was greatest when duct was
≥13 mm, tumor was ≥35 mm, and main duct disease
was involved
Nagai 2008 72 15 15 (100) 49 25 (51.0) In univariate analysis, CA19-9, MPD diameter, tumor size,
and presence of mural nodules were significantly
associated with malignancy in BD-IPMN and mixed
type IPMN. Symptoms of abdominal pain and weight
loss were also significant factors
Jang 2008 138 138 26 (18.8) Cyst size >20 mm was a significant prognostic factor in
MV analysis
Ohno 2009 87 14 11 (78.5) 48 20 (41.7) Type III and IV mural nodules and symptoms significantly
associated with malignancy in MV analysis
Continued

Downloaded for Paul Yoffe (payoffe@montefiore.org) at Montefiore Medical Center - JCon from ClinicalKey.com by Elsevier on May 26, 2018.
For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.
574 Intraductal Papillary Mucinous Neoplasms of the Pancreas

TABLE 1:  Studies Investigating Malignancy in Main Duct and Branch Duct Intraductal Papillary Mucinous
Neoplasms (IPMNs) Before and After Establishment of the Sendai Criteria
Main Duct Branch Duct
First Author Year Total N Malignant* N (%) N Malignant* N (%) Additional Findings
Nara 2009 123 26 26 (100) 59 26 (44.1) Cyst size >40 mm, MD-IPMN or mixed type IPMN, and
presence of mural nodules or a thick septum were
prognostic factors associated with malignancy
Bournet 2009 47 47 10 (21.3)
Hwang 2010 187 28 20 (71.4) 118 19 (16.1) Main duct disease predictive of malignancy. In those with
BD-IPMN, mural nodularity was a prognostic factor
Mimura 2010 82 39 34 (87.2) 43 20 (46.5) Main duct disease and diabetes were prognostic factors of
malignancy in MV analysis
Crippa 2010 389 81 55 (67.9) 159 34 (21.4)
Sadakari 2010 73 73 79 (9.6) MPD ≥5 mm and atypical cytologic features significantly
associated with malignancy
Kanno 2010 159 159 59 (37.1) Mural nodules ≥6.5 mm and CEA >5 ng/mL were
significantly associated with invasive IPMN in MV
analysis
Hodul 2012 105 105 62 (56.9) Cyst size ≥20 mm more likely to harbor invasive cancer
Fritz 2012 287 287 69 (24.0) 24.6% were considered Sendai negative and harbored
malignant features
Sahora 2013 226 226 52 (23.0) MPD dilation and cyst size ≥30 mm significantly associated
with malignancy
Shimizu 2013 310 310 160 (51.6) Cyst size, mural nodule size, and MPD dilation significantly
associated with malignancy
Abdeljawad 2014 52 52 40 (76.9) MPD ≥8 mm significantly associated with malignancy
Kawada 2014 202 202 24 (11.9) Mural nodule size ≥10 mm and positive cytology
significantly associated with malignancy
Lee 2014 84 84 16 (19.0) EUS scoring system including dilation of ducts, size of cyst,
size of mural nodule, septal thickening, and patulous
orifice
Marchegiani 2015 173 173 125 (72.3)
Dortch 2015 66 66 12 (18.2)
Kim 2015 177 177 39 (22.0) Cyst size ≥30 mm and mural nodules significantly
associated with malignancy
*Includes high-grade dysplasia and invasive cancer.
BD-IPMN, Branch duct IPMN; CA19-9, cancer antigen 19-9; CEA, carcinoembryonic antigen; EUS, endoscopic ultrasound scan; MD-IPMN, main duct
IPMN; MPD, main pancreatic duct; MV, multivariate.

subtype is characterized by complex arborizing papillae, oncocytic
cells, and intraepithelial lumina formation. It expresses MUC1,
MUC2, MUC5AC, and MUC6 and tends to form large tissue nodules
with limited mucous production.
CLINICAL PRESENTATION
IPMNs are thought to account for 1% to 3% of all exocrine pancre-
atic neoplasms and 20% to 50% of all cystic pancreatic neoplasms.
They are most commonly diagnosed in the sixth to seventh decade
of life and affect males and females equally. With the increase in
abdominal imaging over the last few decades, most IPMNs are identi-
fied incidentally on imaging done for another purpose, presenting a
management dilemma. When patients do have symptoms, they are
typically nonspecific, such as nausea, vomiting, abdominal pain, back
pain, weight loss, or anorexia. When obstructive jaundice is present,
the concern for malignancy is high. In addition, patients (primarily
those with MD-IPMN) can develop pancreatitis-like symptoms due
to blockage of the pancreatic duct with mucin; others may develop
symptoms of exocrine and endocrine pancreatic insufficiency but
this is usually a sign of advanced disease.
EVOLVING MANAGEMENT
Given the concern for malignant potential, when IPMNs were
first characterized in the late 1990s many surgeons advocated for

Downloaded for Paul Yoffe (payoffe@montefiore.org) at Montefiore Medical Center - JCon from ClinicalKey.com by Elsevier on May 26, 2018.
For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.
TABLE 2:  Histologic Classification of Intraductal Papillary Mucinous Neoplasms (IPMNs) and General
Characteristics Commonly Found Within Each Subtype
Gastric Intestinal Pancreatobiliary Oncocytic
Main duct vs branch duct BD-IPMN MD-IPMN Typically MD-IPMN MD-IPMN
Incidence Most common subtype Most common subtype of Less common Least common
overall (49%-63% of MD-IPMN (18%-36% (7%-18% of (1%-8% of all
all IPMNs) of all IPMNs) all IPMNs) IPMNs)
Location Uncinate process and Pancreatic head, may Pancreatic head
pancreatic periphery involve entire duct
Histologic features Resemble gastric foveolar Villous growth pattern Complex papillae, cells Complex papillae but
cells, form pyloric resemble pancreatic with eosinophilic
gland-like structures at and biliary duct cells cytoplasm, and
the base of the papillae oncocytic cells
Immunohistochemical • Scattered expression of • MUC2 • MUC1 • MUC1
profile MUC2 • MUC5AC • MUC5AC • MUC2
• MUC5AC • CDX2 • MUC5AC
• MUC6 • MUC6
• No expression of MUC1
Progression to invasive 10%-30% 30%-50% >50% Limited, but can be
disease up to 30%
Invasive form Conventional ductal Mucinous (colloid) Conventional ductal
(tubular) carcinoma (tubular)
adenocarcinoma adenocarcinoma
BD-IPMN, Branch duct IPMN; CDX2, caudal-type homeobox 2; MD-IPMN, main duct IPMN; MUC, mucin.

gastric intestinal

A B

pancreatobiliary oncocytic

C D

FIGURE 2  The four histologic classifications of IPMN are (A) gastric, (B) intestinal, (C) pancreatobiliary, and (D) oncocytic. (From Tanaka M, Fernandez-del
Castillo C, Adsay V, et al. International consensus guidelines 2012 for the management of IPMN and MCN of the pancreas. Pancreatology. 2012;12:183-187.)

Downloaded for Paul Yoffe (payoffe@montefiore.org) at Montefiore Medical Center - JCon from ClinicalKey.com by Elsevier on May 26, 2018.
For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.
576 Intraductal Papillary Mucinous Neoplasms of the Pancreas
BOX 1:  Components of High-Risk Stigmata and
Worrisome Features in Patients With Intraductal
Papillary Mucinous Neoplasms That Should Be
Considered for Surgical Resection if Patient
Deemed Surgically Fit
High-Risk Stigmata
Enhancing solid component
Main duct size ≥10 mm
Obstructive jaundice
Worrisome Features
Size >3 cm
Thickened, enhancing cyst walls
Main pancreatic duct size 5-9 mm
Symptoms other than jaundice
aggressive resection to prevent malignant transformation. However,
as we learned more about the anatomic and histologic subtypes, it
has become widely accepted that not all IPMNs need to be resected.
Many incidentally found IPMNs may not pose a threat when consid-
ering a patient’s life expectancy, associated comorbidities, and risk of
postoperative complications. In addition, pancreatic resection is a
high-risk operation with postoperative morbidity as high as 30% to
60%, making risk stratification of the likelihood of malignancy, as
well as operative risk, essential to decision making.
When determining whether a patient should undergo resection
or observation, it is important to use risk stratification to assess the
lesion with regard to the likelihood of existing malignancy and
overall malignant potential. In 2012 the International Consensus
Guidelines (Sendai criteria) were updated and reported on features
considered to be “high risk” and “worrisome” (Box 1). These criteria
helped to establish a detailed algorithm for deciding whether a
patient should undergo resection or surveillance. These guidelines
provide an excellent roadmap for decision making; however, clinical
circumstances (family history of pancreatic cancer, elevated cancer
antigen 19-9 [CA19-9], inability to tolerate surgery, etc.) and patient
preference offer the opportunity for shared decision making in this
complex disease.
Diagnosis and Management
Diagnosis of IPMN should be made using a combination of thor-
ough history and physical examination, noninvasive imaging, and
endoscopy. It is extremely important to accurately differentiate
IPMN from other pancreatic cystic lesions and correctly characterize
it as main duct or branch duct, as this will determine optimal
management.
At the time of presentation to a surgeon, most patients already
will have had cross-sectional imaging with computed tomographic
scan (CT scan) or magnetic resonance imaging (MRI). If this imaging
is not adequate for evaluating pancreatic duct size and the presence
of mural nodules, additional imaging should be performed. An upper
endoscopy that demonstrates an enlarged ampulla or “fish-mouth
sign” is pathognomonic for an IPMN (Figure 3). Once IPMN is in
the differential diagnosis stage, clinicians can refer to the 2012 Inter-
national Consensus Guidelines (Figure 4).
First, patients should be evaluated for the presence of high-risk
stigmata (see Box 1), defined as (1) obstructive jaundice with pres-
ence of a cystic lesion in the head of the pancreas, (2) enhancing solid
component with a cyst, and/or (3) main pancreatic duct 10 mm or
greater in size. If a patient has any of these characteristics, surgery is
recommended if the patient is an appropriate surgical candidate.
If there is no evidence of high-risk stigmata, the next step should
be evaluation for worrisome features (see Box 1). These include (1)
Downloaded for Paul Yoffe (payoffe@montefiore.org) at Montefiore Medical Center - JCon from ClinicalKey.com by Elsevier on May 26, 2018.
For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.
Ampulla
FIGURE 3  Upper endoscopy demonstrating dilated ampulla with
secretion of mucin and “fish-mouth sign,” which is pathognomonic for
intraductal papillary mucinous neoplasm.
cyst 3 cm or greater in size, (2) thickened, enhancing cyst walls, (3)
main duct 5 to 9 mm in size, (4) nonenhancing mural nodule, and/
or (5) abrupt change in caliber of pancreatic duct with distal pancre-
atic atrophy. If a patient meets any of these criteria, an endoscopic
ultrasound scan (EUS) should be done to evaluate for definite mural
nodules, main duct features suspicious for involvement, or cytology
suspicious or positive for malignancy. If EUS demonstrates any of
these findings, surgical resection should be considered strongly if
clinically appropriate. If EUS is inconclusive, close surveillance with
alternating MRI and EUS should be done every 3 to 6 months. If no
worrisome features are identified on EUS, ongoing management is
determined by the size of the largest cyst.
The surveillance protocol differs based on imaging modality and
time interval, depending on the size of the largest lesion. In lesions
smaller than 2 cm, multidetector 3D-CT with pancreatic protocol or
pancreatic-protocol MRI should be the primary imaging modality.
Lesions smaller than 1 cm should be imaged every 2 to 3 years and
lesions that are 1 to 2 cm should be imaged yearly for the first 2 years;
if the lesion remains stable, the surveillance interval may be length-
ened. Both CT and MRI are equivalent with regards to the informa-
tion they provide on tumor type, location, development of additional
lesions, lymph node and organ metastasis, and invasion into adjacent
structures. MRI, however, is able to provide better information
regarding the presence of possible septae and mural nodules (or solid
components) and provides more accurate information on the
involvement of the main pancreatic duct, including detection of
communications between side branch cysts and the main pancreatic
duct that can be missed on CT scan. Ductal communication also may
be evaluated via EUS (Figure 5).
Lesions that measure 2 to 3 cm should undergo EUS 3 to 6
months after diagnosis and then the surveillance interval may be
lengthened, alternating between EUS and MRI, if there is no interval
growth, development of worrisome or high-risk stigmata, or change
in symptoms. If a patient is young and fit and may require prolonged
surveillance, surgery should be considered. In lesions larger than
3 cm, MRI with EUS should be done every 3 to 6 months, and resec-
tion should be considered in young, fit patients.
Patients with two or more affected first-degree relatives should
have more aggressive surveillance, as the risk for malignancy is far
greater in this subset of patients. Imaging should include both MRI/
 T h e Pa n c r e a s 577

Are any of the following high-risk stigmata of malignancy present?

i) Obstructive jaundice in a patient with cystic lesion of the head of
the pancreas
ii) Enhancing solid component within cyst
iii) Main pancreatic duct ≥10 mm in size

Yes No

Consider surgery, Are any of the following worrisome features present?

if clinically appropriate Clinical: Pancreatitis a
Imaging:
i) Cyst ≥3 cm
ii) Thickened/enhancing cyst walls
iii) Main duct size 5–9 mm
iv) Nonenhancing mural nodule
v) Abrupt change in caliber of pancreatic duct with distal
pancreatic atrophy

If yes, perform endoscopic ultrasound No

Are any of these features present? No

i) Definite mural nodule (s) b What is the size of
Yes
ii) Main duct features suspicious for involvement c largest cyst?
iii) Cytology: Suspicious or positive for malignancy Inconclusive

<1 cm 1–2 cm 2–3 cm >3 cm

CT/MRI CT/MRI EUS in 3–6 months, then Close surveillance alternating

in 2–3 years d yearly x 2 years, then lengthen interval alternating MRI with EUS every 3–6 months.
lengthen interval MRI with EUS as appropriate. d Strongly consider surgery in young,
if no change d Consider surgery in young, fit patients.
fit patients with need for
prolonged surveillance.

a. Pancreatitis may be an indication for surgery for relief of symptoms.

b. Differential diagnosis includes mucin. Mucin can move with change in patient position, may be dislodged on cyst lavage, and does not
have Doppler flow. Features of true tumor nodule include lack of mobility, presence of Doppler flow, and fine-needle aspiration of nodule
showing tumor tissue.
c. Presence of any one of thickened walls, intraductal mucin, or mural nodules is suggestive of main duct
involvement. In their absence main duct involvement is inconclusive.
d. Studies from Japan suggest that on follow-up of subjects with suspected BD-IPMN there is increased incidence of pancreatic ductal
adenocarcinoma unrelated to malignant transformation of the BD-IPMN(s) being followed. However, it is unclear if imaging surveillance
can detect early ductal adenocarcinoma, and, if so, at what interval surveillance imaging should be performed.

FIGURE 4  International Consensus Guidelines for surveillance and surgical resection of intraductal papillary mucinous neoplasm (IPMN) depending on
high-risk stigmata, worrisome features, and size. BD-IPMN, branch duct IPMN; CT, computed tomography; EUS, endoscopic ultrasound scan; MRI, magnetic
resonance imaging. (Modified from Tanaka M, Fernandez-del Castillo C, Adsay V, et al. International consensus guidelines 2012 for the management of IPMN and MCN
of the pancreas. Pancreatology. 2012;12:183-187.)

Downloaded for Paul Yoffe (payoffe@montefiore.org) at Montefiore Medical Center - JCon from ClinicalKey.com by Elsevier on May 26, 2018.
For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.
578 Intraductal Papillary Mucinous Neoplasms of the Pancreas
A B
FIGURE 5  Endoscopic ultrasound done for intraductal papillary mucinous neoplasm demonstrating ductal communication between (A) the lesion and
(B) the main pancreatic duct and subsequent fine-needle aspiration biopsy.
magnetic resonance cholangiopancreatography (MRCP) and EUS.
Worrisome features in the setting of a family history certainly warrant
strong consideration for resection. In higher-risk surgical patients
where the decision to operate is more difficult, MRI/MRCP at
3-month intervals and annual EUS should be done for the first 2
years after diagnosis until the risk of malignancy outweighs the risk
of surgical resection.
Certain patients may be deemed unfit for pancreatic resection
given multiple comorbidities, frailty associated with advanced age, or
even patient preference. Given that the risk of morbidity associated
with pancreatic resection is as high as 30%, the risks and benefits of
surgery must be discussed with the patient in order to provide a
tailored treatment approach. In patients who are truly unfit for
surgery, surveillance is not indicated, as it will not change ultimate
management. In higher-risk patients, the number of worrisome or
high-risk stigmata may shift the risk-benefit assessment of surgical
resection; imaging is therefore indicated and shared decision making
is essential.
Finally, some patients may prefer surgical resection to the concern
associated with ongoing surveillance, whereas other patients may
have circumstances that make adequate surveillance unlikely. In both
of these circumstances, surgical resection should be considered in
patients who are good surgical candidates.
Operative Management: Extent of Resection
Preoperatively, it is important to counsel patients on both the risks
and the benefits of pancreatic resection and observation. This
includes discussion of cancer risk, treatment risk, and the need for
long-term surveillance in both settings. Especially in the setting of
multifocal disease, the possibility of total pancreatectomy and its
associated risks should be discussed. Possible postoperative compli-
cations and the risk of endocrine and/or exocrine insufficiency asso-
ciated with even partial pancreatic resection also should be discussed
with patients.
Options for resection include pancreaticoduodenectomy, distal
pancreatectomy, enucleation, and total pancreatectomy. The choice
of procedure depends on the location of the IPMN and pathologic
features. Preoperative planning is essential and largely dependent on
preoperative imaging, as well as intraoperative ultrasound. The entire
gland should be evaluated intraoperatively given the high frequency
of multifocal IPMN. Limited resections include enucleation, which
also may be considered in the setting of benign BD-IPMN; however,
nonanatomic resections may be associated with higher pancreatic
fistula rates or, rarely, leakage of mucin.
Downloaded for Paul Yoffe (payoffe@montefiore.org) at Montefiore Medical Center - JCon from ClinicalKey.com by Elsevier on May 26, 2018.
For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.
H -0.75
0.75
cm/s
In all cases, the goal of the surgery is resection of all invasive
IPMN and IPMN with high-grade dysplasia. At the time of surgical
resection, an intraoperative frozen section of the pancreatic neck
margin should be performed. If the neck margin is positive for high-
grade dysplasia or invasive cancer, re-resection to negative margins
should be performed even if this requires total pancreatectomy.
Current surgical thinking does not recommend total pancreatectomy
in the setting of low-grade or moderate-grade dysplasia at the surgi-
cal margin, as the physiologic implications of total pancreatectomy
outweigh the risks of ongoing surveillance and the risk of developing
another lesion in the remnant pancreas.
Postoperative Care
Perioperative morbidity and mortality is similar to that for any other
patient undergoing a pancreatic resection or major abdominal opera-
tion. Pancreas-specific complications of which the surgeon should be
aware include delayed gastric emptying, pancreatic fistula, and other
anastomotic leaks. Unlike patients with chronic pancreatitis or
aggressive pancreatic adenocarcinoma, most patients with IPMNs
have soft pancreata with normal-size bile ducts and are therefore at
higher risk for pancreatic fistulas and bile leaks. Depending on the
patient and the extent of resection, the patient also may require
pancreatic enzyme supplementation and insulin to help regulate
blood glucose levels.
Postoperative Surveillance
Surveillance protocols can be classified based on whether the patient
underwent resection for BD-IPMN or MD-IPMN. In patients who
undergo resection, surveillance depends on margin status and patho-
logic grade (Figure 6). In patients with a diagnosis of invasive IPMN,
surveillance should mirror that of pancreatic adenocarcinoma, with
CT scans every 3 months for the first year and every 6 months there-
after. In patients who undergo resection for high-grade dysplasia,
MRCP or CT scans should be obtained every 3 to 6 months. For
low-grade and intermediate-grade dysplasia, surveillance depends on
surgical margin status. If margins are negative, surveillance with CT
scans or MRCP at 2 and 5 years is appropriate. However, in those
with low-grade or moderate-grade dysplasia at the resection margin,
MRCP surveillance with history and physical examination should be
done twice a year.
When performing surveillance after resection, it is important
to identify whether a patient has a family history of pancreatic
adenocarcinoma, positive surgical margins, development of new
 T he Pa n c r e a s 579

Resection of MD-IPMN or BD-IPMN

Low- Intermediate- High- Invasive

grade dysplasia grade dysplasia grade dysplasia carcinoma

Margin status
Shorter interval – +
recommended if:
1. Family history of
pancreatic cancer CT every 3 mo
CT 2 and 5 yr MRCP MRCP or CT
2. New symptoms for first 2 yr,
after resection every 6 mo every 3–6 mo
or worsening then every 6 mo
diabetes
3. Positive surgical
margin

Appropriate treatment if recurrent disease

FIGURE 6  Proposed algorithm for management of main duct intraductal papillary mucinous neoplasm (MD-IPMN) and branch duct intraductal papillary
mucinous neoplasm (BD-IPMN) after surgical resection. CT, Computed tomography; MRCP, magnetic resonance cholangiopancreatography.

symptoms, or worsening of diabetes. Such clinical factors may be
worrisome for progression of the disease and these patients should
undergo shorter-interval surveillance. In the case of both MD-IPMN
and BD-IPMN, if additional lesions are identified on surveillance,
management should be based again on the International Consensus
Guidelines (see Figure 4).
Extrapancreatic Malignancy
Extrapancreatic malignancy has been reported in cases of both
MD-IPMNs and BD-IPMNs; the incidence ranges from 10% to 45%.
Frequent sites include breast, colon, and prostate. There are no
current recommendations for screening extrapancreatic malignan-
cies, but initial evaluation with up-to-date screening colonoscopy
and mammogram should be done in addition to continuing recom-
mended screening for these malignancies. Extra attention should be
paid to this both preoperatively and postoperatively.
PROGNOSIS
The overall survival rate for patients with IPMN has been reported
to range between 36% and 77%, with the presence of malignancy
being the strongest predictor of survival. Once invasive disease devel-
ops, the behavior of IPMNs is very similar to that of tubular pancre-
atic ductal carcinoma. For noninvasive neoplasms, the 5-year survival
rate after surgical resection is 70% to 100%; however, for those
patients with invasive disease, the 5-year survival rate ranges from
30% to 60% based on the histologic subtype of invasive malignancy.
Invasive tubular carcinomas occur in the setting of gastric and
pancreaticobiliary histologic subtypes and can be found in both
MD-IPMNs and BD-IPMNs. This type is morphologically indistin-
guishable from pancreatic ductal adenocarcinoma and, as such, has
a greater likelihood of nodal metastasis as well as a higher incidence
of perineural and vascular invasion. Therefore those with invasive
disease are similarly treated with adjunct treatment, including che-
motherapy and/or radiation similar to that used for pancreatic
adenocarcinoma.
The 5-year survival rate for resected tubular type malignancy is
37% to 55%. In contrast, colloid carcinoma (i.e., mucinous carci-
noma) has morphologic characteristics similar to other exocrine
malignancies such as breast and skin. It is associated with the
intestinal subtype, and overall has a better prognosis than the tubular
carcinoma, with an estimated 5-year survival rate of 61% to 87%
after surgical resection.
Other poor prognostic indicators related to survival include inva-
sive disease at the surgical margin and presence of jaundice. Recur-
rence in all patients who undergo resection ranges from 7% to 43%
in the remaining gland. However, good results have been reported in
patients who underwent repeat resection for recurrence in the
remaining gland.
CONCLUSIONS
IPMNs are an important clinical entity that surgeons should be aware
of given the complex management and surveillance associated with
these lesions. Multiple factors must be considered when diagnosing,
treating, and surveilling these patients. Although the Sendai criteria
serve as guidelines in the management of IPMNs, the overall patient
and clinical characteristics of the lesion should be factored in when
deciding on the optimal management plan given that the evolving
management of IPMNs is not a simple one-size-fits-all approach.
The decision to operate versus to surveil these patients is a complex
one and involves frank discussion with the patient regarding opera-
tive risks, benefits, and the risk of malignant progression. The treat-
ment option that is most appropriate should be tailored to the
patient’s needs and specific disease pathology.
SUGGESTED READINGS
Koh YX, Zheng HL, Chok AY, et al. Systematic review and meta-analysis of
the spectrum and outcomes of different histologic subtypes of noninva-
sive and invasive intraductal papillary mucinous neoplasms. Surgery.
2015;57:496-509.
Machado NO, Hani Q, Wahibi K. Intraductal papillary mucinous neoplasm
of pancreas. N Am J Med Sci. 2015;7:160-175.
Ohtsuka T, Tanaka M. Postoperative surveillance of branch duct. In: Tanaka
M, ed. Intraductal Papillary Mucinous Neoplasm of the Pancreas. Japan:
Springer; 2014:189-199.
Ohtsuka T, Tanaka M. Postoperative surveillance of main duct. In: Tanaka M,
ed. Intraductal Papillary Mucinous Neoplasm of the Pancreas. Fukuoka,
Japan: Springer; 2014:181-188.
Tanaka M, Fernandez-del Castillo C, Adsay V, et al. International consensus
guidelines 2012 for the management of IPMN and MCN of the pancreas.
Pancreatology. 2012;12:183-187.

Downloaded for Paul Yoffe (payoffe@montefiore.org) at Montefiore Medical Center - JCon from ClinicalKey.com by Elsevier on May 26, 2018.
For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.