Surgery 174 (2023) 75e82

Contents lists available at ScienceDirect

Surgery
journal homepage: www.elsevier.com/locate/surg

Long-term (10-year) outcomes and prognostic factors in resected

intraductal papillary mucinous neoplasm tumors in Finland: A
nationwide retrospective study
€ Vaalavuo, MDa,b, Martine Vornanen, MDc, Reea Ahola, MD, PhDa,b,
Yrjo
Anne Antila, MD, PhDa,b, Irina Rinta-Kiikka, MD, PhDb,d, Juhani Sand, MD, PhDa,b,
Johanna Laukkarinen, MD, PhDa,b,*
a
Department of Gastroenterology and Alimentary Tract Surgery, Tampere University, Finland
b
Faculty of Medicine and Health Technology, Tampere University, Finland
c
Department of Pathology, Fimlab Laboratories, Tampere University Hospital, Finland
d
Department of Radiology, Tampere University Hospital, Finland

a r t i c l e i n f o a b s t r a c t

Article history: Background: The degree of dysplasia is the most important prognostic factor for patients with resected
Accepted 7 February 2023 intraductal papillary mucinous neoplasms. Intraductal papillary mucinous neoplasms are predominantly
Available online 15 April 2023 premalignant conditions; in most cases, surveillance is an adequate treatment. If worrisome features are
present, surgery should be considered. However, there is limited data on the long-term prognosis of
resected intraductal papillary mucinous neoplasms. We aimed to ascertain the nationwide survival of
patients with resected intraductal papillary mucinous neoplasms and identify factors associated with
survival.
Methods: This is a retrospective nationwide cohort study. All intraductal papillary mucinous neoplasms
operated on in Finland between 2000 and 2008 were identified. Patient records were evaluated, and
original radiologic data and histologic samples were re-evaluated. Survival data were collected after a 10-
year follow-up period.
Results: Out of 2,024 pancreatic resections, 88 were performed for intraductal papillary mucinous
neoplasm. The median age of the patients was 65 years. Histologic diagnoses were main duct intraductal
papillary mucinous neoplasm 47/88 (53,4%), mixed-type intraductal papillary mucinous neoplasm 27/88
(30.7%), and branchduct intraductal papillary mucinous neoplasm 14/88 (15.9%). Of the tumors, 40/88
(45.5%) were low-grade dysplasia, 9/88 (10.2%) high-grade, and 39/88 (44.3%) were invasive cancer. The
median survival was 121 (range 0e252) months. Ten-year survival was 72.5%, 66.7%, and 23.1% in the
low-grade dysplasia, high-grade dysplasia, invasive cancer groups, respectively. Ten-year mortality for
pancreatic cancer was 5%, 9.1%, and 71.8% in the low-grade dysplasia, high-grade dysplasia, invasive
cancer groups, respectively.
Conclusion: Overall, 44.3% of the patients had a malignant tumor, and three-quarters (74.5%) of the main
duct intraductal papillary mucinous neoplasms were malignant or high-grade dysplasia at the time of
surgery. Ten-year survival was significantly better in patients operated on at the stage of a premalignant
tumor (low-grade dysplasia þ high-grade dysplasia) than in patients operated on at the stage of a ma-
lignant tumor.
© 2023 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY license
(http://creativecommons.org/licenses/by/4.0/).

Introduction

The prognosis of patients with resected intraductal papillary

* Reprint requests: Johanna Laukkarinen, MD, PhD, Professor of Surgery, Head,
mucinous neoplasms (IPMN) mostly depends on the degree of
Department of Gastroenterology and Alimentary Tract Surgery, Tampere University
Hospital, El€
am€
anaukio 1, 33520 Tampere, Finland. dysplasia.1 Patients with low malignant potential tumors, such as
E-mail address: johanna.laukkarinen@fimnet.fi (J. Laukkarinen); low-grade (LG) branch duct (BD)eIPMN, have excellent prognoses
Twitter: @TamperePancreas compared to patients with invasive main duct (MD)e or mixed-

https://doi.org/10.1016/j.surg.2023.02.006
0039-6060/© 2023 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
76 Y. Vaalavuo et al. / Surgery 174 (2023) 75e82

Figure 1. A flowchart of resected intraductal papillary mucinous neoplasm tumors in Finland between 2000 and 2008.

type (MT)eIPMN.2 Beyond main pancreatic duct (MPD) involve- histologic glasses by an experienced pancreatic pathologist. The
ment and the degree of dysplasia, other known factors influence findings were classified according to the World Health Organization
long-term prognosis. Tumor size, positive lymph nodes, and posi- classification of pancreatic tumors using hematoxylin and
tive resection margins worsen long-term outcomes.1,2 Also, histo- eosinestained sections.21 The presence of dysplasia was recorded,
logic subtypes are factors for patients’ long-term prognosis; gastric and grading was based on a 2-step grading system (low and high-
and intestinal-type IPMNs are usually associated with superior grade dysplasia). The variables with incomplete data were dis-
outcomes compared to pancreatobiliary type IPMNs, which repre- played with data available; some radiological studies and patho-
sent the more aggressive type, usually associated with high-grade logical slides were not available for this study.
dysplasia (HGD) and invasive adenocarcinoma.3e5 The minimum follow-up time for all patients was 10 years
Since IPMN was included in the WHO classification system in (range 10e21 years). Survival data, including time of death, total
1996, it has been subjected to rigorous scrutiny. Several guidelines mortality, and mortality for pancreatic cancer, was gathered from
on managing IPMN tumors have been issued, such as the European the Finnish Registry Office on November 26, 2020.
Evidence-Based Guidelines on Pancreatic Cystic Neoplasms,6 in
2017, the revised international consensus Fukuoka guidelines on Statistical analysis
the management of IPMN of the pancreas,7 and the American
Gastroenterological Association guidelines.8 The most widely Statistical analyses were performed using SPSS 26.0 for Win-
accepted risk factors for malignancy in these guidelines are main dows (IBM SPSS, Inc, Armonk, NY). Descriptive statistics are re-
duct dilation, cyst diameter, and elevated levels of serum carbo- ported using count, percentage, median, and range unless
hydrate antigen 19-9.9e16 otherwise specified. The c2 analysis was used in univariate ana-
The IPMNs are a fairly new entity; only limited long-term data is lyses. Kaplan-Meier analysis was used to analyze long-term
available, especially in a nationwide setting.17e19 Because IPMN survival.
tumors are optimally operated on during the premalignant phase, it
is probable that the median survival of these patients is excellent. Ethical aspects
Thus, data on long-term outcomes, even beyond 5 years, are
needed to evaluate the actual benefit for the patients undergoing Permission to review patient files and histologic slides was
surgery instead of surveillance. This study aimed to identify obtained from the National Supervisory Authority for Welfare and
nationwide patient characteristics and prognostic factors in a 10- Health (Valvira) (permission 10263/06.01.03.01/2012) and from the
year follow-up period. National Institute for Health and Welfare (permission 1854/
5.05.00/2012).
Methods
Results
This nationwide retrospective study of resected IPMNs with a
10-year follow-up includes all pancreatic lesions operated on in Epidemiology
Finland from 2000 to 2008. The patients were identified by
combining data from the national operations register and hospital Between 2000 and 2008, 2,024 pancreatic resections were
patient archives. Patients with resected IPMNs formed the final performed in Finland. Of those 2,024 resections, 225 operations
study population (Figure 1). were performed for pancreatic cystic neoplasms. Finally, re-
Patient demographics, comorbidities, symptoms, radiological evaluated histology was IPMN in 88/225 (34.5%) cases (Figure 1),
findings, operation details, and histologic findings were obtained and these patients were included in the study database. Resections
from the patient records. Postoperative complications were regis- for IPMN were performed at 12 centers; 49/88 (55.6%) of the re-
tered and graded according to the Clavien-Dindo (CD) classification sections were performed in the 2 largest centers, Tampere and
of surgical complications.20 Helsinki University Hospitals. The population of Finland was
The preoperative imaging studies were reanalyzed for the study roughly 5.25 million during the study period from 2000 to 2008,
by an experienced pancreatic radiologist. Histologic evaluation and thus a resection for an IPMN was performed yearly on 0.19/
with immunohistochemistry was repeated from the original 100,000 patients. For reference, between 2013 and 2018, the rate of
 Y. Vaalavuo et al. / Surgery 174 (2023) 75e82 77

Table I (52.4%), in the body in 19/80 (23.8%), and in the tail in 19/80 (23.8%).
Preoperative findings, radiologic imagining, and operating centers of patients with Parenchymal atrophy in any location of the pancreas was detected
resected IPMNs (2000e2008) in Finland
in 33/56 (58.9%) of the cases. Thickening of the cyst wall >2 mm was
Finding Total, n % of the patients present in 27/54 (50%), septa of the cyst were seen in 39/68 (57.4%),
Sex F 51/88 58% mural nodules of the cyst in 17/64 (26.6%), and calcification of
Sex M 37/88 42% the cyst in 7/73 (6.8%) of the cases. An experienced radiologist
Type 1 diabetes 2/74 2.7% suspected malignancy in re-evaluating 12/55 (21.8%) cases (Tables II
Type 2 diabetes 17/74 22.9%
and III).
Smoking 4/44 9.7%
Previous cancer 9/65 13.8%
Symptomatic 64/88 72.7% Surgery and complications
Duration of symptoms before operation (mo) Mean 6.88 1e37
CT 84/88 95.5%
Pancreaticoduodenectomy was performed on 43/88 (48.9%),
MRI 25/88 28.4%
CT þ MRI 21/88 23.8% distal pancreatic resection (tail resection or body and tail resection)
EUS 5/88 5.7% on 33/88 (37.6%), total pancreatectomy on 11/88 (12.5%), and
Median age at surgery 65.4 (40e87) enucleation on 1/88 (1.15%) of the patients. Overall morbidity (CD
Number of centers 12 1e5) was 43/88 (49.9%), 22/88 (25%) of the patients had serious
Number of cases on 2 high-volume centers 49
Number of cases in other 10 centers 39
postoperative complications (CD 3e5), and 30-day mortality was 2/
88 (2.3%) (Table IV).
IPMN, intraductal papillary mucinous neoplasm; MPD, main pancreatic duct.BD,
branch duct; HGD, high-grade dysplasia; INV, invasive carcinoma; IPMN, intraductal
papillary mucinous neoplasm; LGD, low-grade dysplasia; MD, main duct; MPD, main Histopathologic analysis
pancreatic duct; MT, mixed-type.CT, computed tomography; EUS, endoscopic ul-
trasound; IPMNs, intraductal papillary mucinous neoplasms; MRI, magnetic reso- Histologic diagnoses were MD-IPMN 47/88 (53.4%), MT-IPMN
nance imaging.
27/88 (30.7%), and BD-IPMN 14/88 (15.9%). Overall, 40/88 (45.5%)
of the tumors were LGD, 9 (10.2%) HGD, and 39/88 (44.3%) INV.
resections for IPMN in the Pirkanmaa Hospital District was 0.76/y/ Distributions of dysplasia were for MD-IPMN; LGD 12/47 (25.5%),
100,000 patients (Table I). HGD 6/47 (12.8%), INV 29/47 (61.7%), MT-IPMN; LGD 16/27 (59.3%),
HGD 2/27 (7.4%), INV 9/27 (33.3%), and BD-IPMN; LGD 12/14
(85.7%), HGD 1/14 (7.1%), and INV 1/14 (7.1%). The histological
Patient characteristics and preoperative findings subtypes were analyzed for 23 patients, of whom 11/23 (47.8%) had
intestinal (INT), 8/23 (34.8%) oncocytic, and 4/23 (17.4%) pan-
In 88 patients, the final histology was IPMN. The median age was creatobiliary (PB) subtype of IPMN tumor. For INT, oncocytic, and
65.4 (40e87) years, and 51/88 (58%) were females. Most patients, PB, the respective distributions of dysplasia were LGD 4/11 (36.4%),
64/88 (72.7%), were symptomatic at the time of surgery. Patients HGD 4/11 (36.4%), INV 3/11 (27.3%), LGD 5/8 (62.5%), HGD 0/8
with symptoms had symptoms for a median of 6.9 (1e37) months (0.0%), INV 3/8 (37.5%), LGD 0/4 (0.0%), HGD 0/4 (0.0%), and INV 4/4
before surgery (Table I). (100%) (Table V). Histologic subtype expressions are presented in
The most used (95.5%) preoperative radiologic modality was Supplementary Table S1.
computed tomography. All preoperative used examinations (and
re-examined later in this study) are presented in Table I. The Long-term outcomes
median diameter of the largest cyst was 37.7 mm (7e100); in 68/88
patients (77.3%), there was a single cyst. The median main The minimum follow-up time was 10 years (range 10e21). The
pancreatic duct (MPD) diameter was 5.1 mm (1e17). Distributions median survival was 121 (range 0e252) months. One-year, 5-year,
for MPD calibers were MPD <4.9 mm 40/68 (58.8%), MPD 5 to 9.9 and 10-year survival was 88.6%, 63.6%, and 50.0%, respectively. In
mm 18/68 (26.5%), and MPD >10 mm 10/68 (14.7%). Cysts were the subgroups formed according to the degree of dysplasia, 1-, 5,-
detected to communicate with MPD in 35/68 (51.5%) cases. The and 10-year survival was 97.5%, 87.5%, 72.5% for LGD, 100%, 77.8%,
location of the largest cyst was in the pancreatic head in 42/80 and 72.5% for HGD, and 76.9%, 35.9%, and 23.1% for INV. There was a

Table II
Radiologic findings, type of tumor, and degree of dysplasia of resected IPMN tumors

Radiological findings Total n (%) BD-IPMN n (%) MD-IPMN n (%) MT-IPMN n (%) LGD n (%) HGD n (%) INV n (%)

Diameter of largest cyst, mm, median (range) 37.7 (7e100) 32.4 (12e50) 38.9 (9e86) 38.5 (7e100) 37.9 (7e100) 42.4 (20e64) 36.0 (9e86)
Single cyst 68/88 (77.3) 6/14 (42.9) 41/47 (87.2) 21/27 (77.8) 28/40 (70.0) 8/9 (88.9) 32/39 (82.1)
MPD <4.9 mm 40/68 (58.8) 13/13 (100) 13/32 (40.6.) 14/23 (60.9) 25/34 (73.5) 2/6 (33.3) 13/28 (46.4)
MPD 5e9.9 mm 18/68 (26.5) 0/14 (0.0) 12/32 (37.5) 6/23 (26.1) 7/34 (20.6) 1/6 (16.7) 10/28 (35.7)
MPD >10 mm 10/68 (14.7) 0/14 (0.0) 7/32 (21.9) 3/23 (13.0) 2/34 (5.9) 2/6 (50) 5728 (17.9)
MPD diameter mm, median (range) 5.1 (1e17) 2.58 (2e6) 6.32 (2e15) 4.84 (1e17) 4.1 (1e17) 7.3 (2e11) 5.9 (2e14)
Cyst communicating with MPD 35/68 (51.5) 8/13 (61.5) 17/32 (53.1) 10/23 (43.5) 19/33 (57.6) 4/7 (57.1) 12/28 (42.9)
Location caput - Head 42/80 (52.4) 6/14 (42.9) 22/41 (53.7) 14/25 (56.0) 17/37 (45,9) 5/8 (62.5) 20/35 (57)
Location korpus - Boby 19/80 (23.8) 5/14 (35.7) 8/41 (19.5) 6/25 (24.0) 9/37 (24.3) 2/8 (25) 8/35 (22.9)
Location cauda - Tail 19/80 (23.8) 3/14 (21.4) 11/41 (26.8) 5/25 (20.0) 11/37 (29.7) 1/8 (12.5) 7/35 (20.0)
Parenchymal atrophy 33/56 (58.9) 5/10 (50) 18/29 (62.1) 10/17 (58.8) 15/24 (62.5) 4/7 (57.1) 11/25 (44.0)
Cyst wall >2 mm 27/54 (50) 3/10 (30.0) 16/27 (59.3) 8/17 (47.1) 5/22 (22.7) 6/7 (85.7) 16/25 (64.0)
Septation of cyst 39/68 (57.4) 7/13 (53.8) 20/31 (64.5) 12/24 (50.0) 20/34 (58.8) 4/7 (57.1) 15/27 (55.6)
Mural nodules of the cyst 17/64 (26.6) 1/13 (7.7) 11/32 (34.4) 5/19 (26.3) 3/31 (9.7) 4/7 (57.1) 10/26 (38.5)
Calcification of the cyst 7/73 (6.8) 0/13 80.0) 3/36 (8.8) 2/24 (8.3) 1/35 (2.9) 2/8 (25) 2/30 (6.7)
Suspected malignancy 12/55 (21.8) 1/10 (10.0) 7/28 (25) 4/17 (23.5) 0/23 (0.0) 3/7 (42.9) 9/25 (36)

BD, branch duct; HGD, high-grade dysplasia; INV, invasive carcinoma; IPMN, intraductal papillary mucinous neoplasm; LGD, low-grade dysplasia; MD, main duct; MPD, main
pancreatic duct; MT, mixed-type.
78 Y. Vaalavuo et al. / Surgery 174 (2023) 75e82

Table III specific 5-year survival was 97.1% in the LGD group compared to
Risk factors for malignancy in resected IPMN tumors 40.0 in the INV group. In the LGD group, 2/40 (5%) patients died of
Risk factors for malignancy Benign Malignant P value pancreatic cancer during the 10-year follow-up. Ten-year mortality
Age 65 (40e87) 65 (43e79) .798
from pancreatic cancer was 1/9 (11.1%) in the HGD group and 28/39
Age >60 years 16/22 (73%) 6/22 (27%) .063 (71.8%) in the INV group (Table VI, Figure 2).
Age <60 years 33/66 (50) 33/66 (50)
Symptomatic 31/63 (49) 32/63 (51) .052
Incidental 18/25 (72) 7/25 (28) Discussion
MPD diameter 3 (1e17) 6 (2e14) .129
Single cyst 36/68 (53) 32/68 (47) .340 The degree of dysplasia is considered the most important factor
Multifocal 13/20 (65) 7/20 (35)
MPD <4.9 mm 27/40 (68) 13/40 (32) .212
in determining the patient's survival after pancreatic resection for
MPD 5e9.9 mm 8/18 (44) 10/18 (56) IPMN, but only limited long-term follow-up data are available. We
MPD >10 mm 5/10 (50) 5/10 (50) aimed to investigate the nationwide 10-year survival of all resected
Diameter of largest cyst 40 (7e100) 31 (9e86) .409 IPMN patients and to identify factors associated with survival. We
Cyst not communicating with MPD 17/33 (52) 16/33 (48) .234
found that survival was significantly better when the resection was
Cyst communicating with MPD 23/35 (66) 12/35 (34)
Location head 22/42 (52) 20/42 (48) .724 performed before the malignant transformation. However, 5% to
Location body 11/19 (58) 8/19 (42) 11% of the patients operated on at the stage of LGD and HGD died of
Location tail 12/19 (63) 7/19 (37) pancreatic cancer during the 10-year follow-up.
No parenchymal atrophy 12/23 (52) 11 (23 (48) .689 The IPMN treatment guidelines aim to define tumors with
Parenchymal atrophy 19/33 (58) 14/33 (42)
Cyst wall >2 mm 11/27 (41) 16/27 (59) .056
elevated malignant potential and time the resection before the
Cyst wall <2 mm 18/27 (67) 9/27 (33) malignant transformation.1,22 It is well-established that only a few
No septa 17/29 (59) 12/29 (41) .808 IPMN tumors will present features necessitating surgical resection.
Septation of cyst 24/39 (62) 15/39 (38) Some studies even suggested that small BD-IPMN tumors should
No mural nodules of the cyst 31/47 (66) 16/47) 34 .75
not be followed up at all.23 Once worrisome features are present,
Mural nodules of the cyst 7/17 (41) 10/17 (59)
No calcification of the cyst 40/68 (59) 28/68 (41) .959 resection is recommended, although the predictive value of these
Calcification of the cyst 3/5 (60 2/5 (40) known features is not optimal. The prognosis of resected IPMNs
No suspected malignancy 27/43 (63) 16/43 (37) depends mainly on the degree of dysplasia.1 A BD-IPMN with only
Suspected malignancy 3/12 (25) 9/12 (75) .02 LGD has an excellent prognosis, although the remnant pancreas
IPMN, intraductal papillary mucinous neoplasm; MPD, main pancreatic duct. needs lifelong surveillance. Also, in tumors with HGD, the prog-
nosis is better than in invasive IPMN carcinoma.1 In the early 2000s
statistically significant difference (P < .01) in survival between in Finland, the treatment of pancreatic tumors was not centralized,
invasive cancers and noninvasive tumors (LGD þ HGD). Survival and operations were carried out in many low-volume centers. The
percentages for other subgroups are presented in Table VI. Disease- quality of diagnostics was not always at the level expected today.

Table IV
Complications classified by Clavien-Dindo score and 30-day mortality of resected IPMNs

Clavien-Dindo score N 0 n, (%) 1 n, (%) 2 n, (%) 3 n, (%) 4 n, (%) 5 n, (%) 30-day n, (5)

Pancreaticoduodenectomy 43 14 (32.6) 4 (9.3) 11 (25.6) 10 (23.3) 4 (9.3) 0 (0.0) 1

Distal pancreatectomy 33 24 (72.7) 1(3.0) 4 (12.1) 3 (9.1) 0 (0.0) 1 (3.0) 1
Total pancreatectomy 11 6 (53.5) 0 (0.0) 1 (9.1) 3 (27.3) 1 (9.1) 0 (0.0) 0
Enucleation 1 1 (100) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0
Total 88 45 (51.1) 5 (5.7) 16 (18.2) 16 (18.2) 5 (5.7) 1 (1.1) 2 (2.3)

IPMNs, intraductal papillary mucinous neoplasms.

Table V
Degree of dysplasia and rate of malignancy in IPMN subtypes and histological subtypes

BD-IPMN MD-IPMN MT-IPMN

All 14 (15.9) 47 (53.4) 27 (30.6)

LGD 12/14 (85.8) 12/47 (25.5) 16/27 (59.3)
HGD 1/14 (7.1) 6/47 (12.8) 2/27 (7.4)
INV 1/14 (7.1) 29/47 (61.7) 9/27 (33.3)
Non-INV 13/14 (93) 18/47 (38) 18/27 (67)
P value 0.001
Oncocytic subtype* 3/8 (37.5) 3/8 (37.5) 2/8 (25)
Invasive 0/3 (0.0) 1/3 (33) 2/2 (100)
Non-INV 3/3 (100) 2/3 (66) 0/2 (0.0)
Intestinal subtype* 1/11 (9.1) 5/11 (45.5) 5/11 (45.5)
Invasive 0/1 (0.0) 1/5 (20) 2/5 (40)
Non-INV 1/1 (100) 4/5 (80) 3/5 (60)
Pancreatobiliary subtypey 0/4 (0.09 1/4 (25) 3/4 (75)
Invasive 0/0 (0.0) 1/1 (100) 3/3 (100)
Non-INV 0/0 (0.0) 0/1 (0.0) 0/3 (0.0)

BD, branch duct; HGD, high-grade dysplasia; INV, invasive carcinoma; IPMN, intraductal papillary mucinous neoplasm; LGD, low-grade dysplasia; MD, main duct; MT, mixed-
type.
*
No statistically significant difference (P > .05) in share of patients with malignant disease.
y
All patients with pancreatobiliary cysts had malignant disease.
 Y. Vaalavuo et al. / Surgery 174 (2023) 75e82 79

Table VI
One-year, 5-year, and 10-year survival for IPMN patients

Median (range) IQR 1-year survival% 5-year survival% 10-year survival% 5-year disease-specific survival % Pancreatic cancer mortality n (%)

BD 155 (3e252) 74.75-232.25 92.9 85.7 64.3 100 1/14 (7.1)

MD 87 (1e240) 21-141 85.1 59.6 44.7 65.1 20/47(42.6)
MT 124 (6e240) 17-171 92.6 59.3 51.9 66.7 10/27 (37.0)
LGD 142 (3e252) 101-185.75 97.5 87.5 72.5 97.1 2/40 (5)
HGD 118 (15e229) 58.5-168.5 100 77.8 66.7 87.3 1/9 (11.1)
INV 25 (1e226) 12-109 76.9 35.9 23.1 40.0 28/39 (71.8)
ONC 60.5 (3e242) 5.25-133.5 50 50 50 57.1 3/8 (37.5)
INT 124 (12e240) 15-178 100 72.7 72.7 80.0 2/11 (18.2)
PB 19 (12e34) 13.25-30.75 100 0.0 0.0 0.0 4/4 (100)
All 121 (0e252) 24.25-161.5 88.6 63.6 50.0 70.9 31/88 (35.2)

BD, branch duct; HGD, high-grade dysplasia; INT, intestinal; INV, invasive carcinoma; IPMN, intraductal papillary mucinous neoplasm; LGD, low-grade dysplasia; MD, main
duct; MPD, main pancreatic duct; MT, mixed-type; ONC, oncocytic; PB, pancreatobiliary.

Figure 2. (A) Kaplan-Meier survival curves by intraductal papillary mucinous neoplasm subtypes. (B) Kaplan-Meier survival curves by the degree of dysplasia. (C) Kaplan-Meier
survival curves by immunohistochemical intraductal papillary mucinous neoplasm subtypes.
80 Y. Vaalavuo et al. / Surgery 174 (2023) 75e82

Figure 2. (continued).

For example, magnetic resonance imaging was performed on only Most tumors (53.4%) were classified as MD-IPMN (MT-IPMN
25/88 (28.4%) patients, which obviously affects the ability to eval- 30.6% and BD-IPMN 16.3%). The proportion of resected BD-IPMN is
uate features such as a cyst communicating with the MPD. Sub- low compared to the resection rates in more recent publica-
optimal diagnostics may have led to numerous misdiagnoses both tions.26,27 Although the indications for surgery in BD-IPMN patients
preoperatively and postoperatively. Also, some tumors may have have been tightened, the increased volume of BD-IPMN patients
been diagnosed as PDAC rather than IPMN, which may partially under surveillance has also increased the proportion of resected
explain the low number of resections performed in the earlier years BD-IPMNs compared to other subtypes.
of the study period. Overall, 44.3% of the patients had a malignant operated tumor,
The first IPMN guidelines were issued in 2006; the effect of and three-quarters (74.5%) of the MD-IPMNs were malignant or HGD.
these guidelines on managing patients in this cohort was In comparison, 12/14 (85.8%) patients had an LGD tumor in the BD-
negligible. Decisions to operate likely were based on the clinical IPMN group. The instance of HGD was present more equally in all
judgment of the individual surgeon, with cyst size being the groups (BD-IPMN 7.1%, MT-IPMN 7.1%, and MD-IPMN 12.8%). The rate
most important factor in this decision-making. Currently, sur- of malignancy was high compared to other surgical series. The tu-
geons in Finland rely primarily on the European guidelines. mors were large (median 40.4, range 3.4e120.0 mm), symptomatic
Since applying new guidelines for treating IPMN, the number of (72.7%), and detected late (length of symptoms before operation 6.88
patients undergoing surgery and under surveillance has months). Also analyzed retrospectively, several features (ie, MPD
increased significantly. Although the indications for surgery have diameter, thickened cyst wall, and size of the cyst) were present that
been tightened, the overall increase in the number of patients relate to malignancy. In the current guidelines, MPD dilation, among
on surveillance programs has led to more resections. Population others, is a well-established radiologic feature predicting malignancy.
aging and increased volumes of cross-section imaging have There was a tendency for an elevated risk of malignancy in factors
likewise increased the likelihood of asymptomatic pancreatic such as patients with symptoms, an MPD diameter of 5 mm, and a
cysts being detected.24 cyst wall >2 mm. However, the numbers did not reach statistical
In our nationwide cohort, the patients were predominantly fe- significance. The only preoperative factor of statistical significance in
male, and the median age was slightly above 60. The number of detecting malignancy was if a radiologist suspected malignancy.
incidentally found, asymptomatic cysts was low (27.3%) and can be Based on this finding, the “gut feeling” of the experienced radiologist
explained by the less frequent use of cross-section imaging for should be considered when IPMN cases are discussed in a multidis-
other reasons in the early 2000s. ciplinary setting. Poor quality of preoperative imaging studies also
The rate of complications was high but similar to those found in may have negatively affected the ability to detect any signs of ma-
other studies; after pancreaticoduodenectomy, 29/43 (67.4%) pa- lignant transformation. The number of patients in the histologic
tients had any complication, and for 14/43 (32.5%) of the patients, subtype analyses was low; thus, there was no significant difference
the complications were considered severe (CD 3e5). In compari- among IPMN subtypes (BD, MD, MT) or malignancy, although 75% of
son, in a nationwide register study from Finland from 2012 to 2014, the PB group was MT, and all were malignant.
the rate of severe complications after pancreaticoduodenectomy Survival percentages beyond 5 years in nationwide settings
was 23.3%.25 were not published in large quantities before this study. In our
 Y. Vaalavuo et al. / Surgery 174 (2023) 75e82 81

material, the median survival overall was 121 (0e252) months, not involved in the study design, data collection, data analysis,
and 5- and 10-year survival was 63.6% and 50.0%. The most manuscript preparation, or publication decisions.
important prognostic factor for patients with resected IPMN is the
degree of dysplasia. It is evident that when IPMN is deemed
invasive carcinoma, the prognosis is dismal; for INV compared Conflict of interest/Disclosure
to LGD, 5- and 10-year survival was 35.9 vs 87.5 and 72.5 vs 23.1.
The distinction between LGD and HGD is not so clearcut from The authors have no conflicts of interests or disclosures to
the perspective of survival. Some, although not all, authors have report.
reported differences in survival in these 2 groups.1,18 In this
cohort, there were no statistically significant differences in sur-
Supplementary materials
vival for LGD and HGD in 5 and 10-year surveillance. In the
general population (in this age group) at the time of this study,
Supplementary materials associated with this article can be
10-year survival was roughly 80% compared to 72.5% in the LGD
found in the online version, at https://doi.org/10.1016/j.surg.2023.
group. One factor explaining this difference is the risk of a ma-
02.006.
lignant tumor in the remnant pancreas if total pancreatectomy
has not been performed, even if the histology of the specimen
was benign. References
In this study, mortality from pancreatic cancer was 2/40 (5%) in
the LGD group. Mortalities from pancreatic cancer in the LGD group 1. Blackham AU, Doepker MP, Centeno BA, et al. Patterns of recurrence and long-
included deaths 59 months after the primary operation (MT-IPMN) term outcomes in patients who underwent pancreatectomy for intraductal
and deaths 79 months after primary operation (BD-IPMN). The risk papillary mucinous neoplasms with high grade dysplasia: implications for
surveillance and future management guidelines. HPB (Oxford). 2017;19:
of a malignant tumor in the remnant advocates the surveillance of 603e610.
the remnant pancreas after resection as long as the patient is fit 2. Kang MJ, Jang JY, Lee KB, Chang YR, Kwon W, Kim SW. Long-term prospective
enough for surgery.6 There are probably more factors than described cohort study of patients undergoing pancreatectomy for intraductal papillary
mucinous neoplasm of the pancreas: implications for postoperative surveil-
above that could explain the decrease in survival in benign cases lance. Ann Surg. 2014;260:356e363.
compared to the general population. As expected, mortality from 3. Distler M, Kersting S, Niedergethmann M, et al. Pathohistological subtype
pancreatic cancer was higher in HGD, especially in the INV group. predicts survival in patients with intraductal papillary mucinous neoplasm
(IPMN) of the pancreas. Ann Surg. 2013;258:324e330.
Among MD, MT, and BD, survival was higher in the BD group (which
4. Furukawa T, Hatori T, Fujita I, et al. Prognostic relevance of morphological types
can be explained by the smaller number of malignant tumors in the of intraductal papillary mucinous neoplasms of the pancreas. Gut. 2011;60:
BD group). Still, there was no statistical significance among the 509e516.
groups. Also, there were no significant differences in survival be- 5. Ishida M, Egawa S, Aoki T, et al. Characteristic clinicopathological features of
the types of intraductal papillary-mucinous neoplasms of the pancreas.
tween histologic subtypes (oncocytic, PB, INT), although the small Pancreas. 2007;35:348e352.
number of patients may have affected this analysis. 6. European Study Group on Cystic Tumours of the Pancreas. European evidence-
based guidelines on pancreatic cystic neoplasms. Gut. 2018;67:789e804.
7. Tanaka M, Fern andez-Del Castillo C, Kamisawa T, et al. Revisions of interna-
Strengths and limitations of the study tional consensus Fukuoka guidelines for the management of IPMN of the
pancreas. Pancreatology. 2017;17:738e753.
8. Vege SS, Ziring B, Jain R, Moayyedi P. Clinical Guidelines Committee; American
The strength of this study is the nationwide setting. All resected Gastroenterology Association. American Gastroenterological Association Insti-
patients were included; thus, there was no selection bias. Another tute guideline on the diagnosis and management of asymptomatic neoplastic
pancreatic cysts. Gastroenterology. 2015;148:819e822.
strength of this study is the long follow-up time for all patients. This 9. Han Y, Lee H, Kang JS, et al. Progression of pancreatic branch duct intraductal
is longer than in any other published series of nationwide data. papillary mucinous neoplasm associates with cyst size. Gastroenterology.
However, even though nationwide, the sample size can be 2018;154:576e584.
10. Kolb JM, Argiriadi P, Lee K, et al. Higher growth rate of branch duct intraductal
considered small, and we could not complete radiologic and his-
papillary mucinous neoplasms associates with worrisome features. Clin Gas-
tologic data sets for all patients. This hurt our ability to conduct troenterol Hepatol. 2018;16:1481e1487.
meaningful qualitative analyses and can be considered a limitation 11. 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:
of the study. Also, the poor quality of preoperative radiologic
875e880.
studies (the small number of magnetic resonance imaging studies) 12. Marchegiani G, Andrianello S, Morbin G, et al. Importance of main pancreatic
may have impaired the detection of significant risk factors for duct dilatation in IPMN undergoing surveillance. Br J Surg. 2018;105:
malignancy in our analysis. 1825e1834.
13. Hirono S, Tani M, Kawai M, et al. The carcinoembryonic antigen level in
In conclusion, based on this nationwide cohort with a 10-year pancreatic juice and mural nodule size are predictors of malignancy for branch
survival analysis, we conclude that patients with an IPMN tumor duct type intraductal papillary mucinous neoplasms of the pancreas. Ann Surg.
resected before malignant transformation have a good prognosis. 2012;255:517e522.
14. Ateeb Z, Valente R, Pozzi-Mucelli RM, et al. Main pancreatic duct dilation
In comparison, in the case of invasive carcinoma, 10-year survival greater than 6 mm is associated with an increased risk of high-grade dysplasia
is <25%. Also, the IPMN subtype is an important prognostic factor; and cancer in IPMN patients. Langenbecks Arch Surg. 2019;404:31e37.
in this cohort, mortality due to pancreatic cancer was 7.1% in the 15. Ogura T, Masuda D, Kurisu Y, et al. Potential predictors of disease progression
for main-duct intraductal papillary mucinous neoplasms of the pancreas.
BD-IPMN group compared to around 40% in the MD-IPMN and J Gastroenterol Hepatol. 2013;28:1782e1786.
MT-IPMN groups. The timing of the surgery before malignant 16. Marchegiani G, Andrianello S, Morbin G, et al. Importance of main pancreatic
transformation and, at the same time, avoiding unnecessary op- duct dilatation in IPMN undergoing surveillance. Br J Surg. 2018;105:
1825e1834.
erations remains a challenge for all surgeons working with IPMN
17. Schnelldorfer T, Sarr MG, Nagorney DM, et al. Experience with 208 resections
tumors. for intraductal papillary mucinous neoplasm of the pancreas. Arch Surg.
2008;143:639e646.
18. Aronsson L, Andersson B, Andersson R, Tingstedt B, Bratlie SO, Ansari D.
Funding/Support Intraductal papillary mucinous neoplasms of the pancreas: a nationwide
registry-based study. Scand J Surg. 2018;107:302e307.
19. Oyama H, Tada M, Takagi K, et al. Long-term risk of malignancy in branch-duct
This work was supported by the State Research Funding (VTR), intraductal papillary mucinous neoplasms. Gastroenterology. 2020;158:
Finland and the Sigrid Juselius Foundation, Finland. The funder was 226e237.
82 Y. Vaalavuo et al. / Surgery 174 (2023) 75e82

20. Dindo D, Demartines N, Clavien P. Classification of surgical complications: a 24. Klibansky DA, Reid-Lombardo KM, Gordon SR, Gardner TB. The clinical rele-
new proposal with evaluation in a cohort of 6336 patients and results of a vance of the increasing incidence of intraductal papillary mucinous neoplasm.
survey. Ann Surg. 2004;240:205e213. Clin Gastroenterol Hepatol. 2012;10:555e558.
21. Hruban RH, Adsay NV, Albores-Saavedra J, et al. Pancreatic intraepithelial 25. Ahola R, Sand J, Laukkarinen J. Pancreatic resections are not only safest but also
neoplasia: a new nomenclature and classification system for pancreatic duct most cost-effective when performed in a high-volume centre: a Finnish reg-
lesions. Am J Surg Pathol. 2001;25:579e586. ister study. Pancreatology. 2019;19:769e774.
22. Aronsson L, Andersson B, Andersson R, Tingstedt B, Bratlie SO, Ansari D. 26. Attiyeh MA, Fern andez-Del Castillo C, Al Efishat M, et al. Development and
Intraductal papillary mucinous neoplasms of the pancreas: a nationwide validation of a multi-institutional preoperative nomogram for predicting grade
registry-based study. Scand J Surg. 2018;107:302e307. of dysplasia in intraductal papillary mucinous neoplasms (IPMNs) of the
23. Marchegiani G, Andrianello S, Pollini T, et al. “Trivial” cysts redefine the risk of pancreas: a report from the Pancreatic Surgery Consortium. Ann Surg.
cancer in presumed branch-duct intraductal papillary mucinous neoplasms of 2018;267:157e163.
the pancreas: a potential target for follow-up discontinuation? Am J Gastro- 27. Marchegiani G, Mino-Kenudson M, Ferrone CR, et al. Patterns of recurrence
enterol. 2019;114:1678e1684. after resection of IPMN: who, when, and how? Ann Surg. 2015;262:1108e1114.