Introduction

Intraductal papillary mucinous neoplasms (IPMNs) of the pancreas are characterized by an adenomatous proliferation of pancreatic duct epithelium involving the main pancreatic duct (MD-IPMN type), the branch duct (BD-IPMN type) or a combination of both (mixed type)¹. These tumors were first described in 1982 by Ohhashi et al.² as a distinct entity within the group of cystic pancreatic neoplasms. In 1996 the World Health Organization recognized the malignant potential of IPMNs³. The different guidelines recommend that treatment should be guided by the tumor features on imaging tests⁴.

The incidence of IPMN has increased over the past years due to improved diagnostic tests and is currently estimated to be 2 per 100,000 person-years⁵. The diagnosis is made in 50% of the cystic lesions detected in imaging tests; IPMNs have been found in 24% of autopsy specimens in adults and in 25% of surgical specimens of pancreatic neoplasms^4,6,7. Thirty-seven percent correspond to MD-IPMN type, 27% to BD-IPMN type and 36% to mixed type^4,8.

IPMNs are being increasingly appreciated as important pancreatic ductal adenocarcinoma precursors. For this reason, these lesions and their management according to the current guidelines should be discussed during undergraduate and postgraduate medical education due to the high diagnostic prevalence.

The aim of this study was to describe our experience with the diagnosis, standardized multidisciplinary management and long-term follow-up of patients with IPMNs in a reference center.

Material and methods

We conducted a retrospective and descriptive analysis of patients with pancreatic neoplasms followed up at the Department of General Surgery and Hepato- Biliary Surgery, Hospital Universitario Fundación Favaloro, Buenos Aires, Argentina, between January 2010 and December 2019. Patients who prospectively attended the outpatient clinic were included. Data from the electronic medical records were reviewed, and those patients who did not return to the clinic after the diagnosis was made or were lost to follow-up were excluded. The following variables were considered: age, sex, personal history, clinical picture, results of imaging tests, tumor location, laboratory tests, treatment, complications, length of hospital stay, follow-up, morbidity, mortality, type of duct involvement and histopathologic classification.

With these data, the patients were divided into two groups:

A. Group A: IPMNs diagnosed in the outpatient clinic

B. Group B: IPMNs diagnosed in the surgical specimens of pancreteactomies performed for other reasons.

Surgery was indicated after discussing the case in a multidisciplinary meeting, using the criteria recommended by international guidelines. Group A was divided into two subcategories: AS (patients with indication surgery) and AM (patients with indication of monitoring)^9-11. The complications were reported according to the Clavien-Dindo classification (considering grade > III as complications clinically relevant)¹² and to the definition of the International Study Group of Pancreatic Surgery (ISGPS)^13-15.

Continuous variables are expressed as mean ± standard deviation (SD) or median and range. Categorical variables are presented as frequencies or percentage. A p value < 0.05 was considered statistically significant. Actuarial survival was estimated using the Kaplan- Meier method. All the statistical calculations were performed using IBM SPSS Statistic 25.0 software package.

Results

One-hundred and one patients were diagnosed with IPMN at Hospital Universitario Fundación Favaloro. Eighty-six patients met the inclusion criteria; the 79

diagnosis of IPMN was made in the outpatient clinic in 79 (90%) patients and at the pathological examination in 7; 53 (62%) were women and mean age was 65 ± 14 years.

Thirty-three (38%) patients complaint of unspecific abdominal pain and 10 (12%) had been hospitalized due to acute pancreatitis. In 43 (50%) cases the neoplasm was an incidental finding in imaging tests.

Imaging tests were performed for the diagnosis of the type and characteristics of the lesions, and included computed tomography (CT) scan in 66, magnetic resonance cholangiopancreatography (MRCP) in 79 and endoscopic ultrasound (EUS) in 30 (35%) in whom the diagnosis was unclear.

Fifteen patients (17%) had MD IPMN (2 associated with adenocarcinoma), 60 (70%) had BD IPMN (1 associated with ampullary cancer and 1 with adenocarcinoma) and 11 (13%) presented mixed type [1 associated with neuroendocrine tumor (NET) and 1 with adenocarcinoma]. Thirty-five (41%) tumors were located in the head/uncinate process, 20 (23%) in the body/tail and 31 (36%) had multifocal distribution (Table 1).

In group A, a watchful waiting with monitoring (AM) was decided in 57 of 79 patients (66%) according to the guidelines. This management was decided because:

▪▪53 cases (93%) corresponded to BD-IPMN type without any sign of malignancy.

▪▪2 patients (3.5%) with mixed type-IPMN and 2 patients (3.5%) with BD IPMN decided to undergo clinical management despite medical advice.

Median follow-up was 12 months (range 1-83).

Of the remaining 22 patients:

▪▪18 (29%) underwent surgery (AS) due to:

◦◦MD IPMN in 9

◦◦BD IPMN in 3: 1 with high-risk stigmata, 1 due to the size of the lesion and

◦◦6 mixed type-IPMN

▪▪4 are waiting for surgery: 2 with BD IPMN with elevated CA 19-9 levels in the cyst fluid and 2 with mixed type-IPMN with worrisome features.

Mean follow-up until surgery was 5 months (range 3-106).

Of the 221 pancreatectomies performed in the period analyzed, 142 were due to malignancies and IPMN was found in 7 surgical specimens in the pathological examination (group B): 5 primary neoplasms of the pancreas and 2 ampullary neoplasms (Table 2).

In group AC, the procedures performed were 6 pancreaticoduodenectomies, 5 distal pancreatectomies, 5 total pancreatectomies and 2 bypass surgeries due to infiltration of the superior mesenteric vein in an 83-year-old patient and liver metastasis in another patient (Table 3). Twenty percent of the total pancreatectomies started as a partial pancreatectomy that had to be extended because the surgical margins of the main pancreatic duct sent for frozen section examination were repeatedly positive.

Eight postoperative complications occurred in group AS and three in group B (Table 4).

None of the patients in group AS or B presented recurrent pancreatic cancer in the remnant pancreas. In group B, one patient with a NET associated with mixed-type IPMN with moderate dysplasia and BD IPMN in the head of the pancreas underwent distal pancreatectomy and splenectomy; the IPMN in the head was not resected and monitored was indicated.

In group A, survival at 2 and 5 years was 96% and 89%, respectively, for patients undergoing conservative management (AM), and 95% and 86%, respectively, for those undergoing surgery (AS). In group B, survival at 2 and 5 years was 43% (p = 0.001 group AM vs. AS, Fig. 1. A).

Of the 17 non-invasive IPMNs resected, 13 belonged to group AS (4 MD IPMN, 2 BD IPMN and 7 mixed-type IPMN) and 4 to group B (1 MD IPMN, 1 BD IPMN and 2 mixed-type IPMN). Survival at 5 years for non-invasive IPMNs resected was 83%. Among invasive IPMNs, survival was 38% (p = 0.02. Fig. 1.B). In this group, 4 patients belonged to group AS (2 MD IPMN and 2 mixed-type IPMN) and 3 to group B (2 MD IPMN and 1 BD IPMN).

Two patients in the subgroup AS died due to disease progression (the 2 patients undergoing bypass surgeries). In the subgroup AM, 2 patients died: one patient with cirrhosis during induction of anesthesia for liver transplantation and another due to myocardial infarction (MI). In group B, 4 patients died due to septic shock, MI and disease progression (n = 2).

Discussion

The prevalence of IPMN of the pancreas has increased along with the development of diagnostic imaging tests. A previous history of diabetes, especially with insulin use, chronic pancreatitis, family history of pancreatic ductal adenocarcinoma and smoking habits are all relevant risk factors for the development of IPMN^16,17. IPMNs occur in patients with mean age of 64 years, are more common in men and are usually asymptomatic. The most common sites are head and the uncinate process of the pancreas and are multifocal in about 30-40% of the cases^4,7. In our series, 70% of the tumors corresponded to the BD-IPMN type, 17% were MD-IPMN type and 13% mixed-type IPMNs, which is not consistent with the published literature. However, tumor location in our series is similar to the one reported in international publications.

Most IPMNs are asymptomatic and the diagnosis is made incidentally. When symptoms develop, they are usually unspecific (abdominal pain, flank pain, nausea, vomiting, anorexia and weight loss)¹⁸. Acute pancreatitis and jaundice have also been described^19,20.

Contrast-enhanced magnetic resonance imaging (MRI) with MRCP is recommended as a first step, while CT scan is an alternative. Imaging tests are useful to evaluate the characteristics of the cyst, the presence of mural nodules and the association with the main pancreatic duct. Nowadays, EUS with fine-needle aspiration (FNA) is the most accurate method for the diagnosis of pancreatic cystic lesions and should be performed when the diagnosis is not clear or in case of suspicious of malignancy. The diagnostic performance of the cytological analysis of pancreatic cyst fluid is low due to lack of cellular material. Carcinoembryonic antigen (CEA) levels > 200 ng/mL can accurately suggest mucinous cystic lesions, but do not predict the presence of malignancy²¹. In 2013, Park et al. reported that cyst glucose levels were higher in non-mucinous pancreatic cysts compared with mucinous cysts²².

In 2006 and later in 2017, Tanaka et al. published the guidelines for the management of IPMN of the pancreas, known as the Sendai guidelines and the Fukuoka guidelines, respectively. The latter describes two groups of risk factors for malignant transformation based on findings of imaging tests: “high-risk stigmata” (HRS) and “worrisome features” (WFs) (Table 5); each of them contribute to suspect malingnancy^9,10. In a retrospective study, Izumo et al. reported that an enhancing mural nodule > 5 mm, history of pancreatitis or thickened/enhancing cyst walls were independent predictive factors for high-grade dysplasia. Likewise, a main pancreatic duct diameter > 10 mm, an abrupt change in pancreatic duct caliber with distal pancreatic atrophy or lymphadenopathy were risk factors for intraductal papillary mucinous carcinoma ²³.

Surgery is indicated when HRS develop during surveillance of BD IPMNs. In case of WFs, EUS with FNA biopsy is indicated, and surgery should be considered with positive results for malignancy. If the biopsy is negative, surveillance is indicated at regular intervals depending on the cyst size. For cysts < 1 cm imaging tests should be repeated in 2-3 years, while for cysts between 1 and 2 cm CT/MRI should be repeated yearly for 2 years, and then lengthen the interval if there is no change. For lesions between 2 and 3 cm, perform EUS in 3-6 months, then lengthen interval alternating MRI with EUS, and consider surgery in young, fit patients^9,10,24-26.

In a recent publication, Giuffrida et al. presented their experience in patients with BD IPMN and negative Sendai criteria at the time of diagnosis who converted to positive during follow-up. In this study, 17% of the patients required a change of strategy (surgery) and 5.5% of those undergoing surgery presented malignant lesions. The negative predictive value and positive predictive value of the Sendai criteria for malignancy were 100% and 32.4%, respectively. Thus, they concluded that the absence of positive Sendai criteria excluded malignancy in patients with BD IPMN, but the presence of positive criteria was not conclusive. In patients with positive Sendai criteria, the indication of surgery should be defined according to life expectancy and patients’ risk²⁷.

The risk of high-grade dysplasia and adenocarcinoma is up to 61% for MD-IPMN and mixed-type IPMN, and surgery is indicated. In BD-IPMN, the risk of high-grade dysplasia and adenocarcinoma is 31% and 18%, respectively. In addition to changes in imaging tests, progression to diabetes and elevated CA-19.9 levels should be considered alarming signs.

The type of surgery will depend on the location of IPMN. Laparoscopy is currently the preferred approach. Cephalic pancreaticoduodenectomy is the most common procedure (70%), followed by distal pacreatectomy (25%)²⁸. These figures are similar to those reported in our series. Intraoperative ultrasonography is vital to define management in many cases, with a sensitivity of 95-100%, rule out other lesions and confirm resectability²⁹. Once the pancreas has been resected, frozen section analysis of the resection margins of the main pancreatic duct is recommended, particularly in MD IPMN. In case of high-grade dysplasia, additional resection is recommended. If margins are still positive, total pancreatectomy is indicated⁷. Although enucleation has been proposed, it cannot be recommended due to the lack of long-term outcomes. Segmental pancreatic resections of the lesions with higher risk of malignancy are recommended for multifocal IPMNs, with close monitoring of the lesions left in the remnant³⁰.

The pathological examination of the surgical specimens identifies the histological features of IPMNs and the different subtypes: intestinal (27%), pancreaticobiliary (45%), oncocytic (4%) and gastric (13%). Sometimes, there may be more than one subtype within the same lesion^4,8,31. On occasions, IPMNs may progress to invasive carcinoma, following the sequence of low-grade, moderate-grade or high dysplasia and invasive carcinoma. About 20% of IPMNs found in autopsies present high-grade dysplasia and 30% have invasive carcinoma. Progression from adenoma to carcinoma occurs within 5-6 years and depends on the subtype of IPMN³⁰. In our experience, pure types of IPMNs were identified in only 25% of the cases: gastric type in 9 cases (64%) and intestinal type in 2 patients (14%). Three cases presented more than one subtype within the same lesion: 2 gastric type tumors + 1 intestinal type tumor and 1 gastric type tumor and pancreaticobiliary type tumor.

Uehara et al. reported that pancreatic ductal adenocarcinoma developed in 5 of 60 (8%) BD IPMNs during follow-up³¹. Survival at 5 years is 90-100% for non-invasive IPMNs resected and 31-60% for invasive IPMNs¹¹. In the American College of Surgeons National Surgical Quality Improvement Program, of 478 patients who underwent pancreatic resections, 10% were due to IPMNs, and 23% of these cases showed evidence of high-grade dysplasia or cancer⁷. These figures are similar to those observed in our study in which surveillance was more commonly indicated than surgery, following the current recommendations.

Marchegiani et al. reported that 65 of 381 patients (17%) of their series experienced recurrence after resection; 2 patients developed metachronous pancreatic ductal adenocarcinoma and 9 required reoperation³². There have not been any recurrences after surgical resection in this cohort.

In conclusion, the diagnosis and management of this type of lesions is currently standardized with guidelines that are updated along with advances in imaging tests and treatment options. The natural history of the disease and the events leading to progression to adenocarcinoma are better understood and can be suspected with the results of the imaging tests. When surgery is indicated, long-term survival should be similar to that of the surveillance group.