Introduction

Intraductal papillary mucinous neoplasms (IPMN) are macroscopically detectable epithelial proliferations within the pancreatic ductal system, which may progress to invasive ductal carcinomas. Neuroendocrine tumours (NET) of the pancreas are usually slow-growing neoplasms, which are sometimes associated with symptoms of hormone secretion sydromes. Neuroendocrine microadenomas (NMA) are defined as pancreatic non-functioning neuroendocrine neoplasms of less than 5 mm in diameter. NMA are considered precursor lesions for NET [1].

Some investigators hypothesized that pancreatic NET may preferentially develop in patients with IPMN (or vice versa), so IPMN and NET may be found simultaneously in the pancreata at higher rate than expected. This suggested that IPMN/NET coexistence could be not necessarily accidental [2]. However, literature data on IPMN/NET association are limited [2–15].

Importantly, IPMN and NET may coexist as independent tumours within the pancreas separated by parenchymal tissue or they may form a single lesion [4]. The latter usually develops as a collision tumour, i.e. it consists of topographically related components of most likely independent origin without shared molecular alterations [14]. Recently, several investigators showed that IPMN and NET components within a single IPMN/NET mass may in fact share a molecular profile [13, 14]. This observation served as proof of a common origin of both components in some very rare IPMN/NET cases (so-called composite tumours) [13, 14]. Management of patients with IPMN/NET is not well established. Possibly they should be managed as it is indicated by the nature of each tumour component separately [7].

The purpose of the study was to determine the prevalence of NMA/NET in patients with IPMN in a large single-centre histopathological series of pancreatic resection specimens. For comparative purposes, specimens of patients with conventional pancreatic ductal carcinomas (PDAC) as well as ampullary carcinomas (AMPCA) were taken. The study was designed aiming to exhibit whether IPMN may favour the development of NMA/NET within pancreas. Additionally, the study focused on the microscopical appearance of IPMN/NET lesions/components in the context of potential relatedness of both entities.

Material and methods

Histopathological data on nearly consecutive in-house cases of IPMN of the pancreas diagnosed in resection specimens in the author’s institution between July 2015 and March 2021 were retrieved from a prospective database of pancreatic resection samples. Samples of conventional PDAC (i.e. not derived from IPMN) diagnosed between January 2017 and March 2021 as well as samples of AMPCA diagnosed between July 2015 and March 2021 were taken for comparative purposes. Cases were qualified into particular study groups based on histopathological diagnosis of the primary lesion, i.e. a mass which was the main indication for surgery [3].

A small number of cases encountered in the study period were excluded from the study for several reasons (e.g. specimens obtained following neoadjuvant therapy, rare specimens examined using only representative tissue sections due to technical/billing reasons, enucleation/limited resection specimens, specimens obtained in palliative resections, rare cases dissected by other pathologists).

Importantly, the present study was based on a standardized histopathological examination of surgical specimens. All included cases were macroscopically and microscopically examined by a dedicated pathologist (this author) in a standardized fashion, i.e. entire resected pancreatic tissue was taken for microscopical examination irrespective of its gross appearance. Pancreaticoduodenectomy specimens were processed using Leeds Pathology Protocol.

Histopathological diagnoses were established based on the WHO 2019 criteria [1]. Diagnoses established before 2019 were re-assessed for the purpose of the study. In particular, NMA was defined as a clinically-silent neuroendocrine proliferation less than 5 mm in diameter, and additionally:

An effort was made to distinguish NMA from its mimickers: islet aggregations and (pseudo)hyperplasia. In brief, islet aggregations are typically found in severely atrophic parenchyma, do not show expansive growth and/or trabecular tissue composition and usually retain topographical distribution of expression of pancreatic hormones (albeit an increase of glucagon-producing cells and a reduction of insulin-producing cells are allowed). Islet hyperplasia is a diffuse enlargement of islets (with or without cytological alterations in endocrine cells), which usually involves the entire pancreas and frequently results in clinical symptoms. Size, shapes as well as hormone distribution patterns of pancreatic islets in hyperplasia are altered. Islet pseudohyperplasia involves typically uncinate process and it is asymptomatic. In selected cases, hormone immunostains were helpful for differential diagnosis of NMA. Lack of expression or overexpression of a particular hormone, or abnormal distribution of pancreatic hormones within a lesion favoured NMA over reactive endocrine proliferations [15–17].

For comparison of continuous variables, Kruskal–Wallis ANOVA or Mann–Whitney U tests were used, as appropriate. Ordinal/nominal variables were compared using Chi2 tests. Statistical significance was set up at p ≤ 0.05.

The Institutional Review Board permitted to perform the present observational study without full review, which is necessary for interventional studies on humans, according to national regulations.

Results

Study groups

Some 232 resection specimens were included in the study: 51 cases of IPMN (with or without coexistent invasive carcinomas), 114 cases of PDAC and 67 cases of AMPCA. Invasive adenocarcinomas were found in 51% of samples with IPMN. Gastric and intestinal IPMN subtypes were most prevalent. Conventional tubular adenocarcinoma was the most common histological type among invasive carcinomas associated with IPMN (58%), PDAC (91%), and AMPCA (91%), as expected. All but one invasive cancers coexistent with IPMN were interpreted as derived from IPMN, a single sample was considered equivocal (i.e. it was not clear if invasive cancer was related pathogenetically to IPMN, as described below). Invasive cancers associated with IPMN as well as AMPCA were significantly smaller than conventional PDAC (median 17 mm vs. 15 mm vs. 31.5 mm, respectively, p < 0.001). Not surprisingly, the primary tumour stage and frequency of metastases in regional lymph nodes were lower in carcinomas derived from IPMN in comparison to PDAC. IPMN-related cancers were also enriched in G1 tumours.

There were also some differences related to types of specimens examined within study groups: the majority of conventional PDAC were found in pancreaticoduodenectomy specimens, and total pancreatectomies were performed mainly due to IPMN, as expected. Importantly, median numbers of tissue blocks examined per specimen were similar across study groups (p = 0.199). Details on demographical and histopathological characteristics of study populations were described in table I.

NMA/NET in the study groups

A single case of NET was found in the IPMN group (prevalence of 1.96%). This was a 7-mm incidentally detected, non-functional and presumably sporadic G1 tumour in a 61-year-old man treated with total pancreatectomy due to diffuse low-grade gastric mixed-duct type IPMN of the pancreatic head (formerly IPMN with moderate grade dysplasia). Additionally, a 3-mm focus of invasive G1 adenocarcinoma in close association with non-dilated duct of the pancreatic tail with high-grade pancreatic intraepithelial neoplasia was found. It was not clear if invasive cancer was pathogenetically related to IPMN, as dysplastic lesions of various grades were found in many dilated/non-dilated ducts within the pancreas. NET was seen in the pancreatic head, and it was composed of trabeculae in sclerotic stroma. NET extended focally to the peripancreatic adipose tissue, but perineural/vascular invasion was absent. NET was not associated topographically with IPMN (not shown). Additionally, three IPMN-independent NMA were found in the pancreatic head.

Two NET cases and a single NET case were found in patients with PDAC (2/114; 1.75%) and AMPCA (1/67; 1.49%), respectively. The frequencies of NET did not differ across all three study groups (IPMN vs. PDAC vs. AMPCA; p = 0.981).

These numbers contrasted with a prevalence of NMA, which were found in a larger proportion of IPMN cases (14/51; 27.45%), in comparison to PDAC cases (9/114; 7.89%) and AMPCA cases (5/67; 7.46%). That difference was statistically significant (p < 0.001) and the odds ratio for NMA in IPMN versus NMA in PDAC was 4.41 (95% CI; 1.61–12.49), whereas the odds ratio for NMA in IPMN versus NMA in AMPCA was 3.85 (95% CI; 1.24–13.17).

Enrichment of the IPMN group in NMA also resulted in larger proportion of NMA/NET counted in aggregate in specimens with IPMN (27.45%) in comparison to PDAC (8.77%) and AMPCA samples (8.96%, p = 0.002). Frequencies of NMA/NET across the study populations were summarized in table II.

IPMN with and without coexisting NMA/NET

Patients treated with pancreatic resections due to IPMN with and without NMA/NET (i.e. in essence: with/without NMA) did not differ in terms of their age and sex. Interestingly, NMA/NET in the IPMN group were somewhat frequent in distal pancreatecomy samples (5/16) and total pancreatecomy samples (8/19) in comparison to pancreaticoduodenectomy specimens (1/14 samples), but that difference did not reach statistical significance (p = 0.086). Coexistent NMA/NET was a rare finding when IPMN was localized in the pancreatic head (1/22), but frequent when IPMN diffusely involved the entire pancreas (8/12) (p < 0.001). The histological type of IPMN, the grade of dysplasia, or the presence of invasive cancer did not affect the prevalence of NET/NMA in patients with IPMN. However, NMA/NET were more likely to be found in IPMN samples with limited invasion (p = 0.007). The prevalence of NMA/NET in IPMN samples was not biased by the volume of resected pancreatic tissue (p = 0.527). Histopathological characteristics of IPMN with and without NMA/NET were compared in table III.

NMA in the IPMN group – histopathological characteristics

NMA were found in 14 specimens showing all grades and histological types of IPMN. The number of NMA per specimen ranged from 1 to 6 (median 1). Diameter of NMA ranged from 0.5 mm to 3.8 mm. In 8/14 cases NMA were found in the distal pancreas, in 5/14 – in the pancreatic head, in a single case NMA were found in both segments of the pancreas. NMA were localized within atrophic lobules in a half of IPMN samples. Direct connection of NMA and ducts was found in only two NMA across all the 14 IPMN samples with NMA:

Figure 1. Neuroendocrine microadenoma (NMA) of the pancreas with partial intraductal spread but without direct contact with intraductal papillary mucinous neoplasm (IPMN) (sample no. 10). A – NMA surrounding small duct (lower left of the image, lesion indicated by a rectangle) without direct relationship with IPMN (upper right of the image); B – partial intraductal spread of the NMA is seen at higher magnification; C and D – NMA showed chromogranin-A (not shown) and synaptophysin expression; D – note synaptophysin-negative ductal cells above intraductal spread of neuroendocrine cells. Magnifications: fig. 1A (1.25×), 1B (5×), 1C (5×), 1D (30×)

Figure 2. Neuroendocrine microadenoma (NMA) in the atrophic pancreatic lobule surrounded by intraductal papillary mucinous neoplasm (IPMN) (sample no. 13). A and B – lesion indicated b a rectangle – NMA was found in an atrophic lobule surrounded by gastric-type IPMN with partial intraductal tubular adenoma growth (not shown); C – NMA was glucagon-positive; D – NMA was insulin-negative, and such hormone expression pattern excluded diagnosis of islet aggregation; note insulin-positive cells in residual islets of the atrophic lobule. Magnifications: fig. 2A (3.5×), 2B (10×), 2C (10×), 2D (10×)

Other NMA did not develop within IPMN (as reported previously in composite IPMN/NET tumours [13, 14]), and did not have a direct connection with ductal epithelium. Representative examples of some NMA are depicted in fig. 3. Histopathological details of NMA/NET in samples with IPMN are presented in table IV.

Figure 3. Neuroendocrine microadenomas (NMA) in other pancreatic specimens with intraductal papillary mucinous neoplasm (IPMN). A and B – conventional NMA (indicated by a rectangle) in pancreatic parenchyma without relationship to IPMN; another NMA (indicated by a rectangle) in pancreas with extensive atrophy. C and D – NMA was next to a duct with mucin leakage. NMA was chromogranin A-positive (not shown) and synaptophysin-positive (not shown). Magnifications: fig. 3A (3×), 3B (15×), 3C (2×), 3D (5×)

Type of resection specimens as a potential confounding factor

As described above, NMA/NET in IPMN specimens were found mainly in distal pancreatectomy and total pancreatectomy samples rather than in pancreaticoduodenectomy samples. This may suggest a systematic bias, as distal pancreas was infrequently resected in patients with PDAC/AMPCA (tab. I). However, distal pancreatectomy specimens with IPMN were enriched in NMA (5/16; 31.25%) in comparison to distal pancreatectomy specimens with PDAC (3/30; 10%). The difference was not significant (p = 0.070), possibly due to the limited number of distal pancreatectomy samples. Prevalence of NMA/NET in pancreaticoduodenectomy specimens did not differ across the three study groups (p = 0.964).

Discussion

The present study has shown that pancreata with IPMN may be enriched in neuroendocrine neoplasms. In particular, NMA were found in 27% of IPMN specimens. The majority of NMA were found in distal pancreas, and distal pancreatectomies performed due to IPMN were specifically enriched in NMA in comparison with PDAC specimens. However, NMA in IPMN specimens were usually solitary lesions. Microadenomatosis (i.e. multiple, usually uncountable NMA [1]) was not found in any case. No case showed a histopathological picture suggestive of composite IPMN/NET(NMA), as NMA/NET tissue was not in direct contact with IPMN.

NMA/NET prevalence

The results of the study should be interpreted in the context of “baseline” prevalence of NMA/NET in the population. The real incidence of NMA/NET is difficult to estimate, since only a fraction of them come to clinical attention due to symptoms or as asymptomatic “incidentalomas” found during diagnostic/
/radiologic examinations performed for other reasons. Incidence of clinically-detected pancreatic NET has increased substantially during the last decades [18], but it is still below rates based on autopsy studies [16]. In their autopsy study, Kimura et al. found NMA/NET in 6/60 (10%) totally embedded pancreata and in 12/738 (1.6%) pancreata examined by representative tissue sections [16]. In other autopsy studies, the frequency of pancreatic NET ranged from 0 to 1.4% (as reviewed in [16] and [19]). The prevalence of NMA/NET in retrospective clinical-histopathological studies ranged from 1.4% [15] to 4% [3].

In the present study, the prevalence of NMA/NET in the “control” groups (i.e. PDAC and AMPCA) was between 7 and 9%. These numbers were close to the autopsy-based results in totally embedded pancreata (10%) [16]. Therefore, it can be assumed that the baseline frequency of NET/NMA under the chosen diagnostic approach is probably somewhere around 7–10%. It should be emphasized that the majority of NMA/NET are self-limiting lesions of little clinical significance [3]. However, some rare NMA may have the potential for malignant behaviour [15]. In 2023, the WHO classification of NMA will be renamed as neuroendocrine microtumours as they may very rarely metastasize to the lymph nodes [20].

It should also be kept in mind that the prevalence of NMA in surgical specimens is strongly related to numerous laboratory factors:

NMA/NET in patients with IPMN

NMA/NET coexisting with IPMN are rare, with less than 50 reported cases [7]. Currently, it is not fully clear whether IPMN and NMA/NET coexist more frequently than expected, And whether there is a causal relationship between IPMN and NMA/NET [2, 4]. The only exception are recent reports which have proven that both IPMN and NET components within a single composite IPMN/NET lesion may be clonally related [13, 14]. On the contrary, other investigators did not find any evident molecular relationship between topographically related IPMN and NET [5]. These observations indicate that clonally related IPMN/NET is possible, but remains an exceedingly rare lesion.

Results of the previous studies on IPMN/NET association gave inconsistent results. Sahora et al. noticed that the prevalence of NET in patients with IPMN is similar to the general population; the frequency of NET in specimens with IPMN is similar to the frequencies of other incidental pancreatic neoplasms [8]. In contrast, Marrache et al. [2] and Goh et al. [10] found that the frequency of coexistent IPMN/NET was higher than expected. According to the literature data, IPMN was seen in 2.9% (1/35) to 6.5% (3/46) of NET specimens [2, 7, 10], and NET was found in 1.1% (5/441) to 13.6% (3/22) of IPMN specimens [2, 3, 7, 8, 10, 12, 15]. In this study, the prevalence of NMA/NET among IPMN cases was higher than in previous reports (27%), and this number was related not only to the exact NMA/NET prevalence in the study population, but also to the preconceived diagnostic/screening approach.

Pathogenesis of NMA/NET in patients with IPMN

Little is known about pathogenesis of IPMN/NET coexistence. Some risk factors may be involved in the development of both NET and IPMN. Diabetes, a family history of cancer and chronic pancreatitis increase the risk of both NET [21–23] and IPMN [24].

Molecular profiles of IPMN and NET are different [2], with the exception of composite IPMN/NET [13, 14]. It was hypothesized that IPMN/NET forming a single lesion may develop from a common progenitor, or in transdifferentiation of a single cell type into another cell type [2, 7]. It was also speculated that NET may appear as a result of endocrine differentiation/hyperplasia within IPMN [10], but this hypothesis was not convincingly confirmed [2, 6]. Endocrine/paracrine stimulation of exocrine cells by NET-generated hormones may also play a role [11].

At the moment, it is not possible to indicate the exact reason(s) which can lead to increased prevalence of NMA/NET in patients with IPMN. One may speculate that exocrine-endocrine cross-talk [25] is involved in the pathogenesis of IPMN/NET coexistence, but this requires further study. Another hypothesis which could explain increased prevalence of NET/NMA in pancreata with IPMN is related to tissue remodelling, as seen in obstructive chronic pancreatitis/pancreatic atrophy. The histopathological features of obstructive pancreatitis are frequently seen in pancreata with IPMN (personal observation). Histopathological alterations of neuroendocrine cells have been recognized in pancreatic atrophy for decades [16, 17]. Chronic pancreatitis is considered a risk factor for NET development [22]. In the present study, half of NMA in IPMN specimens was found within atrophic lobules. It is possible that diffuse obstructive atrophy in pancreata with IPMN could promote NET/NMA development (the issue under study).

In summary, IPMN and NET/NMA coexisting in the pancreas may be considered as:

Study strength and limitations

The strength of the study was its design, based on the prospective inclusion of totally embedded pancreatic specimens which were examined histopathologically in a uniform fashion. To the best of our knowledge, this was the first report on IPMN/NET association based on prospectively examined, totally embedded pancreatic resection specimens.

The limitations of the study were the following:

Conclusions

The prevalence of NMA in pancreatic specimens with IPMN was 27% and it was significantly increased when compared to specimens with PDAC/AMPCA. The majority of NMA in IPMN specimens were solitary and localized within the distal portion of the pancreas. Topographical association of IPMN and NMA was rare and no case in the present series showed features suggestive of composite (clonal) IPMN/NMA. IPMN/NMA association may serve as a model for investigation of exocrine-neuroendocrine interaction. The reasons for IPMN/NET coexistence are largely unknown and require further study.

Article information and declarations

Ethics statement

The Institutional Review Board was permitted to perform the present observational study without full review, which is necessary for interventional studies on humans, according to national regulations.

Author contributions

Manuscript prepared by a single author.

Financial support

Internal institutional grant no. PCN-1-177/N/2.

Conflicts of Interest

None declared

Łukasz Liszka

Medical University of Silesia

Department of Pathomorphology and Molecular Diagnostics

ul. Medyków 18

40-754 Katowice, Poland

e-mail: lliszka@sum.edu.pl

Received: 16 Sep 2023

Accepted: 24 Oct 2023

References