Grading was independently assessed by individual participants, as shown in Table 1. The most disagreed case during grading was case #11, which was assessed as noninvasive IPMC/IPMN with high-grade dysplasia by 4 participants, invasive IPMC/IPMN with an associated invasive carcinoma by 4 participants, IPMA/IPMN with low- or intermediate-grade dysplasia by 1 participant, and other by 1 participant (Figs. 1A, B). This case was composed of dilated ducts with eosinophilic cells forming blunt-edged, complex papillae. The eosinophilic cells had enlarged nuclei with loss of polarity. Adjacent to the dilated duct, mucus spilled over in the stroma from a destructed portion of the duct that contained neoplastic cells. We discussed whether the mucus pooling was an invasive lesion. From a general pathological point of view, a mucus lake containing floating neoplastic cells should be assessed as an invasive lesion; however, some participants argued that the mucus pooling with neoplastic cells in this case could be a result of rupture/breakage of the ductal wall with mucus spillover and desquamated neoplastic epithelial cells. This case was finally determined to be an invasive IPMC/IPMN with an associated invasive carcinoma because the clusters of neoplastic cells in the spilled-over mucus could not be ruled out as a result of invasion.
The next-most disagreed cases were cases #6 and #16, both of which were assessed as noninvasive IPMC/IPMN with high-grade dysplasia by 6 participants and as IPMA/IPMN with low- or intermediate-grade dysplasia by 4 participants (Figs. 1C, D for case #6 and Figs. 1E, F for case #16). These cases showed dilated ducts having villous irregular papillae with stratified enlarged cigar-shaped nuclei and abundant mucinous cytoplasm. These might be tumors on the border between intermediate- and high-grade dysplasia. On discussion, the participants agreed that high-grade dysplasia was present, at least focally, in the markedly irregular shapes of the villi with exaggerated stratification of the enlarged, cigar-shaped, atypical nuclei with hyperchromatism and mitoses.
The diagnoses of 5 cases were initially agreed upon by 70% of the participants. Case #2 was assessed as a noninvasive IPMC/IPMN with high-grade dysplasia by 7 participants, as invasive IPMC/IPMN with an associated invasive carcinoma by 2 participants, and as IPMA/IPMN with low- or intermediate-grade dysplasia by 1 participant (Figs. 1G, H). This case showed irregular complex papillae composed of cells with disorientated, enlarged nuclei and scant cytoplasm. There were several epithelial clusters in the stroma adjacent to the dilated duct, which 2 participants considered to indicate invasion (Figs. 1H). On discussion, the participants agreed that these clusters were unlikely to be invasion because the desmoplastic reaction surrounding the clusters of cells was not definitely observed and the basement membrane was recognizable. Therefore, this case was diagnosed as noninvasive IPMC/IPMN with high-grade dysplasia.
Case #7 was assessed as noninvasive IPMC/IPMN with high-grade dysplasia by 7 participants, as invasive IPMC/IPMN with an associated invasive carcinoma by 2 participants, and as IPMA/IPMN with low-grade dysplasia by 1 participant (Figs. 1I, J). This case showed dilated ducts with thick arborizing papillae composed of eosinophilic cells with enlarged disorganized nuclei. Some participants argued that invasion could be assessed because of the fibrovascular stroma of papillae inside the dilated duct (Fig. 1J). A consensus was reached that invasion should be assessed when neoplastic cells were noted beyond confinement of a duct although it should be reported if definite vascular or lymphatic invasions were noted in the stroma inside the duct. The consensus diagnosis of this case was noninvasive IPMC/IPMN with high-grade dysplasia.
Case #14 was assessed as invasive IPMC/IPMN with an associated invasive carcinoma by 7 participants and as other by 3 participants (Figs. 1K, L). This case showed a typical appearance of mucinous colloid carcinoma. An argued point was whether this case was IPMN. On discussion, a ductal wall–like structure with neoplastic cells was noted, which was confirmed as a portion of dilated duct by the presence of elastic fibers in elastica staining and neoplastic epithelial cells consistent with IPMN due to the expression of MUC2 and MUC5AC; thus, the consensus diagnosis was invasive IPMC/IPMN with an associated invasive colloid carcinoma.
Case #17 was assessed as noninvasive IPMC/IPMN with high-grade dysplasia by 7 participants, as invasive IPMC/IPMN with an associated invasive carcinoma by 2 participants, and as other by 1 participant (Figs. 2A, B). This case showed dilated ducts with villous papillae composed of cells with enlarged, cigar-shaped, and disoriented nuclei. There was a mucus lake partially lined with dysplastic epithelia in the stroma adjacent to the dilated duct. Discussion suggested that this lesion might be an existing duct with partially desquamated epithelial cells, which should be examined by staining of the elastic fibers circumventing the existing ducts. An additional examination with elastica staining showed elastic fibers circumventing the mucus lake, which proved that the lesion was actually an existing duct. The consensus diagnosis was noninvasive IPMC/IPMN with high-grade dysplasia.
Case #20 was assessed as IPMA/IPMN with low- or intermediate-grade dysplasia by 7 participants and as noninvasive IPMC/IPMN with high-grade dysplasia by 3 participants (Figs. 2C, D). This case showed thick-fingered, complex papillae partially showing villous architecture in a dilated duct with acellular mucus lakes in the stroma. Cells consisting of papillae showed basally arranged nuclei with some disorientation and abundant mucinous cytoplasm. It was argued that this case might be on the border between intermediate- and high-grade dysplasia; however, most participants thought that the cellular features were organized enough to be assessed as low- or intermediate-grade dysplasia. Therefore, the consensus diagnosis was IPMA/IPMN with low- or intermediate-grade dysplasia.
Types of Papillae
Papillae were typed independently by individual participants, as shown in Table 2. In general, the participants agreed that assessing papillae types only by histological features in hematoxylin and eosin staining was sometimes difficult and that immunohistochemical staining for mucin proteins would be of great help. Also noted were some cases with mixed features of types. In these cases, the consensus agreement was that the most dysplastic papillae type should be used as the representative type for assessment.10 We also found that oncocytic-type papillae were likely to have an inconsistent assessment.
Some disagreements were observed in each case. Agreements as low as 50% among participants were observed in 3 cases. Case #12 was assessed as gastric type by 5 participants, as pancreatobiliary type by 3 participants, and as other by 2 participants (Figs. 2E, F). This case showed dilated ducts with highly dysplastic but relatively low papillae and associated invasive tubular components. On discussion, it was noted that the papillae resembled foveolar glands with high-grade atypia and seemed less complex than those seen in pancreatobiliary-type papillae; hence, the consensus diagnosis for this case was gastric type. The immunohistochemical features of the papillae were MUC1-negative, MUC2-negative, MUC5AC-positive, and MUC6-positive, which were consistent with gastric-type papillae.
Case #14 was assessed as intestinal type by 5 participants and as other by the remaining 5 participants (Figs. 1K, L). This case showed typical histologic features of mucinous colloid carcinoma, and as discussed previously, the main point of disagreement was whether this tumor was associated with IPMN. On discussion, the participants agreed that this case was intestinal IPMN with an associated invasive colloid carcinoma. Immunohistochemical revealed MUC1-negative, MUC2-positive, MUC5AC-positive, and MUC6-negative findings in the dysplastic epithelial lining of a component that seemed to be a ducal wall, which was consistent with intestinal-type IPMN.
Case #15 was assessed as pancreatobiliary type by 5 participants, as intestinal type by 3 participants, as gastric type by 1 participant, and as oncocytic type by 1 participant (Figs. 2G, H). This case showed dilated ducts with eosinophilic papillae in markedly irregular shapes within the intraepithelial lumina, which consisted of cells with atypical round nuclei with marked disorientation. Immunohistochemical features showed MUC1-negative, MUC2-positive, MUC5AC-positive, and MUC6-positive findings. These features indicated that this case was oncocytic-type IPMN, with which the participants agreed.
Case #11 was assessed as oncocytic type by 6 participants, as pancreatobiliary type by 3 participants, and as other by 1 participant (Figs. 1A, B). This case showed dilated ducts with eosinophilic cells forming blunt-edged, complex papillae, as described previously. The consensus was that this case was oncocytic type, as proved by the MUC1-negative, MUC2-positive, MUC5AC-positive, and MUC6-positive immunohistochemical findings.
Case #18 was assessed as pancreatobiliary type by 6 participants, as other by 2 participants, as gastric type by 1 participant, and as intestinal type by another participant (Figs. 2I, J). This case showed dilated ducts with well-developed papillae partially consisting of tubular glands and associated invasive ductal components in the adjacent stroma, which was assessed as invasive IPMC/IPMN with an associated invasive ductal carcinoma. It was noted that this case had mixed features of pancreatobiliary type and gastric type with a pyloric gland variant. The original literature recommended that IPMNs with mixed papillae features should be assessed as the type of the most atypical papillae; thus, because marked atypical features were evident in a part of the pancreatobiliary-type papillae, the consensus was that this case was of pancreatobiliary type. The immunohistochemical features of the pancreatobiliary-type papillae were MUC1-positive, MUC2-negative, MUC5AC-positive, and MUC6-positive, which confirmed the type.
Case #20 was assessed as intestinal type by 6 participants, as gastric type by 2 participants, and as other by 1 participant (Figs. 2C, D). This case showed thick-fingered, complex papillae partially showing villous architecture in a dilated duct with acellular mucus lakes in the stroma, as described previously. The cells comprising the papillae showed basally arranged, cigar-shaped nuclei and abundant cytoplasm. Some participants argued that the papillae seemed to be gastric rather than intestinal type because the villous shapes of papillae were not prominent; however, the consensus diagnosis of this case was intestinal type because some villous papillae were seen and the cigar-shaped nuclei were characteristic of this type. The immunohistochemical features were MUC1-negative, MUC2 -positive, MUC5AC-positive, and MUC6-negative, which was consistent with intestinal-type papillae.
Cases Without Disagreement
Cases with good agreement are shown in the Supplementary Figures 1 and 2, http://links.lww.com/MPA/A712.
This study revealed that the following features might cause disagreements between pathologists:
- Mucus lakes partially lined with neoplastic cells
- Rupture of ductal walls and mucus spillover
- Invasion into the fibrovascular stroma inside the duct
- Mucinous colloid carcinoma without obvious dilated ductal lesions
- Distinction between low/intermediate and high grade in intestinal-type papillae
- Distinction between oncocytic and pancreatobiliary types
- Distinction between gastric type, high-grade, and pancreatobiliary type
- Mixed features of types
We discussed these points of disagreement to develop the following recommendations:
- Mucus lakes containing floating neoplastic cells are noted as evidence of invasion. However, mucus lakes lined with neoplastic cells could be either invasion or existing ducts with partially desquamated epithelial cells. Elastic fiber staining might help to distinguish these conditions because existing ducts are lined with elastic fibers. However, the drawbacks of this recommendation are that the elastic fibers are sometimes too thin to recognize or may be absent in small ducts. Careful examination is necessary in such cases.
- Especially in intestinal-type IPMNs, the ducts are often destroyed and the mucus spills over into the stroma. Such destruction could be a result of either invasion or rupture. Lesions harboring floating neoplastic cells can be assessed as invasion, although there is the possibility that the floating neoplastic cells are desquamated epithelial cells entering the destructed portion. The desquamated epithelial cells may show a sheet-like appearance that could be different from pleomorphic appearances of clusters of invading cells.
- Invasion of IPMNs is usually assessed in the stroma beyond the contour of the duct. However, invasion could occur in the fibrovascular stalks of papillae inside the duct, which may be difficult to identify because of the complex infoldings and outfoldings of the papillae. Careful evaluation of the basement membranes may be necessary to assess such lesions, similar to the invasion of the lamina propria in intramucosal carcinomas of the stomach. In addition, the clinical significance of intraductal invasion is not known, although vascular or lymphatic invasion could be significant. A consensus has been reached that intraductal invasion can be dismissed, except for vascular or lymphatic invasion.
- Mucinous colloid carcinoma is a nodular tumor consisting of clusters of mucus pools containing floating neoplastic cells. Most are believed to be associated with IPMN; however, the characteristics of IPMN, most obviously dilated ducts lined with papillary neoplastic epithelia, are sometimes difficult to recognize in the tumor. Elastic fiber staining may help in the identification of ductal lesions.
- According to the 2010 WHO classification, “IPMNs with low-grade dysplasia are characterized by a single layer of well-polarized cells, small and uniform nuclei with only mild polymorphism, and rare mitosis. IPMNs with intermediate-grade dysplasia have nuclear stratification, crowding, and loss of polarity. The nuclei are enlarged and moderately hyperchromatic. The papillae maintain identifiable stromal cores. IPMNs with high-grade dysplasia are characterized by severe architectural and cytological atypia, with the formation of irregular branching papillae and sometimes cribriform growth. The epithelial cells lack polarity and the nuclei are stratified, hyperchromatic, and polymorphic. Mitoses are frequently observed, can near the luminal surface.”5 Basturk et al6 proposed a 2-tiered grading system consisting of low- and high-grade classifications, in which the former corresponded to low- and intermediate-grade dysplasia, whereas the latter corresponded to high-grade dysplasia. Our consensus is in the same line as that of the 2-tiered grading system. Noninvasive IPMC/IPMN with high-grade dysplasia corresponds to lesions showing the uppermost atypia, in which the nuclear enlargement, stratification, loss of polarity, and irregularity of shapes of papillae were more exaggerated compared with those of low-/intermediate-grade dysplasia (compare panels Figs. 1C, D and Figs. 2C, D). The nucleocytoplasmic ratio (N/C ratio) could be another hallmark of atypia; however, in IPMN, especially intestinal-type papillae, abundant mucinous cytoplasm may cause a relatively small N/C ratio; thus, the N/C ratio is less useful for grading in intestinal-type IPMNs than in other types.
- Oncocytic-type IPMNs are characterized by thick arborizing papillae consisting of cells with disoriented round nuclei and eosinophilic cytoplasm.10,14 Intraepithelial lumina are hallmarks. However, pancreatobiliary-type IPMNs are characterized by complex thin-branching papillae consisting of columnar cells with marked atypical nuclei and neutral or basophilic cytoplasm.5 Some oncocytic-type IPMNs show complex branching papillae with less eosinophilic cytoplasm, which somewhat resemble pancreatobiliary-type IPMNs. However, the disoriented round nuclei and intraepithelial lumina can be at least partially recognized, sometimes in the periphery, which leads to an assessment of oncocytic-type IPMN. Pancreatobiliary-type IPMNs show expression of MUC1, MUC5AC, and MUC6, whereas oncocytic-type IPMNs show consistent expression of MUC5AC and MUC6 and variable expression of MUC1 and MUC2.10,15
- High-grade gastric-type IPMNs show relatively low but complex papillae consisting of cells with marked atypical nuclei and neutrophilic cytoplasm.11 However, pancreatobiliary-type IPMNs show thin-branching complex papillae consisting of cells with marked atypical nuclei and neutral or basophilic cytoplasm. The absence of MUC1 expression can distinguish between high-grade gastric-type and pancreatobiliary-type IPMNs.
- In IPMNs showing mixed features of types, the type of the most dysplastic papillae should be assessed as the representative type of such tumors.10 If a tumor contains mixed types of equivalent dysplastic grades, the most prevalent type should be assessed as the representative type.10
This study has several limitations. First, virtual slide digital images were used to grade and type the IPMNs. All participants felt that the virtual slide digital images were difficult to assess compared with conventional glass slides, which might have caused evaluation disagreements. Second, only hematoxylin and eosin–stained images were used for assessment. In routine practice, difficult cases are assessed using special stains and immunohistochemistry. Indeed, the cases with disagreement were solved using special stains including elastica staining as well as immunohistochemistry analysis of mucin protein expression. The third limitation was the small number of cases with some selection bias.
In conclusion, we revealed points of disagreement among pathologists in assessing IPMN grades and types and developed a consensus and recommendations, which should be of help in making consistent assessments and interpretations for the pathological diagnosis of IPMNs.
The authors are grateful to Dr. Akira Saito, StaGen, Co, Ltd, Tokyo, Japan, for assisting with statistical analysis.
1. Ohhashi K, Murakami Y, Maruyama M, et al. Four cases of mucous secreting pancreatic cancer. Prog Digest Endosc
2. Furukawa T, Takahashi T, Kobari M, et al. The mucus-hypersecreting tumor of the pancreas. Development and extension visualized by three-dimensional computerized mapping. Cancer
3. Yanagisawa A, Ohashi K, Hori M, et al. Ductectatic-type
mucinous cystadenoma and cystadenocarcinoma of the human pancreas: a novel clinicopathological entity. Jpn J Cancer Res
4. Hruban RH, Takaori K, Klimstra DS, et al. An illustrated consensus on the classification of pancreatic intraepithelial neoplasia and intraductal papillary mucinous neoplasms. Am J Surg Pathol
5. Adsay NV, Fukushima N, Furukawa T, et al. Intraductal neoplasms of the pancreas. In: Bosman FT, Hruban RH, Carneiro F, et al, eds. WHO Classification of Tumours of the Digestive System
. 4th ed. Lyon, France: IARC; 2010:304–313.
6. Basturk O, Hong SM, Wood LD, et al. A revised classification system and recommendations from the Baltimore consensus meeting for neoplastic precursor lesions in the pancreas. Am J Surg Pathol
7. Japan Pancreas Society. Classification of Pancreatic Carcinoma
. 4th English ed. Tokyo, Japan: Kanehara Co. Ltd.; 2017.
8. Nakamura A, Horinouchi M, Goto M, et al. New classification of pancreatic intraductal papillary-mucinous tumour by mucin expression: its relationship with potential for malignancy. J Pathol
9. Adsay NV, Merati K, Basturk O, et al. Pathologically and biologically distinct types of epithelium in intraductal papillary mucinous neoplasms: delineation of an “intestinal” pathway of carcinogenesis in the pancreas. Am J Surg Pathol
10. Furukawa T, Klöppel G, Adsay NV, et al. Classification of types of intraductal papillary-mucinous neoplasm of the pancreas: a consensus study. Virchows Arch
11. Furukawa T, Hatori T, Fujita I, et al. Prognostic relevance of morphological types of intraductal papillary mucinous neoplasms of the pancreas. Gut
12. Kim J, Jang KT, Mo Park S, et al. Prognostic relevance of pathologic subtypes and minimal invasion
in intraductal papillary mucinous neoplasms of the pancreas. Tumour Biol
13. Tsutsumi Y, Onoda N, Osamura RY. Victoria blue-hematoxylin and eosin staining: a useful routine stain for demonstration of venous invasion
by cancer cells. J Histotechnol
14. Adsay NV, Adair CF, Heffess CS, et al. Intraductal oncocytic papillary neoplasms of the pancreas. Am J Surg Pathol
15. Xiao HD, Yamaguchi H, Dias-Santagata D, et al. Molecular characteristics and biological behaviours of the oncocytic and pancreatobiliary subtypes of intraductal papillary mucinous neoplasms. J Pathol
IPMN; high-grade; low-grade; invasion; type; pathology
Supplemental Digital Content
Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved.