Intraductal oncocytic papillary neoplasms (IOPNs) are cystic neoplasms with intraductal growth and complex papillae composed of oncocytic cells. IOPNs have been reported both in the pancreas and biliary tree, and are most likely closely related in these 2 locations. In the pancreas, these rare tumors are now considered 1 of the 4 histologic subtypes of intraductal papillary mucinous neoplasm (IPMN). Significant differences in histology, immunophenotype, and molecular genetics have been reported between IOPNs and other IPMN subtypes. However, there are limited data regarding the clinical behavior and prognosis of IOPNs in comparison to other subtypes of IPMN. We review features of pancreatic IOPNs and discuss the differential diagnosis of other intraductal lesions in the pancreas.
Intraductal oncocytic papillary neoplasms (IOPNs) are rare cystic neoplasms composed of oncocytic cells in a complex papillary arrangement.1 They were first described in the pancreas in 1996 by Adsay et al2 and are now classified as 1 of 4 histologic subtypes of intraductal papillary mucinous neoplasm (IPMN) within the pancreas.3 The other 3 histologic subtypes of IPMN are gastric-type, intestinal-type, and pancreatobiliary-type.3,4 Additionally, IOPNs can occur in the intrahepatic and extrahepatic bile duct.5 Although IOPN is a rare entity and there are limited reports and long-term clinical data, significant histologic, immunophenotypic, and potentially biologic differences exist between IOPN and the other subtypes of IPMN.
CLINICAL FEATURES
The clinical features of IOPNs are similar to those of other IPMNs. Estimates of the incidence of IOPNs range from less than 1% to 13% of IPMNs (Table 1).6–9 IOPNs have been reported in adults ranging in age from 39 to 78 years, with a mean age of 67 years, and affect men and women equally.1 In a study of 283 IPMNs, Furukawa et al6 reported a mean age of 60.3 years for 24 patients with IOPN, compared to 65.6, 64.5, and 69.2 years for gastric, intestinal, and pancreatobiliary types, respectively, and suggested an increased likelihood of earlier onset of IOPNs compared with that of the other histologic subtypes. However, other studies9,10 have not reported a significant age difference. For example, a review by Liszka et al10 of 24 pancreatic IOPN cases reported a mean age of 63.9 years with no significant age or sex differences between IOPN and IPMN overall.
Symptoms of IOPN are nonspecific and similar to those of other IPMNs. In the abovementioned review by Liszka et al,10 the most frequent presentation of IOPN was abdominal pain, seen in 13 of 24 patients (54%), followed by pancreatitis in 4 of 24 patients (17%). Their review also found 1 of 24 patients (4%) with weight loss and 1 of 24 patients (4%) with jaundice, and they noted a lower prevalence of these symptoms than that reported in the review by Belyaev et al11 of 3725 patients with pancreatic IPMN of all subtypes, which found weight loss in 27% and jaundice in 16% of cases. Approximately 20% of patients are asymptomatic, similar to IPMNs in general.10 While diabetes mellitus has been associated with IPMNs including IOPN, there are currently no well-established risk factors for IPMN.
GROSS PATHOLOGY
Pancreatic IOPNs are large cystic masses with soft, tan to red-brown, friable, papillary luminal growths.4 By definition, pancreatic IPMNs and therefore also IOPNs are grossly visible and grow within pancreatic ducts. IOPNs generally have less mucin production than other IPMN subtypes.12 Adsay et al2 reported measurements between 1.6 to 15 cm, with a mean of 6.0 cm.
IOPNs arise predominantly in the head of the pancreas but can also arise in other areas in a distribution similar to that of other IPMNs.10 The original report by Adsay et al2 found 8 IOPNs of the pancreatic head compared to 3 of the body and tail. The review by Liszka et al10 found 19 of 28 cases (67.9%) in the head, 3 of 28 cases (10.7%) in the body, 3 of 28 cases (10.7%) in the tail, and 3 of 28 cases (10.7%) in multiple segments. By comparison, the review by Belyaev et al11 of IPMNs of all subtypes found 65% (range, 39%–89%) in the pancreatic head, 24% (range, 13%–31%) in the body, and 11% (range, 5%–21%) in the tail, making the distribution of IOPNs similar to that of IPMNs overall.
The AFIP (Armed Forces Institute of Pathology) Atlas of Tumor Pathology notes that most IOPNs do not involve the main pancreatic duct.1 The review by Liszka et al,10 however, found 10 of 12 cases (83.3%) in the main duct, 1 of 12 cases (8.3%) in the branch duct, and 1 of 12 cases (8.3%) in an accessory pancreatic duct.10 Furukawa et al6 found 8 of 24 main duct–type (33.3%), 12 of 24 branch duct–type (50%), and 4 of 24 mixed-type (16.7%) IOPNs. By comparison, the study of Kang et al9 of IPMNs of all subtypes found 22 of 213 main duct–type (10.3%), 142 of 213 branch duct–type (66.7%), and 49 of 213 mixed-type (23.0%) tumors.
MICROSCOPIC PATHOLOGY
Similar to IPMNs in general, IOPNs display intraductal growth within cystically dilated ducts (Figure, A). Unique microscopic features of IOPNs include complex, thick, arborizing papillae with delicate stroma, lined by multiple layers of cuboidal to columnar epithelium with abundant granular, eosinophilic cytoplasm (Figure, B and C). Intracellular and intraepithelial lumens containing mucin may be seen, which can form a cribriform arrangement or even fuse into solid or sheetlike growth with numerous vessels.4 Scattered goblet cells may be interspersed among the oncocytes. Tumor nuclei are uniform and round, with single eccentric nucleoli (Figure, D), but can also show more prominent cytologic atypia (Figure, E).4
Microscopic images of intraductal oncocytic papillary neoplasms. A, Intraductal oncocytic papillary neoplasm featuring intraductal growth and complex arborizing papillae composed of tumor cells with oncocytic cytoplasm. B, Intraductal oncocytic papillary neoplasms can have complex architecture with cribriform lumen formation. C, Intraductal oncocytic papillary neoplasm showing intraductal growth and florid papillary architecture. D, Intraductal oncocytic papillary neoplasm tumor cells with uniform, round nuclei, single eccentric nucleoli, and intraepithelial lumen formation. E, Intraductal oncocytic papillary neoplasm tumor cells in a papillary arrangement and showing enlarged, hyperchromatic, and pleomorphic nuclei, abundant oncocytic cytoplasm, and intraepithelial lumen formation. F, Hepatocyte paraffin-1 immunohistochemical stain shows nuclear positivity in intraductal oncocytic papillary neoplasm (hematoxylin-eosin, original magnifications ×40 [A], ×100 [B and C], and ×400 [D and E]; original magnification ×100 [F]).
Microscopic images of intraductal oncocytic papillary neoplasms. A, Intraductal oncocytic papillary neoplasm featuring intraductal growth and complex arborizing papillae composed of tumor cells with oncocytic cytoplasm. B, Intraductal oncocytic papillary neoplasms can have complex architecture with cribriform lumen formation. C, Intraductal oncocytic papillary neoplasm showing intraductal growth and florid papillary architecture. D, Intraductal oncocytic papillary neoplasm tumor cells with uniform, round nuclei, single eccentric nucleoli, and intraepithelial lumen formation. E, Intraductal oncocytic papillary neoplasm tumor cells in a papillary arrangement and showing enlarged, hyperchromatic, and pleomorphic nuclei, abundant oncocytic cytoplasm, and intraepithelial lumen formation. F, Hepatocyte paraffin-1 immunohistochemical stain shows nuclear positivity in intraductal oncocytic papillary neoplasm (hematoxylin-eosin, original magnifications ×40 [A], ×100 [B and C], and ×400 [D and E]; original magnification ×100 [F]).
Some authorities believe that IOPNs may automatically warrant categorization as an IPMN subtype with high-grade dysplasia, given their architectural complexity. In the published literature, IOPNs have been reported to be associated with an increased proportion of high-grade dysplasia, compared to IPMNs in general (Table 1). Features of high-grade dysplasia include architectural complexity with irregularly branching papillae and cribriform formation, and cytologic atypia such as increased mitotic figures, loss of nuclear polarity, nuclear pleomorphism, hyperchromasia, and prominent nucleoli.4,12 Liszka et al10 reported high-grade dysplasia in 79% of noninvasive IOPNs (11 of 14 cases). Similarly, both cases of IOPN reported by Kang et al9 showed high-grade dysplasia. By comparison, the study by Kang et al9 of IPMNs in general reported 18 of 213 IPMNs with high-grade dysplasia (8.5%).9
Invasive carcinoma may arise in the setting of IOPNs, similar to other subtypes of IPMNs (Table 1). Carcinomas arising from IOPN may retain or lose oncocytic features.6,10 Purely solid invasive oncocytic carcinoma has been associated with IOPN.10 The study by Furukawa et al6 found invasive oncocytic carcinoma in 4 of 24 IOPN cases (17%), but not in any other histologic subtypes of IPMN. While colloid carcinoma, characterized by mucinous differentiation with neoplastic cells floating in extracellular mucin pools,1 has been reported in roughly 28% to 50% of carcinomas arising from IPMN, pure colloid carcinomas have not been reported arising in the setting of IOPN.6,10
ANCILLARY STUDIES
A unique feature of IOPNs is immunohistochemical expression of markers for mitochondrial antigens, as well as hepatocyte paraffin-1 antibodies (HepPar1) (Figure, F). HepPar1 expression may not signify hepatic differentiation, as more specific markers show negativity, including the use of fluorescence in situ hybridization (FISH) for albumin.1 Furthermore, immunohistochemical studies investigating the expression of MUC, a heterogeneous group of mucin glycoproteins, have revealed different expression profiles between the 4 subtypes of IPMN (Table 1). IOPNs notably express MUC6, are variably positive for MUC1 and focally positive for MUC5AC, but are negative for MUC2 and CDX2 expression.4,13,14 The pancreatobiliary subtype has a similar immunophenotype featuring MUC6 expression and lack of MUC2/CDX2, but may show more consistent MUC1 positivity.14–16 In addition, the degree of MUC6 expression is usually less intense in the pancreatobiliary subtype than in IOPN. The gastric and intestinal subtypes, by contrast, are negative for MUC6 expression. The intestinal subtype instead features MUC2 and CDX2 expression.4,14–17
These different immunophenotypic MUC expression patterns provide clues into developmental pathways in IPMN subtypes. Several studies indicate that the oncocytic and pancreatobiliary subtypes may share a common and distinct pathway of differentiation, so-called pyloropancreatic lineage, which is distinct from the villous/intestinal pathway of differentiation. This is based on common expression of MUC6 seen in IOPN and pancreatobiliary subtype, but not in gastric or intestinal subtypes.14
IOPNs also share keratin expression in common with IPMNs overall.12 Other immunohistochemical stains reported as showing positivity in IOPN include B72.3, mesothelin in most cases, with inconsistent expression of carcinoembryonic antigen and CA 19-9. In addition, weak and focal positivity has been reported with trypsin and chromogranin.1
MOLECULAR PROFILE
Studies investigating the molecular characteristics of IPMNs mainly focus on the more common subtypes, and limited data are available regarding genetic alterations in IOPNs. KRAS mutations are the most common molecular alterations in IPMN, with activating mutations of codons 12 or 13 reported in approximately 30% to 80% of IPMNs.4,5,18 However, KRAS mutations are less frequently found in the IOPN subtype of IPMNs (Table 1). KRAS mutations were reported in 0 of 8 IOPN cases by Chung et al13 ; however, they have recently been reported in 3 of 18 cases (17%) by Xiao et al18 and in 2 of 3 cases (67%) by Amato et al.19
Other genes that have been found to be mutated in IPMN with less frequency than KRAS include BRAF, PIK3CA, GNAS, SMAD4, TP53, and CDKN2A. Limited data are available from studies of IOPNs detecting mutations in these genes.18,19 Notably, the study of 18 IOPNs by Xiao et al18 detected a BRAF mutation in 1 invasive case (6%) with no KRAS mutation, and a recent study by Amato et al19 described 1 of 3 IOPN cases (33%) with mutated GNAS.
Immunohistochemistry has also been used to investigate genetic changes in IOPNs. Findings of p53 overexpression, loss of p16 and SMAD4 expression, and abnormal nuclear β-catenin expression have been reported in IPMNs in general, and in a few studies of the IOPN subtype. Chung et al13 reported p53 staining in at least 25% of cells in 4 of 19 IOPNs (21%), as well as loss of p16 staining in 12 of 19 IOPNs (63%), both similar to IPMNs in general. Xiao et al18 reported loss of SMAD4 in 1 of 18 IOPNs (6%), and abnormal β-catenin expression in 5 of 18 IOPNs (28%). Clearly, more studies are needed to validate the existing data and to further delineate the molecular characteristics of IOPNs, in order to better understand their biologic characteristics, prognostic significance, and natural history in comparison with other IPMN subtypes.
DIFFERENTIAL DIAGNOSIS
The differential diagnosis for IOPN of the pancreas includes other neoplasms with intraductal growth, as well as other oncocytic lesions (Table 2). While some rare entities with oncocytic features can mimic IOPN, including metastatic tumors, in-depth discussion of these rare entities is beyond the scope of this review. Pancreatic neoplasms with intraductal growth include the other histologic subtypes of IPMN. Cases have been reported of multiple histologic subtypes of IPMN coexisting within 1 pancreatic lesion,9,15,18,20 as well as transitional areas between subtypes.1 One entity, the “intraductal tubular adenoma, pyloric gland type,” has been described in the literature, but this is a controversial entity that many authors regard as just a variant of the gastric subtype of IPMN.
Intraductal tubulopapillary neoplasm (ITPN) is an uncommon nodular tumor composed of densely packed glands, which comprises less than 3% of pancreatic intraductal neoplasms.4,12 These share an intraductal growth pattern and propensity for high-grade dysplasia with IOPN, and they can even show papillae, cribriforming areas, and complex architecture.4 However, unlike IOPNs, ITPNs show tubule formation and ductal differentiation as the predominant architectural pattern, and they have modest amounts of eosinophilic to amphophilic cytoplasm rather than the abundant, granular, eosinophilic cytoplasm of IOPNs.4 Most ITPNs express MUC1 and MUC6 by immunohistochemistry4 and are negative for MUC2 and MUC5AC, making them potentially related to IOPNs as part of the proposed pyloropancreatic pathway of differentiation previously discussed above.14 Unlike IOPNs, ITPNs are negative for HepPar1.
IOPNs with solid growth pattern should be differentiated from other pancreatic tumors. Acinar cell carcinomas, neuroendocrine tumors, and ductal adenocarcinomas can rarely grow intraductally,10,12 making diagnosis potentially difficult. To further complicate the picture, many of these tumors can rarely display oncocytic cytology, although these reports are considered incidental variants and not distinct entities.4,10,21 Acinar cell carcinoma, in particular, features abundant zymogen granules, which can mimic granular oncocytic cytoplasm, and can demonstrate intraductal growth and papillae.4 However, in contrast to IOPN, acinar cell carcinoma shows immunohistochemical staining with trypsin and chymotrypsin, as well as zymogen granules that are highlighted by the periodic acid–Schiff stain and are diastase resistant. Pancreatic neuroendocrine tumors (NETs) can show various growth patterns, making confusion possible with solid areas of IOPNs where prominent papillary architecture has been lost. The NET cytologic features include uniform cells with round to oval nuclei, coarse “salt and pepper” chromatin, and finely granular cytoplasm.4 Although these features can mimic IOPNs, particularly in cases of rare oncocytic variants,22 most NETs will express markers of neuroendocrine differentiation such as synaptophysin and chromogranin.
CURRENT TREATMENT
Similar to other types of IPMNs, IOPNs are managed by surgical resection, with international consensus guidelines published in 2012.23 Histologic subtypes of IPMN are discussed, but no specific differences in therapeutic approach are recommended. Surgical resection is recommended with follow-up determined by margin status, as well as radiographic surveillance of small lesions.
PROGNOSIS
Limited case numbers and follow-up studies prevent accurate comparison of prognosis between IOPN and IPMN overall.6,7,9,10,17 The 5-year survival rate of patients with pancreatic IOPN has been reported to be between 83.9%6 and 93%,18 compared to 75% in IPMNs overall (Table 1).11 The main determinant of prognosis for IPMNs in general is the presence of invasive carcinoma, which reduces the 5-year survival rate from 90% to 95% to between 27% and 60%.4 Invasive carcinoma is found in 25% to 50% of IOPNs,1,10 similar to the rate reported in IPMNs overall.4,8,13 However, minimal invasion of less than 0.5 cm is associated with improved survival and may be more frequently seen in IOPNs than other subtypes.6,15 Furukawa et al6 reported 5 of 24 IOPNs (20.8%) with minimal invasion, compared to 2.9%, 8.9%, and 5.3% of gastric-, intestinal-, and pancreatobiliary-type IPMNs, respectively. Colloid carcinoma confers a better prognosis than does ductal or tubular carcinoma4 but has only been reported to arise from the intestinal subtype of IPMNs and not from IOPN. Additional studies with long-term follow-up are warranted to accurately compare prognostic data between pancreatic IOPN and IPMNs overall.
CONCLUSIONS
IOPNs are cystic neoplasms with intraductal growth, complex papillary architecture, and oncocytic cytology. In the pancreas, they are now classified as the rarest of the 4 histologic subtypes of IPMN, and they notably have an increased proportion of high-grade dysplasia. IOPNs express HepPar1 by immunohistochemistry, which may help distinguish them from other entities in the differential diagnosis, including other subtypes of IPMNs and other intraductal neoplasms. Similar to other subtypes of IPMNs, invasive carcinoma may arise in the setting of IOPNs. Additional studies are needed to accurately compare prognostic data and genetic changes between IOPNs and other subtypes of IPMNs.
References
Author notes
The authors have no relevant financial interest in the products or companies described in this article.