Indian Journal of Pathology and Microbiology

: 2021  |  Volume : 64  |  Issue : 4  |  Page : 693--701

MUC1, CK20, and CDX2 immunohistochemical markers can sub-classify periampullary carcinomas into pancreaticobiliary, intestinal, and mixed subtypes

K Murali Manohar1, M Sasikala1, S Anuradha1, C Ramji1, RB Sashidhar2, GV Rao3, R Pradeep4, D Nageshwar Reddy3,  
1 Research Labs, Asian Healthcare Foundation, AIG Hospitals, Somajiguda, Telangana, India
2 Department of Biochemistry, University College of Science, Osmania University, Hyderabad, Telangana, India
3 Research Labs, Asian Healthcare Foundation, AIG Hospitals; GI Surgery Divison, Asian Institute of Gastroenterology, Somajiguda, Telangana, India
4 GI Surgery Divison, Asian Institute of Gastroenterology, Somajiguda, Telangana, India

Correspondence Address:
M Sasikala
Asian Health Care Foundation, Facility Block, AIG Hospitals, Mindspace Rd, Gachibowli, Hyderabad - 500 032, Telangana


Introduction: Pancreaticobiliary subtype of Periampullary carcinoma (PAC) has a poor prognosis in comparison to the intestinal subtype. We assessed the potential of cytokeratins and mucin markers to classify the sub-types of periampullary tumors and compared them with the survival data to identify markers that may predict prognosis. Methodology: PAC tumor tissues were obtained from 94 patients undergoing Whipples Pancreaticoduodenectomy. Paraffin-embedded tissues were immunostained with cytokeratins CK7, CK20), mucins (MUC1, MUC2, MUC5Ac), and CDX2 antibodies. The survival status of patients was obtained as follow-up up to 5-years of surgery. The Receiver Operating Character Curve (ROC) analysis was used for detecting sensitivity and specificity. The survival data were analyzed using the Kaplan-Meier survival curve. Results: Tumors were initially categorized on the basis of histological classification as pancreaticobiliary (n = 46), intestinal (n = 35) and indeterminate (n = 13). Further, using immunohistochemical markers (MUC1, CK20, and CDX2), we gave systematic classification of IHC-PB (n = 51), IHC-Int (n = 30) and IHC-Mixed (n = 13). The interobserver analysis showed good agreement between histologic and IHC type with a kappa value of 0.554. Combined expression of CK20, MUC1 and CDX2 accurately classify the mixed type of tumor. Overall survival rate and duration were 74.4% and 44.95 ± 2.29 months. Survival analysis for subtypes reveal, pancreaticobiliary tumors have low survival (27.9 ± 1.63 months) than mixed type (35.5 ± 0.45 months) and intestinal-type (52.92 ± 2.18 months). Among these, intestinal-type have better survival. Only TNM Stage III (tumor staging as per American Joint Committee on Cancer classification) and perineural invasion have been associated with predicting poor survival in PAC patients. Conclusion: Our results suggest that the combined expression of MUC1, CK20 and CDX2 could serve as markers to diagnose histological inconclusive specimens as mixed subtype tumors.

How to cite this article:
Manohar K M, Sasikala M, Anuradha S, Ramji C, Sashidhar R B, Rao G V, Pradeep R, Reddy D N. MUC1, CK20, and CDX2 immunohistochemical markers can sub-classify periampullary carcinomas into pancreaticobiliary, intestinal, and mixed subtypes.Indian J Pathol Microbiol 2021;64:693-701

How to cite this URL:
Manohar K M, Sasikala M, Anuradha S, Ramji C, Sashidhar R B, Rao G V, Pradeep R, Reddy D N. MUC1, CK20, and CDX2 immunohistochemical markers can sub-classify periampullary carcinomas into pancreaticobiliary, intestinal, and mixed subtypes. Indian J Pathol Microbiol [serial online] 2021 [cited 2023 Jan 31 ];64:693-701
Available from:

Full Text


Periampullary carcinomas (PACs) arise at the margins of the Ampulla of Vater and account for smaller percentage of gastrointestinal malignancies with an estimate of 0.2–0.5% around the world every year.[1],[2],[3] It has curative treatment of oncological resection of tumor together with lymph node dissection is the only gold standard treatment available.[4] Anatomically, these adenocarcinomas could either arise from head of pancreas or distal bile duct or duodenum or ampulla itself. It is often difficult to determine the precise anatomic location or origin of the tumor. Because, identifying the center of the tumor grossly or microscopically in an area where the anatomy can vary and is frequently distorted by the tumor.[5],[6] Even in advanced tumor it is difficult to distinguish tumor origin between ampulla, distal common bile duct, or pancreas. In such situation the diagnosis can be rendered as “carcinoma of the pancreato-biliary-ampullary region.[7] Hence histopathological classification was often preferred over anatomical classification in subtyping of PACs.

Periampullary carcinomas are categorized into three distinct histopathological subtypes on the basis of morphological characteristics.[8] The pancreatobiliary subtype arises from simple mucinous epithelium of the distal common bile duct and terminal pancreatic duct, with simple or branching glands and small solid cell-nests enclosed by desmoplastic stroma. The tumor cells are cuboidal to low columnar arranged in a single layer without nuclear pseudo stratification. The nuclei are rounded with marked variation in size and shape from one cell to another.[9],[10] While intestinal subtype arises in duodenum or intestinal mucosa which extends into ampulla of vater. These tumors are basically composed of well-formed tubular glands, complex cribriform areas, and solid nests. The tumor cells in this subtype are pseudostratified tall columnar with basally located cigar-shaped nuclei and often contain mucin.[11] The mixed subtype is described as a tumor of hybrid differentiation like intestinal architecture with pancreatobiliary cytology.[12] These subtypes also differ in survival duration and in prognosis. Zimmerman et al., in their study reported that, mean survival time for the pancreatobiliary subtype was 52 months, for the intestinal subtype 115 months, and for the mixed subtype 94 months.[13] The 5-year survival rates for pancreatobiliary subtype is 27.5% and 61% for the intestinal subtype for resected patients.[13] This tumor sub-classification brings clinical significance not just in terms of predicting survival and prognosis but also determining therapeutic strategy.[14]

Though morphological heterogeneity characterizes a significant proportion of PACs, at the same time it suffers with some major limitations. These limitations include interobserver variability, low prognostic reliability, and occasionally prone for misdiagnosis. Presence of various other subtypes and unusual types of PAC such as, signet ring type, mucinous type, poorly cohesive type makes it prone for misdiagnosis. This, demands to find precise markers that can sub classify the tumors more accurately.

Towards this, several investigators have attempted to study the role of immunohistochemistry (IHC) in subtyping of these tumors, with cytokeratins (CK7, CK17, CK19 and CK20), apomucins (MUC1, MUC2, MUC4, MUC5AC, and MUC6), and caudal-type homeodomain transcription factor CDX2.[15],[16],[17],[18],[19],[20] However, only few studies have systematically evaluated and added benefit of IHC to histological assessment in classifying periampullary adenocarcinomas. Besides that, there is no complete consensus from reports came from various parts of the world on the type of the markers that can be used for subtyping of periampullary tumors.

The present study aimed to confirm the usefulness of IHC as single or a panel of immunomarkers that included different cytokeratins (CKs), mucins proteins, and transcription factor CDX2 as an aid to the morphologic subtyping of periampullary adenocarcinoma.


To select the precise tumor specimens for study, we followed appropriate definition for periampullary adenocarcinoma. In general, tumors of the ampulla of vater may arise in the ampulla (intra-ampullary) or on the duodenal surface of the papillae (peri-ampullary) or involve both (mixed type). As mentioned above, origin of the tumor may be difficult and occasionally impossible to determine. The differential diagnosis includes carcinoma of the distal common bile duct, main pancreatic duct, and duodenum tumors may be exophytic or ulcerated. By definition, periampullary tumors are those tumors which arise within 2 cm surrounding of the ampulla of vater region in major papilla of the duodenum. Periampullary adenocarcinoma were differentiated from terminal pancreatic duct tumors or distal bile duct tumors with the help of radiological findings, gross findings of Whipples pancreaticoduodenectomy specimen or by microscopy which shows transition from benign ampullary mucosa to malignancy or through IHC.

The study was approved by Institutional Ethics Committee of Asian Institute of Gastroenterology. The study group involved the patients diagnosed for PAC (n = 94) and who underwent surgical resection for Whipples Pancreaticoduodenectomy with curative R0 resection during the years from 2012 to 2018. Informed consent was obtained from all individual participants included in the study. Patients with preoperative death (in-hospital death or death within 30 days postoperatively) were not included in both association analysis and survival analysis. Exclusion criteria included benign tumors or non-invasive (in situ) carcinomas, unusual types of ampullary adenocarcinomas (such as signet ring cell carcinoma, mucinous carcinoma, poorly cohesive carcinoma) and patients who lost for follow up were excluded from the study. The study design was shown as schematic representation in [Figure 1] depicts the inclusion and exclusion criteria along with sample collection details.{Figure 1}

Data on clinical parameters, including sex, age, and preoperative assessment of disease stage, Ca19-9 serum levels, and preoperative jaundice were collected from patient records. Pathological characteristics regarding tumor origin, size, grade, resection status, and TNM classification were retrieved from pathologists' original reports. Follow-up period ranged between 5 and 58 months (mean = 21 months).

Histopathological assessment

The histopathological characteristics such as site of origin of tumor, pathological tumor staging, tumor size, nodal involvement, perineural invasion, vascular involvement, and degree of differentiation were registered for all the patients and re-examined by experienced pathologists. The histological type of differentiation was based on the Albores-Saavedra et al.[8] and Kimura et al.[9] Based on the morphological features, tumor specimens were either assigned as pancreatobiliary or Intestinal histological types. Few histologically ambiguous cases were designated as Indeterminate group.


Paraffin-embedded tissue specimens were retrieved, and blocks were serially sectioned at 4-μm thickness and were fixed on to slide by 10% formalin fixation and incubation at 4°C overnight. The deparaffinization was performed in xylene and alcohol which involves a series of alcohol washes of varying concentrations and final wash with deionized water. Antigen retrieval was done by research grade microwave treatment using citrate buffer at pH 6.0 for MUC1, MUC2, MUC5Ac, Ki-67, and CA19-9; Tris-EDTA buffer at pH 8.0 for CK7, CK20, and CDX2for 30 minutes. Then the slides were subjected to immunostaining with the specific antibodies to form the antigen-antibody complex for the target antigen. Primary antibody was added for overnight at 4°C, followed by secondary antibody for 30 min at room temperature. Staining was visualized by using DAB Detection Kit (K3468, Dako) and sections were counterstained with Mayer's hematoxylin. Sections were photomicrographed with a digital camera mounted on light microscope (Olympus CX31, Japan), digitized, and analyzed. Analysis was performed on of a section at 4×, 10×, and 40× magnification. Positive staining was considered when at least 5% or more tumor cells were stained. Positive and negative control tissues were used to check the efficacy of the antibody reactivity. Details of the antibodies are mentioned in [Table 1].{Table 1}

Statistical analysis

Associations between each immunohistochemical marker and the histological type were evaluated using Fisher's exact test and k statistics (to assess the overall agreement and individual category agreement for histologic typing). A k value of 0.7 was considered to indicate excellent agreement, a value between 0.4 and 0.7 was considered to indicate good agreement, and a value of 0.4 and below was considered to indicate poor agreement. The correlation of immunohistochemically stained intestinal (IHC-IN) or immunohistochemically stained intestinal pancreatobiliary (IHC-PB) with the “final diagnosis” achieved on H and E review was performed by using the Fisher exact test. A 2-sided P value of <0.05 was considered statistically significant.

Pearson's correlation was performed for comparing survival period and expression of IHC markers. Survival data was obtained by the follow up of the patient's up to 5 years of surgery. Only overall survival was considered. By the end of the study, 24 of 94 patients were dead. The Kaplan–Meier method, the log rank test and Cox regression analysis were used to assess overall survival. For all tests, P < 0.05 was considered to be statistically significant. The analysis was carried out by Med Calc software, version 14.8.1 (2014).


Patient characteristics

A total of 94 patients (males 65, age 55.4 ± 14.32; females 29 (age 54.9 ± 18.35) satisfied the inclusion and exclusion criteria and were followed for survival status. Among them 24 patients died within 5 years after surgery during the follow-up period and 74.4% survived to the end of the follow-up period. 31 patients received adjuvant chemotherapy.

Histological classification

Based on histological classification as per Albores-Saavedra et al.,[8] tumors with simple or branching glands and small solid nests of cells surrounded by dense desmoplastic stroma and tumor cells are cuboidal to low columnar arranged in a single layer without nuclear pseudo stratification and the nuclei are rounded with marked variation in size and shape from one cell to another. These tumors are labeled as pancreatobiliary subtype tumors. We found 46 out of 94 periampullary tumors (48.9%) to be H and E PB-type carcinomas [Table 2]. Similarly, intestinal tumors are those tumors consisting of well-formed tubular glands, complex cribriform areas, and solid nests. The tumor cells in this subtype are pseudostratified tall columnar with basally located cigar-shaped nuclei and often contain mucin. We could identify 35 out of 94 (37.3%) tumors to be H and E intestinal-type carcinomas [Table 2]. We found 13 specimens (13.9%) to be indeterminate where histological classification couldn't resolve either the specimens to be PB type or intestinal type. Each subtype Haematoxylin and Eosin-stained representative sections and their morphology are shown in [Figure 2]. The clinicopathological variables were found to be nonsignificant between morphological types of periampullary carcinoma are shown in [Table 3].{Table 2}{Figure 2}{Table 3}

Immunohistochemical expression

A panel of 6 markers CK7, CK20, MUC1, MUC2, MUC5Ac, and CDX2 were used to immunohistochemically classify subtypes of periampullary carcinoma of all locations. Immunohistochemical analysis for all the 94 tumor specimens revealed that, markers CK7 and MUC1 showed positive reactivity in a greater number of specimens that is, CK7 was positive in 68 tumors while MUC1 was positive in 64 specimens out of total 94 tumor specimens. Among these CK7 expression was seen across both the histological-based subtypes and only MUC1 was predominantly positive in PB type. Only 5 out of 56 H and E PB type subtypes showed nonreactivity. Markers that remained largely nonreactive in the study are MUC2 (77 specimens showed negative) and MUC5Ac (60 specimens were nonreactive). CK20 and CDX2 are the only markers showed higher percentage (>60% of specimens) positive reactivity in histological-based intestinal subtype tumors [Supplementary Table 1].[INLINE:1]

To determine individual markers sensitivity and specificity towards substantiating the H and E based classification, we performed receiver operating curve (ROC) analysis. In the analysis we found that none of these markers could exclusively differentiate the subtypes with a high sensitivity and specificity. Only MUC1 showed high sensitivity (91.07%) and moderate specificity (65.79%) for PB type tumors; CK20 and CDX2 showed moderate sensitivity (63.16%, 68.42%) and moderate specificity (69.64%, 71.43%) for intestinal subtype tumors [Table 4].{Table 4}


For immunohistochemical analysis, those specimens with antibody staining with at least more than 5% tumor cells being positive were only considered as positive reactivity and below the 5% were considered as false positives (negative staining). Based on the reactivity of the antibodies the tumors are classified as 0 (no reactivity), 1 (reactivity in 5–10% of tumor cells), 2 (reactivity in >10% but <40% of tumor cells), and 3 (reactivity in >40% of tumor cells). Among the antibodies, cytokeratin 7 showed moderate to strong reactivity in majority of the specimens (59) followed by MUC1 (51). Among the antibodies, MUC2 showed maximum number of nonreactivity (70 specimens). The percentage of tumor cells positivity was shown in [Supplementary Table 2] and [Supplementary Figure 1].[INLINE:2][INLINE:3]

IHC markers combinatorial analysis

To derive more sensitivity and specificity of the markers, we performed the combinatorial analysis of the markers for subtyping of the tumors. For this purpose, we selected MUC1, CK20, and CDX2, as they yielded moderate to high sensitivity and specificity on individual basis. Accordingly, an IHC schema was devised, using this we classified the PACs into 3 subtypes without the bias of histological classification. The IHC schema we used in the study was shown in [Table 5].{Table 5}

Upon comparison of histologically verified H and E PB (n = 46) with IHC-based PB subtype classification (n = 51), we found 41 (80.4%) of the tumors were definite pancreatobiliary subtype (both histologically and immunohistochemically confirmed). Remaining 10 IHC-PB tumors specimens (19.6%) out of 51 IHC-PB, were histologically classified as intestinal type. Similarly, H and E-Intestinal (n = 35) correlated with IHC-Intestinal type (n = 30) classification revealed, 25 IHC-Int specimens (71.4%) truly matched with H and E-Intestinal type, while the remaining 5 IHC-Int specimens were labeled under H and E PB subtype. In H and E Indeterminate specimens (n = 13), all the specimens are exclusively confirmed to be IHC Mixed Type tumors.

Interobserver agreement and correlation of histologic classification and IHC

Interobserver agreement between two pathologists evaluating the histological type by morphological criteria was almost perfect [Κ = 0.90; 95% confidence interval (CI) 0.82, 0.99]. Comparing immunohistochemical and morphological classification of the histological type (pancreatobiliary versus intestinal), was in good agreement with a Κ value of 0.557 (std error, 0.07 and 95% CI 0.4–0.704). The concordance rate between H and E and IHC evaluation was 80.4% for H and E-PB tumors, 71.1% for H and E-Int tumors and 100% for H and E indeterminate/IHC Mixed. The correlation was statistically significant (P < 0.0001).

Accuracies of proposed IHC schema for predicting subtypes

Correspondingly for the above-mentioned patterns the sensitivity and specificity were analyzed with respect to tumor subtypes. Upon analysis, it was identified that, with in the IHC-PB schema which shown in [Table 5] and [Table 6], the IHC-PB1 marker pattern of MUC1+, CK20-Ve, and CDX2-ve was having high sensitivity (80.39%) and specificity (100%). Similarly, for IHC-intestinal group, the IHC-Int1 marker pattern of CK20+, CDX2+ and MUC1- ve was found to have moderate sensitivity (56.6%) and high specificity (100%). For H and E, IHC-mixed type, the combined markers have shown 100% sensitivity and specificity (100%). The proposed IHC schema accuracies were shown in [Table 6]. Illustration of the single representative image of each subtype along with IHC markers expression are shown in [Figure 3]a, [Figure 3]b, [Figure 3]c.{Table 6}{Figure 3}

Survival analysis

The overall survival rate was found to be 44.95 ± 2.29 months (mean ± SD) on the whole for all PAC patients. The survival duration of each subtype; pancreaticobiliary-27.9 ± 1.63, Intestinal- 52.92 ± 2.18 and mixed- 35.5 ± 0.45 months shown in [Figure 4]. Interestingly, although there was a clear trend toward prolonged survival in patients of histologic differentiation of intestinal-type tumors as compared to those with pancreatobiliary type tumors (median = 58 vs 49 months), this difference could reach only marginal statistical significance (P = 0.05). Whereas survival duration in immunohistochemical based subtyping (median = 53 vs 28 vs 36 months) intestinal vs PB vs mixed has shown better statistical significance (P < 0.0001).{Figure 4}

The patients were categorized for statistical purposes: patients with poor and highly differentiated tumors, patients with low- and high-grade tumors and patients with different stages of disease (stage I, II, and III), presence of or absence of perineural invasion and lymphovascular invasion etc., for predicting prognosis. On univariate analysis overall survival in our cohort of patients was adversely influenced by advanced overall (AJCC TNM stage III) stage (P = 0.05) and perineural invasion (P = 0.0012) as shown in [Table 7]. None of the immunohistochemical markers individually showed any correlation with prognosis of the patients.{Table 7}


The two major findings of our study are, first, we propose that combined immuno-positivity of CK20, MUC1, and CDX2 can detect the mixed subtype of periampullary cancers. Second, we confirm the significant relationship between overall survival and clinicopathological prognostic markers, such as TNM staging and perineural invasion.

It is widely known fact that, periampullary adenocarcinoma is not a single entity of tumor but a group of tumors that have its origins and its outcomes differing among each patient. In majority cases, tumor origin cannot be delineated to a precise origin (either pancreas or bile duct or ampulla or duodenum). This is due to the following reasons: either tumors show distorted pattern or tumors invading to surrounding epithelia or tumors growing in size or tumors arising at multiple origins simultaneously. Therefore, for these reasons' determination of the histologic type of differentiation is more useful than anatomical classification. Because, failure to reach a precise diagnosis of tumor origin may lead to false assumptions regarding long-term survival.[8],[11]

Thus, we first attempted to distinguish the tumors based on the histology alone. On the basis of morphology and employing haematoxylin and eosin staining (H and E) we classified 94 tumor specimens into H and E pancreaticobiliary (n = 46) and H and E intestinal subtype (n = 35) and rest 13 tumor specimens were categorized into indeterminate group. The third group was having an admixture morphology of both subtypes and hence labeled as indeterminate. Despite histological classification being a reliable technique for most of the tumor classifications, however, there is still some uncertainty with respect to subtyping in various tumors. Ang et al. in periampullary tumors reported such difficulty in PAC tumors with heterogeneous patterns on histology (i.e., the mixed subtype).[12] We too observed the similar difficulty where third group of 13 specimens requires employing more precise classification criteria in addition to H and E. In this regard various reports have been published underlining the importance of immunohistochemical staining (IHC) using antibodies for cytokeratins, mucin proteins and also CDX2.[15],[16],[17],[18],[19],[20] Herein, we attempted to study whether this subtyping can be achieved on histological grounds alone and whether IHC has any added value.

Many IHC markers have been used to subtype periampullary carcinomas. Several workers have attempted to study the role of immunohistochemistry (IHC) in the subtyping of these tumors, with cytokeratins (CK7, CK17, and CK20), apomucins (MUC1, MUC2, MUC4, and MUC5Ac), and caudal-type homeodomain transcription factor (CDX2). Of these, CK17, MUC4, and MUC1 have been found to be expressed in pancreatobiliary type of periampullary carcinomas.[15],[16],[17],[18] More pertinently, CK17 was reported to be a useful marker in separating pancreatobiliary adenocarcinomas from extra-pancreatobiliary non-mucinous adenocarcinomas, including adenocarcinomas from the colon, breast, gynecologic organs, stomach, lung, prostate, thyroid, kidney, and adrenal gland, and malignant mesothelioma.[15] However, CK17 was not included in present study. CK20, CDX2, and MUC2 positivity and/or MUC1 negativity in ampullary tumors was strongly associated with intestinal morphology.[19],[20],[21]

To our knowledge from the previous studies, the performance of individual markers has not yielded higher sensitivity and specificity to make them utilize as definite markers for subtyping in periampullary tumors. Even we found that individual markers sensitivity and specificity was only optimal to propose it as a reliable marker for subtyping. To address this problem, a few studies have systematically added a panel of immunomarkers and evaluated its performance in combination rather than independent markers. Ang et al. attempted combination of CK7, CK20, CDX2, MUC2, and MUC1, in their study and proposed a combination criterion for sub classification.[12] In this study, we attempted on the similar lines by applying a panel of 6 markers, comprising CK7, CK20, MUC1, MUC2, MUC5Ac, and CDX2 in an Indian context to sub classify the periampullary tumors.

Employing the immunohistochemical markers, we observed that individual markers were having moderate to high sensitivity and specificity. In the analysis of individual markers, CK7 was positive in 51 out of 56 H and E PB tumors, but with an over-all low specificity (42.11%). Though higher percentage (91%) of positivity was seen in pancreatobiliary tumors, CK7 also showed variable positivity in intestinal type. Even with combination analysis of CK7 with other markers, there was no significant improvement in % AUC. Hence, in the study CK7 was ignored for its utility as precise single marker for pancreaticobiliary subtypes. But CK7 in combination with other previously reported pancreatobiliary markers such as CK17 and CK19 may show an improved outcome.[15] Other individual markers of the study such as, MUC2 and MUC5ac were not prominent either in our ROC analysis. MUC2 showed non-reactivity in 77 tumor specimens and MUC5Ac in 60 tumor specimens. In contrast to our result on MUC2 expression, Chu et al., have earlier reported that, MUC2+/CDX2+ can be used as positive markers for the intestinal-type adenocarcinoma of duodenal papillary origin with a positive predictive value of 82%.[15] In our study, we included the exclusion criteria of omitting the unusual type of PAC tumor specimens such as mucinous carcinoma. Most mucinous carcinoma usually exhibit intestinal type. However, we could see that only 17 cases having MUC2 positivity. Perhaps inclusion of specific mucinous carcinoma specimens in the study could have a given better picture on MUC2 positivity in tumors.

Because MUC1, CK20, and CDX2 showed moderate to higher sensitivity and specificity in individual ROCs, and hence, are analyzed in combination. With combinatorial analysis, we were able to classify 51 tumors into IHC-PB type with MUC1+Ve, CK20-Ve or CDX2-Ve combination and 30 tumors as IHC-intestinal type with CDX2+Ve or CK20+Ve and MUC1-Ve. This classification for PB type and intestinal type corroborates with previous reports.[12],[19],[20] Interobserver analysis showed good agreement between histologic and IHC type with a kappa value of 0.554. This suggests that it is feasible to sub-classify periampullary tumors into pancreatobiliary and intestinal by simple IHC technique.

Nevertheless, the IHC criteria for mixed type have not been conclusive previously as well. Our study is first to depict, combination criteria of CK20+ve, MUC1+ve, and CDX2+ve specimens completely matched with 13 tumor specimens that were labeled under indeterminate/mixed morphology group using H and E. All these specimens further upon histological reverification revealed to be mixed subtype of periampullary tumors. Histologic review of these tumors revealed combined morphologic patterns of intestinal architecture with pancreatobiliary cytology.

There are several limitations in this study. First, only 13 tumor specimens represent the true “mixed” subtype which could be of low sample size. Second, we confined the study to a limited set of markers. However, we believe that these limitations did not affect the outcomes of this study significantly because of the striking differences between sub-groups. These limitations warrant further studies such as validating IHC subtyping criteria proposed for mixed type and to screen large number of markers simultaneously by employing tissue microarray, respectively.

Of late we started to look at other markers such as CK17, CK19, apomucin, MUC6, and MUC8 (data were not available for entire study duration). Our preliminary observations indicate that CK19 was more often seen in signet ring carcinoma and in PB type; MUC2 prominence was observed more in mucinous carcinomas and MUC6 positivity in poorly cohesive carcinomas of PACs. The subsequent studies will unravel its significance towards precise criteria for sub-classification.

We also attempted survival analysis in the present study. Our observations corroborate with the previous findings that pancreatobiliary subtype has poor outcome of prognosis than the intestinal and mixed subtype of PAC.[21] The markers that we used in this study also differentiate the patients with poor prognosis.

Notably, the study has been conducted in a tertiary referral hospital like ours and will have the availability of experienced and highly qualified pathologists, and hence it may have contributed to the high concordance rate between the observers and between histologic and IHC diagnoses. This may not be the case with impoverished hospitals where there are chances of high diagnostic misinterpretation. In such a scenario, an IHC criteria that we proposed if validated may augur well in giving results with a high accuracy, which in turn benefits the patients in their outcome of long-term survival.


Our results suggest that the combined immunophenotype expression of MUC1, CK20, and CDX2 could be used as markers to identify mixed subtype tumors where histological verification remains inconclusive.

Declaration of patient consent

The authors certify that they have obtained all appropriate patient consent forms. In the form the patient(s) has/have given his/her/their consent for his/her/their images and other clinical information to be reported in the journal. The patients understand that their names and initials will not be published and due efforts will be made to conceal their identity, but anonymity cannot be guaranteed.


We acknowledge CSIR HRDG for awarding Junior Research Fellowship 09/1147 (0002/2019 EMR-I) to K Murali Manohar.

Financial support and sponsorship

Funding from Asian Health Care Foundation and Council of Scientific and Industrial Research, HRDG acknwoledged.

Conflicts of interest

There are no conflicts of interest.


1Carter JT, Grenert JP, Rubenstein L, Stewart L, Way LW. Tumours of the ampulla of vater: Histopathologic classification and predictors of survival. J Am Coll Surg 2008;207:210-8.
2Westgaard A, Pomianowska E, Clausen OP, Gladhaug IP. Intestinal-type and pancreatobiliary-type adenocarcinomas: How does ampullary carcinoma differ from other periampullary malignancies? Ann Surg Oncol 2013;20:430-9.
3Howe JR, Klimstra DS, Moccia RD, Conlon KC, Brennan MF. Factors predictive of survival in ampullary carcinoma. Ann Surg 1998;228:87-94.
4Donahue TR, Reber HA. Surgical management of pancreatic cancer--pancreaticoduodenectomy. Semin Oncol 2015;42:98-109.
5Frierson HF. The gross anatomy and histology of the gallbladder, extrahepatic bile ducts, Vaterian system, and minor papilla. Am J Surg Pathol 1989;13:146-62.
6Pomianowska E, Grzyb K, Westgaard A, Clausen OP, Gladhaug IP. Reclassification of tumour origin in resected periampullary adenocarcinomas reveals underestimation of distal bile duct cancer. Eur J Surg Oncol 2012;38:1043-50.
7Jagtap SV, Jokhi CD, Patil D, Jagtap SS. Adenocarcinoma of ampulla of vater presented as recurrent jaundice. Ann Pathol Lab Med 2018;5:C82-5.
8Albores-Saavedra J, Hensen DE, Klimstra DS. Tumours of the gallbladder, extrahepatic bile ducts, and ampulla fo Vater. In: Rosai J, Sobin L, editors. Atlas of Tumour Pathology, 3rd Series, Fascicle 27. Washington, DC: Armed Forces Institute of Pathology; 2000.
9Kimura W, Futakawa N, Zhao B. Neoplastic diseases of the papilla of Vater. J Hepatobiliary Pancreat Surg 2004;11:223–31.
10Zhou H, Schaefer N, Wolff M, Fischer HP. Carcinoma of the ampulla of Vater: Comparative histologic/immunohistochemical classification and follow-up. Am J Surg Pathol 2004;28:875-82.
11Westgaard A, Tafjord S, Farstad IN, Cvancarova M, Eide TJ, Mathisen O, et al. Pancreatobiliary versus intestinal histologic type of differentiation is an independent prognostic factor in resected periampullary adenocarcinoma. BMC Cancer 2008;8:170.
12Ang DC, Shia J, Tang LH, Katabi N, Klimstra DS. The utility of immunohistochemistry in subtyping adenocarcinoma of the ampulla of vater. Am J Surg Pathol 2014;38:1371-9.
13Zimmermann C, Wolk S, Aust DE, Meier F, Saeger HD, Ehehalt F, et al. The pathohistological subtype strongly predicts survival in patients with ampullary carcinoma. Sci Rep 2019;9:12676.
14Kim WS, Choi DW, Choi SH, Heo JS, You DD, Lee HG. Clinical significance of pathologic subtype in curatively resected ampulla of vater cancer. J Surg Oncol 2012;105:266-72.
15Chu PG, Schwarz RE, Lau SK, Yen Y, Weiss LM. Immunohistochemical staining in the diagnosis of pancreatobiliary and ampulla of Vater adenocarcinoma: Application of CDX2, CK17, MUC1, and MUC2. Am J Surg Pathol 2005;29:359-67.
16Lau SK, Weiss LM, Chu PG. Differential expression of MUC1, MUC2, and MUC5AC in carcinomas of various sites: An immunohistochemical study. Am J Clin Pathol 2004;122:61-9.
17Chang DK, Jamieson NB, Johns AL, Scarlett CJ, Pajic M, Chou A, et al. Histomolecular phenotypes and outcome in adenocarcinoma of the ampulla of vater. J Clin Oncol 2013;31:1348-56.
18Schueneman A, Goggins M, Ensor J, Saka B, Neishaboori N, Lee S, et al. Validation of histomolecular classification utilizing histological subtype, MUC1, and CDX2 for prognostication of resected ampullary adenocarcinoma. Br J Cancer 2015;113:64-8.
19Kumari N, Prabha K, Singh RK, Baitha DK, Krishnani N. Intestinal and pancreatobiliary differentiation in periampullary carcinoma: The role of immunohistochemistry. Hum Pathol 2013;44:2213-9.
20Perysinakis I, Minaidou E, Leontara V, Mantas D, Sotiropoulos GC, Tsipras H, et al. Differential Expression of β-Catenin, EGFR, CK7, CK20, MUC1, MUC2, and CDX2 in intestinal and pancreatobiliary-type ampullary carcinomas. Int J Surg Pathol 2017;25:31-40.
21Bronsert P, Kohler I, Werner M, Makowiec F, Kuesters S, Hoeppner J, et al. Intestinal-type of differentiation predicts favourable overall survival: Confirmatory clinicopathological analysis of 198 periampullary adenocarcinomas of pancreatic, biliary, ampullary and duodenal origin. BMC Cancer 2013;13:428.