Endoscopic ultrasound-guided fne-needle aspiration cytology in pancreaticobiliary carcinomas: diagnostic effcacy of cell-block immunocytochemistry

Messina, Italy

Endoscopic ultrasound-guided fne-needle aspiration cytology in pancreaticobiliary carcinomas: diagnostic effcacy of cell-block immunocytochemistry

Antonio Ieni, Paolo Todaro, Stefano Francesco Crinò, Valeria Barresi and Giovanni Tuccari

Messina, Italy

BACKGROUND: Endoscopic ultrasound-guided fne-needle aspiration cytology was demonstrated to be a useful tool for the diagnosis and staging of pancreaticobiliary neoplastic lesions. Nonetheless, the diagnostic value of this procedure may be limited by low cellularity of the specimen, contamination of intestinal cells and unfeasibility of ancillary immunocytochemical procedures. The present study was to evaluate its usefulness in the diagnosis of neoplastic lesions.

METHODS: A series of 46 pancreaticobiliary carcinomas with available cell block preparations was submitted to immunocytochemistry against cytokeratins, carcinoembryonic antigen, E-cadherin, CD10 and p53. The sensitivity, specifcity, positive and negative predictive values of the cytological smear in the discrimination of malignant lesions were calculated and compared with those of cell block preparation with the immunocytochemical stains against p53 and CD10.

RESULTS: According to our fndings, the use of cell block preparations together with immunostains against p53 and CD10 allowed to discriminate malignant versus benign specimens with higher sensitivity than the only cytological examination. In detail, CD10 immunostaining was of signifcant help for the discrimination between cytological contaminants, such as benign gastrointestinal cells, and the neoplastic elements of pancreaticobiliary well differentiated adenocarcinomas. Also, intense nuclear immunoreactivity for p53 was encounteredin about 2/3 of the cases and identifed pancreatic malignancy with high sensitivity. CONCLUSIONS: We suggest that immunocytochemistry against both CD10 and p53 could be applied case by case, mainly to differentiate gastrointestinal and pancreatic benign cellular contaminants showing hyperplasia or reactive changes from differentiated pancreaticobiliary adenocarcinomas.

(Hepatobiliary Pancreat Dis Int 2015;14:305-312)

endoscopic ultrasound;fne-needle aspiration cytology; pancreas; cell-block; contaminant cells; immunohistochemistry

Introduction

Endoscopic ultrasound-guided fne-needle aspiration cytology (EUS-FNAC) is a useful tool for the diagnosis of several benign and malignant diseases as well as for the staging of malignant tumors of the gastrointestinal tract and adjacent organs and glands.[1-4]Nevertheless, the diagnostic yield of EUS-FNAC partly depends on the site, size and characteristics of the target tissues as well as on several technical and procedure-related factors and mainly on the expertise, training and interaction between the endosonographer and cytopathologist.[4]

It was shown that cell-block preparation, complementary to cytological smears, may improve the value of EUS-FNAC in the diagnosis of pancreaticobiliary neoplastic lesions; indeed it reduces the false negatives resulting from the interpretation of well-differentiated ductal adenocarcinomas as gastrointestinal contaminant cells.[5-7]Besides, cell block allows the assessment of immunocytochemistry, which may be of use for the differential diagnosis of pancreaticobiliary malignan-cies.[8,9]Although carcinoembryonic antigen (CEA) was proposed as a useful diagnostic marker, the evidence that also duodenal cells may occasionally show CEA immunostaining reduces its specifcity.[8,9]Hence there is the need to identify additional reliable immunocytochemical markers for the differential diagnosis among intestinal contaminant cells, benign neoplasms and well-differentiated pancreaticobiliary carcinomas.

On this premise, the aim of the present study was to evaluate retrospectively the potential value of some immunocytochemical markers for the differential diagnosis between neoplastic cells and normal gastrointestinal cells contaminants in a series of cytological pancreatic samples present in our archive. The availability of an adequate number of serial sections obtained from the cell-blocks was of signifcant aid in order to assess an appropriate diagnostic algorithm.

Methods

The study protocol for collection and use of human samples was approved by the Institutional Ethics Committee of our hospital. A series of 46 pancreaticobiliary neoplastic lesions, obtained from an equal number of patients (21 males, 25 females; age range 40-84 years; mean age 63.8 years), were included in the study. The site of the neoplasms as well as the cytological diagnosis obtained by cytological smear are shown in Table. For each case, the electronic database of the University of Messina was reviewed to verify the corresponding histological diagnosis, which was available in 17 cases. On the other hand, we failed to fnd a histological diagnosis in 29 cases. In detail, histological specimens were unavailable in 18 cases since the patients were not candidates for surgery due to the advanced stage of their disease, while the remaining 12 patients had been surgically treated in a reference center for pancreatic neoplasms in Northern Italy. In these cases cytological diagnosis could be verifed by the comparison with the data retrieved in our Regional Cancer Registry.

EUS-FNAC on pancreaticobiliary neoplastic lesions was performed by using a convex array echoendoscope (EG 3870 UTK, Pentax, Tokyo, Japan) and by making two passes with a 22G needle. The specimens were processed by a cytopathologist in-room and immediately examined for adequate cellularity after staining by hematoxylin and eosin. A second slide was immediately fxed in 98% ethanol and stained with Papanicolaou. In the aim of recovering further material, the needles and syringes used in the procedure were rinsed with 10 mL of 50% ethanol into a special container. All content was centrifuged in a 10 mL disposable centrifuge tube at 4000 rpm for 6 minutes to create 1 or 2 pellets; the supernatant fuid was decanted and the pelleted material was immediately fxed in a freshly prepared solution of 4% neutral buffered formalin for 45 minutes. Then, the cell pellets were placed in a cassette and stored at 80% ethanol until ready for processing in an automatic tissue processor (Leica TP 1020 Leica Microsystems, Wetzlar, Germany). The obtained cell blocks were embedded in paraffn at 56 ℃, cut into 3 µm thick sections and routinely stained with hematoxylin and eosin; parallel serial sections of the same thickness were mounted on poly-lysine-coated glasses and submitted to immunocytochemical procedures, as previously described.[4,10]In each case, a set of the following commercially obtained monoclonal antisera was applied： broad spectrum cytokeratins AE1/AE3 (DakoCytomation, Copenhaghen, Denmark; working dilution 1：300), cytokeratin 7 (DakoCytomation; working dilution 1：300), cytokeratin 19 (DakoCytomation; working dilution 1：300), cytokeratin 20 (DakoCytomation; working dilution 1：300), CEA (DakoCytomation; working dilution 1：500), E-cadherin (Zymed, San Francisco, CA, USA; working dilution 1：50), CD10 (Ventana, Tucson, AZ, USA; prediluted), and p53 (Abcam, Cambridge, UK; working dilution 1：250). All of the immunocytochemical reactions were carried out by using a standardized procedure with Ventana Benchmark automatic immunostainer. When required, antigen retrieval procedure was performed by using three cycles of 10 minutes each in microwave oven at 750 W, with citrate buffer (10mmol/L citric acid, 0.05% Tween 20, pH 6.0). Immunocytochemical staining for each antibody was recorded as negative or positive (any reactivity, cytoplasmic, apical or membranous) and as weak, moderate or strong according to the intensity.

The sensitivity (percentage of correctly classifed malignant lesions of the total malignant lesions), specifcity (percentage of correctly classifed benign lesions of the total benign lesions), positive (PPV) (percentage of correctly classifed malignant lesions of the total lesions classifed as malignant) and negative (NPV) (percentage of correctly classifed benign lesions of the total lesions classifed as benign) predictive values of the diagnosis obtained by the cytological smear or by the use of cell block procedure and ancillary immunocytochemistry against p53 and CD10 were calculated and compared.

For this analysis, we considered as “malignant“ cases positive and suspicious for malignancy and as “benign“those negative and with atypical cytology or positive for neoplastic cells.

Results

The data on the immunocytochemical fndings relative

to p53 and CD10 stains, the cytological diagnosis obtained by using cell blocks and immunocytochemistry and the corresponding histological diagnoses are shown in Table.

Table.Clinical and cytological characteristics of studied cases

The smears of all cases exhibited a hemorrhagic background with clusters or small aggregates of epithelial cells, occasionally arranged in glandular or pseudo-papillary structures. Those cells had roundish, hyperchromatic nuclei, with occasional nucleoli and dense chromatin; in addition, numerous gastric and duodenal epithelial cells were found, which made the cytological diagnosis more complex.

On the basis of the cytological smear, we classifed 24 cases as “malignant“ and 22 as “benign“. Diagnosis of malignancy was confrmed histologically in all but one case (Table). On the other hand only 9 cases had been correctly classifed as benign cytologically, while the remaining 13 had a fnal diagnosis of malignancy (Table). The false negative cases were all represented by indeterminate cases, and specifcally by 9 specimens classifed as “neoplastic cells present“ and by 2 “atypical cytology“. The sensitivity of cytological smear for the distinction of benign and malignant lesions was 63%, while the specifcity was 90%. The PPV and NPV were 95% and 41%, respectively.

A morphology equivalent to that of cytological smears was evidence in the cell block preparations (Fig. 1). With reference to the various immunocytochemical stains, neoplastic pancreaticobiliary elements as well as gastrointestinal benign cellular components showed a diffuse and intense cytoplasmic immuno-positivity for broad spectrum cytokeratins (Fig. 2A). Staining for other cytokeratins, namely 7, 19 and 20, was present in contaminant intestinal cells as well as in neoplastic epithelial cells (Fig. 2B). CEA strongly and homogeneously stained the apical cytoplasmic portion of pancreaticobiliary carcinomas (Fig. 3A), although a weak/moderate immunostaining was also occasionally evidenced at the apical border of benign duodenal cells (Fig. 3B); gastric cells were always unreactive.

Fig. 1.Well-differentiated pancreatic carcinoma with a pseudoglandular pattern (A, original magnifcation ×200); poorly differentiated solid anaplastic aggregates (B, original magnifcation ×400) (cell-block procedure, hematoxylin and eosin).

Normal gastric and duodenal epithelia, which were considered as internal positive controls, exhibited strong membranous E-cadherin expression, while decrease or partial loss of immunostaining was found in pancreatic neoplastic cells, mainly in less differentiated cases.

Fig. 2.An intense cytoplasmic immunoreactivity for broad spectrum cytokeratins (A, original magnification ×200); neoplastic extra-hepatic ductal elements exhibited an evident cytokeratin 7 immunostaining (B, original magnifcation ×200) (cell-block procedure, hematoxylin nuclear counterstain).

Fig. 3.CEA monoclonal antibody revealed a linear apical positivity either in papillary neoplastic elements (A, original magnifcation × 200) or in normal intestinal cells (B, original magnification × 400) (cell-block procedure, hematoxylin nuclear counterstain).

Similarly, immunocytochemical staining for CD10 was present at the apical membrane in benign gastrointestinal cells (Fig. 4A), while the great majority of neoplastic elements of pancreaticobiliary adenocarcinomas were consistently negative (Fig. 4B); surprisingly, an intense and uniform, membranous and cytoplasmic positivity was evidenced in a case of undifferentiated pancreatic carcinoma (Fig. 4C).

Nuclear strong immunoreactivity for p53 was present in 20 samples, and mainly in moderately or poorly differentiated pancreatic carcinomas (Fig. 4D) (Table). None of the cases classifed as benign at histology had shown p53 immunostaining at cytology. On the whole p53 immuno-reactivity was evidenced in 55% of malignant specimens according to histology.

By the use of p53 immunostaining, 11 cases classifed as benign (neoplastic cells present) at cytological smear were considered as malignant at cytological diagnosis with cell block preparation. All of these cases proved to be malignant at histology.

By cell-block preparations with p53/CD10 immunocytochemistry we classifed 35 cases as malignant and 11 as benign. All but one malignant cases and all but two benign proved to be correctly classifed in the comparison with histology.

The sensitivity of cell block preparation with ancillary immunocytochemistry against p53 and CD10 for the discrimination of malignant lesions was 94%, while specifcity was 90%. The PPV and NPV were 97% and 81%, respectively.

Fig. 4.The peculiar pattern of CD10 immunostaining showed a localization at the apical cytoplasm in benign gastrointestinal cells (A, original magnification ×160), while well-differentiated neoplastic elements of were always negative (B, original magnifcation ×160); by contrast, an intense, ring-like, membranous cytoplasmic reactivity was found in anaplastic ductal carcinomas (C, original magnification ×200). Finally, a nuclear strong p53 labeling was represented in neoplastic elements (D, original magnification × 200) (cell-block procedure, hematoxylin nuclear counterstain).

Discussion

EUS-FNAC is widely considered to be a useful approach for the diagnosis and staging of pancreaticobiliary malignancies in daily practice. Indeed, this method allows to accomplish the diagnosis of the great majority of cases on cytological samples and to discriminate benign from malignant lesions.[11,12]Nevertheless, in a small subset of carcinomas cytological diagnosis may be hard to achieve, due to the presence of extensive tumor necrosis, associated infammation, limited sampling and, mainly, contamination by intestinal epithelial cells.[13]In those cases, the cell-block technique can provide signifcant help to the pathologist, since it allows microscopical evaluation of histological mimickers and immunocytochemical stains in serial sections.[4,14]Although a crucial diagnostic point is to distinguish neoplastic cells from gastrointestinal contaminating cells with hyperplastic or reactive changes, unfortunately no specifc and reproducible immunocytochemical markers are available for this purpose at the moment.

In the present study we tested the diagnostic value of some immunocytochemical markers in the differential diagnosis of pancreaticobiliary carcinomas on cellular smears and corresponding cell blocks obtained at EUSFNAC. On the whole, staining for broad spectrum cytokeratins was homogeneously present in both neoplastic and normal cells, regardless of their origin. Similarly, the use of more specifc cytokeratins, such as cytokeratin 7, 17, 19 and 20, was useless for the distinction of pancreatic ductal adenocarcinomas from common bile duct and cystic duct epithelia.[15]Although CEA immunostaining was homogeneously present with strong intensity in the pancreaticobiliary carcinomas, immunoreactivity at the apical cell border was also occasionally found in benign duodenal cells. Hence our fndings confrm previous evidence that it cannot be considered as a highly specifc marker for the distinction of well-differentiated ductal pancreatic carcinoma from intestinal contaminants.[8,16]

Partial loss of E-cadherin expression in the neoplastic cells was previously shown to be an independent predictor or poor outcome in patients affected by pancreatic adenocarcinomas.[17-19]In the present analysis we found that E-cadherin staining in the samples of pancreatic carcinomas progressively decreased from well-differentiated cases to moderately and poorly ones, though statistical signifcance was not reached. Therefore, the progressive reduction of E-cadherin immunoexpression cannot be considered as a reproducible tool in order to distinguishbetween well and less differentiated pancreatic carcinomas; in this instance, only the morphological features can be of help for the cytopathologists. What's more, membranous expression of E-cadherin was reported in more than 90% of neuroendocrine pancreatic neoplasm and tumor cells of solid pseudo-papillary neoplasms of the pancreas may show nuclear rather than cytoplasmic expression of this protein.[20-22]Finally, E-cadherin does not appear to be an appropriate immunocytochemical marker for the discrimination of contaminants, such as normal gastrointestinal epithelia in pancreatic cytological samples obtained by EUS-FNAC, since a strong membranous staining was constantly observed in both nonneoplastic contaminants and pancreatic neoplastic cells.

With reference to CD10, expression of this antigen was documented in several normal tissues and in a number of tumors, mainly in the gastrointestinal and genitourinary tracts.[23-25]Presumably due to its involvement in cellular secretory process,[26]CD10 staining is observed at the apical membrane in normal tissues. Accordingly, we observed CD10 immunostaining at the apical membrane in the benign gastrointestinal cells present within our specimens. Conversely, the large majority of the neoplastic cells of pancreaticobiliary adenocarcinomas were consistently negative for this antigen. Similarly to that observed in our cell blocks of pancreatic carcinomas, CD10 expression was absent in histological surgical samples of pancreatic adenocarcinomas.[22,26]Interestingly, strongly intense immunostaining for CD10 was found at the cytoplasm and cell membrane in one undifferentiated pancreatic carcinoma, however differential diagnosis towards intestinal contaminating cells was straightforward in this case. Of note, while CD 10 was expressed in 100% of solid pseudo-papillary pancreatic neoplasms, only 30% of endocrine pancreatic tumors displayed focal staining for this protein in previous studies.[19,22,24]Thus, the CD10 immunocytochemical staining alone may not be of use to solve differential diagnosis between different kinds of pancreatic neoplasms.

With reference to p53 immunostaining, this was found in 55% of cytological specimens which proved to be malignant at histology. This percentage is in accordance with previous literature.[27]However, differently from other studies,[28]none of the benign cases showed p53 immunoreactivity in our series. It is well known that p53 tumor suppressor gene encodes for a nuclear protein that plays a crucial role in cell cycle regulation and in major early events in pancreatic cancer.[29,30]In particular, p53 protein expression increased progressively from normal pancreatic duct epithelium to pancreatic intraepithelial neoplasia and to invasive pancreatic adenocarcinomas.[29,30]Recently, p53 overexpression was reported in pancreatic adenocarcinomas,[31]although the mechanisms of crosstalk between p53 status and other transcription factors still remain unclear and require further investigations. It has been suggested that if the degree of damage to the genome is excessive, p53 can call a halt to further cell-cycle progression until these conditions have been normalized,[32]on the other hand, if the damage is irreparable, p53 can trigger apoptosis.[32]In any case, it can be argued that p53 activation is a complex phenomenon, varying by cell type as well as by the severity of genomic damage.

Immunocytochemical positivity for p53 which we observed in our cases is to be considered as indicative for a mutant p53 protein with a long half-life. Immunoreactivity for p53 was already recorded as the inactivation of the p53 tumor suppressor gene, and found in 50%-70% of pancreatic carcinomas, but not in chronic pancreatitis,[27]which suggests a close association between p53 mutation and malignant progression.[31]Interestingly, a previous study showed that pancreatic malignancies could be detected with 51% sensitivity and 100% specifcity by combination of p53 immunocytochemistry with cytological examination, compared to 41% sensitivity and 100% specifcity of the cytology alone.[28]In the present study we showed that the use of immunocytochemistry against p53 and CD10 in cell-block preparations allows to identify malignant pancreaticobiliary lesions with 94% sensitivity, 91% specifcity, 87% PPV and 81% NPV, compared to 63% sensitivity, 90% specifcity, 95% PPV and 41% NPV of cytological smear. The apparent contradiction with in p53 diagnostic value may fnd a possible explanation in a selection bias of our cases as refected by the high percentage (72%) of malignant lesions. Similarly, the straightforward diagnostic improvement of cell-block/p53/CD10 may be due to the high number (41%) of cases with indeterminate cytology at smear.

In conclusion, in routinary cytopathology the defnitive diagnosis of pancreaticobiliary adenocarcinomas is based on the examination of conventional smears, although the complementary use of one or more immunocytochemical markers may be required for controversial cases. In accordance, the guidelines for pancreaticobiliary cytology of the Papanicolaou Society of Cytopathology[33]not only defne diagnostic categories and criteria, but also provide details on the immunohistochemical ancillary tests, which may improve the diagnostic value of the cytological evaluation. The most challenging and controversial categories in pancreaticobiliary lesions are “atypical, suspicious and neoplastic“ with the widest range of interpretation of malignancy risk. In the present study, 11 cases which were classifed as indeterminate (neoplastic cells present) at cytological smear, could be correctly classifed by using cell-block preparation andCD10/p53 immunostains. Therefore, we suggest the use of immunocytochemistry against CD10 and p53 in order to distinguish gastrointestinal or pancreatic cellular contaminants with hyperplastic or reactive changes from differentiated pancreaticobiliary neoplastic elements. The cell-block method seems to be the most appropriate to carry out this immunocytochemical panel, since it allows having serial sections and it avoids the background of contaminant cells compared to the traditional cytological techniques. However, further studies are needed to identify additional markers for more precise typization of pancreaticobiliary malignancy.

Contributors:IA, TP, BV and TG are responsible for execution and analysis of the work and drafting the manuscript. TP and CSF are responsible for planning the study, managing the patients and performing EUS-FNAC. IA and TG are responsible for conceiving and coordinating the whole work, execution and analysis of the work and fnal approval of the version to be published. Each author acknowledges that the fnal version was read and approved. TG is the guarantor.

Funding:None.

Ethical approval:The study protocol for collecting and using human samples was approved by the Institutional Ethics Committee of our hospital.

Competing interest:No benefts in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.

1 Jenssen C, Dietrich CF. Endoscopic ultrasound-guided fne-needle aspiration biopsy and trucut biopsy in gastroenterology-An overview. Best Pract Res Clin Gastroenterol 2009;23：743-759.

2 Itoi T, Tsuchiya T, Itokawa F, Sofuni A, Kurihara T, Tsuji S, et al. Histological diagnosis by EUS-guided fne-needle aspiration biopsy in pancreatic solid masses without on-site cytopathologist： a single-center experience. Dig Endosc 2011;23：34-38.

3 Yoshinaga S, Suzuki H, Oda I, Saito Y. Role of endoscopic ultrasound-guided fne needle aspiration (EUS-FNA) for diagnosis of solid pancreatic masses. Dig Endosc 2011;23：29-33.

4 Todaro P, Crinò SF, Pallio S, Fazzari C, Consolo P, Tuccari G. Gastrointestinal stromal tumours of the stomach： Cytological and immunocytochemical diagnostic features of two cases diagnosed by endoscopic ultrasound-guided fne needle aspiration. Oncol Lett 2013;5：1862-1866.

5 Mitsuhashi T, Ghafari S, Chang CY, Gu M. Endoscopic ultrasound-guided fne needle aspiration of the pancreas： cytomorphological evaluation with emphasis on adequacy assessment, diagnostic criteria and contamination from the gastrointestinal tract. Cytopathology 2006;17：34-41.

6 Layfeld LJ, Jarboe EA. Cytopathology of the pancreas： neoplastic and nonneoplastic entities. Ann Diagn Pathol 2010;14：140-151.

7 Kopelman Y, Marmor S, Ashkenazi I, Fireman Z. Value of EUSFNA cytological preparations compared with cell block sections in the diagnosis of pancreatic solid tumours. Cytopathology 2011;22：174-178.

8 Toll AD, Bibbo M. Identifcation of gastrointestinal contamination in endoscopic ultrasound-guided pancreatic fne needle aspiration. Acta Cytol 2010;54：245-248.

9 Liu H, Shi J, Anandan V, Wang HL, Diehl D, Blansfeld J, et al. Reevaluation and identifcation of the best immunohistochemical panel (pVHL, Maspin, S100P, IMP-3) for ductal adenocarcinoma of the pancreas. Arch Pathol Lab Med 2012;136：601-609.

10 Barresi V, Branca G, Vitarelli E, Tuccari G. Micropapillary pattern and poorly differentiated clusters represent the same biological phenomenon in colorectal cancer： a proposal for a change in terminology. Am J Clin Pathol 2014;142：375-383.

11 Varadarajulu S, Wallace MB. Applications of endoscopic ultrasonography in pancreatic cancer. Cancer Control 2004;11：15-22.

12 Pitman MB, Deshpande V. Endoscopic ultrasound-guided fne needle aspiration cytology of the pancreas： a morphological and multimodal approach to the diagnosis of solid and cystic mass lesions. Cytopathology 2007;18：331-347.

13 Zhao H, Mandich D, Cartun RW, Ligato S. Expression of K homology domain containing protein overexpressed in cancer in pancreatic FNA for diagnosing adenocarcinoma of pancreas. Diagn Cytopathol 2007;35：700-704.

14 Hecht SA, McCormack M. Comparison of three cell block techniques for detection of low frequency abnormal cells. Pathol Lab Med Int 2013;5：1-7.

15 Basturk O, Farris AB, Adsay NV. Immunohistology of the pancreas, biliary tract, and liver. In： Dabbs, Diagnostic Immunohistochemistry, Churchill Livingstone (ed), 2010：541-592.

16 Bellevicine C, Malapelle U, Iaccarino A, Schettino P, Napolitano V, Zeppa P, et al. Foamy gland pancreatic ductal adenocarcinoma diagnosed on EUS-FNA： a histochemical, immunohistochemical, and molecular report. Diagn Cytopathol 2013;41：77-80.

17 Hong SM, Li A, Olino K, Wolfgang CL, Herman JM, Schulick RD, et al. Loss of E-cadherin expression and outcome among patients with resectable pancreatic adenocarcinomas. Mod Pathol 2011;24：1237-1247.

18 Jamieson NB, Carter CR, McKay CJ, Oien KA. Tissue biomarkers for prognosis in pancreatic ductal adenocarcinoma： a systematic review and meta-analysis. Clin Cancer Res 2011;17：3316-3331.

19 Gao L, Antic T, Hyjek E, Gong C, Mueller J, Waxman I, et al. Immunohistochemical analysis of E-cadherin and zeste homolog 2 expression in endoscopic ultrasound-guided fne-needle aspiration of pancreatic adenocarcinoma. Cancer Cytopathol 2013;121：644-652.

20 El-Bahrawy MA, Rowan A, Horncastle D, Tomlinson I, Theis BA, Russell RC, et al. E-cadherin/catenin complex status in solid pseudopapillary tumor of the pancreas. Am J Surg Pathol 2008;32：1-7.

21 Burford H, Baloch Z, Liu X, Jhala D, Siegal GP, Jhala N. E-cadherin/beta-catenin and CD10： a limited immunohistochemical panel to distinguish pancreatic endocrine neoplasm from solid pseudopapillary neoplasm of the pancreas on endoscopic ultrasound-guided fne-needle aspirates of the pancreas. Am J Clin Pathol 2009;132：831-839.

22 Kim MJ, Jang SJ, Yu E. Loss of E-cadherin and cytoplasmicnuclear expression of beta-catenin are the most useful immunoprofles in the diagnosis of solid-pseudopapillary neoplasm of the pancreas. Hum Pathol 2008;39：251-258.

23 Chu P, Arber DA. Paraffn-section detection of CD10 in 505nonhematopoietic neoplasms. Frequent expression in renal cell carcinoma and endometrial stromal sarcoma. Am J Clin Pathol 2000;113：374-382.

24 Notohara K, Hamazaki S, Tsukayama C, Nakamoto S, Kawabata K, Mizobuchi K, et al. Solid-pseudopapillary tumor of the pancreas： immunohistochemical localization of neuroendocrine markers and CD10. Am J Surg Pathol 2000;24：1361-1371.

25 Salla C, Konstantinou P, Chatzipantelis P. CK19 and CD10 expression in pancreatic neuroendocrine tumors diagnosed by endoscopic ultrasound-guided fne-needle aspiration cytology. Cancer 2009;117：516-521.

26 Erhuma M, Köbel M, Mustafa T, Wulfänger J, Dralle H, Hoang-Vu C, et al. Expression of neutral endopeptidase (NEP/ CD10) on pancreatic tumor cell lines, pancreatitis and pancreatic tumor tissues. Int J Cancer 2007;120：2393-2400.

27 Itoi T, Takei K, Sofuni A, Itokawa F, Tsuchiya T, Kurihara T, et al. Immunohistochemical analysis of p53 and MIB-1 in tissue specimens obtained from endoscopic ultrasonography-guided fne needle aspiration biopsy for the diagnosis of solid pancreatic masses. Oncol Rep 2005;13：229-234.

28 Jahng AW, Reicher S, Chung D, Varela D, Chhablani R, Dev A, et al. Staining for p53 and Ki-67 increases the sensitivity of EUS-FNA to detect pancreatic malignancy. World J Gastrointest Endosc 2010;2：362-368.

29 Hermanova M, Trna J, Nenutil R, Dite P, Kala Z. Expression of COX-2 is associated with accumulation of p53 in pancreatic cancer： analysis of COX-2 and p53 expression in premalignant and malignant ductal pancreatic lesions. Eur J Gastroenterol Hepatol 2008;20：732-739

30 Norfadzilah MY, Pailoor J, Retneswari M, Chinna K, Noor LM. P53 expression in invasive pancreatic adenocarcinoma and precursor lesions. Malays J Pathol 2011;33：89-94.

31 Sheng W, Dong M, Zhou J, Li X, Liu Q, Dong Q, et al. The clinicopathological signifcance and relationship of Gli1, MDM2 and p53 expression in resectable pancreatic cancer. Histopathology 2014;64：523-535.

32 Hanahan D, Weinberg RA. Hallmarks of cancer： the next generation. Cell 2011;144：646-674.

33 Layfeld LJ, Ehya H, Filie AC, Hruban RH, Jhala N, Joseph L, et al. Utilization of ancillary studies in the cytologic diagnosis of biliary and pancreatic lesions： the Papanicolaou Society of Cytopathology guidelines for pancreatobiliary cytology. Diagn Cytopathol 2014;42：351-362.

Received April 30, 2014

Accepted after revision November 19, 2014

AuthorAffliations:Department of Human Pathology “Gaetano Barresi“, Section of Pathological Anatomy (Ieni A, Todaro P, Barresi V and Tuccari G) and Department of Internal Medicine, Digestive Endoscopy Unit (Crinò SF), Azienda Ospedaliera Universitaria “Policlinico G. Martino“, University of Messina, Messina 98125, Italy

Giovanni Tuccari, MD, Department of Human Pathology “Gaetano Barresi“, Azienda Ospedaliera Universitaria “Policlinico G.Martino“, Via Consolare Valeria 1, Messina 98125, Italy (Tel： +39-90-2212539; Fax： +39-90-2928150; Email： tuccari@unime.it)

© 2015, Hepatobiliary Pancreat Dis Int. All rights reserved.

10.1016/S1499-3872(15)60367-8

Published online May 6, 2015.