Adenosquamous Of The Pancreas: A Clinicopathologic Series Of 25 Cases David E. Kardon, M.D., Lester D.R. Thompson, M.D., Ronald M. Przygodzki, M.D., Clara S. Heffess, M.D. Department of Endocrine and Otorhinolaryngic-Head & Neck Pathology (DEK, LDRT, CSH), and Department of Cellular Pathology (RMP), Molecular Division, Armed Forces Institute of Pathology, Washington, DC

months after diagnosis (range, 1 to 33 months). Background: Adenosquamous carcinoma is a Conclusions: Adenosquamous carcinoma of the rare aggressive subtype of pancreatic adenocar- pancreas represents a distinct clinical and cinoma. We describe the clinical, pathologic, pathologic entity, demonstrating the expected and molecular characteristics of 25 of these le- immunoprofile and k-ras oncogene mutation of sions, the largest series to date. Methods: a ductal origin, with a worse prognosis than Twenty-five cases of adenosquamous carcinoma ductal . of the pancreas diagnosed between 1961 and 1994 were retrieved from the files of the Endo- KEY WORDS: Pancreas, Adenosquamous carci- crine Registry of the Armed Forces Institute of noma, , Adenoacan- Pathology. Histologic features were reviewed, thoma, , Immunohistochemistry, K-ras. histochemical, immunohistochemical, and mo- Mod Pathol 2001;14(5):443–451 lecular (k-ras) studies were performed, and pa- tient follow-up was obtained. Results: The pa- Carcinoma of the pancreas has an incidence of 7.5 tients included 17 men and eight women, aged to 10 per 100,000 person years and is one of the 28 to 82 years (mean, 65.4 y). The patients usu- more lethal in humans, accounting or painless (17 ؍ ally experienced weight loss (n while also presenting with for the fourth highest cause of deaths in the ,(11 ؍ jaundice (n other abdominal symptoms. The tumors af- United States (1). Adenosquamous carcinoma, fol- which has been called (2, 3) and ,(17 ؍ fected the head most frequently (n ,Five mucoepidermoid carcinoma (4), is a rare subtype .(4 ؍ or body (n (9 ؍ lowed by the tail (n cases involved more than one anatomic region accounting for 1 to 4% of exocrine malignancies of the pancreas. Microscopically, all tumors (5–7), and is characterized by the histologic pat- demonstrated dual differentiation toward ade- terns of both ductal adenocarcinoma and squa- nocarcinoma and squamous carcinoma. All mous carcinoma within the same tumor. Despite cases tested were immunoreactive with keratin isolated case reports or small series published on (AE1:AE3 and CK1), whereas other keratin this tumor type (3, 5, 8, 9) and statistical inclusion markers were variably expressed: CK5/6 (88%), in general studies of pancreatic exocrine CK7 (68%), Cam5.2 (41%), and CK20(26%). CA- (2, 6), there is no large clinicopathologic series to 19–9 (84%) and CEA (74%) were positive in the date, and no series that specifically addresses the majority of the cases. K-ras oncogene mutations histochemical, immunohistochemical, and molec- were identified in seven of 13 cases. All patients ular profile of these uncommon tumors. Further- died from their disease an average of 5.8 more, there is no correlation of these results with the clinical presentation and patient outcome. We describe a series of 25 patients with adenosqua- Copyright © 2001 by The United States and Canadian Academy of mous carcinoma to specify the clinical findings as- Pathology, Inc. VOL. 14, NO. 5, P. 443, 2001 Printed in the U.S.A. sociated with these tumors, illustrate their patho- Date of acceptance: October 10, 2000. logic features, document their immunophenotype, The opinions or assertions contained herein are the private views of the authors and are not to be construed as official or as reflecting the views of present their K-ras oncogene profile, and analyze the Department of Defense. these data as they relate to patient outcome in a Address reprint requests to: Clara S. Heffess, M.D., Department of Endocrine and Otorhinolaryngic-Head & Neck Pathology, Building 54, single comprehensive study. Our results are ana- Room G066-09, Armed Forces Institute of Pathology, 6825 16th Street, lyzed in comparison with a review of the English N.W., Washington, DC 20306-6000; e-mail: [email protected]; fax: 202-782-3130. literature (MEDLINE 1966–2000).

443 MATERIALS AND METHODS was considered sufficient for the diagnosis of ade- nosquamous carcinoma. Squamous differentiation Twenty-five cases of adenosquamous carcinoma was defined as solid nests or sheets of cells with of the pancreas were selected from the files of the abundant cytoplasm, distinct cellular borders, and Endocrine Tumor Registry of the Armed Forces In- intercellular bridges, with or without extracellular stitute of Pathology, Washington, DC, between 1961 keratin production. Mucicarmine stains were used and 1994. The cases were recovered from a search to confirm mucin production by the adenocarci- of 8,372 (0.3%) benign and malignant primary pan- noma fraction of the tumors. Fifteen of the consults creatic neoplasms seen in consultation during this were autopsies, and 10 were surgical specimens, time. Twenty-four of the specimens were from hos- either resections or . pitals within the United States, and one was from a Immunophenotypic analysis was performed in 19 military base in Europe. Ten cases were obtained cases with suitable material. The standardized from military hospitals, 10 from Veterans Adminis- avidin-biotin method of Hsu et al. (10) was em- tration medical centers, and five from civilian ␮ sources including four community hospitals and ployed, using 4- m thick, formalin-fixed, paraffin- one university hospital. Four additional cases were embedded sections taken from a single representa- identified by the search, but were excluded because tive block. In five cases immunophenotypic of a significant component of anaplastic carcinoma expression was determined on the primary tumor within the primary lesion. as well as the metastatic deposit. Table 1 docu- Adequate follow-up was a prerequisite for inclu- ments the panel of commercially available immu- sion in the study. Materials within the AFIP files nohistochemical antibodies used. When required, were supplemented by a review of the patient de- proteolytic antigen retrieval was performed by pre- mographics, symptoms at presentation, medical digestion for 3 minutes with 0.05% protease VIII history including radiation exposure, surgical pa- (Sigma Chemical Co., St. Louis, MO) in a 0.1 M thology and operative reports, cancer registry phosphate buffer at a pH of 7.8 at 37°C. Antigen records, and by written questionnaires or oral enhancement (recovery) was performed, as re- communication with the treating physician(s). quired, using formalin-fixed, paraffin-embedded Follow-up data included the exact location of tissue treated with a buffered citric acid solution the primary, the pattern of metastatic spread, the and heated for 20 minutes in a calibrated micro- specific treatment modalities used, and the wave oven (11). After this, the sections were allowed length of survival. We were unable to obtain ac- to cool at room temperature in a citric acid buffer curate staging information on the cases due to solution for 45 minutes before the procedure was the referral nature of our practice. This clinical continued. Standard positive controls were used investigation was conducted in accordance and throughout, with serum used as the negative con- compliance with all statutes, directives, and trol. A positive reaction was determined by moder- guidelines of the Code of Federal Regulations, ate to heavy chromogen deposition within the cy- Title 45, Part 46, and the Department of Defense toplasm of the tumor cells. Directive 3216.2 relating to human subjects in Point mutations in K-ras-2 were sought in 15 research. cases. In five of the cases, metastatic lesions were Hematoxylin and eosin-stained slides from all analyzed as well as the primary . Muta- cases were reviewed to confirm that histopatho- tional analysis was performed by topographic mi- logic criteria for the diagnosis of adenosquamous crodissection as previously described (12–14). This carcinoma were met. Any degree of definitive squa- type of sampling specifically selects areas of tumor mous differentiation within an adenocarcinoma from 5-␮m thick paraffin-embedded tissue sec-

TABLE 1. Immunohistochemical Panel

Primary Antigen/Antibody Company Dilution Antigen Recovery Antibody Cytokeratin (AE1/ mm Boehringer Mannheim Protease treatment AE3 and CK1) Biochemicals, Indianapolis, IN, 1:50 and Dako, Carpinteria, CA 1:200 Cam 5.2 (CK 8/18) mm Becton Dickinson, San Jose, CA 1:50 Protease treatment CK 5/6 mm Dako Microwave CK 7 mm Dako 1:40 Protease treatment CK 20 mm Dako 1:100 Protease treatment CA 19-9 mm Signet Labs, Dedham, MA Neat Enzyme digestion pCEA rp Dako 1:2650 — Chromogranin rp Dako 1:640 — mm, mouse monoclonal; rp, rabbit polyclonal

444 Modern Pathology tions, which will yield the highest concentration of ranged from 1 to 27 months, with a mean duration tumor cells and the lowest number of “back- of 5.4 months. The patients with the longest dura- ground” normal cells. No differentiation between tion of symptoms had nonspecific symptoms of the glandular and squamous components was bloating and nausea. made in the sampling of the tumors for molecular analysis, as the frequent mingling of the tumor elements precluded this type of separation. PCR Pathology amplification for the K-ras-2 exon 1 gene was per- Macroscopic formed using flanking intron primers (12–14). Cycle sequencing was performed using the BigDye termi- The tumors occurred in the head of the pancreas ϭ ϭ ϭ nator cycle sequencing approach (Perkin-Elmer, (n 17), the tail (n 9), and the body (n 4). Foster City, CA) with one of the amplifying primers, Among these, five tumors encompassed multiple and subsequently run on an ABI Prism 377 auto- anatomic segments of the pancreas, with one case mated sequencer (Perkin-Elmer) using a 6% dena- involving the entire pancreas. The pancreatic tu- turing polyacrylamide gel. Suspect mutations were mors ranged in size from 2 to 12 cm in greatest subsequently re-amplified and confirmed by se- dimension, with a mean size of 6.3 cm. Information quencing the opposite strand. about the size of the primary was unavailable for A review of English journal publications (1966 to four of the cases. Macroscopically, the tumors were 2000) was performed, and all cases of adenosqua- described as infiltrative yellow-white, gray, or green mous carcinoma (adenoacanthoma and mucoepi- firm masses with areas of softening or liquefaction. dermoid carcinoma were also reviewed) were in- Three of the masses were multinodular and two cluded in the review, even if part of a large review were cystic. The cut surface was described as fi- series. However, our summary of the literature is brous or gritty, and a minor number extruded mu- confined to reports that provide sufficient clinical cinous material. and pathology description to confirm the diagnosis. No foreign language articles were included. Microscopic All tumors exhibited a biphasic malignant growth identified as well to poorly differentiated adenocar- RESULTS cinoma and well to poorly differentiated squamous Clinical cell carcinoma. The adenocarcinoma component contained ductal or glandular structures with focal The demographic data of the patients are sum- to abundant intracellular or extracellular mucin marized in Table 2. Eighteen patients reported (Fig. 1A–B). Rare tumors contained foci of clear cell weight loss, 13 presented with jaundice, and 13 adenocarcinoma or signet cell carcinoma. Squa- complained of abdominal symptoms, which in- mous differentiation was characterized by irregular cluded pain, nausea, vomiting, bleeding, bloating, and infiltrating nests or sheets of polygonal cells and a palpable mass. The duration of symptoms with distinct cellular borders, intercellular bridges, opaque eosinophilic cytoplasm, and varying de- grees of keratinization (Fig. 1C–D). These two dif- TABLE 2. Patient Demographics and Clinical Presentation ferent patterns could be seen separated topograph- ically within the substance of the tumor or Characteristic Result intimately amalgamated with one another (Fig. 2). Gender Men 17 Benign squamous metaplasia in the surrounding Women 8 pancreatic tissue was absent in all of the cases. No Age “intermediate” cells like the type seen in mucoepi- Range 28–82 years Mean 65.4 years dermoid carcinoma of the salivary gland were evi- Racea dent in any of the cases. Perineural or neural inva- Caucasian 23 sion was identified in 20 of the tumors (Fig. 3) and African-American 1 Presenting symptoms both large vessel and lymphatic invasion were Weight loss 18 common. Jaundice 13 The two types of carcinoma were present in vari- Abdominal symptoms 13 Duration of symptoms able proportions, with the minority of tumors ex- Range 1–27 months hibiting almost exclusive squamous cell phenotype. Mean 5.4 months However, adenocarcinoma was always present Location of primary Head 17 somewhere within the tumor, and was confirmed in Body 4 all cases by a positive mucicarmine stain (Fig. 4). Tail 9 We reviewed the metastatic deposits in 13 cases a Race unknown in one patient. (pathology slides or blocks of the metastatic foci

Pancreas adenosquamous carcinoma (D.E. Kardon et al.) 445 FIGURE 1. The morphologic spectrum of glandular and squamous components. Adenocarcinoma similar to conventional ductal carcinoma (A), but rare cases were clear cell (B). Squamous cell carcinoma composed of irregular nests of polygonal cells with no gland formation. A well- differentiated SCC with abundant keratin (C). Keratin was absent in a moderate to poorly differentiated tumor (D).

FIGURE 2. The adenocarcinoma and squamous cell carcinoma could FIGURE 3. Glandular (left) and squamous cell (right) perineural be topographically separated (left) or intimately admixed (right). invasion.

were not available in the remaining cases), nine of collections of loosely cohesive sheets of markedly which were exclusively adenocarcinoma. Interest- enlarged cells with bizarre nuclei, prominent nucle- ingly, three of the carcinomas underwent a dedif- oli, and multinucleated forms (Fig. 5). Squamous ferentiation on metastasizing and exhibited the gi- cell carcinoma was only identified focally (less than ant cell or pleomorphic cell-type anaplastic 5% of the volume) in the metastatic lesions of two carcinoma. These anaplastic metastatic foci were cases.

446 Modern Pathology (16%) yielded focal reactivity to anti-CK20. All cases that were immunoreactive with anti-CK20 were more diffusely and intensely reactive for CK7. Im- munoreactivity for CK7 and CK20 was restricted to the adenocarcinoma component. CK5/6 was reac- tive (88% of cases tested) predominantly in the squamous cell component (Fig. 6). Cam 5.2 showed a slightly greater predilection for adenocarcinoma and was positive in 47% of the cases. Whereas CA 19–9 stained both components (89% of cases test- ed), the adenocarcinoma reacted with a greater in- tensity and distribution than the squamous cell car- cinoma, which tended to react in the central areas FIGURE 4. Mucicarmine staining of the tumor shows squamous of the squamous nests (Fig. 7). Likewise, CEA had differentiation (polygonal cells and intercellular bridges) along with an almost identical pattern of distribution as glandular differentiation (a mucicarmine positive mucin droplet in the center). CA19–9, while reactive in 79% of cases. All cases were negative for chromogranin.

FIGURE 5. An anaplastic morphology of a metastatic deposit in a lymph node. The cells are large and discohesive with bizarre nuclear features (inset).

FIGURE 6. The squamous component (left) was strongly Immunohistochemistry immunoreactivity with CK5/6, while the glandular component was Tissue blocks were available in 19 cases. In five of negative (right). the cases, tissue from metastatic lesions was avail- able, and immunohistochemistry was performed on the metastatic lesions as well as the primary lesions. All cases were positive for the pan- cytokeratin (Table 3). The majority of the tumors (68%) expressed CK7, whereas only a few cases

TABLE 3. Immunohistochemical Findings

Positive Reactions/Number of Immunohistochemical Antibody Cases Tested Primary Lesion Metastatic Focia

Cytokeratin (AE1/AE3 and CK1) 19/19 6/6 Cam 5.2 7/17 5/6 CK 5/6 15/17 4/6 CK7 13/19 4/6 CK20 5/19 1/6 CA 19-9 16/19 5/6 CEA 14/19 4/6 FIGURE 7. The glandular elements (left) were strongly CA19–9 Chromogranin 0/7 0/6 immunoreactive, while the squamous elements (right) were positive a The metastatic lesions were taken from five cases. In one case, two only in the center of the nests. A similar pattern of staining was seen separate lesions were evaluated. with CEA.

Pancreas adenosquamous carcinoma (D.E. Kardon et al.) 447 TABLE 4. Molecular Findings as Related to Survival and Proportional Tumor Morphology

Case K-ras Genotype Survival (Months) Tumor Morphology Number (Amino Acid Change) 1 Wild-type 1 Squamous Ͼ adenocarcinoma 2 Heterozygous (gly-val) 11 Squamous ϭ adenocarcinoma 3 Wild-type 6 Intermingled 4 Wild-type 24 Squamous Ͼ adenocarcinoma 5 Wild-type 5 Squamous Ͼ adenocarcinoma 6 Heterozygous (gly-val) 5 Intermingled 7 Heterozygous (gly-asp) 33 Intermingled 8 Heterozygous (gly-val) 3 Squamous ϾϾ adenocarcinoma 9Aa Wild-type 2 Squamous ϭ adenocarcinoma B Wild-type Adenocarcinoma C Wild-type Anaplastic 10 A Wild-type 1 Squamous ϭ adenocarcinoma B Homozygous (gly-asp) Anaplastic 11 A Heterozygous (gly-asp) 2 Squamous ϭ adenocarcinoma B Wild-type Adenocarcinoma 12 A Wild-type 3 Adenocarcinoma ϾϾ squamous B Wild-type Adenocarcinoma 13 A Heterozygous (gly-asp) 2 Intermingled B Heterozygous (gly-asp) Anaplastic

aA, primary tumor; B and C, .

Molecular Pathology none. Radiation therapy was not used for any K-ras oncogene mutational analysis was per- patients. formed on 13 cases with amplifiable DNA. Six cases The presence of K-ras oncogene mutations was showed a heterozygous mutation at codon 12, of correlated with survival, as presented in Table 4. which two cases had metastases: one had a het- The seven patients whose tumors had codon 12 erozygous mutation identified in both the primary mutations survived for an average of 8 months and and metastasis, and the other had a heterozygous the patients whose tumors expressed a wild-type mutation in the primary but not in the metastasis. phenotype survived for an average of 7 months. One additional case contained a homozygous mu- All of the patients died as a result of their disease, tation in the metastatic focus alone. The mutations irrespective of the type of therapy instituted. Sur- caused a change in the amino acid product of vival after diagnosis ranged from 1 to 33 months, codon 12 from glycine to valine in three of the cases with a mean of 6 months. Only three patients sur- and from glycine to aspartate in the other four, vived beyond 1 year, all of whom had submitted to including the homozygous mutation. No K-ras on- a Whipple procedure. Widely disseminated meta- cogene mutation was present in six cases (wild static disease was the most common cause of type), including two cases with metastatic lesions. death; other causes included disseminated intra- The different morphologies of the tumors are com- vascular coagulopathy, acute gastrointestinal pared with the oncogene mutations in Table 4. bleeding, pulmonary embolism, pneumonia, sep- There was no appreciable correlation between the tic shock, and end-stage liver disease. Metastatic presence of K-ras mutations and the proportion of disease was identified in the local lymph nodes (n squamous and adenocarcinomatous elements. ϭ 13 cases), distant lymph nodes (n ϭ 6 cases), and other solid organ sites (n ϭ 14), including liver, , adrenal, brain, stomach, kidney, Treatment and Follow-Up , and peritoneum. Direct extension by Twelve patients received no surgical therapy. Five the tumor into the common bile duct, spleen, patients received palliative bypass surgery to alle- porta hepatis, and gallbladder was documented viate symptoms of bile duct obstruction, whereas clinically or microscopically. eight patients had a partial or total pancreatectomy. Four patients had an initial of the pancreatic TABLE 5. Treatment and Relationship to Survival ϭ mass, followed by bypass surgery (n 2), resection Number Mean Patient (n ϭ 1), or no additional procedure (n ϭ 1). Only Surgical Treatment of Survival (Range) four patients accepted : two after Patients in Months partial pancreatectomy, one after a choledochoje- None 12 3.0 (1–6) Chemotherapy only 1 3.0 junostomy bypass, and one who underwent no Resection 8 11.3 (1–33) other treatment (Table 5). The overall survival was Adjuvant chemotherapy 2 20.0 (7–33) 12.5 months for patients treated with chemother- Palliation other than resection 5 3.8 (1–7) apy and 4.5 months for the patients who received Adjuvant chemotherapy 1 7.0

448 Modern Pathology DISCUSSION identify mutations that are more significantly cor- related to tumor biology (unpublished data). Adenosquamous carcinoma is included among All of the mutations in this series occurred at the histologic variants of pancreatic carcinoma that codon 12, the principle locus of K-ras mutations in have been delineated in several studies (6, 15, 16). ductal adenocarcinoma. Moreover, the mutation As mentioned above, a number of synonyms have does not correlate with tumor morphology because been applied to adenosquamous carcinoma since K-ras abnormalities were noted in tumors that were its initial description as “cancroide.” (17) Adeno- intermingled, contained a dominant squamous cell squamous carcinoma is distinctive, separable from differentiation, or had a predominantly glandular the more common ductal adenocarcinoma, as it differentiation. Therefore, based on these findings, demonstrates both malignant squamous cell and the presence of K-ras oncogene mutation suggests a glandular differentiation. ductal histogenesis of adenosquamous carcinoma Malignant tumors that contain both squamous and does not have specific prognostic or diagnostic cell and glandular elements are described in a num- implications. ber of different anatomic sites, including the respi- The exact proportion of squamous cell differen- ratory tract, , genitourinary tiation required to diagnose adenosquamous carci- tract, pancreaticobiliary tract, salivary glands, and noma is variable, although arbitrarily set by a few thyroid (3, 5, 18). Whereas a few sites, such as the authors at 30% of the routinely sectioned tumor lung and esophagus, may develop either adenocar- volume (5, 7). The percentage of squamous cell cinoma or squamous cell carcinoma as a primary carcinoma within this series varied, from predom- tumor, most of the other locations almost exclu- inantly squamous cell carcinoma to mostly adeno- sively give rise one type or the other. Based on the carcinoma. Whereas we did not prosect the speci- findings of this study, we subscribe to the theory mens and so could not meticulously map the tumor that adenosquamous carcinomas in the pancreas topography as others have (20), we did not find a occur as a result of malignant squamous metaplas- consistent topical distribution of the two compo- tic change of an adenocarcinoma (3). nents. We believe that accurate evaluation of the Squamous metaplasia of the pancreatic ductal percentages is too subjective to be reliably repro- occurs most commonly in the setting of duced based on the amount of tumor sectioned, the chronic pancreatitis, but is noted in the adjacent specific area selected, and the method used to de- ducts of only about 4% of (15). termine the tumor proportions. It is our opinion Similar to the literature, we were unable to demon- that primary tumors of the pancreas that show any strate either squamous metaplasia in the unin- degree of definitive malignant squamous cell differ- volved pancreas or a transition from squamous entiation on routine sectioning should be consid- metaplasia to squamous cell carcinoma. By con- ered adenosquamous carcinoma. Furthermore, it trast, a spectrum of ductal hyperplasia to carci- has been suggested in the medical literature that noma in situ was easily identified in cases of ade- squamous cell carcinoma occurs de novo within the nocarcinoma, both in this series and in reports in pancreas without an identifiable ductal component the literature (15). The propinquity and often inti- (16, 21–23). Based on our review of these publica- mate intermingling of the squamous cell and ductal tions, there was either limited diagnostic material elements of the cases in this series morphologically or no resection specimen, or the cases were pre- dispels the notion of a collision tumor. K-ras onco- sented as part of a large series of , gene mutations, present in ductal adenocarcinoma, which did not specify the details of the individual have also been used to prove the ductal histogene- cases. In our experience, we have yet to identify a sis of anaplastic carcinoma (19). Therefore, by in- primary pure squamous cell carcinoma of the pan- ference, the presence of K-ras oncogene mutations creas, and if such an entity exists it is vanishingly in over 50% of adenosquamous carcinomas we an- rare. With adequate tissue processing we believe a alyzed, even within tumors that were almost exclu- ductal component is almost always present even in sively squamous, suggests a ductal origin for both tumors that show a dominant squamous cell elements of adenosquamous carcinomas. This in- morphology. cidence of mutation is lower than that reported in Perineural infiltration is a common feature of the literature for adenocarcinoma. However, the adenosquamous carcinoma, occurring in 80% of method we employed for detection does not in- the cases in this series, which is a higher incidence clude a restriction step to enhance recovery of K-ras than has been previously reported (21). As with mutations, and only yields a positive result when conventional adenocarcinoma, this feature is help- approximately 20% or more of the cells analyzed ful in diagnosing on core biopsies, but harbor a mutation. Based on preliminary work at shows no prognostic significance. Although we our institute, this less sensitive detection of K-ras noted both large vessel and lymphatic spread in the mutations in pancreatic adenocarcinoma may tumors, we did not quantify the percentage of cases

Pancreas adenosquamous carcinoma (D.E. Kardon et al.) 449 with vascular invasion because of its lack of impact safety, and increasing popularity of intraoperative on the prognosis of these aggressive neoplasms and or transabdominal fine-needle aspiration biopsy, because of interobserver variability in interpreta- adenosquamous carcinoma may be diagnosed by tion, particularly in small biopsy specimens. cytology before resection (8, 28). As such, this may Immunophenotypically, the glandular and squa- introduce another variable in deciding whether to mous cell areas were distinctive. Keratin expression undergo major abdominal surgery when there is no differed, with the adenocarcinoma reacting with chance of cure and a more rapid clinical course anti-CK7 and the squamous cell carcinoma reacting based on current data (5, 8). In this series, the with anti-CK5/6. The CK7 immunoreactivity of the patients who survived the longest had all under- adenocarcinoma component duplicates that of gone a Whipple resection, suggesting that this pal- normal pancreatic ducts and conventional adeno- liative treatment may extend the patient’s life, and carcinoma of the pancreas (24, 25). The tumor may improve the quality of life by reducing chronic markers CA19–9 and pCEA are thought to represent pain, biliary obstruction, or duodenal obstruction markers of adenocarcinoma, and predominantly (26). Alternatively, the longer survival may merely stained the glandular component, but also demon- reflect the earlier stage of the primary in these strated focal immunoreactivity in squamous cell areas. This differential antigen expression should cases, rendering them amenable to surgical resec- not be construed as proof of separate clonality or tion. The use of chemotherapy as an adjunct to origin, but rather as different phenotypic expres- surgery may also increase the duration of survival, sion within a single neoplastic proliferation. as is suggested in this series. However, owing to the The distinction of adenosquamous carcinoma rarity of adenosquamous carcinoma, the number of from ductal adenocarcinoma may have certain clin- cases is too small to statistically support this claim. ical relevance, even though the clinical presenta- tion and epidemiology of the patients in this series are similar to conventional ductal adenocarcinoma Acknowledgment: The authors thank Ms. Serena Lei (26). Whereas the prognosis for ductal adenocarci- for her editorial review of the article. noma of the pancreas is grim, witha3to5%5-year survival rate (26, 27), the overall survival for adeno- squamous carcinoma appears to be worse. Only three patients in this series survived for at least 1 REFERENCES year after diagnosis, with an overall average survival of less than 6 months. Specifically, the 5-year sur- 1. Ahlgren J. Epidemiology and risk factors in pancreatic can- vival rate of resectable adenocarcinomas is 10% cer. Semin Oncol 1996;23:241–50. with a median survival of 18 to 20 months (26), 2. Sommers SC, Meissner WA. Unusual carcinomas of the pan- creas. Arch Pathol Lab Med 1954;58:101–11. while the eight patients in this series who received 3. Cihak RW, Kawashima T, Steer A. Adenoacanthoma (adeno- complete surgical resection had an average survival squamous carcinoma) of the pancreas. Cancer 1972;29: of only 11 months. Although the number of cases in 1133–40. this series is too small to demonstrate statistical 4. Frantz WK. Adenosquamous carcinoma. In: Frantz WK, ed- significance to this discrepancy, the data at least itors. Tumors of the pancreas. Atlas of tumor pathology. suggest a more rapid clinical course. The aggressive Series 1 Fascicle 8. Washington, DC: Armed Forces Institute behavior of adenosquamous carcinoma is analo- of Pathology; 1959. p. 66. 5. Madura JA, Jarman BT, Doherty MG, Yum MN, Howard TJ. gous to that of anaplastic carcinoma. These two Adenosquamous carcinoma of the pancreas. Arch Surg 1999; entities are likely related, and represent variant de- 134:599–603. grees of differentiation of ductal adenocarcinoma. 6. Cubilla AL, Fitzgerald PJ. Morphological patterns of primary The three metastatic tumors in this series that ex- nonendocrine human pancreas carcinoma. Cancer Res 1975; hibited anaplastic carcinoma, two of which demon- 35:2234–48. strating a K-ras oncogene mutation, support this 7. Solcia E, Capella C, Klöppel G. Adenosquamous carcinoma. In: Solcia E, Capella C, Klöppel G, editors. Tumors of the hypothesis. Similarly, we rejected four additional pancreas. Atlas of tumor pathology. Series 3 Fascicle 20. cases from our search because they contained a Washington, DC: Armed Forces Institute of Pathology; 1995. large proportion of anaplastic carcinoma in the pri- p. 90–1. mary tumor. The overall survival of the patients is 8. Wilczynski SP, Valente PT, Atkinson BF. Cytodiagnosis of not affected by quantifiable molecular mutations. adenosquamous carcinoma of the pancreas. Use of intraop- The longest surviving patient had a tumor with a erative fine needle aspiration. Acta Cytol 1984;28:733–6. heterozygous K-ras mutation, whereas many pa- 9. Aranha G, Yong S, Olson M. Adenosquamous carcinoma of the pancreas. Int J Pancreatol 1999;26:85–91. tients with carcinomas expressing a wild-type ge- 10. Hsu SM, Raine L, Fanger H. Use of avidin-biotin-peroxidase nome died rapidly from metastatic disease. complex (ABC) in immunoperoxidase techniques: a compar- The usual aggressive clinical course could affect ison between ABC and unlabeled antibody (PAP) proce- decisions regarding therapy. With the reliability, dures. J Histochem Cytochem 1981;29:577–80.

450 Modern Pathology 11. Tacha DE, Chen T. Modified antigen retrieval procedure: 20. Yamaguchi K, Enjoji M. Adenosquamous carcinoma of the calibration technique for microwave ovens. J Histotech 1994; pancreas: a clinicopathologic study. J Surg Oncol 1991;47: 17:365–6. 109–16. 12. Przygodzki RM, Finkelstein SD, Keohavong P, Zhu D, Bakker 21. Matsuya S, Pour PM. Adenosquamous carcinoma. In: Pour A, Swalsky PA, et al. Sporadic and Thorotrast-induced an- PM, Konishi Y, Klöppel G, Longnecker DS, editors. Atlas of giosarcomas of the liver manifest frequent and multiple exocrine pancreatic tumors. Tokyo: Springer-Verlag; 1994. p. point mutations in K-ras-2. Lab Invest 1997;76:153–9. 159–68. 13. Przygodzki RM, Koss MN, Moran CA, Langer JC, Swalsky PA, 22. Sears HF, Kim Y, Strawitz J. Squamous cell carcinoma of the Fishback N, et al. Pleomorphic (giant and spindle cell) car- pancreas. J Surg Oncol 1980;14:261–5. cinoma is genetically distinct from adenocarcinoma and 23. Brayko CM, Doll DC. Squamous cell carcinoma of the pan- squamous cell carcinoma by K-ras-2 and analysis. Am J creas associated with hypercalcemia. Gastroenterology 1982; Clin Pathol 1996;106:487–92. 83:1297–9. 14. Przygodzki RM, Finkelstein SD, Langer JC, Swalsky PA, Fish- 24. Osborn M, van Lessen G, Weber K, Klöppel G, Altmanns- back N, Bakker A, et al. Analysis of p53, K-ras-2, and C-raf-1 berger M. Differential diagnosis of gastrointestinal carci- in pulmonary neuroendocrine tumors. Correlation with his- nomas by using monoclonal antibodies specific for indi- tological subtype and clinical outcome. Am J Pathol 1996; vidual keratin polypeptides. Lab Invest 1986;55:497– 148:1531–41. 504. 15. Chen J, Baithun SI, Ramsay MA. Histogenesis of pancreatic 25. Tot T. Adenocarcinomas metastatic to the liver: the value of carcinomas: a study based on 248 cases. J Pathol 1985;146: cytokeratins 20 and 7 in the search for unknown primary 65–76. 16. Baylor SM, Berg JW. Cross-classification and survival char- tumors. Cancer 1999;85:171–7. acteristics of 5,000 cases of cancer of the pancreas. J Surg 26. DiMagno EP, Reber HA, Tempero MA. AGA technical re- Oncol 1973;5:335–58. view on the epidemiology, diagnosis, and treatment of 17. Herxheimer G. Über heterologe Cancroide. Bietr Pathol Anat pancreatic ductal adenocarcinoma. Gastroenterology 1907;41:348–412. 1999;117:1464–84. 18. Wenig BM, Adair CF, Heffess CS. Primary mucoepidermoid 27. Tabuchi T, Itoh K, Ohshio G, Kojima N, Maetani Y, Shibata T, carcinoma of the thyroid gland: a report of six cases and a et al. Tumor staging of pancreatic adenocarcinoma using review of the literature of a follicular epithelial derived tu- early- and late-phase helical CT. AJR Am J Roentgenol 1999; mor. Hum Pathol 1995;26:1099–108. 173:375–80. 19. Hoorens A, Prenzel K, Lemoine NR, Klo¨ppel G. Undifferen- 28. Lozano MD, Panizo A, Sola IJ, Pardo-Mindan FJ. FNAC tiated carcinoma of the pancreas: analysis of intermediate guided by computed tomography in the diagnosis of primary filament profile and Ki-ras mutations provides evidence of a pancreatic adenosquamous carcinoma. A report of three ductal origin. J Pathol 1998;185:53–60. cases. Acta Cytol 1998;42:1451–4.

Pancreas adenosquamous carcinoma (D.E. Kardon et al.) 451