Morphological Patterns of Primary Nonendocrine Human Pancreas Carcinoma'
Total Page:16
File Type:pdf, Size:1020Kb
[CANCER RESEARCH 35, 2234-2248, August 1975] Morphological Patterns of Primary Nonendocrine Human Pancreas Carcinoma' Antonio L Cubifla and Patrick J. Fitzgerald2 Department of Pathology, Memorial Hospital, Memorial Sloan-Kettering Cancer Center, New York, New York UX@21 Summary the parenchymal cells. In the subsequent 5 decades terms such as mucous adenocarcinoma, colloid carcinoma, duct The study of histological sectionsof 406 casesof nonen adenocarcinoma, pleomorphic cancer, papillary adenocar docrine pancreas carcinoma at Memorial Hospital mdi cinoma, cystadenocarcinoma, and other variants, such as cated that morphological patterns of pancreas carcinoma epidermoid carcinoma, mucoepidermoid cancer, giant-cell could be delineated as follows: duct cell adenocarcinoma carcinoma, adenoacanthoma, and acinar cell carcinoma, (76%), giant-cell carcinoma (5%), microadenocarcinoma have appeared (7, 18, 23, 47, 62). Subtypes of islet-cell (4%), adenosquamous carcinoma (4%), mucinous adeno tumors have been defined (27). As pointed out by Baylor carcinoma (2%), anaplastic carcinoma (2%), cystadenocar and Berg (5) in discussing the limitations of their study of cinoma ( 1%), acinar cell carcinoma (1 %), carcinoma in 5000 patients with pancreas cancer from 8 cancer registries, childhood (under 1%), unclassified (7%). few pathologists precisely characterize the microscopic In 195 cases of patients with pancreas carcinoma, search features of their cases. was made for changes in the pancreas duct epithelium and We have reviewed cases of cancer of the pancreas at these were compared to duct epithelium in a control group Memorial Hospital to determine whether there are defina of 100 pancreases from autopsies of patients with nonpan ble morphological subgroups and to indicate their relative creatic cancer. The following incidences were found for distribution in our material. The delineation of morphologi pancreas cancer and nonpancreatic cancer, respectively: cal patterns of pancreas cancer may eventually permit a mucous cell hypertrophy, 39 versus 28%; pyloric gland correlation with some genetic, biochemical, endocrinologi metaplasia, 28 and 17%; epidermoid metaplasia, 6 and 12%; cal, clinical, epidemiological, prognostic, or other parame papillary hyperplasia, 42 and 12%; atypical duct hyper ter of significance and thereby further an understanding of plasia, 14% and none; carcinoma in situ in 19% and none in this neoplasm that appears to be increasing and has such a the control group. poor prognosis for the patient at the present time. Mucin in the majority of pancreas cancers suggested that the cell type of origin of the common pancreas cancer is the mucin-producing duct epithelium. The association of Materials and Methods atypias and carcinomas in situ in the patients with pancreas carcinoma implies, by analogy to other organs, that there From a survey of 757 cases listed as cancer of the may be a significant latent period between the appearance of pancreas in the Cancer Registry, Clinical Information carcinoma in situ and the grossly recognizable pancreas Center, Memorial Hospital, during the years 1949 through cancer. 1972, we have been able to examine histological sections of tissue from 503 cases. Sixty-seven were eliminated from the study for the following reasons: in 42 cases biopsy material Introduction did not show any cancer; 3 cases were reclassified as periampullary carcinoma; 7 cases were reclassified as The morphological classification of primary pancreas metastatic cancer (2 casesof oat-cell carcinoma of lung; 1 carcinomas of the nonendocrine variety would seem to pose case each of adenocarcinoma of the lung, breast, and colon; no problem to the pathologist because most of these cancers I malignant lymphoma; and 1 embryonal rhabdomyo@ar are adenocarcinomas and they span the spectrum com coma of the retroperitoneum); in 13 cases aspiration biop monly found with adenocarcinomas of other organs. Origi sies did not furnish enough tissue for classification; 1 de nally, Ewing ( 18), from this laboratory, described only 2 batable case w@sof either gastric or pancreas origin; and 1 types of pancreascancers,cylindrical cell adenocarcinoma tumor was believed to be of common bile duct origin rather arising from the ducts and carcinoma simplex arising from than of pancreatic duct. Thirty cases of malignant islet-cell tumor, recently reported from this laboratory (14), were 1 Presented at the Symposium “Current Progress in Pancreatic Car also excluded. cinogenesis Research,― February 4, 1975, Orlando, Fla., as part of the Cases coded as ampullary or periampullary cancer, Third Annual Carcinogenesis Conference. Aided by Contract NOl CP-43232 from the National Cancer Institute, NIH, Bethesda, Md. duodenal cancer, islet-cell tumors, or retroperitoneal tu 2 Presenter. mors involving the pancreas were not examined, and 2234 CANCER RESEARCH VOL. 35 Downloaded from cancerres.aacrjournals.org on September 27, 2021. © 1975 American Association for Cancer Research. Patterns of Pancreatic Carcinoma possibly a few pancreas cancers were thereby lost to our interpretation made on the small sampling of a tumor, study. which was sometimes restricted to only a few slides. There In 406 cases there was an operative biopsy ofa pancreatic were some cases (7%) where admixtures oftypes occurred so tumor; a confirmation of a pancreas cancer in the surgically that they were listed as unclassified. A much larger resected pancreas, or autopsy specimen; or, in association sampling might change the subtyping of cancers or the with a lesion clearly defined in the operative notes as a incidence of changes in duct epithelium. pancreatic tumor, there was a biopsy taken of a regional lymph node, or of a metastatic lesion, which revealed a neoplasm consistent with pancreas cancer. In 29 1 cases Results tissue diagnostic of pancreas cancer was obtained at opera tion, and in 115 casesthe tissue was obtained at autopsy. In Nonendocrine Pancreas Neoplasms 99 cases surgically resected (partial or total) pancreas tissue was studied. Histological slides of the primary tumor varied Benign Epithelial Nonendocrine Neoplasms. We were able from 4 to 9, but often only 1 or a few, slides of the to examine histological material from only 7 cases of benign metastasis were available. Additional slides, particularly epithelial nonendocrine neoplasms (3 serous and 4 mucinous for the demonstration of mucin, were cut from the paraffin cystadenomas) and from I case containing a microscopic embedded block in 53 cases. focus of adenomatous hyperplasia of acinar cells found Recently, we have obtained for electron microscopy incidentally at autopsy. ultrastructural studies pancreas cancer tissue removed at Malignant Epithelial Neoplasms. We have classified pri operation in 28 cases and from autopsy in 3 cases (not mary malignant lesions of the nonendocrine pancreas under included in the morphological classification study). the general headings of carcinomas (Table 1), sarcomas, In 214 cases (99 cases of partial or total pancreatectomy primary malignant lymphomas, a miscellaneous group, and and 115 autopsies), additional study of the pancreas cancer metastatic cancer. We estimate that over 95% of the material for associated hyperplasia, metaplasia, atypias, or malignant lesions of the nonendocrine pancreas are adeno carcinoma in situ changes in the duct epithelium was carcinomas and the following appear to be distinctive types. attempted. Nineteen cases were rejected because of autol Duct Cell A denocarcinoma [Adenocarcinoma, Cylindri ysis ofduct epithelium or insufficient ductal tissue present in cal-Cell Adenocarcinoma (18), Papillary Adenocarcinoma the histological slides, leaving 195 cases with adequate (23), Pancreas Cancer (5)]. In the normal duct system of the material. Pancreases from 100 consecutive autopsies (per human pancreas there is variation from the flat to cuboidal formed in 1973) of patients with nonpancreatic cancers, epithelium of the smallest ductules to the high columnar similar in age and sex to the pancreascancer cases,were mucouscell ofthe epithelium in the headofthe pancreas(9, used as controls. Patients were subjected to different 24). There is absence of mucin in the smallest ductules, as regimensof chemotherapy or radiotherapy in the 2 groups, indicated by the negative mucin stains, and the presence of and no corrections were made for these factors. Four to 9 mucin, indicated by a positive mucin stain, in the apical histological sections from each pancreas of the 2 groups region of the columnar cell of the duct epithelium, particu were examined. larly in the larger ducts (24, 46). By electron microscopy it Standard fixation in 10% formalin solution and staining can be demonstrated that droplets of mucigen are present in with hematoxylin and eosin were used for all tissues. Meyer's the duct epithelium (9) (Fig. 7) and not in the smallest mucicarmine stain (45) wasusedto demonstratemucin, and epithelial cells of the ductules. the periodic acid-Schiff (43) and Alcian blue stains (44) Our cases of duct adenocarcinomas were similar to the were used to indicate pyloric gland metaplasia (57, 58, 68). adenocarcinomas seen elsewhere in the biliary tract, gas The tissue for electron microscopy was fixed overnight in trointestinal tract, lung, and other organs (Figs. I and 2). Karnovsky fixative, rinsed in buffer (s-collidine), and post There was a mixture of large and small glands, often with I fixed for 1 hr in 1% osmium tetroxide. The specimens were embedded in epoxy