Cancer. Introduction

This supplement presents the study of various histologic types of cancers diagnosed in the populations covered by the Surveillance, Epidemiology, and End Results (SEER) Program. It describes the SEER program and the coding of histologic type by the International Classification of Diseases for Oncology. Each of the 19 articles deals with the histologic types of cancer found in major sites or with specific histologic types, such as lymphomas or melanomas. Histologic types have been grouped based on those developed by Dr. John Berg. Data presented in this supplement are based on more than one million microscopically proven invasive cancers and 98,000 in situ cancers diagnosed during the period 1973-1987 in areas covered by the SEER Program.

Gastrointestinal cancer

BACKGROUND

There are great differences in incidence and prognosis of cancer among the component sites of the digestive tract. Furthermore, a number of the histologic types of tumors at these sites differ in their incidence and prognosis.

METHODS

The SEER data from 1973 to 1987 on frequency, incidence, staging, and survival for the various histologic types of 194,452 cancers of the esophagus, stomach, and intestines were studied.

RESULTS

There was an increased incidence of adenocarcinomas of esophagus, gastric cardia, small intestine, and colon; incidence rates for rectal adenocarcinoma and esophageal squamous cell carcinoma remained stable; and rates for gastric adenocarcinoma decreased. Blacks had higher incidence rates than whites for gastric adenocarcinoma; squamous cell carcinoma of the esophagus; carcinoids of the small intestine, colon, and rectum; and gastric sarcomas. Whites, especially males, had higher incidence rates for adenocarcinomas of the esophagus and gastric cardia than did blacks. Blacks had lower rates for rectal adenocarcinomas than did whites, but slightly higher rates than whites for rectal mucinous carcinomas and colonic adenocarcinomas. Data on race, sex, survival, and time trends indicate that mucinous colorectal carcinomas are biologically different from other adenocarcinomas. Males had poorer overall survival than did females for most histologic types of digestive tract tumors. Appendiceal mucinous adenocarcinomas with distant metastases were associated with a high (50%) 5-year survival. The poor prognosis of cases diagnosed as colonic carcinoid raises the possibility that some of these actually may be carcinomas.

CONCLUSION

Studies of digestive tract cancer require analysis by histologic type for adequate assessment.

Expression of proliferating cell nuclear antigen in hyperplastic polyps, adenomas and inflammatory cloacogenic polyps of the large intestine

AIMS

To compare proliferating cell nuclear antigen (PCNA) immunoexpression in hyperplastic polyps, adenomas, and inflammatory cloacogenic polyps of the human colon and rectum using paraffin wax embedded tissue.

METHODS

The monoclonal antibody PC10 was used to demonstrate PCNA immunoreactivity in 88 polypoid lesions from 68 patients. Cases in which immunoexpression was completely absent were excluded, leaving 32 hyperplastic polyps, 31 adenomas, and seven inflammatory cloacogenic polyps for analysis. Labelling indices for the upper and lower third of each lesion and for adjacent normal mucosa were calculated.

RESULTS

The upper third labelling indices for adenomas were substantially higher than those for hyperplastic polyps or normal mucosa, whereas those for the upper thirds of hyperplastic polyps and normal mucosa did not differ greatly. The differences between the lower third samples were not significant. In 16 (50%) hyperplastic polyps positive cells persisted onto the luminal surface. Some adenomas showed the most intense staining and the highest labelling indices in the upper third, with strong staining of surface cells; this pattern was not seen in the other lesions. The inflammatory cloacogenic polyps did not show a consistent pattern of immunoexpression.

CONCLUSIONS

Differences in cell kinetics between adenomas, hyperplastic polyps, and normal mucosa may be shown in formalin fixed, paraffin wax embedded tissue using PC10 as a marker of proliferative activity. PCNA expression also persists into the upper portions of hyperplastic polyps. Assuming that hyperplastic polyps are hypermature lesions with a slower rate of cell migration, this finding suggests that there may be an alteration in PCNA protein metabolism.

CD34 immunoexpression in stromal tumours of the gastrointestinal tract and in mesenteric fibromatoses

The aim of this study was to explore whether CD34 immunoreactivity can distinguish between different types of gastrointestinal stromal tumour, i.e. smooth muscle and neurogenic. We studied 87 stromal tumours from different sites in the gastrointestinal tract, as well as the omentum and mesentery, using a monoclonal antibody to CD34 (QBEND10). We also determined the immunoexpression of smooth muscle and muscle specific actins, S-100 protein, cytokeratin, desmin and vimentin. In addition, 15 cases of mesenteric fibromatosis were tested for CD34. Immunoexpression of CD34 was observed in 40 of the 87 stromal tumours and correlated with evidence of differentiation towards a smooth muscle phenotype. Large intestinal stromal tumours were less likely than gastric lesions to be CD34 positive. None of 15 cases of mesenteric fibromatosis was positive for CD34. We conclude that CD34 immunoexpression is seen in a proportion of stromal tumors of the gastrointestinal tract, mesentery and omentum, particularly those of smooth muscle type, and it may be useful as part of an immunohistochemical panel in the differential diagnosis of these neoplasms.

Proliferating-cell nuclear antigen immunocytochemistry in the evaluation of gastrointestinal smooth-muscle tumors

Current methods do not provide a way always to distinguish benign from malignant gastrointestinal smooth-muscle tumors. We compared immunocytochemical assessment of proliferating-cell nuclear antigen expression with flow cytometry and mitotic figure-counting as a prognostic marker in 85 gastrointestinal smooth-muscle tumors. Although proliferating-cell nuclear antigen expression was associated with poor prognosis in univariate statistical analysis, it was not significant in multivariate proportional hazards models that included either the mitotic index or aneuploidy of the flow-cytometric G2M peak. We conclude that proliferating cell nuclear antigen assessment is not warranted in the routine evaluation of gastrointestinal smooth-muscle tumors.

Squamous cell papillomas of the esophagus: a study of 23 lesions for human papillomavirus by in situ hybridization and the polymerase chain reaction

This study assessed squamous cell papillomas of the human esophagus for the presence of human papillomavirus (HPV) and correlated the results with histological features. Twenty-three lesions obtained by endoscopic biopsy from 17 patients were studied, first by in situ hybridization (ISH) for HPV types 6-11, 16-18, 18, and 31-33-51, and second by the polymerase chain reaction (PCR) with amplification of multiple HPV types and demonstration of amplified product by ethidium bromide staining and Southern blot hybridization for HPV types 6-11, 16, and 18 in each case. Evidence of HPV DNA was found in only one lesion, which showed HPV type 6-11 by ISH and HPV positivity by Southern blotting of the amplified product after the PCR. This case exhibited histological features suggestive of HPV infection, although no morphological changes specific to the lesion were identified. The remaining 22 lesions, including those from cases in which multiple papillomas were present, were negative for HPV. The results show that HPV DNA is frequently not detectable in esophageal squamous cell papillomas, even when highly sensitive techniques are used. These findings are consistent with the hypothesis that other pathogenetic mechanisms, such as mucosal injury and repair, are important in the etiology of these lesions.

Gastric carcinoids. An immunohistochemical and clinicopathologic study of 104 patients

BACKGROUND

Gastric carcinoids are uncommon, and are unlike carcinoids at other gastrointestinal sites, clinically and pathologically.

METHODS

The authors studied specimens from 104 patients with gastric carcinoid, with study emphasis being placed on pathologic features, immunohistochemistry, clinical associations, and prognostic factors.

RESULTS

The average age of the 47 male patients and 57 female patients was 61 years. Twenty-seven patients had chronic atrophic gastritis, 12 had pernicious anemia, and 6 had hypergastrinemia; no patient had carcinoid syndrome. Most of the tumors were confined to the mucosa and submucosa. Lymph node metastases were present in only one patient. The tumors were argyrophilic in 84% and argentaffin in 14%. Chromogranin tested positive in all patients; serotonin was detected in one-third; other hormones were much less common. Gastrin-positive tumors were antral. Of the 62 patients with follow-up, 44 were alive without disease, 4 were alive with disease, and 14 were dead (4 died of carcinoid-related disease). None of the deceased had pernicious anemia or hypergastrinemia. The tumors in patients with a fatal outcome were 2 cm or larger.

CONCLUSION

Gastric carcinoids generally are indolent tumors, particularly when associated with pernicious anemia or hypergastrinemia or when smaller than 2 cm. Chromogranin is the most sensitive marker.

Nonsteroidal anti-inflammatory drug-induced diaphragm disease arising in a bypassed ileal segment

The formation of mucosal ileal diaphragms has been previously reported in patients receiving nonsteroidal anti-inflammatory drugs over extended periods of time. Whether this effect is the result of a local or systemic action by the medication has remained unknown. We report the first known instance of diaphragm disease arising in a segment of ileum that had been bypassed for many years and therefore was not exposed normally to luminal contact with the drug. This case supports at least a partial systemic mechanism in the production of the lesion.

Body iron stores and risk of colonic neoplasia

BACKGROUND

Body iron stores and dietary iron intake have both been shown to be positively associated with subsequent risk of colon cancer. This finding comes from a cohort study involving 14,000 men, but the positive association occurred in only 12 cases.

PURPOSE

We performed a case-control study of 264 men and 98 women to test for an association between serum ferritin levels and the presence of adenoma of the colon that would be independent of other known risk factors.

METHODS

Serum ferritin levels were determined in this study from sera, frozen at -80 degrees C for 5-8 years, that had been originally obtained between 1984-1987 at the Walter Reed Army Medical Center from adult male and postmenopausal female patients undergoing routine colonoscopic examination and previously enrolled in a case-control study that assessed the potential dietary and environmental risk factors for colonic neoplasia. The presence of fecal occult blood in the stool or the suggestion of colonic polyps seen on barium enema defined eligibility for the study. Patients with known preexisting colonic disease were excluded. Eligible patients had their blood drawn and serum prepared. Following colonoscopy and histologic review, the patients were classified into three groups: normal (without neoplastic disease), 159 subjects; adenoma, 145 subjects; and colon cancer, 29 subjects. Body iron stores were determined by measuring serum ferritin levels by a competitive-binding radiometric immunoassay. Ferritin levels categorized into quintiles for adenoma were defined. Crude and adjusted odds ratios (ORadj) with 95% confidence intervals (CIs) for cancer and adenoma related to ferritin were calculated, controlling for known or suspected risk factors including sex, age, race, body mass index, family history, tobacco use, and alcohol consumption.

RESULTS

Statistically significant associations of adenoma risk were seen in the third ([ORadj] = 3.8; 95% CI = 1.5-9.5) and fourth (ORadj = 5.1; 95% CI = 2.0-12.7) quintiles of ferritin relative to the first quintile, for smoking history (ORadj = 2.4; 95% CI = 1.3-4.3), for male sex (ORadj = 1.9; 95% CI = 1.0-3.7), and for family history of polyps or cancer (ORadj = 1.8; 95% CI = 1.0-3.4). From a second set of analyses that excluded 36 patients with serum ferritin of greater than or equal to 399 ng/mL, the greatest effect of ferritin on adenoma risk by anatomic subsite was seen in the right colon.

CONCLUSION

The apparent dose-response for serum ferritin level and adenoma risk suggest that exposure to iron may be related to adenoma formation.

Ganglioneuromas of the gastrointestinal tract. Relation to Von Recklinghausen disease and other multiple tumor syndromes

We studied 43 patients with ganglioneuromas of the gastrointestinal tract accessioned at the Armed Forces Institute of Pathology (AFIP) from 1940 to 1990 in order to determine their relation to von Recklinghausen's disease and other multiple tumor syndromes. They fell into three groups: polypoid ganglioneuroma (28 patients); ganglioneuromatous polyposis (7 patients); and diffuse ganglioneuromatosis (8 patients). Follow-up (1-24 years, average 8 years) for 16 of 28 patients with polypoid ganglioneuroma showed that none of these patients developed von Recklinghausen's disease or evidence or multiple tumor syndromes. Three of seven patients with ganglioneuromatous polyposis were alive and well but were reported to have multiple cutaneous lipomas and one reported a family history of multiple intestinal polyps. For seven of eight patients, diffuse ganglioneuromatosis was associated with other tumors, namely multiple endocrine neoplasia type IIb, multiple ganglioneuromas and neurofibromas limited to the gastrointestinal tract, von Recklinghausen's disease and neurogenic sarcoma. We conclude that the solitary polypoid ganglioneuroma of the gastrointestinal tract is not associated with the subsequent development of von Recklinghausen's disease or multiple endocrine neoplasia. All three forms of gastrointestinal ganglioneuromatous disease appear to be largely centered in the colon and rectum, unlike neurofibromas and neurofibromatosis, which, in our experience, occur more commonly in the small intestine and stomach.

Squamous cell papilloma of the esophagus: a clinicopathologic and follow-up study of 25 cases

OBJECTIVES

The aim of this study was to correlate clinical and pathologic features of squamous cell papillomas of the human esophagus and investigate their pathogenesis and malignant potential.

METHODS

Clinical and pathologic data on 25 patients were studied.

RESULTS

All patients in whom the race was known were white, and the majority of lesions occurred in the distal esophagus. There were multiple lesions in five cases. Many cases showed endoscopic evidence of hiatus hernia, gastroesophageal reflux, or esophagitis. No examples of Barrett esophagus were found. No esophageal malignancies occurred, and no dysplasia was seen histologically. Evidence of recurrence was found in only two patients, but several synchronous or metachronous carcinomas of the oro-respiratory tract occurred.

CONCLUSIONS

Although our findings do not exclude a viral etiology, they suggest a role for mucosal injury and regeneration in the pathogenesis of these lesions. The association with other malignancies may be significant.

Predicting prognosis of gastrointestinal smooth muscle tumors. Role of clinical and histologic evaluation, flow cytometry, and image cytometry

Both flow cytometry (FCM) and morphometry have been proposed as techniques for predicting the prognosis of gastrointestinal (GI) smooth muscle tumors (SMTs). In particular, DNA aneuploidy by FCM has been associated with high histologic grade and shortened survival, whereas the DNA index determined by image cytometry has been proposed as a criterion for the diagnosis of malignancy. To further define the potential roles of these two techniques, we performed a variety of morphometric and FCM measurements on paraffin blocks from 122 patients with GI SMTs, with a median follow-up period of 6 years, together with assessments of tumor size and mitotic activity. None of the morphometric measurements (nuclear perimeter, area, form factor, longest diameter, average ferret diameter, equivalent diameter, and DNA index) was a significant prognostic factor when analyzed using a univariate Cox model. In contrast, the flow cytometric mean channel number, the fraction of cells in G2M, aneuploidy of the G0/G1 peak, aneuploidy of the G2M peak, tumor size, and mitotic activity index were statistically significant in univariate models, together with the patient age and sex, and whether or not the patient presented with metastases. In a multivariate model, > 10 mitotic figures per 50 high-power fields and metastases indicated a poor prognosis. If metastasis was not allowed to enter the model, the mitotic index and aneuploidy of the G2M peak portended a poor prognosis.

The World Health Organization histological classification of tumours of the upper respiratory tract and ear. A commentary on the second edition

The second edition of Histological Typing of Tumours of the Upper Respiratory Tract and Ear, in the World Health Organization series International Histological Classification of Tumours, provides a comprehensive classification of tumors and tumor-like lesions occurring in (1) the nasal cavity and paranasal sinuses; (2) the nasopharynx; (3) the larynx, hypopharynx, and trachea; (4) the external ear; and (5) the middle and inner ear. The classification includes several newly recognized entities and many that were omitted from the first edition on account of their rarity. The terminology and definitions of several tumor types have been revised in the light of experience gained during the past 12 years. This article discusses the revised classification with special reference to some of the major changes in the listings, terminology, and definitions.

Association of mucinous tumors of the ovary and appendix. A clinicopathologic study of 25 cases

Twenty-five patients with mucinous tumors of the ovary and appendix were studied. The average age of the patients was 52 years, and the ovarian and appendiceal tumors were discovered synchronously in all but two cases. The majority had either a pelvic mass or abdominal or pelvic pain. A high frequency of bilateral ovarian tumors (11/25), and right-sided predominance for the unilateral ovarian tumors (nine right, five left) were found. Four patients had ovarian mucinous carcinomas, 10 had mucinous tumors of low malignant potential, 10 had mucinous cystadenomas, and one had a mucinous cyst. Pseudomyxoma ovarii was present in 22 cases. Twenty-two of 24 appendices were grossly abnormal. There were six appendiceal mucinous adenocarcinomas, 10 mucinous tumors of uncertain malignant potential, seven mucinous cystadenomas, one hyperplastic polyp, and one mucocele. Twelve patients had ovarian and appendiceal tumors of similar malignant potential, nine had appendiceal tumors with more aggressive morphologic features than the corresponding ovarian tumor, and four had ovarian tumors with more aggressive morphologic features than the appendiceal tumor. Eighteen patients had peritoneal involvement by mucinous epithelium admixed with mucus (nine localized, nine diffuse). Immunoperoxidase reactions for four epithelial antigens in 15 cases showed complete concordance between ovarian and appendiceal lesions in only five cases and were not helpful in determining the site of origin of the peritoneal tumor. Our findings suggest an independent origin of the ovarian and appendiceal tumors in most cases and do not favor an origin in a single site. Furthermore, it is proposed that the peritoneal lesions may arise de novo as part of a multifocal neoplastic process.

Serum gastrin is not higher in subjects with colonic neoplasia

Two previous studies have shown higher circulating gastrin levels in subjects with colonic neoplasia than in colonoscopy-negative controls. In this much larger study, sera were collected from fasting subjects undergoing colonoscopy. Colonoscopy with biopsy classified participants as having colonic adenomas (N = 139), colon carcinoma (N = 29), or controls without colonic neoplasia (N = 150). Frozen, stored sera were later analyzed for gastrin by radioimmunoassay. Serum gastrin values were no higher in subjects with colonic adenomas or carcinoma than in colonoscopy-negative controls. We conclude that elevated serum gastrin levels play little, if any, role in the initiation of colonic neoplasia.

The WHO Histological Classification of Tumors of the Gallbladder and Extrahepatic Bile Ducts. A commentary on the second edition

The second edition of the WHO Histological Classification of Tumors of the Gallbladder and Extrahepatic Bile Ducts is more comprehensive and detailed than the previous one. Advances in our understanding of dysplasia, carcinoma in situ, various lines of differentiation among the carcinomas, and the recognition of a variety of tumor-like lesions have resulted in more than three times as many entities in the current classification as in the previous one. The new edition should facilitate pathologic, epidemiologic, and therapeutic comparisons.

The World Health Organization's Histological Classification of Salivary Gland Tumors. A commentary on the second edition

The second edition of the World Health Organization's Histological Classification of Salivary Gland Tumors is more extensive and detailed than the previous edition published 20 years ago. The new edition is based on data regarding newly described tumor entities and the behavior and prognosis of the previously classified tumors. The distinct morphologic features of monomorphic adenomas justify their separation for purposes of identification. Among the carcinomas, various types were distinguished for purposes of recognition, prognosis, and treatment. The term tumor was replaced by carcinoma in the following two entities: acinic cell carcinoma and mucoepidermoid carcinoma. The tumor-like lesions were described in more detail.

Cancer staging: future directions for the TNM classification

The major accomplishment of the 1987 edition of the TNM classification was the unification of criteria for all site classifications, namely, the elimination of all variations that had developed over the years. The main directions that TNM is taking for the future are 1) verification of published classifications, 2) classification of new sites and tumor types, 3) addressing the integration of nonanatomic factors with TNM to achieve prognostic grading, and 4) application of TNM beyond pure clinical-pathological aspects to assess methods of early detection, quality of care, and population trends.

Peutz-Jeghers syndrome

The Peutz-Jeghers polyp is an unusual type of hamartomatous polyp; its characteristic feature is a smooth muscle core arising from the muscularis mucosae and extending into the polyp. Peutz-Jeghers polyps vary in size and shape; are found in the stomach, small bowel, and colon; and are usually multiple. Peutz-Jeghers syndrome is an inherited condition that often remains undiagnosed until after the polyps are identified, despite mucocutaneous pigmented lesions on the lips and mouth of children or young adults. In the past, standard therapy involved removal of the polyps that produced intussusception, but now endoscopic removal of all polyps is recommended. The polyps are not premalignant, but a definite association exists between Peutz-Jeghers syndrome and gastrointestinal carcinoma. Evidence shows that the syndrome is associated with an increased risk of extraintestinal malignancy, especially carcinomas of the pancreas, breast, and reproductive organs.

Dysplasia of the stomach and Barrett esophagus: a follow-up study

Seventy-seven patients with biopsy-proven glandular dysplasia of the esophagus (39 patients) or stomach (38 patients) were followed for a mean of 44 mo or until resection. Of 34 patients with low-grade dysplasia at initial biopsy, 29 had no evidence of high-grade dysplasia or carcinoma on follow-up, three died of other causes, and two had severe dysplasia diagnosed on subsequent biopsy. Of 43 patients with high-grade dysplasia, 28 had no evidence of dysplasia on follow-up (four died of unrelated causes and two of postoperative complications). Fifteen were shown to have invasive carcinoma. Thirteen of the 15 diagnoses of invasive carcinoma were made within 12 mo of original endoscopy. Ulcerated high-grade dysplasias were more likely associated with carcinoma than were nonulcerated high-grade dysplasias (P = 0.001). The presence of ulceration and short time interval to the progression of carcinoma suggest that carcinoma may have been present at the time of initial biopsy in many cases. The frequency of adenocarcinoma's arising in severe dysplasia was slightly less in the stomach (29%) than in the esophagus (41%). We conclude that low-grade dysplasia is often indolent and that ulcerated high-grade dysplasia is often a marker for adjacent invasion.

Polypoid and pseudopolypoid manifestations of inflammatory bowel disease

The polypoid manifestations of the idiopathic inflammatory bowel diseases (ulcerative colitis and Crohn disease) are often confusing. Inflammatory polyps project above the level of the surrounding mucosa. Pseudopolyposis (in ulcerative colitis) or a cobblestone appearance (in Crohn disease) results when extensive ulceration develops and only scattered islands of relatively normal mucosa remain; thus, the ulcerated areas may be falsely perceived as the baseline and the islands as polyps. Postinflammatory (filiform) polyps--fingerlike projections of submucosa covered by mucosa on all sides--reflect healing of undermined mucosal and submucosal remnants and ulcers and are almost always multiple. Patients with ulcerative colitis or Crohn disease are at increased risk for developing adenocarcinoma. Occasionally, dysplasia occurs as a polypoid lesion. Dysplasia of the colon (mucosal atypia) is a histologic marker highly associated with adenocarcinoma. Because differentiating adenocarcinoma and dysplasia from inflammatory or postinflammatory polyps is sometimes difficult or impossible, endoscopy and biopsy are usually recommended for definitive diagnosis of suspicious lesions.