Search

Intestinalized epithelium in traditional long-segment Barrett's esophagus (BE) shows increased proliferative activity, which is postulated to be an early step in the metaplasia-dysplasia-carcinoma sequence. The aim of this study was to evaluate the proliferative activity of intestinalized epithelium of the distal esophagus and gastroesophageal junction (IMEGEJ). Tissue sections from 78 consecutive patients (20 with IMEGEJ, 58 without IMEGEJ) who had elective upper gastrointestinal endoscopy over a 6-month period were immunohistochemically stained with MIB-1, the Ki-67 proliferation-antigen-associated marker, for evaluation of the crypt MIB-1 proliferation index (PI), size of the proliferative zone (PZ), and the presence of surface epithelial staining. Data from the IMEGEJ and non-IMEGEJ groups, and from 15 age-matched patients with traditional long-segment BE (>3.0 cm), were compared statistically. IMEGEJ patients showed a statistically significant increase in the mean crypt PI compared with non-IMEGEJ controls (21.9+/-19.5 v 14.3+/-9.3; P=.01). In addition, IMEGEJ cases showed an increase in the mean crypt PZ (52.3+/-16.4 v 45.2+/-17.2; P=.05), and a trend toward an increase in the percentage of cases with MIB-1-positive surface epithelial cells (50% v 33%, P=.18). Patients with IMEGEJ did not differ from patients without IMEGEJ with respect to any other clinical or histological feature, including signs or symptoms of gastroesophageal reflux disease and presence or absence of esophagitis or carditis. The MIB-1 results of the patients with long-segment BE (MIB-1 PI = 22.6+/-20.5, MIB-1 PZ = 51.8+/-19.6, proportion of cases with MIB-1-positive surface cells = 66%) were similar to those with IMEGEJ. Intestinalized epithelium in the distal esophagus or gastroesophageal junction shows increased proliferative activity in comparison with patients without intestinalized epithelium. This finding supports an increased risk of carcinogenesis in patients with IMEGEJ.

To explore the potential contributions of gastroesophageal reflux disease, as opposed to Helicobacter pylori infection, to the development of gastric carditis, we evaluated gastric carditis (using the criteria of the updated Sydney system for the classification of gastritis), clinical and morphologic features of esophagitis, and H. pylori infection (evaluation of Steiner stains) in biopsy specimens from the gastroesophageal squamocolumnar junction. We correlated clinical, endoscopic, and histologic features in an unselected group of 116 patients. Some degree of carditis was found in 107 (92%) of the patients. The mean age of the patients increased with increasing severity of carditis (P < .05). The various groups of patients with different degrees of carditis did not differ significantly in sex ratio, ethnic background, presence of obesity, percentage having symptoms of gastroesophageal reflux disease (such as heartburn, regurgitation, dysphagia, or odynophagia), endoscopic evidence of esophagitis and columnar epithelium in the distal esophagus, or histologic evidence of active esophagitis. The presence, however, of active gastritis and H. pylori infection in the distal stomach and/or in the cardia was significantly associated with carditis. In patients without carditis, H. pylori was not detected in any cardiac or distal gastric biopsy specimen. In contrast, H. pylori was demonstrated in gastric tissue samples (either from the cardia or distally) of patients with carditis, with the prevalence rate increasing with greater degrees of cardiac inflammation. The H. pylori prevalence rate was 12% in the group with mild carditis, 40% in those with moderate carditis, and 57% in patients with marked carditis (P = .0001). In summary, carditis is commonly found in patients with symptoms related to upper gastrointestinal diseases. From analysis of our study cohort, we concluded that carditis was significantly associated with H. pylori infection and active gastritis but not with symptoms or signs of gastroesophageal reflux disease. These findings suggest that carditis with histologic features similar to those of gastritis in the distal stomach was a sequel of H. pylori infection and represented a part of an H. pylori--associated gastric inflammation.

Objective: Periodic endoscopic surveillance is generally recommended for patients with Barrett's esophagus. The optimal follow-up strategy for uncomplicated Barrett's esophagus is controversial, in part because of limited data on the rate of neoplastic progression (through the sequence of metaplasia-dysplasia-carcinoma) during endoscopic surveillance. This study aims to quantify the development of dysplasia in patients with uncomplicated Barrett's esophagus and to explore clinical risk factors associated with the development of dysplastic lesions., Methods: We identified 102 patients with endoscopic evidence of Barrett's esophagus and the presence of specialized columnar epithelium who had received endoscopic surveillance for adenocarcinoma at our medical center between 1970 and 1994. We abstracted endoscopic and histologic data from the medical record. All specimens that showed any degree of atypia (per report) were reexamined in blinded fashion by a team of study pathologists who indicated the grade of dysplasia. Time to first diagnosis of dysplasia was plotted using Kaplan-Meier survival curves, and risk factors for development of dysplasia were assessed using Cox regression., Results: During 563 patient-yr of endoscopic follow-up, three patients developed adenocarcinoma at least 4 yr after initial diagnosis (one developed adenocarcinoma of the cardia, which was incidentally detected during surveillance for Barrett's esophagus). At some point during follow-up, 19 patients developed new onset, low grade dysplasia and four developed high grade dysplasia. None of the patients who had received antireflux surgery developed dysplasia., Conclusion: If confirmed by larger follow-up studies, our results suggest that surveillance endoscopy can be safely deferred for at least 2 yr following an initial biopsy that is negative or indeterminate for dysplasia. Adoption of this approach would substantially reduce the cost of surveillance for adenocarcinoma. Future trials should explore the role of antireflux surgery in protecting against neoplastic transformation of Barrett's esophagus.

Purpose: The purpose of this study was to determine the utility of spiral CT in the diagnosis of brain death., Methods: Spiral CT was evaluated prospectively in 14 brain-dead patients and in 11 healthy subjects. A two-phase protocol was used. Twenty seconds after intravenous injection of a nonionic iodinized contrast medium, the CT table was drawn through the gantry at a rate of 10 mm/s while scanning was in progress. The second scanning phase was started automatically a mean of 54 seconds later, using the same parameters. Opacification or absence of opacification of carotid, vertebral, and basilar arteries and intracerebral veins was ascertained for each image in both phases. The diagnosis of brain death was confirmed by elecroencephalography (n = 7), angiography (n = 5), or both (n = 2). Statistical analysis with the Fisher exact test enabled us to compare the brain-dead patients with the healthy control subjects., Results: In brain death, the pericallosal and terminal arteries of the cortex did not opacify during the two phases of spiral CT, whereas the superficial temporal arteries were always visible. The internal cerebral veins, the great cerebral vein, and the straight sinus did not opacify, whereas the superior ophthalmic veins were visible on both sides 13 times. For each vessel type, specificity was 100% for nonvascular opacification criteria on the right and left sides., Conclusion: Two-phase spiral CT can demonstrate the absence of intracerebral blood flow in brain death.

Dysplasia in inflammatory bowel disease (IBD) is categorized as either flat or associated with a raised lesion or mass (dysplasia-associated lesion or mass [DALM]). One specific subtype of a dysplasia-associated lesion or mass consists of isolated discrete nodules or polyps that are difficult to distinguish from sporadic adenomas. Because the clinical management of these two lesions is different, we performed this study to (1) evaluate the clinical presentation, pathologic features, and natural history of polypoid dysplastic lesions and sporadic adenomas in patients with IBD and (2) determine whether there are clinical, endoscopic, or pathologic findings useful in differentiating between these two lesions. The morphologic features of 89 benign polypoid epithelial neoplasms from 59 patients with IBD (51 with ulcerative colitis, 8 with Crohn's colitis) were evaluated and correlated with the clinical, endoscopic, and follow-up data. In a separate analysis, patients were categorized arbitrarily as having (1) a probable sporadic adenoma if the polypoid epithelial neoplasm was not located within areas of histologically proven colitis, (2) a probable IBD-associated polypoid dysplasia if the lesion developed within an area of colitis, and associated flat dysplasia or an adenocarcinoma was detected during follow-up evaluation or (3) an indeterminate polyp, which was seen in the remainder of the cases. The clinical, endoscopic, and histologic data were compared among these three patient and polyp subgroups. There were 35 males and 24 females (median age, 57 years; range, 27-85 years). Median duration of disease was 10 years. Forty-nine percent of the patients had pancolitis; 66% had histologically active disease at the time of presentation. Nearly 70% of patients had only one polyp; the majority occurred in either the left colon or the rectum (66%). Most polyps were described as a sessile nodule, whereas only 7 (7.8%) were pedunculated. Polyps ranged from 2 mm to 50 mm (median, 5 mm); most had a tubular architecture (84.3%) and contained low-grade dysplasia (64%). In addition, most polyps had mildly increased lamina propria and intraepithelial neutrophilic and mononuclear inflammation. At follow-up evaluation (40 patients; median follow-up time, 13 months; range, 1-78 months), a further neoplastic lesion developed in 20%; low-grade flat dysplasia was seen in 5 (12.5%), and adenocarcinoma developed in 3 (7.5%). However, dysplasia or adenocarcinoma did not develop in the patients who had polyps located outside of areas of histologically proven colitis. In addition, at least one more benign polypoid epithelial neoplasm developed in 15 of 40 patients (37.5%). Patients with probable IBD-associated polypoid dysplasia had a statistically significant (p < 0.05), longer disease duration than patients with probable sporadic adenoma. A statistically significant, higher proportion of polyps with tubullovillous or villous architecture, an admixture of normal and dysplastic epithelium at the surface of the polyps, and increased lamina propria mononuclear inflammation was noted in probable IBD-associated polypoid dysplastic lesions compared with those considered to be sporadic adenomas. Several clinical and pathologic features may be useful to help categorize a polypoid dysplastic lesion as a sporadic adenoma or an IBD-related neoplasm in a patient with IBD. This distinction is important because the natural history of these two lesions (as shown by the results of this study) and their subsequent management are quite different.

This article has reviewed the morphologic features of BE and its complications of dysplasia and adenocarcinoma in routinely processed endoscopic biopsy and resection specimens. In addition, the problems related to the diagnosis of dysplasia in tissue samples have been analyzed, and the role of esophageal brush cytology in the diagnosis of these lesions has been reviewed.

Tissue sampling is essential for detecting intestinal metaplasia in the distal esophagus (Barrett's esophagus), because symptoms and endoscopy are not reliable in making this diagnosis. The utility of cytology in this process is unknown. All adult patients having elective upper gastrointestinal endoscopy over a 6-month period were invited to participate in a prospective study whose aim was to determine the prevalence of intestinal metaplasia in the distal esophagus in an adult population with diverse upper gastrointestinal symptoms. Clinical data and endoscopic findings were recorded. Brush cytology and biopsy specimens were obtained from both sides of the apparent squamocolumnar junction. The cytology specimens were processed routinely, stained with the Papanicolaou technique, and reviewed blinded to the clinical information and the histological findings in the corresponding biopsy specimens. One hundred fifty-five patients (81 women, 74 men; 137 whites, 11 blacks, 7 others; mean age, 52 years) were included. Glandular epithelium/cells were present on both histology and cytology in 147 specimens. Thirty-two patients (22%) showed intestinal metaplasia on histology. Of the cytology specimens from these 32 patients, 6 contained definite goblet cells (19%), 7 probable goblet cells, and 19 no goblet cells. Goblet cells and probable goblet cells were observed on cytology in 7 and 11 additional specimens, respectively. One was from a patient known to have intestinal metaplasia in the esophagus. Follow-up endoscopy with biopsy was performed in two of these latter 18 patients and did not show intestinal metaplasia. One case of high-grade dysplasia, two of low-grade dysplasia, and three indefinite for dysplasia were diagnosed on histology. All three cases of dysplasia were also identified on cytology. The three indefinite cases on histology were considered reactive in two and unremarkable in one on cytology. Low-grade dysplasia was diagnosed on cytology alone on two cases. Follow-up endoscopy with biopsy was performed in one patient, and low-grade dysplasia was found. Cytology using the Papanicolaou stain is not as sensitive and specific as histology for detecting intestinal metaplasia in the distal esophagus. However, it may be at least as useful as tissue sampling in detecting dysplasia.

Background & Aims: The cell of origin for Barrett's epithelium is unknown. A multilayered epithelium within Barrett's epithelium has been noted recently. To investigate the hypothesis that this multilayered epithelium may be a transitional stage between squamous and Barrett's epithelium, cytokeratin immunocytochemistry was used to examine normal squamous, Barrett's, and multilayered epithelium., Methods: Seventeen endoscopic biopsy specimens taken from the squamo-Barrett's junction of 8 patients with Barrett's epithelium and 3 biopsy specimens from the gastroesophageal junction of 3 patients without Barrett's epithelium were investigated. Antibodies to cytokeratins 4 and 13 were used as markers for squamous differentiation, and antibodies to cytokeratins 8 and 19 were used as markers for glandular differentiation. Coded samples were evaluated by immunocytochemistry., Results: In patients with Barrett's epithelium and control patients, staining with columnar markers was confined to either the Barrett's or the gastric columnar epithelium. Staining with squamous markers was confined to the adjacent squamous epithelium. In contrast, focal areas of multilayered epithelium amidst Barrett's epithelium stained with cytokeratin antibodies for both squamous and columnar epithelium., Conclusions: A focal multilayered epithelium within Barrett's epithelium that expresses concurrently both squamous and glandular cytokeratin markers is described. These findings suggest a multipotential cell as the cell of origin of Barrett's epithelium.

Pancreatic acinar metaplasia (PAM), defined as nodules of glandular tissue forming acini composed of cells with coarse apical eosinophilic granules, with or without mucous cells, was recently recognized in gastric mucosa, but its significance is not known. As part of a study on intestinal metaplasia at the gastroesophageal junction (GEJ), we evaluated the prevalence and clinical and histologic correlates of PAM in biopsy specimens from the gastroesophageal squamocolumnar junction. All adult patients having elective upper gastrointestinal endoscopy over a 6-month period were invited to participate. Clinical data and endoscopic findings were recorded. Biopsy specimens, obtained from both sides of the apparent squamocolumnar junction, were processed routinely and reviewed (without knowledge of the clinical data) to evaluate types of epithelium, types and degree of inflammation, and the presence of PAM. The presence or absence of PAM was then correlated with the other histologic findings and with the clinical and endoscopic data. The study comprised 155 patients (79 women, 76 men; 139 white patients, nine black patients, and seven patients of other ethnic groups). Their mean age was 52 years (range: 18-89 years). PAM was present in 37 patients (24%). PAM was not associated with any of the reported symptoms, endoscopic evidence of esophagitis or columnar epithelium in the distal esophagus, or any of the histologic features evaluated, including active esophagitis, intestinal metaplasia at the GEJ, active and chronic gastritis, intestinal metaplasia in the stomach, and Helicobacter infection. Although PAM is present in a considerable proportion (24%) of patients with mucosal biopsy specimens from the squamocolumnar junction, it appears to be an incidental finding unrelated to clinical or histologic abnormalities. We therefore suggest a congenital, rather than an acquired, origin for this entity.

Celiac disease (CD) is associated with marked mononuclear cell inflammation in the small intestinal mucosa. This study was performed to evaluate analogous changes in the gastric and esophageal mucosa of pediatric patients with CD, with emphasis on epithelial lymphocytosis. We evaluated intraepithelial lymphocytes (IELs) in 23 gastric (no.IELs/100 epithelial cells) and 14 esophageal mucosal biopsy specimens (IELs/hpf) from 23 pediatric cases of CD and 10 nonceliac matched controls. Four patients had postgluten withdrawal biopsy specimens reviewed, and one of these had further postgluten challenge biopsy specimens evaluated as well. Gastric specimens from the CD cases showed a significantly increased IEL count (20.5 +/- 14.4; range, 4-50) compared to controls (3.4 +/- 1.9; range, 1-8; p < 0.001), which also correlated directly with the histologic severity of the small intestinal disease as assessed by the degree of villous shortening. Sixteen (69.5%) of 23 gastric specimens showed > 8 IELs, which was the highest value obtained in control specimens. The four posttreatment specimens showed a significant reduction in the gastric IEL counts from a mean of 19.8 to 3.5 IELs/100 epithelial cells (p < 0.001). The single case that had a further postgluten challenge biopsy showed a return to the pregluten withdrawal IEL count. However, the degree of gastric intraepithelial lymphocytosis did not correlate with any of the clinical data, such as age, gender, presenting symptoms, or serum antibody levels (antigliadin, antireticulin, or antiendomysium). Furthermore, no differences were observed in the IEL count in CD esophageal specimens (5.3 +/- 2.6; range, 2-10) compared to controls (5.2 +/- 1.5; range, 3-8; p = 0.935). These findings suggest that an immune-mediated lymphocytic response linked to gluten occurs in the gastric epithelium, similar to that seen in the small intestine of pediatric patients with CD. Therefore, gastric intraepithelial lymphocytosis may represent a concurrent manifestation of CD rather than a separate entity in the pediatric population.

Barrett's esophagus is a metaplastic condition in which the normal stratified squamous epithelium of the distal esophagus is replaced by columnar epithelium. Our group has previously characterized a unique surface cell (the "distinctive cell") at the junction of squamous and Barrett's epithelium. This cell is notable for the simultaneous presence on its surface of both squamous and columnar cell features. The aims of our present study were, first, to evaluate prospectively the frequency with which Barrett's patients have the distinctive cell at the squamo-Barrett's junction; second, to further elucidate the characteristics of the distinctive cell; and third, to perform a combined morphological study of the squamo-Barrett's junction using scanning electron microscopy followed by transmission and light microscopy. We divided study patients into two groups: Group I consisted of Barrett's patients and group II of non-Barrett's control patients. Of eight group I Barrett's patients with junctional biopsies, three were noted to have the distinctive cell (37.5%). In contrast, this cell was not observed in any of the group II control patients. Biopsies in control patients as well as Barrett's patients without the distinctive cell revealed abrupt squamogastric or squamo-Barrett's junctions by scanning electron microscopy and light microscopy. In contrast, biopsies from the Barrett's patients with the distinctive cell revealed junctions that were not abrupt and had the distinctive cells overlying normal squamous epithelium. By scanning electron microscopy, the distinctive cells were flattened, polygonal cells with surface microvilli (a columnar cell feature) and were demarcated from one another by shallow depressions, or by intercellular ridges (a squamous cell feature). By transmission electron microscopy, the distinctive cells were cuboidal in shape with abundant apical microvilli and secretory vesicles. We have confirmed that distinctive cells are present in some Barrett's patients. This cell is a morphologic hybrid, sharing features of both squamous and columnar cells, and may be analogous to hybrid cells identified in other locations that undergo metaplasia (eg, the human cervix). Its origin may be the result of transformation of multipotential basal cells of squamous epithelial origin. We hypothesize that the distinctive cells may represent an intermediate stage in the development of Barrett's epithelium.

We present the case of a 4-month-old infant with a prolonged illness due to an occult intestinal stricture. Diagnosis of the lesion escaped conventional radiologic and surgical methods and required intraoperative enteroscopy. Resection of the stricture lead to prompt clinical improvement. Our patient illustrates a primary diagnostic use of intraoperative enteroscopy in the detection of a partial intraluminal small bowel obstruction. As further experience accumulates with small bowel endoscopy, roles for therapeutic enteroscopy (such as polypectomy, photocoagulation, and perhaps balloon dilatation) will certainly arise. Such endeavors will depend on the continued productive collaboration between members of the surgical and gastroenterological teams.

Hepatic microvesicular fat (MVF) has been associated with clinical syndromes of liver failure (such as acute fatty liver of pregnancy and Reye's syndrome). We performed this pilot study to investigate the specificity and potential risk factors of MVF. Sections of the liver from 17 adult autopsies were snap-frozen, and Oil Red O (ORO)-stained sections were obtained. The presence of MVF in the ORO stains was analyzed in a semiquantitative manner: 3+ = > 90% liver parenchyma with MVF; 2+ = 25-90%; 1+ = 1-24%. H&E-stained liver sections were reviewed independently to determine the presence and extent of inflammation, fibrosis, and necrosis. Autopsy reports were reviewed for the gross findings; medical records were reviewed for demographic data, medical history, and, in particular, medication usage. The study consisted of 10 males and 7 females (15 white, 2 black), with a median age of 76 years (range: 50-88). MVF was identified in 16 patients (94%); it was infrequent (1+, 2+) in the majority, but 4 patients (24%) had 3+ MVF. MVF was not associated with the age or sex of the patients, liver weight, postmortem interval to examination, or other histologic features in the liver. None of the patients, including those with 3+ MVF, had a history of liver disease. All 4 patients with extensive MVF were using salicylates, as opposed to 2 of 5 with 2+ MVF and 1 of 8 with 0 to 1+ MVF (P = 0.015). In this autopsy study, the prevalence of MVF was high, with 24% of patients having extensive MVF.(ABSTRACT TRUNCATED AT 250 WORDS)

Specialised columnar epithelium (SCE), a form of intestinal metaplasia usually found in Barrett's oesophagus, cannot be distinguished endoscopically from normal gastric epithelium. Endoscopists seldom obtain biopsy specimens from a normal-appearing gastro-oesophageal junction, and therefore short segments of SCE in this region may go unrecognised. We studied patients who had short segments of SCE at the gastro-oesophageal junction. All patients scheduled for elective endoscopic examinations in our general endoscopy unit, irrespective of indication, were questioned for symptoms of gastro-oesophageal reflux disease. At endoscopy, severity of oesophagitis was graded, and biopsy specimens obtained from the squamocolumnar junction, irrespective of its appearance or location in the oesophagus. Among 142 patients without endoscopically apparent Barrett's oesophagus, 26 (18%) were found to have SCE. All patients with SCE were white, and the male/female ratio was 1.9. In contrast, non-whites accounted for 14% of the 114 patients without SCE and the male/female ratio was 0.8. The groups did not differ significantly in the frequency of symptoms and endoscopic signs of gastrooesophageal reflux. We conclude that adults frequently have unrecognised segments of SCE at the gastro-oesophageal junction; this may underlie the rising frequency of cancer of the gastrooesophageal junction in the USA and Europe.

Primary adenocarcinoma of the esophagus, previously considered a rare neoplasm, has shown a dramatic increase in its incidence rate among White men in the United States since 1970. The reason for this increase is unknown. Since the presence of Barrett's esophagus is essential for the development of most esophageal adenocarcinomas, the increasing incidence of esophageal adenocarcinoma may be related to an increasing prevalence of Barrett's esophagus, and its precursor, gastroesophageal reflux. An association between this increasing incidence and an increasing use of pharmaceutical agents that relax the lower esophageal sphincter is proposed. The data on the dollar amount and approximate quantity in milligrams purchased per capita through retail pharmacies and hospitals in the United States from 1957 to 1986 are presented for four categories of such agents. An upward trend is observed for all four categories.

Gastric adenocarcinoma is among the most common malignancies worldwide. Its etiopathogenesis is complex and, as yet, incompletely understood; however, diet, infection with Helicobacter pylori, and genetic factors are involved. It may be classified into two main types, intestinal and diffuse. The intestinal type has decreased in incidence, whereas the diffuse tumors as well as those confined to the cardia are increasing. Numerous conditions, such as gastritis, gastric atrophy, and intestinal metaplasia (IM), are associated with intestinal type gastric cancer in retrospective studies, but only epithelial dysplasia has a positive predictive value for malignancy. These precursor conditions and lesions are analyzed for their clinicopathological significance in this review, which concludes with a brief summary of curable (early) forms of gastric cancer.

Although Helicobacter pylori (H. pylori) is considered by many to be the major cause of primary antral gastritis (PAG), several important questions concerning its pathogenetic role remain unanswered. The most basic unresolved issue relates to the low prevalence of H. pylori in children in developed countries. If H. pylori is the cause of PAG, the prevalence of PAG should also be low, but previous studies have not provided data on this issue. To answer this question, we prospectively studied 408 children who underwent esophagogastroduodenoscopy and biopsy from January 1, 1988, to December 31, 1990, for symptoms consistent with peptic disease or immunocompromise. Although the prevalence of PAG was similar (about 20%) in the four age groups of patients studied (< 5, 5-9, 10-14, and 15-20 years), the prevalence of H. pylori infections was significantly greater in the oldest age group (67%, P < 0.0001). Only four of 39 children < 10 years old with PAG had evidence of H. pylori. H. pylori is an uncommon finding in our population of young American children with PAG, indicating that it does not play an important role in the pathogenesis of this disorder in this age group.

We performed a clinicopathologic analysis of 74 periampullary and duodenal adenomas from 30 patients with familial adenomatous polyposis to evaluate the growth pattern, mucin histochemical profile, and degree and significance of Paneth cell and endocrine cell differentiation in these lesions. A trend toward more numerous adenomas was seen in patients with the longest intervals since their previous colectomies. Adenomas were most often clustered in the periampullary region and were numerous (> 20) in 40% of patients. Periampullary and duodenal adenomas were morphologically similar; both exhibited the same size-villous shape-dysplasia relationship as sporadic and familial adenomatous polyposis-associated colonic adenomas. Adenomas showed a heterogenous pattern of mucin production (mixed small intestinal-large intestinal types), with a shift toward the colonic type (sulfomucin) in lesions that were larger, villous, or severely dysplastic. Ninety-two percent of adenomas showed Paneth cell differentiation (average n Paneth cells/high power field = 24.5), and 99% showed endocrine cell differentiation (average n endocrine cells/high power field = 64.4). In general, the proportion of these specialized cell types was inversely related to size, villous shape, and degree of epithelial dysplasia. These results suggest that periampullary and duodenal adenomas in familial adenomatous polyposis arise from the neoplastic transformation of an undifferentiated stem cell that retains its ability for multidirectional differentiation.

We previously demonstrated that in esophageal mucosal biopsies, intraepithelial cells with irregular nuclear contours (CINC) as well as mononuclear cells with round nuclei are T-lymphocytes and we suggested that they may be an independent marker of esophagitis. To investigate this hypothesis, we evaluated Hematoxylin and eosin-stained biopsy specimens from 201 consecutive patients (115 female, 86 male; mean age: 52 years) that showed either no accepted features of esophagitis (n = 122) or changes typical of esophageal reflux (n = 79). The number of intraepithelial CINC and mononuclear cells in the most densely populated high power field were counted. Twenty-two of these biopsies were also stained with UCHL-1 (T-lymphocyte marker) and L-26 (B-lymphocyte marker). All medical records were reviewed to determine clinical and endoscopic findings of esophagitis, which were then compared with the histologic results for correlations. The immunohistochemical results confirmed that the majority (> 95%) of CINC and mononuclear cells were T-lymphocytes, and the sum of CINC and mononuclear cells represents best the number of T-lymphocytes. The number of T-lymphocytes was significantly correlated with that of eosinophils (R = 0.24, P = 0.0008) but not with neutrophils (R = 0.069, P = 0.33). They were greater in number, although not significantly so, in the specimens with histologic evidence of esophagitis than those without (12.7 +/- 8.9 versus 11.3 +/- 7.7, P = 0.23).(ABSTRACT TRUNCATED AT 250 WORDS)

Psychogeriatrics is the general psychiatry of old age. It is closely related with the biological and human fundamental sciences and with the medical disciplines concerned with the care of old patients. Its issues are ubiquitous and also interest non psychiatrists, particularly geriatricians, internists, neurologists and psychologists. Therefore there is a risk for research to be spread out. As a medicine of synthesis, it must assert itself as a convergence point for the related branches. Its very vast field of research is confronted with particular methodological problems, like the chronological and biological age, the heterogeneity of the population, the polymorbidity. Psychogeriatrics is young and still requires an important involvement in the clinical tasks, in the medico-social coordination and in the teaching to medical and paramedical circles. Rare are centers having the critical mass for developing good research with complete peace of mind, Switzerland is an example of this. An analysis of the 1992 international literature shows a certain scattering of the psychogeriatric themes with a 50% predominance of papers devoted to dementias. In the future, psychogeriatrics must well limit its territory, particularly by developing services where care, teaching and research are harmonized.

Classic teaching emphasizes that chronic ulcerative colitis is characterized morphologically by the presence of fixed architectural and cellular mucosal changes that categorize the process as chronic. To examine the effect of topical 5-aminosalicylic acid (5-ASA) enemas on the presence of six histological features of chronicity in established chronic ulcerative colitis, 123 mucosal biopsies were taken prospectively at 1-month intervals, all from the same anatomic location (10 cm), from 14 patients treated with either 5-ASA or placebo enemas. The biopsies were evaluated for the presence of mixed inflammation in the lamina propria, crypt architectural abnormalities, basally located lymphoid aggregates, basal plasmacytosis, villiform surface epithelial configuration, and Paneth cell metaplasia. Overall, 29% of biopsies from 64% of patients were histologically normal (no chronic features, no active disease). Compared with patients treated with placebo enemas, patients treated with 5-ASA enemas showed a significantly higher percentage of normal biopsies (36% ASA group vs. 12% placebo group; p = 0.005) and a lower percentage occurrence of each individual histological feature of chronicity. In addition, patients treated with 5-ASA had a higher average number of normal biopsies per patient (3.0) than those treated with placebo enemas (1.3). Therefore, histologically normal-appearing mucosal biopsies do occur in established cases of chronic ulcerative colitis, and this finding is enhanced by treatment with 5-ASA enemas. Awareness of these results should prevent the presence of normal rectal mucosal biopsy findings in chronic ulcerative colitis patients from being misinterpreted as either evidence against this diagnosis or as representing focal skip areas characteristic of Crohn's disease.

The pathogenesis of esophageal squamous papilloma is not known, but chronic mucosal irritation and infection with human papillomavirus are two proposed etiologies. To investigate these hypotheses, we analyzed the clinical data and histological features of 38 esophageal squamous papillomas from 33 patients and performed the polymerase chain reaction technique for detection of several common human papilloma virus types on a subset of cases (n = 26) with sufficient available material. Clinically, males were affected more often than females (M:F ratio = 24:9); average age was 50 (range, 2-86 years). Most papillomas occurred singly (85%) and were located in the distal esophagus (70%). Patients with esophageal squamous papillomas, especially those with lesions in the distal esophagus, commonly had an associated chronic and often severe form of esophageal mucosal irritation such as esophagitis or Barrett's esophagus. Esophageal squamous papillomas were small polyps (average size, 0.5 cm) that we classified histologically into three types (exophytic, 50%; endophytic, 37%; spiked, 13%) based on the predominant shape of the squamous papillae. Fifty percent of the papillomas (13 of 26) tested, from 57% of patients (12 of 21), were positive for human papilloma virus, most commonly type 16 (nine of 13), less often type 16 and 18 together (3/13), and rarely type 6b/11 (1 of 13). We propose a multifactorial etiology in which the synergistic action of mucosal irritation and human papilloma virus may be necessary for the development of esophageal squamous papillomas.

Allergic proctocolitis is a major cause of rectal bleeding in infants, but its clinical features and laboratory results are often nonspecific. Our previous retrospective study demonstrated that large numbers of eosinophils in colonic mucosal biopsy specimens were highly associated with cases of allergic proctocolitis. Therefore, we prospectively examined 60 colonic mucosal biopsy specimens from the same sites (4, 8, and 12 cm from the anal verge) in 20 infants with clinically confirmed allergic proctocolitis to validate this morphologic feature, to characterize its distribution, and to correlate these data with the clinical information. The patients (age range, 4 to 304 days) were fed breast milk or a variety of formulas and all presented with rectal bleeding. Sigmoidoscopic examination was abnormal in 19 cases, typically characterized by focal areas of mucosal erythema. The major histologic finding was a strikingly focal increase in the number of eosinophils in all mucosal compartments, with a predilection to aggregate in close association with lymphoid nodules. Eosinophilic infiltration varied not only between biopsies at different sites, but also within individual biopsy specimens. Only 12 of 20 patients (60%) had all three of their biopsy specimens categorized as abnormal; in the remainder, only one (four patients) or two (four patients) of the three biopsy specimens were abnormal. The average number of eosinophils per high-power field of lamina propria for all cases was 15.6. No significant correlation was identified between the number of eosinophils in the mucosa and the patient's age, length of illness, endoscopic appearance, or type of inciting formula. In summary, eosinophils appear to be an excellent marker for infantile allergic proctocolitis. Given the focal distribution of the eosinophils, multiple mucosal biopsy specimens should be obtained and several levels of each examined to confirm the diagnosis.

Objective: Axonal necrosis was first described in samples of small intestine from patients with Crohn's disease (A.M. Dvorak et al. Hum Pathol 1980; 11:620-634). Clinically evident inflammation of continent ileal reservoirs (pouches) has clinical features that resemble Crohn's disease. Possible similarities in the pathogenesis of Crohn's disease and pouchitis were sought using ultrastructural and microbiologic tools to identify damaged enteric nerves and tissue bacteria., Methods: An encoded ultrastructural and microbiologic study of replicate biopsies from 114 samples of human intestine was done. Biopsies from ileum, colon, conventional ileostomy or continent pouch were obtained from patients with ulcerative colitis, Crohn's disease, or familial polyposis and grouped into three clinical study groups (control, normal pouch, pouchitis), based on clinical and endoscopic criteria. Biopsies were prepared for electron microscopy with standard methods; replicate biopsy samples were washed extensively before preparing cultures designed to identify aerobic as well as facultative and obligate anaerobic bacteria (Onderdonk et al. J Clin Microbiol 1992; 30:312-317). The ultrastructural diagnosis of damaged enteric nerves was based on previously published criteria for axonal necrosis (A.M. Dvorak and W. Silen. Ann Surg 1985; 201:53-63). Intergroup comparisons were tested for significance using Chi-square analysis., Results: The highest incidence of axonal necrosis was present in Crohn's disease control biopsies (53%), regardless of whether bacteria were present (or not) in cultures of replicate biopsies. Axonal necrosis also occurred in more ulcerative colitis and familial polyposis biopsies (regardless of biopsy site) that had positive bacterial cultures than in those that did not (p < 0.001). In addition, axonal necrosis was documented in 42% of the pouch biopsies from ulcerative colitis and familial polyposis patients, particularly in those pouches that were found to be inflamed by clinical criteria and that also had positive bacterial cultures of the biopsied tissues. Control biopsies from patients with ulcerative colitis and familial polyposis had significantly less nerve damage than pouch biopsies in the presence of positive cultures (p < 0.01). Among the clinically inflamed pouches biopsied in ulcerative colitis or familial polyposis patients, we found that none had damaged enteric nerves when bacterial cultures were negative (p < 0.005). If the presence of axonal necrosis alone was compared with the presence of undamaged enteric nerves in all biopsies from patients with ulcerative colitis, a highly significant number of ulcerative colitis biopsies with axonal necrosis occurred in pouches (72%) compared with controls (p < 0.001)., Conclusions: The ultrastructural finding of axonal necrosis in Crohn's disease confirms previous studies. The presence of damaged enteric nerves in patients with pouchitis provides ultrastructural support to the clinical impression of similarities between pouchitis and Crohn's disease. The association of damaged nerves and invasive bacteria in pouchitis suggests mechanistic similarities for the pathogenesis of Crohn's disease that requires further investigation.

Metaplastic columnar epithelium replaces the normal squamous epithelium in Barrett's esophagus. We characterized the surface epithelial cells of the junction between squamous and Barrett's epithelium using scanning electron microscopy and light microscopy. In four biopsy specimens from the squamous-Barrett's junction in three patients, we found a distinctive cell type having features intermediate between those of squamous and columnar epithelium. Its distinguishing characteristic is the presence on its surface of two disparate structures not normally present on the same cell in the gastrointestinal tract: microvilli (a scanning electron microscopy feature of glandular epithelium) and intercellular ridges (a scanning electron microscopy feature of squamous mucosa). The surface characteristics of this newly recognized cell were strikingly similar to those of cells found in the transformation zone of the uterine cervix, an area in which squamous epithelium physiologically replaces columnar epithelium. We also examined 28 biopsies of the gastroesophageal junction area from 14 patients with and without a history of heartburn but with no evidence of Barrett's esophagus. None of these biopsies showed the distinctive cell. We hypothesize that this distinctive cell represents an intermediate step in either the development or the healing of Barrett's epithelium, during which surface characteristics of two different cell types, columnar and squamous, coexist on the same cell.

Tubal metaplasia of the endocervix (TME), a condition that may be confused morphologically with glandular neoplasia, is frequently found in cone or hysterectomy specimens. To determine the frequency of detecting TME in cytologic smears, we retrospectively reviewed 28 Papanicolaou (Pap) smears from 22 women (mean age 39.1 yr; range 25-60 yr) with histologically proven TME. Our criteria for TME were the presence of two cell types in addition to endocervical secretory cells, i.e., peg cells (cells with dark and granular cytoplasm and elongate nuclei) and ciliated cells. All women had cervical cytology specimens obtained with an endocervical brush shortly before the procedures in which TME was diagnosed, and five also had at least one post-procedure smear. Of 20 smears with an adequate, non-neoplastic endocervical component, TME was found in 2 (10%). In these two, TME cells constituted 10% and < 5% of all the glandular cells, respectively, and the percentage of ciliated cells in the TME was approximately 25% and 75%. In conclusion, TME was noted infrequently (10%) on the cervical cytosmears of women with histologically-proven TME. This result corresponds to the histologic finding that TME typically involves the upper endocervix and glandular epithelium, with only 13% of the women having TME on the surface of the lower endocervix. Atypical glandular cells on cervical cytology are a problem for clinicians and pathologists alike. The differential diagnosis of such atypia, including TME, cells of the lower uterine segment, squamous intraepithelial lesion in glands and glandular neoplasia, is discussed.

Twenty-two percent of 117 biopsies of human intestinal tissues had ultrastructural images of classical regulated secretion from eosinophils in vivo i.e. eosinophil granule extrusion (EGE). Replicate intestinal biopsies that were positive for bacteria had EGE more often than not (p < 0.05); 77% of the isolates were Staphylococci. Some of the intestinal biopsies also had damaged nerves; all that had EGE and damaged enteric nerves also had positive bacterial cultures. The EGE that we observed could not account for all enteric nerve damage, suggesting multifactorial mechanisms for nerve damage in gut tissues. Among the possibilities are release of neurotoxic eosinophil granule proteins by an alternate secretory route, i.e., piecemeal degranulation, direct toxicity of tissue invasive bacteria and/or damaged nerves of unknown etiology such as those that are regularly present in uninvolved tissues of patients with Crohn's disease.

The etiology of ulcerative colitis (UC) and Crohn's disease (CD) remains enigmatic. Infiltrating intestinal macrophages are capable of producing the proinflammatory cytokines tumor necrosis factor-alpha (TNF-alpha), interleukin-1 beta (IL-1 beta), and interleukin-6 (IL-6). We investigated the presence of IL-6, TNF-alpha and IL-1 beta mRNA transcripts in inflammatory bowel disease (IBD), normal, and other inflammatory intestinal specimens utilizing the polymerase chain reaction (PCR). TNF-alpha mRNA levels did not very between inflammatory bowel disease and control specimens. IL-1 beta mRNA levels were highest in active UC and noninflammatory bowel disease inflammatory specimens while IL-6 mRNA levels were highest in active IBD specimens. Infiltrating T cells, macrophages, and B cells were identified as sources of IL-6 protein in inflammatory bowel disease specimens by immunofluorescent staining. IL-6 transcripts were elevated only in active inflammatory bowel disease specimens, suggesting that IL-6-mediated immune processes are ongoing in the inflammatory mucosal environment of CD and UC.

Esophageal mucosal biopsies from patients with upper gastrointestinal symptoms often contain intraepithelial cells with irregular nuclear contours (CINC). To investigate their nature and significance, we retrospectively studied biopsies and resection specimens from symptomatic patients that showed the following histologies: unremarkable mucosa (11 cases), active esophagitis (11 cases), infection (herpes: two cases; fungus: five cases), Barrett's esophagus (five cases), adenocarcinoma (three cases), and squamous cell carcinoma (two cases). H&E sections were reviewed to determine the number of intraepithelial CINC, neutrophils, eosinophils, and mononuclear cells with small, round nuclei typical of lymphocytes in five high-power fields. All specimens were also stained for S-100 protein, and six additional biopsies with numerous CINC were stained for leukocyte common antigen (LCA), L-26 (a B-cell marker), and Leu-22 and UCHL-1 (both T-cell markers). Clinical charts were reviewed to analyze endoscopic findings. The CINC had the immunohistochemical profile of T-lymphocytes: LCA (+), Leu-22 (+), UCHL-1 (+), L-26 (-), and S-100 protein (-). Their average number per high-power field was low in benign squamous epithelium away from tumors (average 0.8/high-power field) but much higher in biopsies with the other categories of histology, especially those with active esophagitis and infection (6.3/high-power field). The difference between the mean of the two tumor groups (0.8) and that of esophagitis (6.3) was statistically significant (p = 0.005). Of 11 patients with unremarkable biopsies, there was endoscopic evidence of esophagitis in five. The biopsies of these five showed an average number of CINC of 6.4, whereas the number was only 3.5 in the six with normal endoscopy (p = 0.15).(ABSTRACT TRUNCATED AT 250 WORDS)

Tubal metaplasia (TM) of endocervical epithelium may be confused with endocervical dysplasia or adenocarcinoma in histologic or cytologic specimens. However, the potential magnitude of this problem is unknown, because the prevalence of endocervical TM in routine surgical pathology material is undefined. To determine the prevalence, distribution, and associated features of endocervical TM, we retrospectively reviewed a consecutive series of cone biopsy specimens (n = 82) and hysterectomy specimens in which the entire cervix had been sectioned (n = 26), obtained over 28-months. Tubal metaplasia was defined as epithelium containing all three cell types (ciliated, secretory, and intercalary) found in the normal fallopian tube. The mean patient age was 41 years (range, 21-79). Endocervical TM was present in 33 of 108 patients (31%) and was evenly distributed among all age groups. Its prevalence was related to the number of sections examined (23% of cases with less than or equal to 12 blocks of the cervix; 52% of cases with greater than 12 blocks; p less than 0.01) and was greater in hysterectomy than in cone specimens (62 vs. 21%, p less than 0.001). Although most frequent in the upper endocervix and in deep portions of glands, TM involved the surface in 36%, the superficial parts of glands in 64%, and the lower endocervix in 30% of positive cases. There was no association with phase of the menstrual cycle, inflammatory changes, or low-grade cervical intraepithelial neoplasia (CIN), but TM was inversely related to high-grade CIN in glands, presumably due to replacement of metaplastic cells by neoplastic cells.(ABSTRACT TRUNCATED AT 250 WORDS)

One hundred and seventeen coded intestinal biopsies were examined by electron microscopy and evaluated for morphological evidence of mast cell and basophil secretion in situ. Sixty percent of the biopsies had evidence of secretion. Mast cell secretion was evident in control biopsies, many of which were obtained from uninvolved tissues of patients with inflammatory bowel disease. Biopsies of inflamed continent pouches from ulcerative colitis (UC) patients showed more mast cell secretion than noninflamed UC pouch biopsies. This evidence of mast cell secretion supports recent work that documents high constitutive levels of histamine in jejunal fluids of Crohn's disease patients and suggests a proinflammatory role for mast cells in inflammation associated with pouchitis.

One hundred and seventeen coded intestinal biopsy specimens were examined by electron microscopy. All surgical biopsies were obtained from uninvolved sites of patients with two inflammatory bowel diseases (ulcerative colitis or Crohn's disease) and from patients with preneoplastic and neoplastic diseases (adenocarcinoma, rectal polyp, familial polyposis). Biopsy sites included normal ileum, colon, and rectum as well as conventional ileostomies and continent pouches constructed from the ileum. The data reported here describe the ultrastructural anatomy of human gastrointestinal tract mucosal mast cells in vivo and their anatomic associations with enteric nerves.

Although the main risk factors for atherosclerosis have been clearly identified and their correction has demonstrated its efficacy, many subjects do not receive the corrective measures appropriate for their case. The authors describe a hospital service for the prevention of atherosclerosis, which is intended for both primary and secondary prevention as well as for epidemiological monitoring and above all the evaluation of the preventive strategies employed. The results after the service has been running for one year are reported. They appear to confirm the initial options. Cooperation with other people involved in prevention both inside and outside the hospital, is also a fundamental aspect of this task.

In the last 10 years, new observations have been made of the biologic behavior of clear-cell adenocarcinoma of the cervix and vagina arising in young women exposed to diethylstilbestrol in utero. Of particular note is the tumor's capacity to recur after an extended disease-free interval following initial therapy. We report the case of a woman who had her first recurrence 17 years after initial therapy, presenting with metastatic disease to the lungs and cerebellum. This case represents the longest reported interval between primary therapy and recurrence and supports the conclusion made by others that women who have been treated for clear-cell adenocarcinoma of the cervix and vagina may remain at risk of disease progression for many years after initial therapy. Therefore, these women should continue to be monitored. We also recommend periodic chest x-rays for two reasons: 1) The lungs are the most common site of distant spread of disease; and 2) metastatic nodules may be amenable to curative surgical intervention.

To evaluate the diagnostic utility of rectal mucosal biopsies in infants with proctocolitis, we compared the clinical and histologic features of an allergy-related group (N = 36) with those of normal (N = 12) and inflammatory (N = 8) controls. Clinical features were nondiscriminatory among the three groups of patients, except for an increased absolute peripheral eosinophil count in the allergic group. Similarly, morphologic evidence of proctitis (cryptitis and crypt abscesses) and small or moderate numbers of eosinophils (less than or equal to 60 per ten high power fields) in the lamina propria of the biopsies did not discriminate among the three groups. However, large numbers of eosinophils (greater than 60 per ten high power fields) and eosinophils located in the muscularis mucosae or as the predominant cell in crypt abscesses were significantly associated with allergy-related disease. No histologic features of chronic colitis were noted in the allergy-related group. Thus, in tandem with the remainder of the clinical data, rectal mucosal biopsy is a useful adjunct in the diagnosis of allergic proctocolitis.

Involucrin, a protein subunit of keratinocyte cross-linked envelopes, is a distinctive marker for suprabasal differentiation in stratified squamous epithelium. Immunoperoxidase staining for involucrin was used to evaluate paraffin sections of tissue obtained by colposcopically directed biopsies of infectious, metaplastic, and dysplastic lesions of the cervix and vagina. Areas of normal squamous epithelium, papillary and flat condyloma acuminatum, and mature and immature squamous metaplasia showed positive staining in 99 per cent of samples lacking significant inflammation and in 60 per cent of those with moderate or severe inflammation. In contrast, only 19 per cent of the squamous cell dysplasias, even those without much inflammation, showed positive staining, and no area with moderate or severe inflammation showed positive staining. These findings indicate that expression of involucrin is modulated by cellular pathologic features and microenvironment. We suggest that immunoperoxidase staining for involucrin may be useful in distinguishing mild dysplasia from immature metaplasia and flat condyloma in some biopsy specimens in which routine histologic examination yields an indeterminate diagnosis.

A 78-year-old woman with extramammary Paget's disease presented with vaginal bleeding and vulvar pruritus. Paget's cells were demonstrated initially in a cervical cytosmear and, subsequently, in the epithelia of the vagina and cervix, in addition to those of the vulva, periurethral area, and perineum. Such extensive intraepithelial spread of Paget's disease in the genital tract has rarely been reported previously. This is the first case, to our knowledge, in which the cells of cervicovaginal Paget's disease were detected by Papanicolaou smear.

Two hundred and fifty patients were examined because of a history of in utero exposure to diethylstilbestrol (DES) or because of the presence of physical findings suggesting such exposure. One thousand biopsies were examined for the presence of neoplasia and then compared to the colposcopic findings. There were no cases of glandular or squamous cell carcinoma. Fifteen (6 per cent) of the patients had squamous cell dysplasia. The degree of dysplasia was mild in 11 and moderate in only 4 (1.6 per cent) of the women. The majority of the cases of dysplasia involved the cervix, whereas the vagina was involved in only four cases, with simultaneous cervical dysplasia in three of these. Patients with cervical mosaic and white epithelium had dysplasia on biopsy much more frequently as compared with patients with similar colposcopic appearances in the vagina. Our results suggest a low incidence of significant squamous precancerous change in the DES-exposed population and provide evidence that colposcopic data concerning dysplasia pertinent to the cervix cannot be applied without modification to the evaluation of dysplasia in vaginal adenosis.

Patients with distal ulcerative colitis may be refractory to standard drug therapy. 5-Aminosalicylate, the active moiety of sulfasalazine, is effective topically in new onset distal colitis. Six patients with distal colitis unresponsive to sulfasalazine and corticosteroids had a prompt and dramatic response to the administration of 5-aminosalicylate enemas with clinical, sigmoidoscopic, and histological improvement. Topical 5-aminosalicylate should be a useful addition to the current available therapies for some patients with refractory distal ulcerative colitis.

A case of barbiturate-induced submassive hepatic necrosis is presented and the literature is reviewed. The patient's clinical presentation, histologic features on liver biopsy and favorable course after stopping the drug suggest that barbiturates can be added to the list of agents which can cause submassive hepatic necrosis.

Biopsy-proven vaginal adenosis was found in 88% of a group of 100 asymptomatic young women with known or suspected to have histories of intrautering diethylstilbestrol (DES) exposure. Histologic features were roughly quantitated and were found to vary markedly from patient to patient, but the following proved to be the most consistent features: (1) variation in glandular configuration (simple, complex, cystic); (2) the presence of epithelium recapitulating all parts of the müllerian duct system (endocervical, endometrial and tubal); (3) limited involvement of the vaginal mucosa (no surface involvemtn in 23.8% of the patients); (4) extensive squamous metaplasia; (5) submucosal inflammation. There was no glandular dysplasia or coexistent adenocarcinoma. The squamous epithelium in many specimens contained little or no glycogen, and in some manifested cellular activity consistent with a regenerative response to inflammation. Sufficient atypicality to implicate vaginal adenosis as a premalignant lesion was not found in any biopsy specimen.

The potential value of biopsy surveillance of patients with Barrett's esophagus for dysplasia is diminished by a lack of agreement on the diagnostic criteria for dysplasia. In a preliminary consensus conference, experienced gastrointestinal pathologists from four medical centers agreed on criteria for a five-tiered histologic classification of dysplasia in Barrett's esophagus--negative for dysplasia, indefinite for dysplasia, low-grade dysplasia, high-grade dysplasia, and intramucosal carcinoma. Eight morphologists in the four centers tested the criteria for interobserver agreement by examining a set of coded slides that had been chosen to include some especially difficult interpretative problems in all five histologic classifications. Interobserver agreement of 85 and 87% was achieved in successive reviews when the combined group of high-grade dysplasia and intramucosal carcinoma was compared with the combined group of low-grade dysplasia, indefinite for dysplasia, and negative for dysplasia. Comparison of other groups yielded less agreement. For example, negative for dysplasia could be distinguished from all other diagnoses with an interobserver agreement of 72%. We conclude that experienced gastrointestinal morphologists can diagnose high-grade dysplasia and intramucosal carcinoma with a high degree of agreement and thus can detect those patients who may need immediate rebiopsy or esophageal resection. Either further refinement of histologic criteria or alternate diagnostic methods will be needed to achieve the reproducible diagnosis of indefinite changes and low-grade dysplasia. This is important because patients with such changes theoretically merit closer endoscopic surveillance.