Despite increasing interest during the past few decades,

Transcription

1 Review Article Putative Precursors of Gallbladder Dysplasia A Review of 400 Routinely Resected Specimens Sanjay Mukhopadhyay, MD; Steve K. Landas, MD Context. Dysplasia is thought to be a precursor of invasive gallbladder carcinoma, but it is unsettled whether dysplasia arises from other precursor lesions. Objective. To ascertain the presence and nature of precursors of dysplasia in the gallbladder. Design. Four hundred consecutive cholecystectomy specimens were processed and stained routinely for diagnosis. We retrospectively reviewed these cases to look for the presence of epithelial changes, including antral-type metaplasia, intestinal metaplasia, and dysplasia. Results. Antral-type metaplasia, intestinal metaplasia, and dysplasia were found in 238 (59.5%), 39 (9.8%), and 20 (5.0%) cases, respectively. The mean patient age was 47.7 years (range, years). The mean ages for patients with antral-type metaplasia, intestinal metaplasia, and dysplasia were 49.4, 50.9, and 52.6 years, respectively. Statistically significant associations were found between antraltype metaplasia and intestinal metaplasia (P.007, 2 test) and between intestinal metaplasia and dysplasia (P.001, 2 test). Conclusion. These associations, along with the age gradient from antral-type metaplasia to dysplasia, suggest a progression from antral-type metaplasia to dysplasia via intestinal metaplasia. (Arch Pathol Lab Med. 2005;129: ) Despite increasing interest during the past few decades, the precise relationship between gallbladder cancer and its precursors remains nebulous. In other organs, well-defined entities such as cervical intraepithelial neoplasia and prostatic intraepithelial neoplasia have been delineated, and a well-defined progression from low-grade to high-grade lesions is widely accepted. Although previously proposed, such a sequence of progression for gallbladder carcinoma is neither clearly defined nor widely appreciated. Although a definite epidemiologic parallel exists between gallbladder carcinoma and cholelithiasis, the pathogenetic relationship remains obscure. 1 Several studies 2 8 have shown that invasive carcinoma of the gallbladder is preceded by dysplasia, but the sequence of progression from cholelithiasis to dysplasia is less well delineated. A few investigators have suggested that antral-type or intestinal metaplasia may be the pathogenetic link between cholelithiasis and dysplasia. 2,5,8,9 The objectives of this study were to examine routinely resected gallbladders to ascertain the frequency of antraltype metaplasia, intestinal metaplasia, and dysplasia (individually and in combination); to determine the age distribution among patients with these lesions; and to delineate a sequence of progression of precancerous changes in the gallbladder. Accepted for publication October 20, From the Department of Pathology, State University of New York Upstate Medical University, Syracuse. The authors have no relevant financial interest in the products or companies described in this article. Reprints: Steve K. Landas, MD, Department of Pathology, State University of New York Upstate Medical University, 750 E Adams St, Syracuse, NY ( landas@upstate.edu). MATERIALS AND METHODS The study was performed on 400 routinely resected gallbladders retrieved from the surgical pathology archives of the State University of New York Upstate Medical University. Each gallbladder was processed in paraffin and stained with hematoxylineosin. In most cases (359/400 cases), only one block was submitted for histologic examination (3 standard sections per block). Multiple blocks were submitted in the remaining 41 cases (2 blocks in 24 cases, 3 blocks in 12 cases, and 4 blocks in 5 cases), depending on the presence of unusual gross abnormalities or the need for further sampling based on microscopic findings. The presence of the following lesions was recorded: (1) antral-type (pseudopyloric) metaplasia, according to criteria proposed by Laitio 10 ; (2) intestinal metaplasia, using the presence of goblet cells as a marker; and (3) dysplasia, as defined by Albores-Saavedra et al. 2 These authors defined dysplasia (or atypical hyperplasia ) as loss of architecture, disorganization of the epithelium, and nuclear atypia when occurring on a background of pseudostratification, nuclear crowding, increased height of columnar cells, and occasional normal mitotic figures. We did not divide dysplasia into low and high grades. Although some lesions were subtle, no difficulty was encountered in differentiating the sharply demarcated lesions of dysplasia from the more diffuse lesions of reactive atypia. Histologic examination was performed without knowledge of the age of the patients. Consecutive specimens were reviewed, with extensively autolyzed gallbladders being excluded during the accrual of our 400 cases. Statistical analysis of association of variables was done using the 2 test. The study was performed in accord with the guidelines of the institutional review board for protection of human subjects. RESULTS Pathologic Features Antral-type metaplasia (Figure 1) was manifested in the form of branching glands composed of columnar cells with abundant pale cytoplasmic mucin compressing the nucleus to the abluminal aspect of the cell. The resultant 386 Arch Pathol Lab Med Vol 129, March 2005 Putative Precursors of Gallbladder Disease Mukhopadhyay & Landas

2 Figure 1. Antral-type metaplasia. Normal surface epithelium is to the right (hematoxylin-eosin, original magnification 100). Figure 2. Intestinal metaplasia. Note conspicuous goblet cells (hematoxylin-eosin, original magnification 100). Figure 3. Dysplasia. Note presence of mitotic figures (arrows) and relative scarcity of inflammatory cells (hematoxylin-eosin, original magnification 400). Figure 4. Reactive atypia. Inflammatory granulation tissue is present in the subepithelial region (hematoxylin-eosin, original magnification 400). Figure 5. Antral-type metaplasia (short arrow) and intestinal metaplasia (long arrow) in close proximity in the same specimen (hematoxylin-eosin, original magnification 100). Figure 6. Intestinal metaplasia adjacent to a segment of dysplastic epithelium. Goblet cells identify areas of intestinal-type epithelium (arrows) (hematoxylin-eosin, original magnification 100). Arch Pathol Lab Med Vol 129, March 2005 Putative Precursors of Gallbladder Disease Mukhopadhyay & Landas 387

3 Table 1. Number and Age Distribution of Patients With Various Epithelial Changes No. of Patients Epithelial Change (N 400)* None 152 (38.0) Antral-type metaplasia 238 (59.5) Intestinal metaplasia 39 (9.8) Dysplasia 20 (5.0) * Data are given as number (percentage) of patients. Age, Mean (Range), y 45.6 (15 90) 49.4 (17 93) 50.9 (19 93) 52.6 (19 82) Table 3. Intestinal Metaplasia Association of Intestinal Metaplasia and Dysplasia* Dysplasia Absent Present Total Absent Present Total * A statistically significant association was found between intestinal metaplasia and dysplasia (2-tailed P.001, 2 1 test). Table 2. Antral-Type Metaplasia Association of Antral-Type Metaplasia and Intestinal Metaplasia* Intestinal Metaplasia Absent Present Total Absent Present Total * A statistically significant association was found between antral-type metaplasia and intestinal metaplasia (2-tailed P.007, 2 1 test). morphologic appearance was identical to that of epithelial cells of the gastric pyloric antrum. Glands similar in appearance to antral-type metaplasia but with flatter epithelial cells, wider lumina, and less cytoplasmic mucin were regarded as mucous neck glands Intestinal metaplasia (Figure 2) appeared as patchy zones of sharply delimited intestinal-type epithelium, alternating with gallbladder-type columnar cells. Goblet cells and enterocytes were the predominant intestinal cell types, although in some cases Paneth cells were also observed. Dysplastic epithelium (Figure 3) was seen as sharply demarcated foci of variable mitotic activity and cytologically atypical nuclei, with the degree of atypia being greatly out of proportion to any regional inflammation. In contrast, epithelium showing reactive atypia (Figure 4) lacked sharp demarcation from adjacent normal epithelium and was generally observed in the vicinity of ulceration or intense inflammation. The sharp circumscription of dysplasia was a helpful diagnostic feature. The patients age distribution is shown in Table 1. Epithelial lesions were found in 248 (62.0%) of 400 cases. Of these, 208 showed only one type of epithelial lesion (antral-type metaplasia in 200, intestinal metaplasia in 6, and dysplasia in 2). Of the remaining 40 cases, 31 showed 2 epithelial changes in the same specimen. Specifically, 22 cases showed both antral-type and intestinal metaplasia (Figure 5), 7 showed both antral-type metaplasia and dysplasia, and 2 showed both intestinal metaplasia and dysplasia (Figure 6). The remaining 9 cases showed the coexistence of antral-type metaplasia, intestinal metaplasia, and dysplasia in the same specimen. In 152 cases (38.0%), neither metaplasia (antral-type or intestinal) nor dysplasia was found. There was only one case with invasive carcinoma. The scant residual epithelium in this case showed foci of antral-type metaplasia. Intestinal metaplasia and dysplasia were not identified. Statistically significant associations were found between antral-type metaplasia and intestinal metaplasia (Table 2) and between intestinal metaplasia and dysplasia (Table 3). These associations were reflected in the anatomic relationships of the changes observed. In most cases, antral-type Table 4. Histologic Continuity or Blending Observed in Cases With Multiple Epithelial Changes in the Same Section* Epithelial Changes Observed AM, IM, D IM, D AM, D AM, IM 388 Arch Pathol Lab Med Vol 129, March 2005 Putative Precursors of Gallbladder Disease Mukhopadhyay & Landas No. of Cases Epithelial Changes in Direct Continuity With Each Other AM with IM, IM with D, and AM with D IM with D only IM with D AM with D No continuity of changes AM with IM No continuity of changes No. of Cases * AM indicates antral-type metaplasia; IM, intestinal metaplasia; and D, dysplasia. metaplasia was continuous with intestinal metaplasia, and intestinal metaplasia, in turn, was continuous with dysplasia (Table 4). On the other hand, a statistically significant association between antral-type metaplasia and dysplasia was not present (P.06). This lack of association was also substantiated by histologic examination, because in most cases (6/7 cases) in which antral-type metaplasia and dysplasia were the only changes seen, they were not seen in contiguity with each other (Table 4). The distribution of the various epithelial changes did not follow any pattern with regard to occurrence in the neck, body, or fundus of the gallbladder. Clinical Findings The patients ages ranged from 15 to 93 years (mean, 47.7 years; median, 46 years). The mean ages showed a progression that correlated with advancement of the histologic lesion from reactive to preneoplastic (Table 1). The youngest mean age (45.6 years) was seen in patients with no epithelial changes. The mean ages of the patients increased according to type of lesion, from antral-type metaplasia (49.4 years) to intestinal metaplasia (50.9 years) to dysplasia (52.6 years). However, these differences were not statistically significant. COMMENT Our findings, based on a large number of routinely resected gallbladders, indicate that dysplasia in gallbladder epithelium is strongly associated with intestinal metaplasia, which in turn is statistically associated with antraltype metaplasia. We also demonstrate an age gradient progressing from antral-type metaplasia (youngest mean age) to intestinal metaplasia (intermediate mean age) to dysplasia (oldest mean age). These findings suggest that antral-type metaplasia and intestinal metaplasia are precursors of dysplasia in the gallbladder.

4 Several investigators have studied the incidence of dysplasia in gallbladder epithelium. 2,8,9 Albores-Saavedra et al 2 studied 200 consecutive cholecystectomy specimens from Mexican patients and found dysplasia in 13.5%. Duarte et al 9 reported a similar figure (13.6%) in a study from Chile. In contrast, a study from Canada found only one case of dysplasia (0.4%) among 277 cholecystectomy specimens. 4 In our series, the incidence of dysplasia (5.0%) was similarly low. These differences in the incidence of dysplasia may be explained by a combination of reasons, including extent of sampling, varying definitions of precursor lesions, and, perhaps most important, geographic or racial differences in the incidence of cancer of the gallbladder and its precursors. The incidence of gallbladder carcinoma is highest in Latin America, high among Native Americans, intermediate in Japan, and lowest among white and black Americans. 14 This distribution may explain the lower figures obtained by studies of North American populations (Dowling and Kelly 4 and the present study) compared with figures among Latin American patients. 2,9 Similarly, the mean age of individuals with dysplasia in our series (52.6 years) differs from that reported from Chile (46.3 years). 15 This difference is presumably due to regional differences in genetic susceptibility to neoplasia or environmental factors that affect the incidence of gallstones, chronic cholecystitis, and carcinoma. Our finding of an association between intestinal metaplasia and dysplasia is consonant with observations made by other investigators. 8,9 In a Chilean study of 162 extensively sampled gallbladders, Duarte et al 9 demonstrated a significant association between intestinal metaplasia and dysplasia. A similar association was found in a Japanese study of 1000 resected gallbladders. 8 Several studies, mostly from Japan, have demonstrated similarities in the histochemical profiles of antral-type metaplasia, intestinal metaplasia, and carcinoma, supporting the theory that gallbladder carcinoma may be derived from metaplastic epithelium. In a study of 22 surgically removed adenocarcinomas of the gallbladder, Kushima et al 18 used antibodies specific to gastric and intestinal mucins, as well as chromogranin A and pepsinogen II, to determine whether gallbladder cancers showed evidence of gastric or intestinal differentiation. Evidence of gastric or intestinal differentiation was found in 12 of 16 tumors. Sasaki et al 20 used immunohistochemical staining for MUC2, MUC5AC, and MUC6 apomucins on cholecystectomy specimens from patients with carcinoma (55 patients), dysplasia (20 patients), and metaplasias (15 patients). Staining for all 3 apomucins was seen in most carcinomas, as well as in most metaplasias and dysplasias. Mucin glycoprotein antigen 6, for example, was expressed in 100% of cases with pseudopyloric (antral-type) metaplasia, 91% of dysplasias, and 90% of carcinomas. Additional support for gastric intestinal derivation of gallbladder cancer comes from reports of gastric or intestinal differentiation in dysplastic gallbladder epithelium, in the mucosa adjacent to invasive cancer, and within neoplastic cells in gallbladder carcinoma. 17,23 25 In an attempt to trace the pathogenetic sequence back to its origin, some investigators have attempted to demonstrate that metaplastic changes of gallbladder epithelium are associated with gallstones. In an Indian study 26 involving 150 cholecystectomies, the prevalence of gallstones was higher in gallbladders with metaplastic and dysplastic changes. In a Japanese study, 8 the incidence of intestinal metaplasia and antral-type metaplasia showed a close relationship to the presence of gallstones and inflammation. The demonstration of an association is not, in itself, proof of causality. A hypothesized pathogenetic connection between lesions is strengthened if such an association has biologic plausibility. Antral-type and intestinal metaplasia may represent local responses to chronic injury secondary to the presence of gallstones and inflammation. Although the exact relationship between intestinal differentiation and gallbladder carcinoma is unknown, it is well known that metaplastic epithelium is more susceptible to malignant transformation than normal epithelium. 9,27 If the process of intestinal metaplasia in the gallbladder is analogous to that known to occur in the stomach, 28,29 it is reasonable to consider the possibility that intestinal metaplasia also predisposes to gallbladder carcinoma. 30 The route for the development of invasive carcinoma of the gallbladder would then be the following: gallstones chronic inflammation antral-type metaplasia intestinal metaplasia dysplasia carcinoma. If metaplasia indeed progresses to dysplasia, patients with metaplasia should, on average, be younger than those with dysplasia. Our results demonstrate such an age gradient, increasing from patients with normal mucosa to those with antral-type metaplasia, intestinal metaplasia, and dysplasia. However, the time required for progression from antral-type metaplasia to dysplasia (approximately 3 years) seems short compared with the long interval that has been shown by other investigators for gallbladder dysplasia to progress to invasive carcinoma (approximately years). 15,31 It is possible that antral-type metaplasia progresses rapidly to intestinal metaplasia and dysplasia, with the subsequent development of invasion being a far slower process. We must also add the caveat that, although the mean ages of the various groups were different, these differences did not attain statistical significance. An alternative explanation for the observed coexistence of antraltype metaplasia, intestinal metaplasia, and dysplasia could be that these lesions are manifestations of inflammation, not following an orderly sequence of progression. If this were the case, dysplasia should have been equally frequent regardless of the presence or absence of the other lesions. This was not the case. Of the 20 cases of dysplasia, a coexistent metaplastic lesion was seen in 18 cases, suggesting that dysplasia arises from precursor lesions, not directly from an inflammatory background. Similarly, of 33 cases of intestinal metaplasia, this lesion was seen in isolation in only 6 cases. In contrast, the first lesion in the sequence we propose (antral-type metaplasia) occurred much more frequently in isolation (200/238 cases) than in combination with another lesion (38/238 cases), which is expected if antral-type metaplasia is indeed the basis for the subsequent lesions. This study was designed to detect (or refute) a statistical association between dysplasia and its putative precursor lesions, not between dysplasia and carcinoma. Therefore, we did not attempt to address the issue of whether dysplasia is a precursor to invasive carcinoma, a question that has been answered in the affirmative by several other investigators. 2 4,6 8 Another issue requiring comment is the use of routine histologic sections in this study and the attendant limitation of sampling. The detection of epithelial changes that are nonglobal (not diffuse) is a reflection of the number of sections examined, 32 and this is mirrored in estimates of the incidence of these lesions. A study in- Arch Pathol Lab Med Vol 129, March 2005 Putative Precursors of Gallbladder Disease Mukhopadhyay & Landas 389

5 volving fewer patients but using far more extensive sampling yielded results similar to ours. 9 Armed with data on the true incidence of epithelial lesions based on comprehensive mapping, these investigators analyzed the sensitivity of single random sections. They found that a single random histologic section detected fewer than one third of metaplasias and dysplasias in the gallbladder. This may, in part, explain why our incidence figures are approximately one third of those obtained by whole-gallbladder mapping. It may be argued, however, that our figures, being based on routine histologic material, reflect more closely the incidence that pathologists may expect to see in everyday practice. Furthermore, although our incidence figures are lower than the true values by a factor of one third, this skew applies to all epithelial lesions and does not invalidate the demonstrated associations. In summary, we have described the precursors of gallbladder dysplasia using histologic examination of a large number of routinely resected cases from a North American population. Our results are in agreement with earlier data (based on studies involving Japanese 8 and Latin American 9 populations) that dysplasia is strongly associated with intestinal metaplasia, which in turn is associated with antral-type metaplasia. These associations are supported by an age gradient, with younger patients at the benign end of the sequence (antral-type metaplasia) and older patients at the dysplastic end. Other investigators have substantiated the proposed relatedness of these epithelial changes with carcinoma at a molecular level. 33 If future studies are able to demonstrate accrual of molecular abnormalities correlating with the statistically substantiated morphologic sequence demonstrated herein, considerable evidence would be added in support of such precursor lesions. References 1. Rosai J. Gallbladder and extrahepatic bile ducts. In: Rosai J, ed. Rosai and Ackerman s Surgical Pathology. 9th ed. Edinburgh, Scotland: Elsevier; 2004: Albores-Saavedra J, Alcantra-Vazquez A, Cruz-Ortiz H, Herrera-Goepfert R. The precursor lesions of invasive gallbladder carcinoma: hyperplasia, atypical hyperplasia and carcinoma in situ. Cancer. 1980;45: Chang HJ, Kim SW, Kim YT, Kim WH. Loss of heterozygosity in dysplasia and carcinoma of the gallbladder. Mod Pathol. 1999;12: Dowling GP, Kelly JK. The histogenesis of adenocarcinoma of the gallbladder. Cancer. 1986;58: Laitio M. Histogenesis of epithelial neoplasms of human gallbladder, I: dysplasia. Pathol Res Pract. 1983;178: Ojeda VJ, Shilkin KB, Walters MN. Premalignant epithelial lesions of the gall bladder: a prospective study of 120 cholecystectomy specimens. Pathology. 1985;17: Sasatomi E, Tokunaga O, Miyazaki K. Precancerous conditions of gallbladder carcinoma: overview of histopathologic characteristics and molecular genetic findings. J Hepatobiliary Pancreat Surg. 2000;7: Yamagiwa H, Tomiyama H. Intestinal metaplasia-dysplasia-carcinoma sequence of the gallbladder. Acta Pathol Jpn. 1986;36: Duarte I, Llanos O, Domke H, Harz C, Valdivieso V. Metaplasia and precursor lesions of gallbladder carcinoma: frequency, distribution, and probability of detection in routine histologic samples. Cancer. 1993;72: Laitio M. Goblet cells, enterochromaffin cells, superficial gastric-type epithelium and antral-type glands in the gallbladder. Beitr Pathol. 1975;156: Laitio M. Intestinal, gastric body and antral-type mucosal metaplasia in the gallbladder. Beitr Pathol. 1976;159: Laitio M. Morphology and histochemistry of non-tumorous gallbladder epithelium: a series of 103 cases. Pathol Res Pract. 1980;167: Frierson HF Jr. The gross anatomy and histology of the gallbladder, extrahepatic bile ducts, Vaterian system, and minor papilla. Am J Surg Pathol. 1989; 13: Albores-Saavedra J, Henson DE, Klimstra DS. Gallbladder cancer: epidemiology, etiology, clinical manifestations, and laboratory studies. In: Tumors of the Gallbladder, Extrahepatic Bile Ducts and Ampulla of Vater. Washington, DC: Armed Forces Institute of Pathology; 2000:37. Atlas of Tumor Pathology; 3rd series, fascicle Roa I, Araya JC, Villaseca M, et al. Preneoplastic lesions and gallbladder cancer: an estimate of the period required for progression. Gastroenterology. 1996;111: Kijima H, Watanabe H, Iwafuchi M, Ishihara N. Histogenesis of gallbladder carcinoma from investigation of early carcinoma and microcarcinoma. Acta Pathol Jpn. 1989;39: Kozuka S, Kurashina M, Tsubone M, Hachisuka K, Yasui A. Significance of intestinal metaplasia for the evolution of cancer in the biliary tract. Cancer. 1984; 54: Kushima R, Lohe B, Borchard F. Differentiation towards gastric foveolar, mucopeptic and intestinal goblet cells in gallbladder adenocarcinomas. Histopathology. 1996;29: Laitio M, Terho T. Polysaccharides of metaplastic mucosa and carcinoma of the gallbladder. Lab Invest. 1975;32: Sasaki M, Yamato T, Nakanuma Y, Ho SB, Kim YS. Expression of MUC2, MUC5AC and MUC6 apomucins in carcinoma, dysplasia and non-dysplastic epithelia of the gallbladder. Pathol Int. 1999;49: Tatematsu M, Furihata C, Miki K, et al. Complete and incomplete pyloric gland metaplasia of human gallbladder. Acta Pathol Jpn. 1987;37: Tsutsumi Y, Nagura H, Osamura Y, Watanabe K, Yanaihara N. Histochemical studies of metaplastic lesions in the human gallbladder. Arch Pathol Lab Med. 1984;108: Nishihara K, Takashima M, Furuta T, Haraguchi M, Tsuneyoshi M. Adenosquamous carcinoma of the gallbladder with gastric foveolar-type epithelium. Pathol Int. 1995;45: Sakaki M, Hirokawa M, Sano T, Horiguchi H, Wakatsuki S, Ogata S. Gallbladder adenocarcinoma with florid neuroendocrine cell nests and extensive Paneth cell metaplasia. Endocr Pathol. 2000;11: Xeropotamos N, Skopelitou AS, Batsis C, Kappas AM. Heterotopic gastric mucosa together with intestinal metaplasia and moderate dysplasia in the gall bladder: report of 2 clinically unusual cases with literature review. Gut. 2001;48: Gupta SC, Misra V, Singh PA, Roy A, Misra SP, Gupta AK. Gall stones and carcinoma gall bladder. Indian J Pathol Microbiol. 2000;43: Azadeh B, Parai SK. Argentaffin cells, intestinal metaplasia and antral metaplasia in carcinoma of the gall bladder. Histopathology. 1980;4: Filipe MI, Munoz N, Matko I, et al. Intestinal metaplasia types and the risk of gastric cancer: a cohort study in Slovenia. Int J Cancer. 1994;57: Iida F, Kusama J. Gastric carcinoma and intestinal metaplasia. Significance of types of intestinal metaplasia on development of gastric carcinoma. Cancer. 1982;50: Albores-Saavedra J, Nadji M, Henson DE, Ziegels-Weissman J, Mones JM. Intestinal metaplasia of the gallbladder: a morphologic and immunocytochemical study. Hum Pathol. 1986;17: Smok G, Cervilla K, Bosch H, Csendes A. Precancerous lesions of invasive carcinoma of the gallbladder. Rev Med Chil. 1986;114: Albores-Saavedra J, Henson DE, Klimstra DS. Dysplasia and carcinoma in situ of gallbladder. In: Tumors of the Gallbladder, Extrahepatic Bile Ducts and Ampulla of Vater. Washington, DC: Armed Forces Institute of Pathology; 2000: 51. Atlas of Tumor Pathology; 3rd series, fascicle Wistuba II, Sugio K, Hung J, et al. Allele-specific mutations involved in the pathogenesis of endemic gallbladder carcinoma in Chile. Cancer Res. 1995; 55: Arch Pathol Lab Med Vol 129, March 2005 Putative Precursors of Gallbladder Disease Mukhopadhyay & Landas