Atypical Regenerative Hyperplasia of the Esophagus in Endoscopic Biopsy. A Mimicker of Squamous Esophagic Carcinoma

Transcription

1 Atypical Regenerative Hyperplasia of the Esophagus in Endoscopic Biopsy A Mimicker of Squamous Esophagic Carcinoma Julián Arista-Nasr, MD; Irene Rivera, MD; Braulio Martinez-Benitez, MD; Leticia Bornstein-Quevedo, MD; Héctor Orozco, MD; Yadira Lugo-Guevara, MD Context. Atypical regeneration can mimic carcinoma in various epithelia. On endoscopic biopsies, atypical regenerative hyperplasia of the esophagus may show pleomorphism and atypia, simulating esophageal squamous cell carcinoma. Objective. To establish the most useful histologic features to distinguish atypical regenerative hyperplasia from esophageal carcinoma in endoscopic biopsies. Design. To study the frequency and histologic appearance of atypical regenerative hyperplasia, which simulate carcinoma, we reviewed 600 endoscopic biopsies (555 with chronic esophagitis and 45 with carcinomas of the esophagus). We selected those cases in which the differential diagnosis included regenerative atypical hyperplasia versus esophageal carcinoma and cases of atypical regenerative hyperplasia that were mistaken for carcinoma. For comparative purposes, we studied 10 cases of esophageal carcinoma from endoscopic biopsies that were confirmed by esophagectomy. Results. Among the cases with chronic esophagitis, we found 10 biopsies (1.8%) in which atypical regenerative hyperplasia mimicked carcinoma. In 7 cases, there were 4 to 12 years of follow-up, and no patient developed esophageal neoplasm. The remaining 3 patients were submitted to esophagectomy. None of these patients had carcinoma or dysplasia in the esophageal resection (false-positive biopsies). The most useful architectural changes in squamous carcinoma included stromal infiltration by nests, cords, or thin prongs of neoplastic keratinocytes, palisading desmoplasia, and in situ carcinoma in the adjacent epithelium. Malignant keratinocytes showed variable degrees of differentiation with differently shaped and sized cells, squamous epithelial pearls, individual keratinization, and atypical mitosis. In contrast, biopsies with atypical hyperplasia showed detached nests or irregular fragments without stroma and were made up of immature and relatively monotonous medium or small keratinocytes that were intermixed with inflammatory cells. Individual keratinization was rare, and no squamous pearls were seen. Other features of atypical hyperplasia included granulated tissue with atypical endothelial cells, nonatypical mitosis, lymphoid hyperplasia, and the absence of dysplasia or carcinoma in situ. Two biopsies showed stromal pseudoinfiltration as a result of tangential sectioning and were characterized by thick, round prongs composed of keratinocytes that penetrated regions with granulation or the inflamed tissues of esophageal ulcers. Conclusions. Atypical esophageal regenerative hyperplasia may mimic carcinoma in a small percentage of esophageal biopsies. If the histologic changes are not sufficient to establish an accurate diagnosis, medical treatment and subsequent biopsies should be performed, particularly if there are no endoscopic or radiologic data to support the presence of a neoplasm. (Arch Pathol Lab Med. 2005;129: ) It has been widely recognized that atypical regenerative hyperplasia (ARH) of diverse epithelia can mimic neoplasm; ARH of the gallbladder, 1 stomach, 2 bladder, 3 pros- Accepted for publication March 11, From the Department of Pathology, Instituto Nacional de Ciencias Médicas y Nutrición, Salvador Zubirán (INCMNSZ), Mexico City, Mexico, Vasco de Quiroga No. 15, Tlalpan, México D.F., Mexico (Drs Arista-Nasr, Martinez-Benitez, Bornstein-Quevedo, and Orozco); and the Department of Pathology, Hospital Central Sur de Petróleos Mexicanos (PEMEX), Mexico City, Mexico, Periférico Sur. No. 3146, México D.F., Mexico (Drs Rivera and Lugo-Guevara). The authors have no relevant financial interest in the products or companies described in this article. Reprints: Julian Arista-Nasr, MD, Department of Pathology, Instituto Nacional de Ciencias Médicas y Nutrición SZ, Vasco de Quiroga No. 15, Tlalpan, 14000, México D.F., Mexico ( pipa5@hotmail.com). tate, 4 and skin, 5 among others, can simulate carcinoma; also, atypical mesothelial hyperplasia can mimic mesothelioma. 6,7 During the past decade, we have studied several cases in which esophageal epithelial cells showed regenerative changes with marked atypia and pleomorphism that mimicked squamous cell carcinoma (SCC) of the esophagus in endoscopic biopsies. The purpose of this study was to analyze the histologic characteristics and the evolution of 10 endoscopic biopsies with ARH that simulated SCC. MATERIALS AND METHODS From the records of the Hospital Regional de Pemex and the Instituto Nacional de la Nutricion, both in Mexico City, we retrieved 600 esophageal biopsies taken between 1980 and Of these, 45 were SCC, and 555 were benign lesions. When the be- Arch Pathol Lab Med Vol 129, July 2005 Atypical Regenerative Hyperplasia Arista-Nasr et al 899

2 Table 1. Clinical and Endoscopic Features in Atypical Regenerative Hyperplasia and Squamous Cell Carcinoma Atypical Regenerative Hyperplasia (10 Cases) Clinical Age: y Substernal pain Dysphagia Nausea/vomiting 6moto10y Endoscopic features Superficial ulcers (4) Esophageal epithelium erosion (3) Erythematous mucosa with granular appearance (3) Squamous Cell Carcinoma (10 Cases) Age: y Substernal pain Increasing dysphagia Vomiting/bleeding 8moto1y Fungating masses (6) Ulcers, irregular edges (4) nign lesions were graded according to atypia, 502 were classified as chronic esophagitis without atypia or with slight atypia, 43 were classified as chronic esophagitis with regenerative changes and moderate atypia, and 10 were classified with marked atypia that mimicked carcinoma. The last group represented 1.6% of all biopsies examined and 1.8% of the biopsies with esophagitis. Seven of these 10 patients had 4 to 12 years of clinical follow-up, in which there was no clinical, endoscopic, or histopathologic evidence of neoplasm. Three cases were misdiagnosed as SCC, and these patients underwent esophagectomy. For comparison, 10 biopsies of SCC that had been diagnosed by endoscopic biopsy and confirmed after surgical resection were included. In both groups, the following data were recorded: age, sex, radiologic studies, macroscopic endoscopic appearance, main clinical symptoms, and original diagnosis. Biopsies were examined for the following histologic changes: growth pattern; stromal infiltration by atypical keratinocytes; stromal pseudoinfiltration (tangential sections that were characterized by thick, round prongs composed of immature, atypical keratinocytes surrounded by granulation or the inflamed tissue of esophageal ulcers); lymphatic infiltration; squamous epithelial pearls; atypical keratinocytes with nucleomegaly, prominent nucleoli, or both; individual keratinization; high-grade dysplasia/in situ carcinoma of esophageal epithelium; acute and chronic inflammation; frequent mitoses (3 or more in 10 high-power fields); granulated tissue with atypical endothelial cells; and marked lymphoid hyperplasia. Immunohistochemical studies were performed on formalinfixed, paraffin-embedded tissue using the standard avidin-biotin complex technique. In cases in which granulated tissue with atypical endothelial cells was identified, immunohistochemical studies with CD34 (1:100 dilution; Dako Corporation, Carpinteria, Calif) were performed to confirm its endothelial nature. To exclude the presence of viral infection and confirm the epithelial nature of atypical cells, immunohistochemical studies with cytokeratin AE1/AE3 (1:10 dilution; Dako), herpes simplex virus type 1 (1: 500 dilution; Dako) and type 2 (1:200 dilution; Dako), and cytomegalovirus (1:10 dilution; Dako) were performed. RESULTS In the SCC group, ages ranged from 50 to 81 years, and there were 6 women and 4 men. In the ARH group, ages ranged from 49 to 80 years, and the group included 2 women and 8 men. In both groups, symptoms included dysphagia, substernal pain, upper gastrointestinal bleeding, and a history of chronic esophagitis of months or years. The most frequent endoscopic appearances in biopsies with carcinoma were fungating masses that partially occluded the esophageal lumen or ulcers with irregular edges. In cases with ARH, 4 showed superficial ulcers, 3 showed esophageal epithelium erosion, and 3 showed erythematous mucosa with a granular appearance of the epithelium (Table 1). The original histologic diagnosis in the 10 endoscopic biopsies with carcinoma was SCC, which was confirmed Table 2. Settings Where Atypical Regenerative Hyperplasia Can Be Confused With Squamous Cell Carcinoma Tangential sectioning Detached epithelium without stroma Dense inflammatory infiltrate obscuring epithelial/stromal infiltrate Atypical keratinocytes intermixed with inflammatory cells by resection of the neoplasm. Six tumors were classified as moderately differentiated, 2 as well-differentiated, and 2 as poorly differentiated carcinomas. The original histologic diagnosis in 3 patients with ARH was SCC; after esophagectomy, neither neoplastic nor dysplastic changes of the esophageal epithelium were found. In 7 patients, the differential diagnosis between ARH and SCC was established. In the absence of neoplasm in endoscopic studies, radiologic studies, or both, the 7 patients were treated with motility-promoting drugs, H2 receptor antagonists, or gastric proton pump inhibitors. At 3 to 5 weeks after treatment, additional biopsies were taken. In all of them, ARH involution was found; mucosal appearance was normal or showed slight or moderate esophagitis; and there was no histologic evidence of residual atypical hyperplasia. After a follow-up of 4 to 12 years, there was no evidence of neoplasm in any of these 7 patients. The histologic results of biopsies with ARH and SCC are summarized in Tables 2 and 3. The most frequent growth pattern observed in the SCC cases was nests of cohesive keratinocytes that infiltrated the stroma. Another pattern of infiltration included irregular, jagged, thin projections of malignant cells and trabecular growth. In 5 cases, there was collagenized desmoplasia, and in another 2 cases, there was lymphatic infiltration. In some fields, neoplastic cells showed palisades of malignant keratinocytes. Neoplastic cells showed pleomorphism, individual keratinization, or squamous epithelial pearls of variable grade. Two biopsies showed poorly differentiated carcinomas with small, atypical basophilic cells diffusely infiltrating the stroma. In 6 biopsies, the neoplasm had different degrees of differentiation in diverse areas (Figure 1), in contrast with ARH, in which the proliferating keratinocytes showed a monotonous appearance (Figures 2 through 4). In carcinomas, nucleomegaly and prominent nucleoli were common, as were atypically appearing mitoses. In 5 cases, high-grade dysplasia or in situ carcinoma in superficial epithelium was observed. Mixed inflammatory infiltrate, granulated tissue with or without prominent endothelial cells were found in some biopsies, but to a lesser extent than in the biopsies with ARH. Other changes, such as 900 Arch Pathol Lab Med Vol 129, July 2005 Atypical Regenerative Hyperplasia Arista-Nasr et al

3 Table 3. Pattern of growth Cytologic features Histologic Changes in Squamous Carcinoma Versus Atypical Regenerative Hyperplasia in Endoscopic Biopsy Squamous Carcinoma (10 Biopsies) Stromal infiltration by nests or irregular thin cords of malignant keratinocytes Occasional desmoplastic stroma Occasional palisades Pleomorphic cells with varied sizes and shapes Poorly differentiated carcinomas with small, atypical cells infiltrating the stroma Different degrees of differentiation Individual keratinization 8 3 Squamous epithelial pearls 6 0 Inflammatory background Moderate or marked acute and chronic inflammation Lymphoid hyperplasia mimicking lymphoma or pseudolymphoma Dysplasia/in situ carcinoma of adjacent 5 0 esophageal epithelium Frequent mitoses* 8 (atypical) 5 (nonatypical) * Three or more mitoses in 10 high-power fields. Atypical Regenerative Hyperplasia (10 Biopsies) Detached nests or irregular fragments without stromal infiltration Pseudoinfiltration by thick prongs of immature keratinocytes in granulated or inflamed tissues of ulcer bed No palisades No desmoplastic stroma Immature, medium-sized proliferating keratinocytes with relatively monotonous appearance intermixed with inflammatory cells Hyperchromatic noncohesive small keratinocytes without stromal infiltration basal cell hyperplasia, elongation of esophageal epithelial papilla, and recent hemorrhage, were also found with less frequency than in biopsies with ARH. Biopsies with atypical regenerative changes showed detached nests or irregular fragments of atypical keratinocytes without esophageal stroma (Figures 2 and 3) that exhibited immature, atypical squamous cells with nucleomegaly, nucleoli, or both that simulated moderately differentiated SCC or noncohesive groups of small keratinocytes with hyperchromatism that resembled small, poorly differentiated SCC (Figure 4). Reactive, atypical cells were frequently seen intermixed with inflammatory cells that included lymphocytes, neutrophils, eosinophils, and plasma cells (Figures 2, 3, 5, and 6). A pseudoinfiltrative pattern was seen in 2 cases of ulcerated esophagitis and was characterized by thick, round prongs of atypical hyperchromatic immature keratinocytes (Figures 5 and 6). In both cases, the proliferating, atypical keratinocytes were surrounded by stroma with a marked inflammatory infiltrate that, in some fields, was seen around the atypical cells (Figure 6). The stroma did not show a desmoplastic reaction in any case. In 5 biopsies, frequent mitoses were observed (3 or more in 10 highpower fields); nevertheless, they did not show an atypical appearance. In contrast with some biopsies of SCC, no biopsy with ARH showed lymphatic infiltration, superficial carcinoma in situ, or squamous epithelial pearls, but in 3 cases, individual keratinization was observed (Figure 3). In some fields, the granulated tissue showed endothelial cells with atypia, which contributed to the malignant appearance of the biopsy (Figure 7). However, an immunohistochemical analysis with the CD34 antigen confirmed the endothelial nature of the cells. In 2 biopsies, tissue fragments were found with chronic and dense inflammatory infiltrate that simulated pseudolymphoma. In biopsies with ARH, esophagitis, basal cell hyperplasia, and elongation of papillae were common. The immunohistochemical study for herpes virus and cytomegalovirus was negative in all cases. COMMENT Regeneration is a common change observed in patients with esophagitis, and it can easily be recognized as benign because reactive keratinocytes show slight or moderate atypia. However, in a small percentage of biopsies, regeneration may show unusual atypia and mimic SCC. In this series in which 555 esophageal biopsies with esophagitis were reviewed, 1.8% of the biopsies showed marked ARH, and 3 were diagnosed as carcinoma. The erroneous interpretation of ARH as SCC can result in unnecessary surgical resections or aggressive treatment, such as radio- and chemotherapy, which may have considerable morbidity. 8,9 Even though ARH of the esophageal epithelium may mimic carcinoma and occasionally lead to false-positive results, this lesion has been only briefly mentioned and illustrated in pathology textbooks; additionally, to our knowledge, there are no reports that analyze the histologic differences between ARH and esophageal SCC. Lewin and Appelman, 10 in the Atlas of Tumor Pathology of the Armed Forces Institute of Pathology, mention that the most significant differential diagnosis with SCC in endoscopic biopsies is the squamous epithelial proliferation that occurs at the edges and bases of chronic reflux induced ulcers. These appear as uniformly thick, parallel prongs of primitive squamous cells that penetrate the granulated tissue of the ulcer bed. Early in the healing process, cells in these prongs are cuboidal with little cytoplasm, large, often misshapen, hyperchromatic nuclei, and a very high nuclear-cytoplasmatic ratio. Mitotic figures are often prominent. If these foci are sectioned perfectly, diagnosis is not difficult because the parallel arrangement of the prongs is obvious. However, in bias-cut Arch Pathol Lab Med Vol 129, July 2005 Atypical Regenerative Hyperplasia Arista-Nasr et al 901

4 Figure 1. Esophageal squamous carcinoma. Atypical keratinocytes show several degrees of differentiation. Some keratinocytes are of medium size and alternate with others in which marked pleomorphism and individual keratinization are recognizable (hematoxylin-eosin, original magnification 500). Figure 2. Atypical regenerative hyperplasia. Esophageal epithelial fragments are arranged in nests and are composed of monotonous mediumsized keratinocytes. Many reactive keratinocytes show atypical nuclei with nucleolus and are intermixed with inflammatory cells. Other fragments of the same biopsy showed granulated tissue, hemorrhage, and fibrin (hematoxylin-eosin, original magnification 575). Figure 3. Irregular tissue fragment with atypical regenerative hyperplasia. Keratinocytes show hyperchromatism and individual keratinization. Reactive lymphocytes and eosinophils can be seen intermixed with atypical keratinocytes (hematoxylin-eosin, original magnification 575). Figure 4. Atypical regenerative hyperplasia. Noncohesive reactive small keratinocytes resemble small poorly differentiated carcinoma. Although the inflammatory infiltrate is scant, the absence of stromal infiltration led the pathologist to suspect atypical regenerative hyperplasia (hematoxylineosin, original magnification 575). samples, the prongs are often sectioned across, so that they appear trapped and separated within the stroma and resemble nests of invasive carcinoma. Chandrasoma 11 has also pointed out that ulcer-related changes can produce nuclear enlargement, hyperchromasia, nuclear irregularity, prominent nuclei, and mitotic activity that closely mimic malignancy. The irregularity of epithelial proliferation in cellular regeneration and the tangential sectioning of biopsies can closely resemble carcinoma; thus, careful attention to architectural and cytologic changes is necessary to avoid a false-positive diagnosis. 11 Stromal infiltration by atypical keratinocytes is a strong marker for the diagnosis of SCC; however, it must be remembered that areas of pseudoinfiltration may be present in biopsies of ARH. In both of the cases in this series that showed a pseudoinfiltrative pattern, the parallel arrangement of the prongs described and illustrated by Lewin and Appelman 10 were not found, probably because these tissue fragments were not sectioned perfectly, and the size of the fragments was too small to recognize the parallel array. When a pseudoinfiltrative pattern that is due to tangential cuts is suspected, the monotonous appearance of the atypical keratinocytes and the presence of long and thick prongs of primitive atypical squamous cells that penetrate the granulated tissue or stroma with abundant inflammatory infiltrate may be useful points to consider toward resolving the diagnosis. The diagnosis of esophageal carcinoma in an endoscopic biopsy includes nests, cords, or thin prongs of malignant keratinocytes that infiltrate the esophageal stroma, which may display a desmoplastic reaction and peripheral palisades. Individual keratinization and squamous epithelial pearls are common. Neoplastic cells may be oval, polyhedral, spindle-shaped, or giant, although in an undifferentiated carcinoma, they present as small cells without keratinization. Different degrees of differentiation are frequently seen. In some biopsies, dysplasia or carcinoma in situ is identified in the superficial esophageal epithelium. 902 Arch Pathol Lab Med Vol 129, July 2005 Atypical Regenerative Hyperplasia Arista-Nasr et al

5 Figure 5. Pseudoinfiltration in atypical regenerative hyperplasia. A thick, round, long prong is extensively surrounded by stroma with abundant inflammatory infiltrate. Atypical prongs are trapped and surrounded by esophageal stroma in tangential sections that should not be confused with infiltrating carcinoma (hematoxylin-eosin, original magnification 75). Figure 6. Detail of Figure 5. Keratinocytes are relatively monotonous with moderate atypia. Marked stromal inflammation is seen around atypical cells, and its presence may point to a benign process (hematoxylin-eosin, original magnification 500). Figure 7. Granulated tissue in atypical regenerative hyperplasia. Some vessels have atypical endothelial cells, which should not be confused with atypical epithelial cells (hematoxylin-eosin, original magnification 650). In contrast, the biopsies with atypical regenerative changes showed detached nests or irregular fragments without stroma composed of atypical immature, mediumsized keratinocytes with nucleomegaly, nucleoli, or both. The immature proliferating keratinocytes had a relatively monotonous appearance in contrast with malignant keratinocytes, in which differently shaped and sized cells are common. Occasionally, the reactive keratinocytes may show individual keratinization and nonatypical mitosis. Other significant differences with carcinoma include the absence of palisading in the nest and cords, atypical mitosis, desmoplastic stroma, and superficial carcinoma in situ. Although granulated tissue and inflammation could be seen in biopsies with SCC, these changes were more frequent and apparent in biopsies with ARH, and the immature keratinocytes in all cases were intermixed with inflammatory cells. In some biopsies, groups of atypical noncohesive small hyperchromatic keratinocytes were also found. These groups resembled poorly differentiated carcinoma, but they never were associated with stromal infiltration and frequently were also associated in the same fragment or in other fragment tissues with granulation, inflammation, or both. Consequently, the presence of cells with pleomorphism, atypia, or a high nuclear-cytoplasmatic ratio by itself is not enough to establish a diagnosis of carcinoma. If the biopsy shows small or medium, relatively monotonous atypical keratinocytes associated with moderate or marked inflamed or granulated tissue that is consistent with the bed of the ulcer and the other criteria of ARH summarized in Table 2, a diagnosis of atypical regeneration must be favored. An infrequent anomaly that can be disconcerting in endoscopic biopsies is the presence of a chronic, dense lymphocytic infiltrate. This tissue response seems to represent an exaggerated inflammatory reaction in chronic esophagitis, and it is important to recognize this variation to avoid confusion with pseudolymphomas 12,13 or lymphoproliferative processes of the gastroesophageal junction. 14,15 In conclusion, we believe that there are sufficient architectural and cytologic criteria to establish a distinction between esophageal carcinoma and ARH in most endoscopic biopsies. If the histologic changes are not absolutely convincing or the surgical pathologist does not feel comfortable establishing a definite diagnosis, it is better to recommend medical treatment and repeat the biopsy, particularly if there are no endoscopic or radiologic data to support the presence of a neoplasm. In our experience, atypical regenerative changes can decrease notably or even disappear within 3 to 5 weeks. However, it must be mentioned that this is not always a definitive solution, because the absence of a well-defined lesion leaves open the possibility that the region of concern will not be accurately sampled on a second biopsy. 11 For this reason, clinical follow-up and subsequent endoscopic studies should be recommended for these patients. References 1. Albores-Saavedra J, Henson DE, Klimstra DS. Tumors of the Gallbladder, Extrahepatic Bile Ducts, and Ampulla of Vater. Washington, DC: Armed Forces Institute of Pathology; 1998: Atlas of Tumor Pathology; 3rd series, fascicle Martinez-Madrigal F, Ortiz-Hidalgo C, Torres-Vega C, et al. Atypical regenerative changes, dysplasia, and carcinoma in situ in chronic gastritis associated with. Helicobacter pylori. Rev Gastroenterol Mex. 2000;65: Epstein JI, Amin MB, Reuter VE. Flat urothelial lesions. In: Bladder Biopsy Interpretation. Philadelphia, Pa: Lippincott Williams & Wilkins; 2003:30. Arch Pathol Lab Med Vol 129, July 2005 Atypical Regenerative Hyperplasia Arista-Nasr et al 903

6 4. Epstein JI, Yang XJ. Prostate Biopsy Interpretation. 3rd ed. Philadelphia, Pa: Lippincott Williams & Wilkins; 2002: Civatte J. Pseudo-carcinomatous hyperplasia. J Cutan Pathol. 1985;12: McCaughey WTE, Al-Jabi M. Differentiation of serosal hyperplasia and neoplasia in biopsies. Pathol Annu. 186;21: Rosai J, Dehner LP. Nodular mesothelial hyperplasia in hernia sac: a benign reactive condition simulating a neoplastic process. Cancer. 1975;35: Iizuka AN, Ide H, Ishida K, et al. Surgery plus chemotherapy compared with surgery alone for localized squamous cell carcinoma of the thoracic esophagus: a Japan Clinical Oncology Group study JCOG9204. J Clin Oncol. 2003;21: Aroori S, Parshad R, Kapoor A, Gupta SD, Kumar A. Neoadjuvant chemotherapy in squamous cell carcinoma of the esophagus using low dose continuous infusion 5-fluoracil and cisplatin: results of a prospective study. Indian J Cancer. 2004;4: Lewin KJ, Appelman HD. Tumors of the Esophagus and Stomach. Washington, DC: Armed Forces Institute of Pathology; 1995: Atlas of Tumor Pathology; 3rd series, fascicle Chandrasoma P. Gastrointestinal Pathology. Stanford, Conn: Appleton & Lange; 1999: Rosai J. Rosai and Ackerman s Surgical Pathology. 9th ed. St Louis, Mo: The CV Mosby Co; 2004: Sheahan DG, West AB. Focal lymphoid hyperplasia (pseudolymphoma) of the esophagus. Am J Surg Pathol. 1985;9: Hosaka S, Nakamura N, Akamatsu T, et al. A case of primary low grade mucosa associated lymphoid tissue (MALT) lymphoma of the esophagus. Gut. 2002;51: Wotherspoon AC. Extragastric MALT lymphoma. Gut. 2002;51: Arch Pathol Lab Med Vol 129, July 2005 Atypical Regenerative Hyperplasia Arista-Nasr et al