Printer Friendly

Morphological Features and Prognostic Significance of ARIDIA-Deficient Esophageal Adenocarcinomas.

Esophageal adenocarcinomas (EACs) are morphologically heterogeneous tumors. (1-3) This phenotypic diversity is reflected in the heterogeneous molecular pathways described in EACs. Mutations of tumor suppressor genes TP53 (4) and pl6 (5); dysregulation of proto-oncogenes EGF, (6) EGFR, (7) TGF, (8) and HER2 (9); and loss of DNA mismatch repair function10,11 have all been described. There is also evidence of epigenetic alteration by promoter hypermethylation involving APC, CDHl, CDKN2A, MGMT, and TMEFF2/ HPP1. (12) Recently, whole exome analyses have revealed several novel, recurrent mutations in EACs, including lossof-function mutations in ARIDlA, with a prevalence of approximately 10%. (13,14)

ARIDlA (AT-rich interactive domain 1A), also known as BAF250A and SMARCFl, is a member of the switch/sucrose nonfermentable, ATP-dependent family of chromatin-restructuring genes that are involved in epigenetic regulation and are collectively mutated in approximately 20% of all malignancies. (15) The chromatin-remodeling function of these switch/sucrose nonfermentable complexes is essential for the regulation of gene expression, cell proliferation, cell-fate determination, and DNA damage repair. (16) The ARID-DNA interactions are essential for the tumor-suppressor function of ARIDlA, and inactivation of the gene by somatic mutation or other epigenetic mechanisms leads to tumorigenesis. (17)

Loss-of-function somatic mutations in ARIDlA have been described in carcinomas from multiple sites, including the esophagus, stomach, colon, liver, biliary tract, endometrium, ovary, urinary bladder, and pancreas. (15,18) ARIDlA mutations in ovarian cancer have been described in approximately 50% of ovarian clear cell carcinomas (19) and 30% of endometrioid endometrial carcinomas. (20) Moreover, loss of the ARID1A protein expression in Mullerian adenocarcinomas correlates with DNA mismatch-repair enzyme deficiency (4,21,22) and is mutually exclusive of TP53 mutations, (23) suggesting a distinct pathway of carcinogenesis. The aims of our study were to analyze the morphological and immunophenotypic associations of ARID1A loss in EAC and to determine the prognostic significance of ARID1A-deficient EACs.


Study Group

We searched our archives for all esophagectomies and esophagogastrectomies performed between 1989 and 2011 for esophageal adenocarcinoma. Of the 863 resections, 521 (60%) had received neoadjuvant therapy and were excluded from the study. Of the remaining 342 primary resections, 164 tumors (48%) were centered in the esophagus with or without involvement of the gastroesophageal junction. Complete clinical information and archival slides and paraffin blocks were available for the final 120 tumors (73%) in the study group. Patient demographics and clinical outcome data were obtained by medical chart review, and overall survival was obtained by querying the social security death index. The study was approved by the institutional review board.

Morphological and Immunohistochemical Evaluation

All cases were evaluated for a predetermined set of parameters, including tumor location, tumor size and grade, maximum depth of invasion, lymphovascular and perineural invasion, presence of Barrett esophagus (BE), and the grade of any BE-associated dysplasia. In addition, the presence or absence of tumor-infiltrating lymphocytes and peritumoral lymphoid aggregates was noted. Tumors were classified according to the most recent World Health Organization classification (1) and reviewed and restaged using the American Joint Committee on Cancer, 7th edition, staging criteria. (24)

Representative tumor tissue blocks were selected from each case for immunohistochemical analysis to include EAC as well as background BE and any BE-associated dysplasia, whenever possible. Immunoperoxidase studies were performed on 5-[micro]m sections with antigen retrieval to assess expression of p53 (mouse monoclonal 1767, 1:1200; Beckman Coulter, Atlanta, Georgia), MLH1 (mouse monoclonal, MCL-L-MLH1; 1:75, Leica, Buffalo Grove, Illinois), and ARID1A (rabbit polyclonal, HPA 005456; 1:500, Sigma-Aldrich, St. Louis, Missouri) with automated immunostainers.

ARID1A and MLH1 expression in the tumor cells was scored as deficient when complete loss of expression was present in the entire tumor with intact staining in adjacent stromal cells, or when a discrete, confluent focus of complete loss, compatible with a distinct tumor clone, was present anywhere in the tumor. Similarly, ARID1A and MLH1 were scored as deficient in background BE and any BE-associated dysplasia only when complete loss of staining involved the entire crypt from the crypt base to the surface epithelium, with retained nuclear staining in the adjacent stromal cells. P53 expression, in EAC and in BE-associated dysplasia, was scored as a wild-type pattern when only scattered nuclei were positive and as a mutant pattern when confluent, strong overexpression or complete absence of staining (null pattern) was seen in the neoplastic cells. (25,26)

Data Analysis

Fisher exact test was used to compare differences in categorical variables between ARID1A-deficient and -intact EAC. Differences in overall survival were analyzed with a Kaplan Meier analysis and a Cox proportional hazards model computed with the R statistical programming language (version 2.15.1, R development core 2012; R Foundation for Statistical Computing, Wien, Austria, http://cran. P < .05 was considered statistically significant.


Study Group

The final study group consisted of 97 men (81%) and 23 women (19%), with a mean age of 66 years (range, 30-87 years). The clinical and pathologic features of the study group are summarized in Table 1. Barrett esophagus, defined as columnar metaplasia in the distal esophagus with goblet cells, was seen in 77 of 120 tumors (64%) and BE-associated dysplasia adjacent to carcinoma was present in 88 of 120 cases (73%). In the tumor tissue blocks selected for immunohistochemical analysis, 43 of 120 (36%) contained adjacent, nondysplastic BE, and 17 of 120 (14%) and 47 of 120 (39%) contained foci of adjacent BE-associated low- and high-grade dysplasia, respectively.

Morphological and Immunohistochemical Data

Loss of ARID1A staining was seen in 12 of 120 EACs (10%), with complete loss of staining in the entire tumor in 10 of 12 cases (83%) and in a discrete, confluent focus in 2 of 12 (17%) cases. In 1 of these latter 2 cases, loss of ARID1A staining was restricted to the poorly differentiated component, with intact staining in the conventional glandular component. Loss of ARID1A in EAC was associated with a distinctive morphology reminiscent of microsatellite instability high (MSI-H) colorectal adenocarcinomas. A medullary phenotype, with solid architecture and increased tumor-infiltrating lymphocytes; a conventional gland-forming carcinoma, with increased tumor-infiltrating lymphocytes or an exophytic, tubulovillous growth pattern with copious, extracellular mucin, consistent with mucinous adenocarcinoma; or prominent, peritumoral, lymphoid aggregates (Figure 1, A through C) were characteristic of ARID1A-deficient EAC. At least one of those features was present in 6 of 12 ARID1A-deficient tumors (50%), with loss of staining throughout the tumor, compared with only 5 of 108 of the ARID1A-intact tumors (5%; P < .001). The remaining 4 tumors with complete ARID1A loss were conventional gland-forming adenocarcinomas (Figure 1, D), with no tumor-infiltrating lymphocytes or peritumoral lymphoid aggregates.

Loss of MLH1 expression was present in 2 of 12 cases (17%) with ARID1A loss but in none of the ARID1A-intact tumors. Interestingly, in one case, the MLH1 loss was noted in both the carcinoma and in the adjacent dysplasia, whereas the ARID1A loss was seen in the carcinoma component only (Figure 2). The other case with concordant MLH1 and ARID1A loss showed diffuse loss of MLH1 in the entire tumor and a discrete, confluent focus of ARID1A-deficient tumor cells.

ARID1A and MLH1 staining was patchy in nondysplastic and dysplastic BE, and negative surface epithelial staining was quite frequently associated with intact staining in the crypt base (Figure 3, A and B). Loss of ARID1A staining was not seen in any of the 43 foci of nondysplastic BE adjacent to the carcinoma. In the 64 cases (53%) with BE-associated low- or high-grade dysplasia, loss of ARID1A was seen in only one focus (1.5%) of low-grade dysplasia (Figure 3, C and D) separate from the main tumor mass. Thus, in most cases, loss of ARID1A was restricted to the carcinoma component only.

A mutant p53 staining pattern was seen in 52 of 120 EACs (43%), as well as in 3 of 17 foci (18%) of low-grade and 15 of 47 foci (32%) of high-grade dysplasia. A mutant p53 pattern of staining was not seen in any focus of nondysplastic BE. Six of 12 (50%) of the ARID1A-deficient tumors also showed a mutant staining pattern for p53 (Figure 4, A through D).

Clinicopathologic Correlates and Prognostic Significance of ARID1A Loss in EAC

As mentioned above, loss of ARID1A expression in EAC showed a significant association with a medullary and mucinous histology. All 3 EACs with a medullary phenotype and 3 of 7 mucinous EACs (43%) were ARID1A-deficient compared with only 6 of 110 conventional gland-forming adenocarcinomas (5.4%; Table 2). A trend was noted for increased risk of nodal metastasis in ARID1A-deficient tumors but was not statistically significant (P = .06). The overall survival of patients with ARID1A-deficient EAC was similar to that of those with intact ARID1A expression on both a Kaplan Meier (Figure 5; P = .10) and Cox proportional hazards analysis (Table 3).


ARIDlA functions as a tumor suppressor and has an important role in carcinogenesis in many organs. (27) There is a strong correlation between loss of ARID1A expression and DNA mismatch repair deficiency in carcinomas involving the stomach and colon, suggesting that ARID1A loss may be a secondary phenomenon related to impaired DNA repair and not a primary driver of carcinogenesis. (22,28,29) In our study, loss of ARID1A expression was present in 10% (12 of 120) of EACs and the majority of ARID1A-deficient EACs showed morphologic features similar to microsatellite unstable colorectal adenocarcinomas. Only 2 of 12 tumors (17%) with ARID1A loss showed a concurrent loss of MLH1, whereas a mutant p53 pattern was seen in 6 of 12 (50%) of the ARID1A-deficient tumors. These findings suggest that ARID1A is a primary driver of carcinogenesis in a subset of EACs and, unlike Mullerian carcinomas, abnormal ARID1A and p53 expression are not mutually exclusive in EAC.

Loss of DNA mismatch repair proteins with MSI-H occurs in only 3% to 6% of all EACs. (3) The MSI-H esophageal adenocarcinomas are morphologically heterogeneous but may show medullary or mucinous differentiation or marked tumor-infiltrating lymphocytes, similar to their colorectal counterparts. The association between ARID1A loss and DNA mismatch repair protein deficiency has been reported in gastric, colorectal, and Mullerian adenocarcinomas and shows some interesting site-specific correlations. (29) ARID1A mutation or loss of protein expression in gastric cancer shows a strong correlation with Epstein-Barr virus infection and microsatellite instability. (29,30) ARID1A is predominantly mutated by indels involving short mononucleotide repeats in MSI-H gastric carcinomas, but ARID1A mutations in microsatellite-stable gastric cancer seldom involve similar indels and are most often single-nucleotide variations, comprising nonsense or missense mutations. (31) Similar findings have also been reported in colorectal carcinoma. (22) Knockdown of ARID1A leads to inhibition of Fas-mediated apoptosis, (32) and it has been speculated that selection for ARIDlA mutations in MSI-H tumors may be due to its ability to inhibit apoptosis and thereby promote immune evasion from the abundant tumor-infiltrating lymphocytes that are a feature of microsatellite-unstable gastric carcinomas. (31) In contrast, MSI-H endometrial carcinomas do not show indels of short repeats within ARID1A. (33) The association between ARIDlA and MSI is not surprising because the ARIDlA coding region contains many short, 4 to 7 mononucleotide repeats. However, the mutation rate of ARIDlA in MSI-H gastric cancer is 12- to 60-fold higher than the global background somatic indel rate in mononucleotide repeats of similar length in other genes. (31) This finding indicates preferential selection of a driver gene in MSI-H gastric carcinomas and is similar to other genes, such as TGFBR2, which are selected in MSI-H colon cancer. (34) In addition, mutations in the switch/ sucrose nonfermentable family of chromatin-remodeling genes ARIDlA, SMARCA4, and ARID2 also occur in about 20% of microsatellite-stable EACs, supporting the role of ARIDlA as an independent tumor-suppressor pathway in esophageal carcinogenesis. (13) This is further supported by the reports of ARIDlA mutations not only in EACs but also in a subset of nondysplastic BE and BE-associated dysplasia. (35)

Our findings support the role of switch/sucrose nonfermentable chromatin regulators as important tumor suppressor genes (36) and highlight some features that distinguish ARID1A-deficient carcinomas in the esophagus from similar tumors at other sites. An inverse relationship has been reported between ARIDlA and TP53 mutations in gastric and colorectal carcinomas. (22,30,31) Similarly, TP53 mutations are uncommon in ovarian clear cell carcinomas and endometrioid endometrial carcinomas that carry ARIDlA mutations, whereas ovarian serous carcinomas that typically harbor TP53 mutations seldom show concurrent mutations in ARIDlA. (23) Interestingly, there was no difference in prevalence of a mutant p53 pattern in ARID1A-proficient (52 of 108; 48%) and ARID1A-deficient (6 of 12; 50%) tumors (P = .55). Similar findings were reported in a prior small series of EACs using next-generation sequencing, in which 10 of 13 ARIDlA-mutant EACs (77%) demonstrated concurrent TP53 mutations. (13)

Loss of ARID1A has been shown in precursor lesions adjacent to carcinomas at various sites, (37,38) suggesting that it is an early event in carcinogenesis. Similarly, loss of ARID1A expression was previously reported in 5% of nondysplastic BE, with increasing prevalence in BE with low- and high-grade dysplasia. (35) However, in our study, loss of ARID1A was restricted to foci of carcinoma, and only one focus of low-grade dysplasia showed complete, confluent loss of ARID1A staining. The reason for this discrepancy may be the strict definition used in our study to define ARID1A loss in nondysplastic and dysplastic epithelium. As illustrated in Figure 4, ARID1A staining in nondysplastic and dysplastic BE consistently shows a heterogeneous pattern and a significant reduction in the intensity of staining toward the surface epithelium. The latter finding is similar to that seen in DNA mismatch repair immunohistochemistry in healthy colonic mucosa. The similarity in reported prevalence of ARIDlA mutations (13,14) and the loss of protein expression in our cohort suggests that the strict criteria used in our study to define ARID1A loss were appropriate. A precise correlation of ARIDlA mutational status and protein expression in metaplastic and dysplastic background epithelium was beyond the scope of this study in which our primary aim was to determine the morphological correlates and prognostic significance of ARID1A loss in EAC.

There are some limitations to our study. We did not perform any mutational analysis in EAC cases that were stained for ARID1A. Although loss of protein expression certainly implies loss of function, there may be additional tumors with retained protein expression that harbor mutations or epigenetic changes with deleterious functional consequences. Secondly, we used primary resections to ensure we had adequate tissue material to evaluate ARID1A expression in EAC and in background columnar mucosa. The lack of prognostic significance for ARID1A-deficient tumors may be due to the small number of EACs with ARID1A loss in our study. Moreover, the prognostic findings may not be applicable to all patients with EAC because most cases are now treated with neoadjuvant therapy followed by surgery.

In summary, loss of ARID1A expression is seen in about 10% (12 of 120) of EACs, and most of those tumors show features similar to those described in MSI-H colorectal carcinomas. The ARID1A-deficient EACs are also more likely to show loss of MLH1 compared with ARID1A-intact tumors. The prevalence of a mutant p53 pattern is similar in EAC with and without the loss of ARID1A. Finally, in primary resections for EAC, loss of ARID1A expression carries no prognostic significance.

Please Note: Illustration(s) are not available due to copyright restrictions


(1.) Bosman FT, Carneiro F, Hruban RH, Theise ND, eds. WHO classification of Tumours of the Digestive System. 4th ed. Lyon, France: IARC Press; 2010. World Health Organization Classification of Tumours; vol 3.

(2.) Farris AB III, Demicco EG, Le, LP, et al. Clinicopathologic and molecular profiles of microsatellite unstable Barrett esophagus-associated adenocarcinoma. Am J Surg Pathol. 2011; 35(5):647-655.

(3.) Falkenback D, Johansson J, Halvarsson B, Nilbert M. Defective mismatchrepair as a minor tumorigenic pathway in Barrett esophagus-associated adenocarcinoma. Cancer Genet Cytogenet. 2005; 157(1):82-86.

(4.) Bian YS, Osterheld MC, Bosman FT, Benhattar J, Fontolliet C. p53 gene mutation and protein accumulation during neoplastic progression in Barrett's esophagus. Mod Pathol. 2001; 14(5):397-403.

(5.) Mokrowiecka A, Wierzchniewska-Lawska A, Smolarz B, Romanowicz-Makowska H, Malecka-Panas E. p16 gene mutations in Barrett's esophagus in gastric metaplasia--intestinal metaplasia--dysplasia--adenocarcinoma sequence. Adv Med Sci. 2012; 57(1):71-76.

(6.) Menke V, Pot RG, Moons LM, et al. Functional single-nucleotide polymorphism of epidermal growth factor is associated with the development of Barrett's esophagus and esophageal adenocarcinoma. J Hum Genet. 2012; 57(1):26-32.

(7.) Marx AH, Zielinski M, Kowitz CM, et al. Homogeneous EGFR amplification defines a subset of aggressive Barrett's adenocarcinomas with poor prognosis. Histopathology. 2010; 57(3):418-426.

(8.) Lagarde SM, ten Kate FJ, Richel DJ, Offerhaus GJ, van Lanschot JJ. Molecular prognostic factors in adenocarcinoma of the esophagus and gastroesophageal junction. Ann Surg Oncol. 2007; 14(2):977-991.

(9.) Prins MJ, Ruurda JP, van Diest PJ, van Hillegersberg R, ten Kate FJ. Evaluation of the HER2 amplification status in oesophageal adenocarcinoma by conventional and automated FISH: a tissue microarray study. J Clin Pathol. 2014; 67(1):26-32.

(10.) Kulke MH, Thakore KS, Thomas G, et al. Microsatellite instability and hMLH1/hMSH2 expression in Barrett esophagus-associated adenocarcinoma. Cancer. 2001; 91(8):1451-1457.

(11.) Muzeau F, Flejou JF, Belghiti J, Thomas G, Hamelin R. Infrequent microsatellite instability in oesophageal cancers. Br J Cancer. 1997; 75(9):1336-1339.

(12.) Sato F, Meltzer SJ. CpG island hypermethylation in progression of esophageal and gastric cancer. Cancer 2006; 106(3):483-493.

(13.) Dulak AM, Stojanov P, Peng S, et al. Exome and whole-genome sequencing of esophageal adenocarcinoma identifies recurrent driver events and mutational complexity. Nat Genet. 2013; 45(5):478-486.

(14.) Chong IY, Cunningham D, Barber LJ, et al. The genomic landscape of oesophagogastric junctional adenocarcinoma. J Pathol. 2013; 231(3):301-310.

(15.) Shain AH, Pollack JR. The spectrum of SWI/SNF mutations, ubiquitous in human cancers. PloS One. 2013; 8(1):e55119.

(16.) Wilson BG, Wang X, Shen X, et al. Epigenetic antagonism between polycomb and swi/snf complexes during oncogenic transformation [published correction appears in Cancer Cell. 2011; 19(1):153]. Cancer Cell. 2010; 18(4): 316-328.

(17.) Zhang X, Sun Q, Shan M, et al. Promoter hypermethylation of ARID1A gene is responsible for its low MRNA expression in many invasive breast cancers. PloS One 2013; 8(1):e53931.

(18.) Shain AH, Giacomini CP, Matsukuma K, et al. Convergent structural alterations define switch/sucrose nonfermentable (SWI/SNF) chromatin remodeler as a central tumor suppressive complex in pancreatic cancer. Proc Natl Acad Sci U S A. 2012; 109(5):E252-E259.

(19.) Jones S, Wang TL, Shih IeM, et al. Frequent mutations of chromatin remodeling gene ARID1A in ovarian clear cell carcinoma. Science. 2010; 330(6001):228-231.

(20.) Guan B, Mao TL, Panuganti PK, et al. Mutation and loss of expression of ARID1A in uterine low-grade endometroid carcinoma. Am J Surg Pathol. 2011; 35(5):625-632.

(21.) Allo G, Bernardini MQ, Wu RC, et al. ARID1A loss correlates with mismatch repair deficiency and intact p53 expression in high-grade endometrial carcinomas. Mod Pathol. 2014; 27(2):255-261.

(22.) Chou A, Toon CW, Clarkson A, et al. Loss of ARID1A expression in colorectal carcinoma is strongly associated with mismatch repair deficiency. Hum Pathol. 2014; 45(8):1697-1703.

(23.) Bosse T, Ter Haar NT, Seeber LM, et al. Loss of ARID1A expression and its relationship with PI3K-Akt pathway alterations, TP53 and microsatellite instability in endometrial cancer. Mod Pathol. 2013; 26(11):1525-1535.

(24.) Edge SB, Byrd DR, Compton CC, Fritz AG, Greene FL, Trotti A, eds. AJCC Cancer Staging Manual. 7th ed. New York, NY: Springer; 2010.

(25.) Lassus H, Leminen A, Lundin J, Lehtovirta P, Butzow R. Distinct subtypes of serous ovarian carcinoma identified by p53 determination. Gynecol Oncol. 2003; 91(3):504-512.

(26.) Yemelyanova A, Vang R, Kshirsagar M, et al. Immunohistochemical staining patterns of p53 can serve as a surrogate marker for TP53 mutations in ovarian carcinoma: an immunohistochemical and nucleotide sequencing analysis. Mod Pathol. 2011; 24(9):1248-1253.

(27.) Wu RC, Wang TL, Shih IeM. The emerging roles of arid1a in tumor suppression. Cancer Biol Ther. 2014; 15(6):655-664.

(28.) Inada R, Sekine S, Taniguchi H, et al. ARID1A expression in gastric adenocarcinoma: Clinicopathological significance and correlation with DNA mismatch repair status. World J Gastroenterol. 2015; 21(7):2159-2168.

(29.) Abe H, Maeda D, Hino R, et al. ARID1A expression loss in gastric cancer: pathway-dependent roles with and without Epstein-Barr virus infection and microsatellite instability. Virchows Arch. 2012; 461(4):367-377.

(30.) Zang ZJ, Cutcutache I, Poon SL, et al. Exome sequencing of gastric adenocarcinoma identifies recurrent somatic mutations in cell adhesion and chromatin remodeling genes. Nat Genet. 2012; 44(5):570-574.

(31.) Wang K, Kan J, Yuen ST, et al. Exome sequencing identifies frequent mutation of arid1a in molecular subtypes of gastric cancer. Nat Genet. 2011; 43(12):1219-1223.

(32.) Luo B, Cheung HW, Subramanian A, et al. Highly parallel identification of essential genes in cancer cells. Proc Natl Acad Sci USA. 2008; 105(51):20380-20385.

(33.) Cancer Genome Atlas Research Network; Kandoth C, Schultz N, Cherniack AD, et al. Integrated genomic characterization of endometrial carcinoma [published correction appears in Nature. 2013; 500(7461):242]. Nature. 2013; 497(7447):67-73.

(34.) Markowitz S, Wang J, Myeroff L, et al. Inactivation of the type II TGF-beta receptor in colon cancer cells with microsatellite instability. Science. 1995; 268(5215):1336-1338.

(35.) Streppel MM, Lata S, DelaBastide M, et al. Next-generation sequencing of endoscopic biopsies identifies arid1a as a tumor-suppressor gene in Barrett's esophagus. Oncogene. 2014; 33(3); 347-357.

(36.) Huang J, Zhao YL, Li Y, Fletcher JA, Xiao S. Genomic and functional evidence for an arid1a tumor suppressor role. Gene Chromosomes Cancer. 2007; 46(8):745-750.

(37.) Yamamoto S, Tsuda H, Takano M, Tamai S, Matsubara O. Loss of ARID1A protein expression occurs as an early event in ovarian clear-cell carcinoma development and frequently coexists with PIK3CA mutations. Mod Pathol. 2012; 25(4):615-624.

(38.) Ayhan A, Mao TL, Seckin T, et al. Loss of ARID1A expression is an early molecular event in tumor progression from ovarian endometriotic cyst to clear cell and endometrioid carcinoma. Int J Gynecol Cancer. 2012; 22(8):1310-1315.

Michael G. Drage, MD, PhD; Mingkhwan Tippayawong, MD; Agoston T. Agoston, MD, PhD; Yifan Zheng, MD; Raphael Bueno, MD; Jason L. Hornick, MD, PhD; Robert D. Odze, MD; Amitabh Srivastava, MD

Accepted for publication October 5, 2016.

Published as an Early Online Release April 25, 2017.

From the Departments of Pathology (Drs Drage, Tippayawong, Agoston, Hornick, Odze, and Srivastava) and Surgery (Drs Zheng and Bueno), Brigham and Women's Hospital, Boston, Massachusetts. Dr Drage is now with the Department of Pathology, University of Rochester Medical Center, Rochester, New York.

The authors have no relevant financial interest in the products or companies described in this article.

Presented in part at the 103rd Annual Meeting of the United States and Canadian Academy of Pathology; March 1-7, 2014; San Diego, California.

Reprints: Amitabh Srivastava, MD, Department of Pathology, Brigham and Women's Hospital, 75 Francis St, Boston, MA 02115 (email:

Caption: Figure 1. A through C, The morphological spectrum of ARID1A-deficient esophageal adenocarcinoma is similar to microsatellite-unstable colon cancer. A, Tubulovillous precursor lesions with conventional gland-forming adenocarcinoma and prominent peritumoral lymphoid aggregates (arrowheads). B, Medullary carcinoma with solid architecture and marked intratumoral lymphoid infiltration. C, Mucinous adenocarcinoma with abundant extracellular mucin. D, A minor subset of ARID1A-deficient tumors has conventional gland-forming adenocarcinomas that do not show any of the features described above. All tumors illustrated showed diffuse loss of ARID1A staining with intact nuclear positivity in the stromal cells (hematoxylin-eosin, original magnifications x12 [A], x100 [B and C], and x200 [D]).

Caption: Figure 2. In concurrent MLH1- and ARID1A-deficient tumors, the loss of MLHI expression appears to precede the loss of ARIDIA. In this example of dysplastic Barrett esophagus (left half of image) and intramucosal adenocarcinoma (right half; A), MLH1 expression is lost in both the dysplastic and carcinoma components (B), whereas ARID1A loss is limited to the carcinoma component only, with intact staining in the dysplastic epithelium (C) (hematoxylin-eosin, original magnification x100 [A]; MLH1, original magnification x100 [B]; ARID1A, original magnification x100 [C]).

Caption: Figure 3. A and B, ARID1A expression in nondysplastic Barrett esophagus is typically heterogeneous with consistent staining of the crypt bases and reduced or absent expression toward the surface epithelium. C and D, Loss of ARID1A staining was not seen in any focus of nondysplastic Barrett esophagus in our study and in only one focus of low-grade dysplasia (hematoxylin-eosin, original magnification x100 [A and C]; ARID1A, original magnification x100 [B and D]).

Caption: Figure 4. Loss of ARID1A expression is not exclusive to the p53 mutant pattern of staining in esophageal adenocarcinomas. A poorly differentiated adenocarcinoma with tumor infiltrating lymphocytes (A) shows loss of ARID1A expression (B), retained MLH1 expression (C), and diffuse, strong positivity for p53 immunohistochemistry (D) (original magnification x40 [A through D]).

Caption: Figure 5. Kaplan-Meier analysis of overall survival among patients with esophageal adenocarcinoma with retained or deficient ARID1A expression shows no difference in prognosis.
Table 1. Patient Demographics and Tumor
Characteristics of Study Group, n = 120

Characteristic                               Result

Age, y, mean (range)                       66 (30-87)
M:F ratio                                     4.2
  Tumor size, cm, mean (range)            3.1 (0.2-18)
Histologic type, No. (%)
  NOS                                     108 (90.0)
  Mucinous                                  7 (5.8)
  Medullary                                 3 (2.5)
  Signet ring cell                          2 (1.7)
Histologic grade, No. (%)
  Low                                      41 (34.2)
  High                                     79 (65.8)
Other features, No. (%)
  IELs, >3/HPF                             20 (16.7)
  LVI                                      46 (38.3)
  PNI                                      34 (28.3)
Tumor stage, No. (%)
  1a                                       31 (25.8)
  1b                                       39 (32.5)
  2                                        12 (10.0)
  3                                        38 (31.7)
Lymph node stage, No. (%)
  0                                        80 (66.7)
  1                                        22 (18.3)
  2                                        12 (10.0)
  3                                         6 (5.0)
Duration of follow-up, y, mean (range)    5.2 (< 1-22.4)
Overall mortality, No. (%)                 67 (55.8)

Abbreviations: HPF, high-power field; IEL, intraepithelial
lymphocyte; LVI, lymphovascular invasion; NOS, not otherwise
specified; PNI, perineural invasion.

Table 2. Clinicopathologic Correlates and Prognostic Significance
of ARID1A Loss in Esophageal Adenocarcinoma

Variable                    Total, No. (%),   ARID1A Deficient,
                               (n = 120)      No. (%), (n = 12)

Age, [greater than            72 (60)              7 (58)
  or equal to] 65 y
Male                          97 (81)             11 (92)
Mid/proximal location          4 (3)               2 (17)
Tumor size,                   68 (57)              7 (58)
  [greater than or
  equal to] 2.0 cm
High tumor grade              79 (66)             11 (92)
Histologic type
  NOS                        110 (92)              6 (50)
  Mucinous                     7 (6)               3 (25)
  Medullary                    3 (3)               3 (25)
  Signet ring cell             2 (2)                0
  IELs, >3/HPF                19 (16)              3 (25)
PTL                           28 (23)              5 (42)
LVI                           45 (38)              3 (25)
PNI                           33 (28)              4 (33)
pT stage, [greater than       88 (73)             10 (83)
  or equal to] 1b
pN stage, [greater than       39 (33)              7 (58)
  or equal to] 1
High stage, [greater than     52 (43)              8 (67)
  or equal to] IIB
MLH1 deficient                 2 (2)               2 (17)
p53 mutant pattern            52 (43)              6 (50)
Overall mortality             66 (55)              9 (75)

Variable                     ARID1A Retained,    OR (a)   P (a)
                            No. (%), (n = 108)

Age, [greater than              65 (60)            0.93   >.99
  or equal to] 65 y
Male                            86 (80)            2.81    .46
Mid/proximal location            2 (2)             9.8     .06
Tumor size,                     61 (56)            1.1     .24
  [greater than or
  equal to] 2.0 cm
High tumor grade                68 (63)            5.34    .10
Histologic type
  NOS                          104 (96)            0.04   <.001
  Mucinous                       4 (4)             8.67    .02
  Medullary                       0              >99       .001
  Signet ring cell               2 (2)             0       .81
  IELs, >3/HPF                  16 (15)            1.92    .40
PTL                             23 (21)            2.64    .15
LVI                             42 (39)            0.47    .36
PNI                             29 (27)            1.22    .75
pT stage, [greater than         78 (72)            1.92    .51
  or equal to] 1b
pN stage, [greater than         32 (30)            3.33    .06
  or equal to] 1
High stage, [greater than       44 (41)            2.91    .12
  or equal to] IIB
MLH1 deficient                    0              >99       .10
p53 mutant pattern              46 (43)            1.35    .76
Overall mortality               57 (53)            2.68    .22

Abbreviations: HPF, high-power field; IEL, intraepithelial
lymphocytes; LVI, lymphovascular invasion; NOS, not otherwise
specified; OR, odds ratio; PNI, perineural invasion; PTL, peritumoral
lymphoid aggregates.

(a) Fisher exact test for univariate comparison between
ARID1A-deficient and -intact esophageal adenocarcinoma.

Table 3. Prognostic Significance of Clinicopathologic Features and
Immunohistochemical Findings in Esophageal Adenocarcinoma, n = 120

Risk Factor                 Dead, No./Risk Factor
                              Present, No. (%)

Age, [greater than            43/72 (60)
  or equal to] 65 y
Male                          52/97 (54)
Location (mid/proximal)         2/3 (67)
Tumor size, [greater than     48/66 (73)
  or equal to] 2.0 cm
Poor differentiation          27/36 (75)
Histologic type
  Mucinous                      3/7 (43)
  Medullary                     3/3 (100)
  Signet ring cell              2/2 (100)
LVI                           37/46 (80)
PNI                           29/34 (85)
lELs, >3/HPF                  11/19 (58)
PTLs                          18/29 (62)
pT, [greater than or          34/38 (89)
  equal to] 3
N stage, [greater than        16/17 (94)
  or equal to] 2
High stage, [greater than     41/50 (82)
  or equal to] IIB
ARID1A deficient               9/12 (75)
MLH1 deficient                  0/2 (0)
p53 mutant pattern            31/52 (60)

Risk Factor                  HR       95% CI        P

Age, [greater than          1.75   1.043-2.930     .03
  or equal to] 65 y
Male                        0.75   0.418-1.334     .32
Location (mid/proximal)     1.16   0.283-4.761     .84
Tumor size, [greater than   3.31    1.93-5.70     <.001
  or equal to] 2.0 cm
Poor differentiation        2.56   1.560-4.209    <.001
Histologic type
  Mucinous                  0.95   0.297-3.037     .93
  Medullary                 3.93   1.224-12.588    .02
  Signet ring cell          3.66   0.881-15.202    .07
LVI                         3.05   1.910-5.120    <.001
PNI                         4.34   2.625-7.176    <.001
IELs, >3/HPF                0.93   0.472-1.834     .84
PTLs                        1.02   0.585-1.771     .95
pT, [greater than or        4.01   2.448-6.576    <.001
  equal to] 3
N stage, [greater than      3.83   2.144-6.841    <.001
  or equal to] 2
High stage, [greater than   2.88   1.749-4.747    <.001
  or equal to] IIB
ARID1A deficient            1.82   0.893-3.686     .10
MLH1 deficient              0.05    0.00-193.84    .47
p53 mutant pattern          1.24   0.760-2.020     .39

Abbreviations: HPF, high-power field; HR, hazard ratio; lELs,
intraepithelial lymphocytes; LVI, lymphovascular invasion; PNI,
perineural invasion; PTL, peritumoral lymphoid aggregates.
COPYRIGHT 2017 College of American Pathologists
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2017 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:AT-rich interactive domain 1A
Author:Drage, Michael G.; Tippayawong, Mingkhwan; Agoston, Agoston T.; Zheng, Yifan; Bueno, Raphael; Hornic
Publication:Archives of Pathology & Laboratory Medicine
Article Type:Report
Date:Jul 1, 2017
Previous Article:A New Kind of Autopsy for 21st Century Medicine.
Next Article:Solid Pseudopapillary Neoplasm of the Pancreas: A Rare Entity With Unique Features.

Terms of use | Privacy policy | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters