Clinical and pathologic features of young endometrial cancer patients with loss of mismatch repair expression

Gynecologic Oncology 126 (2012) 408–412

Contents lists available at SciVerse ScienceDirect

Gynecologic Oncology journal homepage: www.elsevier.com/locate/ygyno

Clinical and pathologic features of young endometrial cancer patients with loss of mismatch repair expression☆ Kassondra S. Grzankowski ⁎, David M. Shimizu, Chieko Kimata, Michael Black, Keith Y. Terada Department of Obstetrics, Gynecology, and Women's Health, John A. Burns School of Medicine, University of Hawaii, Honolulu, Hawaii, USA Department of Pathology, John A. Burns School of Medicine, University of Hawaii, Honolulu, Hawaii, USA

a r t i c l e

i n f o

Article history: Received 29 February 2012 Accepted 14 May 2012 Available online 20 May 2012 Keywords: Endometrial cancer Mismatch repair Body mass index MSH6/MSH2 MLH1/PMS2 Microsatellite instability

a b s t r a c t Objective. This study examines premenopausal and early menopause patients in a unique population with endometrial cancer and loss of mismatch repair (MMR) gene expression. The purpose is to compare clinical and pathologic differences in patients with loss of expression (LOE) to those with normal expression (NE). Methods. Endometrial cancer patients under age 60 in-between 1998 and 2008 were identified from a single tumor registry. Clinical and pathologic data were abstracted from records. Staining for expression of MSH6, MSH2, MLH1, and PMS2 were performed on archived tissue blocks. Statistical analysis was performed. Results. 158 patients were analyzed; 58% Asian, 34% Pacific Islander, and 8% Caucasian. 31 demonstrated LOE of at least one MMR gene; 127 retained NE. 50% Caucasian, 21.9% Asian, and 12.5% Pacific Island populations had LOE of one or more MMR genes. LOE was found to have a higher incidence of Grade III (p = 0.0013) and stage 3–4 tumors (p = 0.0079), mean depth of myometrial invasion (p = 0.0019), lymphovascular space invasion (p = 0.0020), nodal metastases (p = 0.0157), and a lower incidence of Grade I (p = 0.0020) and stage 1A tumors (p = 0.0085). LOE had a significantly lower mean BMI (p = 0.0001). 35% of patients in the NE vs zero in the LOE group had a BMI greater than 40. Conclusion. Younger patients with LOE endometrial cancer appear to represent a clinically significant subgroup of patients without features characteristically found in classic type 1 endometrial cancer generally demonstrating lower BMI and tumors associated with poor prognostic characteristics. It is unclear if the distinctive ethnicity found in Hawaii has a significant impact on outcome. Further investigation is necessary to identify appropriate treatment strategies. © 2012 Published by Elsevier Inc.

Introduction Endometrial carcinoma is the most common gynecologic malignancy in the United States with approximately 43,470 cases annually. American women have an estimated 2.5% lifetime risk of developing endometrial cancer. Well recognized risk factors for endometrial cancer in younger women include obesity, nulliparity, diabetes, anovulation, and hyperestrogenic conditions. There is an evolving appreciation, however, that loss of function of the mismatch repair (MMR) genes is associated with endometrial cancer in a small but distinct subpopulation of young patients [1]. Loss of MMR function may arise due to a germline mutation (Lynch syndrome or hereditary nonpolyposis colorectal cancer) or an acquired epigenetic defect.

☆ Funding: Funded in part by a grant from the Queen Emma Foundation. ⁎ Corresponding author at: John A. Burns School of Medicine, University of Hawaii, Department of Obstetrics, Gynecology, and Women's Health, 1329 Lusitana St. #703, Honolulu, HI 96813, USA. Fax: + 1 808 528 3671. E-mail address: [email protected] (K.S. Grzankowski). 0090-8258/$ – see front matter © 2012 Published by Elsevier Inc. doi:10.1016/j.ygyno.2012.05.019

Hereditary nonpolyposis colorectal cancer (HNPCC) is the most common hereditary cancer syndrome in the United States. The mutations occur in genes that encode for MLH1 on chromosome 3, PMS2 on chromosome 7, MSH2 on chromosome 2, and MSH6 on chromosome 2. The role of MMR genes is to maintain genomic integrity by correcting base substitution mismatches and small insertion–deletion mismatches that are generated by errors in base pairing during DNA replication. These function as pairs; MSH6/MSH2 and MLH1/ PMS2. Acquired defects may also occur, and are generally associated with hypermethylation of the MLH 1 gene promoter region. It is estimated that approximately 20–30% of endometrial cancers have loss of MMR gene expression. Five percent appear to be attributable to germline mutations, HNPCC; the remainder appears to arise due to hypermethylation of the MLH1 promoter region. There is some data to suggest that loss of MMR function is associated with a distinct phenotype. Such patients lack the usual characteristics of obesity and hyperestrogenism [2]. These malignancies also appear to be associated with a high risk histopathologic profile [3]. Shih et al. reported that for a group of women under the age of 40 with endometrial cancer, loss of MMR expression was noted in

K.S. Grzankowski et al. / Gynecologic Oncology 126 (2012) 408–412

16% of patients who were more likely to have high grade and advanced stage malignancies [4]. Endometrial cancer has historically been subdivided into type 1 and type 2 phenotypes. Type 1 cancers are generally found in younger women and associated with obesity and hyperestrogenic conditions. These malignancies are typically low grade endometrioid cancers with a favorable outcome. Type 2 cancers arise in older women, are often associated with papillary serous or clear cell histology, and typically have a poor prognosis. The current study was a retrospective examination of endometrial cancer in a group of younger women. This population includes premenopausal women and women in early menopause that would generally be associated with type 1 cancers. The purpose of the study was to examine within this population the incidence of MMR gene associated loss of protein expression. Methods After obtaining the Institutional Review Board approval, the tumor registry for Queens Medical Center was used to identify patients less than age 60 (18–59) with endometrial cancer. All patients were treated by a single Gynecologic Oncologist (KYT) and all pathologic specimens were reviewed by a single gynecologic pathologist (DMS). An 11 year period was reviewed (January 1, 1998–December 31, 2008). Patients were excluded if there was inadequate tissue available for immunohistochemical analysis. Patients with pure sarcomas were also excluded. All adenocarcinomas, including serous cancers, undifferentiated cancers, and carcinosarcomas were included. 158 patients were identified for inclusion in the study. Immunohistochemistry was performed using antibodies to MLH1 (clone G168-728; BD Biosciences, San Jose, CA; 1:40 dilution), MSH2 (clone FE11; Calbiochem, San Diego, CA; 1:20 dilution), MSH6 (clone 44; BD Biosciences, San Jose, CA; 1:400 dilution), and PMS2 (clone A16-4; BD Biosciences, San Jose, CA; 1:50 dilution). In brief, detection was obtained using diaminobenzidine after a polymer based amplification step (Envision Plus, Dako, Carpinteria, CA) for MLH1 and MSH2 and Mach 2 polymer (Biocare Medical, Concord, CA) for PMS2 for antibodies run on the Dako Autostainer (Dako, Carpinteria, CA). Detection for MSH6 (run on the Ventana BenchMark autostainer, Ventana, Tucson, AZ) was obtained using diaminobenzidine and avidin-biotin complex methodology. Nuclear staining of normal lymphocytes and/or stromal cells in each slide served as a positive internal control. A retrospective review of medical records for all patients was performed, and pertinent demographic, clinical, and pathologic data were abstracted. All patients were staged using Federation of Gynecology and Obstetrics (FIGO) criteria prior to 2009. Appropriate statistical analysis of the data was performed. T‐Test was used for continuous variables including age, average tumor size, and body mass index (BMI). Mann Whitney Wilcoxon test was utilized to analyze depth of myometrial invasion. Fisher's Exact Test was used for nominal data including sample size, histological grade, stage, positivity of lymph nodes, lymphovascular space invasion, endometrioid histology, positive cytology, estrogen and progesterone receptor status, family history of cancer, smoking history, nulliparity, hypertension history, and diabetes mellitus history. Statistical significance is defined as p b 0.05.

Results Table 1 demonstrates the clinical and pathologic characteristics of the 158 patients. Table 2 summarizes the ethnic distribution of the patient population by NE and LOE. Thirty-one patients demonstrated loss of MMR expression in the primary tumor and 127 demonstrated normal MMR expression. The mean age was 48 years for the cohort with normal MMR expression and 50 for the cohort with loss of MMR expression;

409

Table 1 Clinical and pathologic characteristics.

Sample sizea Age at diagnosis Histological grade Grade I Grade II Grade III Stage Stage 1A Stage 1B–2 Stages 3–4 Depth of myometrial invasion (%) BMI Under weight b 18.5 18.5 ≤ Normal range b 25 25 ≤ Overweight b 30 30 ≤ Class 1 obesity b 35 35 ≤ Class 2 obesity b 40 40 ≤ Class 3 obesity Average tumor size (cm) Positive lymph node metastasisb Positive lymphovascular space Endometrioid histologyc Positive cytology ER positive tumor PR positive tumor Smoking history Family history of breast, colon, ovarian, endometrial cancer Nulliparity HTN history DM history

MMR loss of expression

MMR normal expression

p Value

31/158a 50.39 (SD 7.74)

127/158 48.09 (SD 7.98)

0.0001 0.1499

10/31 7/31 14/31

82/127 24/127 21/127

0.0020 0.6212 0.0013

9/31 9/31 13/31 44.74 (median 46, range 0–100, with 8/31 = 0%) 27.00 (SD 6.61) 1 (3%) 14 (45%)

72/127 32/127 23/127 22.21 (median 0, range 0–100, with 68/126 = 0%) 36.19 (SD 11.53) 0 (0%) 23 (18%)

0.0085 0.6534 0.0079 0.0019

4 (13%) 9 (29%)

24 (19%) 19 (15%)

3 (10%)

17 (13%)

0 (0%) 3.94 (SD 2.41) 11/29

44 (35%) 4.05 (SD 2.57) 12/80

0.8291 0.0157

13/31

19/127

0.0020

23/31 3/31 21/24 20/24 5/31 8/31

110/127 13/127 67/71 63/71 17/127 21/127

0.1028 1.0000 0.3637 0.4911 0.7724 0.2990

14/31 9/31 3/31

60/89 57/127 30/127

0.0336 0.1545 0.1368

0.0001

Bold = statistical significance. a 21 demonstrated loss of MLH1/PMS2 and 10 patients demonstrated loss of MSH6/MSH2. b Pelvic and/or para-aortic lymph nodes. c MMR loss of expression: 2/31 carcinosarcoma, 4/31 serous, 2/31 undifferentiated; MMR normal 2/127 carcinosarcoma, 6/127 mucinous, 3/127 serous, 6/127 undifferentiated, 1/127 mixed.

this difference was not statistically significant. The LOE population demonstrated a statistically significant difference from the NE population with greater incidence of Grade III tumors (p = 0.0013) and stage 3–4 cancers (p = 0.0079), lower incidence of Grade I tumors (p = 0.0020) and stage 1A tumors (p = 0.0085), greater depth of myometrial invasion (p = 0.0019), greater incidence of lymph node metastases (p = 0.0157), and positive lymphovascular space invasion (p = 0.0020). There was no significant difference in histologic type, peritoneal cytology, or estrogen/progesterone receptor status. There was also no difference in the incidence of hypertension, diabetes, or family history of colon or gynecologic malignancy. Lower BMI was noted in the LOE group (p = 0.0001). At the time of diagnosis 48%of study population had an underweight to normal BMI in the LOE group versus 23% in the NE group. A BMI greater than 40 was observed in zero of the LOE group versus 35% of the NE group. A sub-analysis was performed comparing characteristics between patients with loss of individual MSH6, MSH2, MLH1, and PMS2. These results are summarized in Table 3. No significant difference was noted in these four groups. If sub-analysis was performed as two groups, MLH1/PMS2 and MSH6/MSH2, no difference was

410

K.S. Grzankowski et al. / Gynecologic Oncology 126 (2012) 408–412

Table 2 Ethnicity in study population (158 total patients). Ethnicity Caucasian (8%) Asian (58%) Japan China Korea Vietnam Taiwan Other Pacific Islands (34%) Hawaiian Islands Philippines Samoa Micronesia Mariana Islands Tonga Other Total

Normal expression

MLH1/PMS2 loss of expression

MSH6/MSH2 loss of expression

7

4

3

39 5 1 0 1 4

4 1 0 1 0 1

4 2 0 0 0 1

27 21 10 7 2 1 2 127

3 6 0 1 0 0 0 21

0 0 0 0 0 0 0 10

50% of Caucasian, 21.9% of Asian, and 12.5% of Pacific Island population had loss of protein expression of one or more MMR genes.

seen other than patients with MLH1/PMS2 LOE appeared to have a larger tumor size of borderline significance (p = 0.0476).

Discussion Historically endometrial cancer has been subdivided into two broad subtypes: type 1 and type 2. Type 1 tumors are generally associated with younger patients, obesity, nulliparity, late menopause, and excess estrogen states. The tumors are characteristically endometrioid histology, low grade, and associated with favorable prognostic features. Type 2 cancers typically arise in older women and are associated with clear cell and serous histology and typically exhibit poor prognostic features. The current study defines a

unique group of patients with endometrial cancer. These women are pre-menopausal or peri-menopausal with tumors characterized by loss of MMR gene expression. The women tend to be nonobese and have high grade endometrioid tumors that are associated with deep myometrial invasion and lymphatic invasion. This group comprised 17% of pre- and peri-menopausal patients with endometrial cancer in our study. The DNA MMR mechanism primarily involves two functional pairs of genes. MSH 6 and MSH 2 recognize and bind to mismatched DNA sequences. Following this MLH 1 and PMS 2 excise and repair the mismatched nucleotides. Loss of function of DNA MMR genes has been reported in up to 20–30% of patients with endometrial cancer [1,5]. Inability to repair DNA errors may result in impaired function for critical regulatory pathways. Loss of these pathways, integral for regulating cell growth and division, results in carcinogenesis. Loss of MMR function may result from germline mutations in MMR genes, commonly referred to as Lynch syndrome or hereditary nonpolyposis colorectal cancer/HNPCC. These individuals are at increased risk for a variety of malignancies. It is well recognized that endometrial cancer is often the sentinel cancer; and it is reported that approximately 9% of patients under the age of 50 with endometrial cancer harbor one of these germline mutations [6]. Loss of MMR function may also result from epigenetic alterations resulting in LOE. This involves methylation of the MLH 1 promoter region, with subsequent LOE. Loss of MMR expression may also result from spontaneous mutations in MMR genes, but this is probably an uncommon event. It has been reported that the majority of gynecologic tumors with MLH 1 LOE are due to acquired hypermethylation, rather than germline mutation [7]. Our study did not include testing for hypermethylation or germline mutations; therefore we did not specifically address the underlying cause for loss of MMR function. The data did not demonstrate any difference in phenotype between MLH 1/PMS 2 losses and MSH 6/MSH 2 losses. The current study did not demonstrate a significant difference in mean age between patients with MLH 1/PMS 2 losses and MSH

Table 3 Comparison of patient characteristics with MLH1, PMS2, MSH6, and MSH2 gene loss of protein expression (31 patients).

Sample size (n = 55) Age at diagnosis Histological grade Grade I Grade II Grade III Stage Stage 1A Stage 1B–2 Stages 3–4 Average depth of myometrial invasion (%) BMI Average tumor size (cm) Positive lymph node metastasis Positive lymphovascular space Endometrioid histology Positive cytology ER positive tumor PR positive tumor Smoking history Family history of breast, colon, ovarian, endometrial cancer Nulliparity HTN history DM history a b c d

Presumed defect in MLH1 (loss in MLH1/PMS2). Presumed defect in PMS2 (loss in PMS2 only). Presumed defect in MSH6 (loss in MSH6 only). Presumed defect in MSH2 (loss in MSH6/MSH2).

MLH1a Loss of expression

PMS2b Loss of expression

MSH6c Loss of expression

MSH2d Loss of expression

18 47.9 (SD 8.59)

3 55 (SD 1.73)

4 53.5 (SD 4.65)

6 53.3 (SD 6.47)

6 4 8

2 1 0

2 2 0

0 0 6

5 4 9 49.7 27.4 (SD 6.69) 4.76 (SD 2.26) 9/17 10/18 13/18 2/18 12/14 13/14 5/18 3/18 11/18 6/18 3/18

2 0 1 31.3 32.2 (SD 6.97) 3.09 (SD 4.02) 1/3 1/3 3/3 1/3 1/1 1/1 0/3 3/3 0/3 3/3 0/3

2 2 0 36.5 30.1 (SD 5.56) 3.93 (SD 1.84) 0/4 1/4 4/4 0/4 3/3 3/3 0/4 1/4 1/4 0/4 0/4

0 3 3 42 21.1 (SD 2.45) 1.88 (SD 0.99) 1/5 1/6 3/6 0/6 5/6 3/6 0/6 1/6 2/6 0/6 0/6

K.S. Grzankowski et al. / Gynecologic Oncology 126 (2012) 408–412

6/MSH 2 loss. Patients with MLH 1/PMS 2 loss were generally younger with a mean age of 48.9 versus a mean age of 53.4 for patients with MSH 6/MSH 2 loss; this was not a statistically significant difference. Zauber et al. reported that 62% of patients with endometrial cancer and loss of MMR expression under the age of 50 had unmethylated tumors, suggesting a germline mutation. Conversely only 17% of patients over the age of 50 had unmethylated tumors, suggesting an acquired epigenetic etiology for the majority of older patients [8]. Intuitively, one might expect the incidence of acquired defects to increase with increasing age. Their study examined patients up to age 90; the upper age limit for the current study was 59 since the focus was on pre-menopausal and perimenopausal patients. It is likely that acquired LOE increases in the 60–90 year age range, thereby explaining the difference in findings. Earlier studies on tumors arising in individuals with Lynch syndrome indicate that these cancers tend to be of endometrioid histology, higher grade, and with a higher risk of lymphatic invasion and tumor infiltrating lymphocytes [9]. Shih et al. reported that patients under the age of 40 with loss of MMR gene expression demonstrated poor prognostic features and had a worse survival than patients with intact MMR expression [4]. The current study confirms and expands these findings. Patients with MMR loss were found to have higher grade, stage, increased incidence of deep myometrial invasion, increased incidence of lymphovascular space invasion, and increased incidence of lymph node metastases. The majority of tumors were of endometrioid histology, similar to tumors with intact MMR function in this younger population. Interestingly, there was no difference in the incidence of positive cytologic washings between the two groups. The current study did not have sufficiently long follow-up to examine survival, and only two patients are deceased from unrelated causes. The current study also compared body mass index (BMI) for patients with tumors expressing MMR function versus patients with loss of MMR expression. Patients with LOE had a significantly lower BMI with an average BMI of 27 (normal) versus an average BMI of 36 (obese) in patients with normal MMR expression. Win et al. reported similar findings in patients with germline mutations in MMR genes [2]. They suggested that tumors arising from loss of MMR function were independent of estrogen stimulation. Our study, however, notes an equal incidence of estrogen receptor positivity in both groups of patients. Also, unlike type 2 tumors which are predominantly serous and clear cell histology and receptor negative, these are predominantly endometrioid tumors. Therefore the role of estrogen in endometrial carcinogenesis for patients with loss of MMR function remains unclear. This study is unique regarding the ethnic distribution of the study population. This is largely a reflection of the ethnic diversity present in the state of Hawaii. Some authors have noted ethnic differences in MMR expression among Hispanic with decreased incidence of MLH1 methylation [11] and among Middle Eastern population [12]. 50% of Caucasian, 21.9% of Asian, and 12.5% Pacific Island population had loss of protein expression of one or more MMR genes. Similarities in histopathologic characteristics of the tumors in this small subset resemble the findings in larger studies of non-Hawaiian populations. More patients are needed in these ethnic categories to clearly define any differences in tumor characteristics, disease progression, or survival. The current study clearly demonstrates that pre-menopausal and peri-menopausal patients with loss of MMR gene expression represent a distinct subgroup of patients. These patients present with high grade tumors and with poor prognostic features. Presumably, loss of MMR function will eventuate in other mutations in critical regulatory pathways. Whether specific genetic pathways are involved, however, remains unclear. Although the PTEN pathway has been implicated in the development of endometrial cancer, it appears that mutations in the PTEN pathway are not associated with MMR

411

mutations [10]. Furthermore, while BRAF mutations are associated with MLH1 loss in colorectal cancer, this appears not to be the case with endometrial cancer [7]. Therefore it remains unclear whether there are additional specific regulatory pathways involved in carcinogenesis for this group of patients, or whether loss of the MMR machinery results in a more global loss of cell regulation. Our study did not specifically examine outcome or survival. Therefore we cannot determine whether MMR loss might be independently associated with prognosis. Nonetheless, the association of MMR loss with poor prognostic features clarifies the empiric observation that a small but distinct group of younger women with endometrial cancer does not exhibit the typical type 1 features. Currently our institution routinely screens patients under the age of 50 with endometrial cancer for loss of mismatch repair expression. There is considerable data to suggest that this more accurately identifies individuals at risk for Lynch syndrome than the Amsterdam or Bethesda criteria. Expanding within the unique population found in the Hawaiian Islands offers insight into a population not usually studied. Further utilization of data from multiple sites with different ethnic composition will ultimately provide data that can be used to stratify populations by risk and potentially be used as a screening tool. Screening modalities are regularly trying to be identified. Since endometrial biopsy can be readily obtained, one may ask if MMR LOE of NE on EMB can be utilized to stratify patients into ‘high risk’ and ‘low risk’ groups. Our data suggests that loss of MMR expression is associated with a high risk profile. Nonetheless, normal MMR expression should not be seen as ‘low risk.’ Lymph node positivity in this group was 12/80 (15%). This incidence of positive lymph nodes is similar to that seen in GOG, ASTEC, and PORTEC trials. There is insufficient data at this time to identify a truly low risk group. In conclusion, our data indicates that patients with endometrial cancer with loss of MMR expression represent a clinically significant subgroup of patients. These tumors arise in young patients and appear to be associated with poor prognostic features. Further study is needed to determine whether prognosis or outcome is different for these individuals and what type of treatment strategies are effective in managing these patients. Role of funding source A grant from the Queen Emma Foundation helped supply the immunohistochemical staining materials. Conflict of interest statement The authors have no conflict of interest to declare.

References [1] Matthews KS, Estes JM, Conner MG, Upender M, Whitworth JM, et al. Lynch syndrome in women less than 50 years of age with endometrial cancer. Obstet Gyncol 2008;111:1161–6. [2] Win AK, Dowty JG, Antill YC, English DR, Baron JA, et al. Body mass index in early adulthood and endometrial cancer risk for MMR gene mutation carriers. Obstet Gynecol 2011;117:899–905. [3] Broaddus R, Lynch H, Chen L, Daniels M, Conrad P, et al. Pathologic features of endometrial carcinoma associated with HNPCC, a comparison with sporadic endometrial carcinoma. Cancer 2006;106:87–94. [4] Shih KK, Garg K, Levine DA, Kauff N, Abu-Rustum NR, et al. DNA MMR protein defects and endometrial cancer in women age 40 and younger. J Clin Oncol 2009:27 15 s, abstract 5579. [5] Hampel H, Frankel W, Panesc J, Lockman J, Sotamaa K, et al. Screening for Lynch Syndrome (hereditary nonpolyposis colorectal cancer) among endometrial cancer patients. Cancer Res 2006;66:7810–7. [6] Lu KH, Schorge JO, Rodabaugh KJ, Daniels MS, Sun CC, et al. Prospective determination of prevalence of Lynch syndrome in young women with endometrial cancer. J Clin Oncol 2007;25:5158–64. [7] Kitayama J, Longacre TA, Liou S, Ma L, Arber D, et al. MLH-1 deficient ovarian and endometrial carcinomas most often result from epigenetic silencing of MLH 1 promoter hypermethylation and do not harbor BRAF V600E mutations:

412

K.S. Grzankowski et al. / Gynecologic Oncology 126 (2012) 408–412

Implications for identifying patients with Lynch Syndrome. Mod Pathol 2011;24(supplement 1):252A. [8] Zauber NP, Denehy TR, Taylor RR, Ongcapin EH, Marotta SP, et al. Microsatellite instability and DNA methylation of endometrial tumors and clinical features in young women compared with older women. Int J Gynecol Cancer 2010;20: 1549–56. [9] Walsh MD, Cummings MC, Buchanan DD, Dambacher WM, Arnold S, et al. Molecular, pathologic, and clinical features of early-onset endometrial cancer: identifying presumptive Lynch syndrome patients. Clin Cancer Res 2008;14: 1692–700.

[10] Mackay H, Gallinger S, Tsao M, McLachlin CM, Kieser K, et al. Microsatellite instability and loss of PTEN expression in early versus late-stage endometrial cancer: results from studies of the National Cancer Institute of Canada Clinical Trials Group (NCIC-CTG). J Clin Oncol 2009:27 15 s, abstract 5534. [11] De Jesus-Monge WE, Gonzalez-Keelan C, Zhao R, Hamilton SR, Rodriguez-Bigas M, Cruz-Correa M. Mismatch repair protein expression and colorectal cancer in Hispanics from Puerto Rico. Fam Cancer Jun 2010;9(2):155–66. [12] Chan AO, Soliman AS, Zhang Q, Rashid A. Differing DNA methylation patterns and gene mutation frequencies in colorectal carcinomas from the Middle Eastern countries. Clin Cancer Res Dec 2005;1(11(23)):8281–7.