Colonic carcinoma, one of the leading causes of cancer-related death worldwide (Jemal et al., 2006), is a biologically heterogeneous disease, and this heterogeneity affects its prognosis and clinical management (Kocarnik et al., 2015).
Whereas the majority of colorectal carcinomas (CRCs) follow the classical adenoma–carcinoma sequence, which is commonly accompanied by underlying chromosomal instability (Ionov et al., 1993; Lengauer et al., 1997; Lindor et al., 2002; Boland et al., 2008; Sinicrope et al., 2011), a subset of about 15% of CRCs acquire genetic alterations as a consequence of defective DNA mismatch repair (MMR) system (Ionov et al., 1993; Boland and Goel, 2010). MMR-proficient (pMMR) CRCs are mostly microsatellite stable or exhibit low-frequency MSI (Ionov et al., 1993). Conversely, MMR-deficient (dMMR) tumours are mostly chromosomally stable but exhibit numerous insertion or deletion mutations at short repetitive microsatellite DNA stretches (Boland and Goel, 2010), because of an inability to repair single-nucleotide DNA mismatches, resulting in inactivating mutations in multiple genes (Sinicrope et al., 2011), a phenotype termed high-level microsatellite instability (MSI-H) (Kloor et al., 2014).
MSI-H CRCs develop as either sporadic tumours (about 12% of all CRCs) (Boland and Goel, 2010; Jasperson et al., 2010) or in a hereditary context (about 3% of all CRCs) (Boland and Goel, 2010; Jasperson et al., 2010; Kaur et al., 2011). The DNA MMR deficiency in sporadic MSI-H CRCs is due to epigenetic inactivation – hypermethylation of the MLH1 promoter – of the MLH1 gene resulting in reduction or loss of MLH1 expression (Lynch and de la Chapelle, 2003; Poynter et al., 2008; Boland and Goel, 2010; Jasperson et al., 2010). CRC in younger patients with the MSI-H phenotype is closely related to Lynch syndrome [hereditary nonpolyposis cancer of the colon (HNPCC)] (Boland and Goel, 2010; Jasperson et al., 2010). In HNPCC the MMR deficiency is due to germline inactivation in any of the MMR genes (Lynch and de la Chapelle, 2003; Poynter et al., 2008) followed by somatic inactivation of the remaining functional allele (Kloor et al., 2014).
Defining MSI-H CRCs has its clinical implications, as these tumours are less aggressive, present at a lower stage, and characteristically respond to certain adjuvant chemotherapy (Kaur et al., 2011).
As MMR gene mutations usually lead to the absence of a detectable gene product with loss of immunohistochemical (IHC) protein expression, IHC – which is easier to perform and less expensive compared with MSI testing – provides gene-specific information that can highlight the need for further genetic analysis (Kheirelseid et al., 2013).
Unfortunately, no data are available regarding the prevalence of MMR gene defects among Egyptian CRCs, and the incidence of HNPCC is not documented. The current study was undertaken to study the MMR gene status, by testing the IHC expression of the MMR proteins MSH2, MSH6, MLH1 and PMS2, and to determine the frequency of MMR gene protein expression defects among CRC patients in our hospital, which is a referral centre in the city of Alexandria, Egypt. A second aim was to study the association of abnormal MMR protein expression with CRC clinicopathological characteristics.
Materials and methods
This retrospective study included 56 consecutive colectomies of pathologically confirmed colorectal adenocarcinomas submitted to the Pathology Department, Faculty of Medicine, Alexandria University, from December 2012 to March 2015. Cases that received preoperative downstaging chemotherapy were excluded, and only cases with available tumour tissue specimens sufficient for IHC studies were enroled in our study. The Alexandria University Faculty of Medicine Research Ethics Committee approved the study.
Clinical and pathological data were abstracted from the pathological and clinical records. The primary tumour site was categorized as proximal colon if the tumour was located above the splenic flexure or distal colon if it was located at or below the splenic flexure (Sinicrope et al., 2011).
All specimens were from Egyptian patients (21 men and 35 women), ranging in age from 19 to 95 years (M=52.9, SD=14.2). Of the 56 tumours, 18 (32.1%) were proximal to the splenic flexure, and 38 (67.9%) were distal. The rectosigmoid was the most common site of cancer (22/56, 39.3%), followed by the descending colon (16/56, 28.6%), ascending colon (8/56, 14.3%), caecum (7/56, 12.5%) and transverse colon (3/56, 5.3%).
Haematoxylin and eosin (H&E)-stained sections were retrieved and evaluated for different histological features, including tumour grade following the WHO criteria based on the percentage of glands in the tumour (>95% well, 50–95% moderate and <50% poor differentiation), and morphological pattern (glandular, mucinous or solid). Tumours were categorized as glandular (adenocarcinomas) when gland formation was noted without excessive mucus production, as mucinous when more than 50% of the tumour was composed of extracellular mucin pools containing malignant epithelium, and as adenocarcinoma with mucinous component if less than 50% of the lesion was composed of mucin. Solid differentiation was assigned when sheets, cords or trabeculae of malignant cells in the absence of glandular formation were observed. To score the morphology, an average of three H&E-stained slides with tumour per case were examined. Moreover, all resected lymph nodes were examined for metastatic deposits with observation of the morphologic pattern according to the previous definitions following the latest WHO criteria.
Sections from the primary tumour were also assessed for additional pathological features according to published data and without knowledge of the status of the specimen. These features included the following: (a) tumour margin classified as either expanding (pushing) or infiltrating (Jass et al., 1996); (b) presence of features of poor differentiation (whether or not this was the major component), including poor glandular formation with epithelial cells being arranged in small and irregular clusters, or as single cells as in signet ring cell carcinoma, or as solid sheets or in the form of trabeculae or islands (Bosman et al., 2010); (c) mucinous carcinoma in which at least 50% of the tumour comprised lakes of mucin (Bosman et al., 2010); (d) peritumoral lymphocytes noted as a cap or mantle of chronic inflammatory cells at the deepest point of direct spread (Bosman et al., 2010); (e) Crohn’s-like infiltrate scored on the basis of finding within a single ×4 field of at least three nodular aggregates of lymphocytes deep to the advancing margin of the tumour (Graham and Appelman, 1990); and (f) presence of tumour heterogeneity defined as detecting two or more distinct subclones identified on the basis of tumour type, grade of differentiation, or with a distinctive architecture (Alexander et al., 2001).
Amsterdam II criteria (Ward et al., 2005) and Bethesda guidelines (Hampel et al., 2005) were not fully implemented because of several limitations such as lack of family history, particularly for cancers that are potentially related to HNPCC other than CRC.
After reviewing the H&E-stained sections of the primary tumours, one paraffin block showing full thickness of the tumour with adjacent normal colonic mucosa was selected for IHC studies.
Immunohistochemical analysis of mismatch repair expression
The selected tumour tissue blocks were cut into 4-μm-thick sections on Superfrost/Plus slides (Thermo Scientific, Fremont, California, USA). For antigen retrieval the slides were heated in 0.01 mol/l citrate buffer (pH 6.0) in a microwave for 20 min. Endogenous peroxidase activity was blocked using 3% H2O2. Primary antibodies were then applied and incubated with tissue sections in a humidity chamber overnight. Staining was performed using ready-to-use mouse monoclonal primary antibodies; MSH2 [Ab-1 (clone 25D12); Lab Vision Corporation, Fremont, California, USA], PMS2 (clone A16–4), MSH6 (clone BC/44) and MLH1 (clone G168–15) were supplied by Biocare Medical Inc. (Concord, California, USA). Antigen visualization was done using the Thermo Scientific UltraVision LP Detection System (Thermo Scientific). IHC reactions were developed with diaminobenzidine, and sections were counterstained with Harrris haematoxylin. All immunostains were manually processed. Appropriate positive (adjacent normal colonic mucosa) and negative (omission of the primary antibody) controls were included for each IHC run.
Scoring of the immunostained slides
Evaluation of the immunostained slides was performed blindly and independently by two pathologists (S.E. and S.N.). Loss of MMR protein expression (indicating underlying abnormal gene expression) was defined as complete absence of nuclear staining within the carcinoma cells, whereas MMR protein expression was defined as the presence of nuclear staining within tumour cells regardless of its intensity or the number of positive nuclei (Kheirelseid et al., 2013). The adjacent non-neoplastic colon and stromal inflammatory cells served as internal positive control.
Data were analysed using the statistical package for social sciences (version 20; SPSS Inc., Chicago, Illinois, USA). Quantitative data were described using mean and SD. Qualitative data were described using number and percentage. The association between two qualitative variables in 2×2 tables was determined using Pearson’s χ2-test. If more than 20% of cells had an expected cell count less than 5, we used Fisher’s exact significance test (in 2×2 table) and Monte Carlo significance test (in >2×2 table). Multiple response was used for data that had more than one response as tumour pattern and nodal pattern. In all statistical tests, a level of significance of 0.05 was used, below which the results were considered statistically significant.
Most tumours (n=44, 78.6%) were classified as conventional adenocarcinomas (eight well, 34 moderate, and four poorly differentiated). Eight (14.3%) tumours were of the mucinous type, and four (9.1%) were signet ring carcinomas. Seven (12.5%) cases were T2, 41 (73.2%) were T3, and eight (14.3%) cases were T4. Regarding the nodal stage, 34 (60.7%) cases were N0, 14 (25%) cases were N1, and eight (14.3%) were N2.
Protein expression of MLH1, MSH2, MSH6 and PMS2 antigens was evaluated in all 56 studied CRCs. MLH1 revealed heterogeneous staining pattern. MSH2, PMS2 and MSH6 produced good nuclear staining, which was associated with weak cytoplasmic staining in some of the cases. Normal colon and stromal inflammatory cells were used as internal positive controls.
Eight (14.3%) of the 56 studied CRCs revealed an absence of protein expression of one or more of the four MMR genes. Out of those eight cases, isolated loss of a single MMR protein was noted in five cases (MLH1 protein expression was lost in a single case, MSH2 in two cases, and PMS2 in two cases). Isolated loss of MSH6 expression was not noticed in any of our studied cases. Three cases revealed concurrent loss of MSH2 and MSH6. Table 1 summarizes the MMR staining and its interpretation in the studied cases.
MMR defects due to germline mutation of MSH2 were the most common among our cases (n=5, 62.5%), as combined loss of MSH2 and MSH6 was noted in 37.5% of cases (n=3), (Fig. 1) and isolated loss of MSH2 was noted in 25% (n=2). Loss of PMS2 expression was detected in 25% of dMMR cases (n=2) (Fig. 2) and the least common was the isolated loss of MLH1 expression (n=1, 12.5%) (Fig. 3). The relation between the clinicopathological features of the studied cases and the associated MMR expression status is shown in Table 2.
Statistical analysis showed that abnormal MMR protein expression associated significantly with proximal tumour location (P=0.001), as abnormal MMR protein expression was noted in 38.9% (7/18 cases) of proximal colonic tumours compared with only 2.7% (1/38) of distal colon carcinomas. Three (42.9%) of seven caecal carcinomas and four of eight (50%) carcinomas in the ascending colon showed abnormal MMR protein expression compared with only one of 16 carcinomas in the descending colon (6.3%). None of the tumours in the rectosigmoid colon (100%) was MMR deficient.
A significant association was also noted between abnormal MMR protein expression and tumour histological type (P=0.021). Four (50%) of eight mucinous carcinomas were MMR deficient, whereas only 10% (4/40) of adenocarcinomas (not otherwise specified) demonstrated absent MMR protein expression. All four signet ring carcinomas revealed normal MMR expression (pMMR). Tumour grade, T-stage and lymphovascular invasion did not associate with abnormal MMR protein expression.
Tumour heterogeneity in the primary tumour and lymph node metastasis
Five (62.5%) out of the dMMR tumours revealed more than one morphological growth pattern, with combinations of glandular, mucinous, signet ring and/or solid differentiation. This mixed morphology (tumour heterogeneity) was observed in only 39.6% (19/48) of the pMMR tumours (P=0.268).
The glandular pattern was the most frequent growth pattern in the dMMR heterogeneous tumours (four out of five), followed by mucinous and solid differentiation. Of the five dMMR cases presenting with morphological heterogeneity, four presented with nodal metastasis. The glandular component was the most common component that metastasized to the lymph nodes in those cases. Table 3 summarizes the primary tumour and metastatic nodal morphologic pattern in the dMMR tumours.
We also tested the association between the MMR protein expression and the histopathologic features that were reported to be identified with increased frequencies in Lynch syndrome patients, and the data are presented in Table 4.
The identification of HNPCC can be lifesaving as it can lead to early detection of cancer (Järvinen et al., 2000). Moreover, several studies revealed that dMMR tumours respond differently to chemotherapeutic agents (Anthoney et al., 1996; Fink et al., 1998). Because there are limitations for depending on clinical criteria to guide the testing for Lynch syndrome and because of the prognostic information that could be provided through the identification of the MMR status, the current study was undertaken. Our study is the first to provide data on the prevalence of MMR gene defects in CRC in our hospital, which is a referral centre in the city of Alexandria, in a trial to assess the frequency of cases suggestive of HNPCC, aiming to introduce the IHC assessment of MMR protein expression as a first-line screening tool for all newly diagnosed CRCs.
We chose IHC as it is a simple, sensitive and inexpensive technique to test the MMR protein expression status that can specify the defective MMR gene, an ability that is not possessed by MSI testing, thereby providing gene-specific information and directing the genetic analysis to avoid performing exhausting, time and material-consuming unnecessary tests. However, the ability of IHC to identify the specific gene defect is dependent on the utilized antibody panel, and it may miss cases with defects in untested genes (Shia, 2008). Moreover, IHC can miss functional loss – that is, protein expression with corresponding antigen positivity in the absence of function (Kheirelseid et al., 2013).
In their functional state, the MMR proteins form heterodimers. MSH2 dimerizes with MSH6, and MLH1 dimerizes with PMS2. The MSH2 and MLH1 proteins are the obligatory partners of their respective heterodimers. IHC with antibodies to only MLH1/MSH2 is not able to detect all MLH1 or MSH2 abnormalities, as certain mutations may be associated with retained protein expression. However, PMS2 and MSH6 antibodies have the ability to detect most abnormalities in MLH1 and MSH2, in addition to detecting mutations in the genes that encode themselves – that is, PMS2 and MSH6. This explains why some studies that used MLH1/MSH2 IHC yielded a lower predictive value than those that included PMS2 and MSH6 (Shia, 2008).
On the basis of the aforementioned data and the lower sensitivity of MLH1/MSH2 IHC compared with MSI testing in predicting gene mutation, we chose an IHC panel that included the two obligatory partners (MLH1 and MSH2) and their respective heterodimers (PMS2 and MSH6) to achieve a higher detection rate and a predictive value for dMMR detection that is nearly equivalent to that of MSI testing.
Among our cases, the incidence of MMR gene defects was 14.3% (8/56), which is in accordance with others who reported dMMR in 15–17% of all primary CRCs (Popat et al., 2005; Imai and Yamamoto, 2008). Conversely, our rate contradicts previous reports that reported lower rates of MSI in the Mediterranean countries, because of dietetic, toxic or other environmental factors that cause epigenetic disruption of hMLH1, such as promoter hypermethylation of the gene, influencing in this manner the proportion of MMR-defective tumours (Slattery et al., 2000).
In accordance with others (Shia et al., 2004), focal weak staining with or without positive internal control was commonly noted with MLH1 immunostaining. This staining pattern has been suggested to reflect certain effects of the tumour microenvironment and tissue preservation, as tissue hypoxia and oxidative stress have been demonstrated to hinder MMR function in genetically pMMR tissues (Chang et al., 2002; Bindra et al., 2007). Thereby, the weak/lost regional staining pattern in cancerous tissue could be attributed to underlying regional hypoxia (Shia, 2008).
Sporadic CRCs often reveal absent MLH1 protein expression due to hypermethylation of the promoter region of the MLH1 gene resulting in transcriptional silencing of the MLH1 gene. Conversely, HNPCC exhibits germline mutations in one of the MMR genes, usually MLH1 (40–45% of cases), MSH2 (40–45%), MSH6 (5–10%) and PMS2 (1%). Thereby, the loss of expression of MSH2, MSH6 or PMS2 in isolation or in combination, and not the MLH1, provides reasonably strong evidence of a germline mutation in the respective gene and is therefore highly suggestive of HNPCC (Kaur et al., 2011).
Mutations of MLH1 or MSH2 often cause concurrent loss of MLH1/PMS2 or MSH2/MSH6, respectively, as seen on IHC, as abnormalities of MLH1 or MSH2 can cause proteolytic degradation of their dimer and consequent loss of both the obligatory and secondary partner proteins. However, mutations of PMS2 or MSH6 often cause isolated loss of PMS2 or MSH6 only. Most mutations in MSH2 are protein truncating; consequently, most MSH2-mutant colorectal tumours are expected to show absent MSH2 expression by IHC. However, more than one-third of the mutations in MLH1 are missense mutations that may result in mutant proteins that are antigenically intact but catalytically inactive (Shia, 2008). These missense MLH1 mutations could manifest immunohistochemically as a loss of PMS2 protein expression in the presence of IHC-detectable MLH1 protein (de Jong et al., 2004).
In agreement with these data, the current study revealed the concurrent loss of MSH2/MSH6 in 37.5% of cases (n=3) and isolated loss of MSH2 in 25% of cases (n=2). The abnormal MMR expression in those five (62.5%) cases was attributed to underlying MSH2 germline mutation. Regarding the single case that showed isolated loss of MLH1 expression, a characteristic of sporadic CRC, further methylation analysis would be of help in the determination of the nature of mutation in this case, as IHC loss of expression of MLH1 alone cannot differentiate sporadic tumours with MLH1 promotor hypermethylation from Lynch syndrome because of germline mutation in hMLH1 (approximately half of the cases) (Kheirelseid et al., 2013). Our two cases that revealed isolated PMS2 loss may indicate isolated germline mutations of PMS2, or could indicate the possible presence of missense mutations in the MLH1 gene with antigenically intact MLH1 protein. We therefore demonstrate that, on IHC, at least five out of the eight dMMR tumours are highly suggestive of being HNPCC, whereas the other three cases await further MSI testing to document/rule out underlying MLH1 hypermethylation or functional missense mutation.
Although the majority of CRCs arising in Lynch syndrome are mostly morphologically similar to sporadic CRCs (Jover et al., 2004), some histopathologic features are identified with increased frequencies in Lynch syndrome patients, and we studied them in detail. We noticed an increased frequency of pushing tumour margin, tumour heterogeneity, mucinous tumour histology, features of poor differentiation, peritumoral lymphocytes, and Crohn’s-like infiltrate at the tumour advancing edge among dMMR compared with pMMR tumours. This increased frequency reached statistical significance only for pushing the tumour margin and mucinous tumour histology.
MMR-defective sporadic and hereditary tumours have special pathologic characteristics. They are large high grade tumors, with signet ring cells and prominent intra/peritumoral lymphocytic infiltration, that are frequently located in the proximal colon. The presence of undifferentiated or medullary carcinoma with a trabecular growth pattern and significant intratumoral lymphocytic infiltrate has been associated with the MSI phenotype (Ruschoff et al., 1997). This histologic type frequently shows loss of hMLH1 expression in nonselected patients (Hinoi et al., 2001), but in those with HNPCC it is frequently associated with germline mutation in hMSH2 (Shashidharan et al., 1999).
In this study, we assessed the MMR protein expression without considering the family history as unfortunately it was lacking from the patients’ records. The eight MMR-defective cases fulfilled some of the revised Bethesda criteria for prediction of an MMR mutation. Six out of those eight defective dMMR CRCs were below 50 years of age (four of those six patients were female, and four cases presented with caecal tumours). Four cases were mucoid carcinomas, seven cases revealed Crohn’s-like lymphocytic reaction, and five possessed features of poor differentiation in patients below 60 years. Immunohistochemically, two of those six cases revealed isolated loss of MSH2, and two other cases revealed concurrent loss of MSH2 and MSH6, thereby reflecting the presence of underlying germline mutations suggestive of Lynch syndrome. A single case revealed isolated loss of MLH1 (discussed previously), and another two cases revealed isolated loss of PMS2 (discussed previously). The hereditary versus sporadic nature of those three CRCs can be uncovered by further MSI testing.
A multivariate logistic regression to establish a predictive model for dMMR tumours was not conducted because of the small sample size as only eight cases were MMR deficient. On the basis of the results of the current study, the minimum required sample size to determine the independent predictors of dMMR using logistic regression was 286. This sample size was estimated on the basis of the prevalence of dMMR among our CRC cases (14.3%) and the number of potential predictors, which was four (the significant variables in the univariate analysis).
Lynch syndrome is not uncommon and represented at least 8.9% of our CRC cases, with MSH2 germline mutations being the most common. With proper panel selection and appropriate interpretation, IHC proved to be a beneficial, inexpensive tool to detect dMMR CRCs, with an ability to specify the defective MMR gene. Our findings augment the results of previous studies and highlight the importance of using IHC an initial screening test for MMR gene defects to identify CRC patients for subsequent genetic testing. We therefore recommend performing IHC for MMR protein expression as a routine test for all newly diagnosed CRCs at our hospital to identify MMR-defective CRC cases who are expected to have a better prognosis and accordingly direct their adjuvant treatment regimens in the future.
Conflicts of interest
There are no conflicts of interest.
Alexander J, Watanabe T, Wu TT, Rashid A, Li S, Hamilton SR (2001). Histopathological identification of colon cancer with microsatellite instability. Am J Pathol 158:527–535.
Anthoney DA, McIlwrath AJ, Gallagher WM, et al (1996). Microsatellite instability, apoptosis, and loss of p53 function in drug-resistant tumor cells. Cancer Res 56:1374–1381.
Bindra RS, Crosby ME, Glazer PM (2007). Regulation of DNA repair in hypoxic cancer cells. Cancer Metastasis Rev 26:249–260.
Boland CR, Goel A (2010). Microsatellite instability in colorectal cancer. Gastroenterology 138:2073–2087.
Boland CR, Koi M, Chang DK, Carethers JM (2008). The biochemical basis of microsatellite instability and abnormal immunohistochemistry and clinical behavior in Lynch syndrome: from bench to bedside. Fam Cancer 7:41–52.
Bosman FT, Carneiro F, Hruban RH, Theise NDHamilton SR, Bosman FT, Boffetta P, Ilyas M, Morreau H, Nakamura SI, et al (2010). Pathology and genetics: tumours of the digestive system. World Health Organisation Classification of Tumours, 4th ed. Lyon, France: IARC Press. 131–182.
Chang CL, Marra G, Chauhan DP, Ha HT, Chang DK, Ricciardiello L, et al (2002). Oxidative stress inactivates the human DNA mismatch repair system. Am J Physiol Cell Physiol 283:C148–C154.
De Jong AE, van Puijenbroek M, Hendriks Y, Tops C, Wijnen J, Ausems MG, et al (2004). Microsatellite instability, immunohistochemistry, and additional PMS2 staining in suspected hereditary nonpolyposis colorectal cancer. Clin Cancer Res 10:972–980.
Fink D, Aebi S, Howell SB (1998). The role of DNA mismatch repair in drug resistance. Clin Cancer Res 4:1–6.
Graham DM, Appelman HD (1990). Crohn’s-like lymphoid reaction and colorectal carcinoma: a potential histologic prognosticator. Mod Pathol 3:332–335.
Hampel H, Frankel WL, Martin E, Arnold M, Khanduja K, Kuebler P, et al (2005). Screening for the Lynch syndrome (hereditary nonpolyposis colorectal cancer). N Engl J Med 352:1851–1860.
Hinoi T, Tani M, Lucas PC, Caca K, Dunn RL, Macri E, et al (2001). Loss of CDX2 expression and microsatellite instability are prominent features of large cell minimally differentiated carcinomas of the colon. Am J Pathol 159:2239–2248.
Imai K, Yamamoto H (2008). Carcinogenesis and microsatellite instability: the interrelationship between genetics and epigenetics. Carcinogenesis 29:673–680.
Ionov Y, Peinado MA, Malkhosyan S, Shibata D, Perucho M (1993). Ubiquitous somatic mutations in simple repeated sequences reveal a new mechanism for colonic carcinogenesis. Nature 363:558–561.
Jasperson KW, Tuohy TM, Neklason DW, Burt RW (2010). Hereditary and familial colon cancer. Gastroenterology 138:2044–2058.
Jass JR, Ajioka Y, Allen JP, Chan YF, Cohen RJ, Nixon JM, et al (1996). Assessment of invasive growth pattern and lymphocytic infiltration in colorectal cancer. Histopathology 28:543–548.
Järvinen HJ, Aarnio M, Mustonen H, Aktan-Collan K, Aaltonen LA, Peltomäki P, et al (2000). Controlled 15-year trial on screening for colorectal cancer in families with hereditary nonpolyposis colorectal cancer. Gastroenterology 118:829–834.
Jemal A, Siegel R, Ward E, Murray T, Xu J, Smigal C, et al (2006). Cancer statistics, 2006. CA Cancer J Clin 56:106–130.
Jover R, Payá A, Alenda C, Poveda MJ, Peiró G, Aranda FI, et al (2004). Defective mismatch-repair colorectal cancer clinicopathologic characteristics and usefulness of immunohistochemical analysis for diagnosis. Am J Clin Pathol 122:389–394.
Kaur G, Masoud A, Raihan N, Radzi M, Khamizar W, Kam LS (2011). Mismatch repair genes expression defects & association with clinicopathological characteristics in colorectal carcinoma. Indian J Med Res 134:186–192.
Kheirelseid EA, Miller N, Chang KH, Curran C, Hennessey E, Sheehan M, et al (2013). Mismatch repair protein expression in colorectal cancer. J Gastrointest Oncol 4:397–408.
Kloor M, Staffa L, Ahadova A, von Knebel Doeberitz M (2014). Clinical significance of microsatellite instability in colorectal cancer. Langenbecks Arch Surg 399:23–33.
Kocarnik JM, Shiovitz S, Phipps AI (2015). Molecular phenotypes of colorectal cancer and potential clinical applications. Gastroenterol Rep (Oxf) 3:269–276.
Lengauer C, Kinzler KW, Vogelstein B (1997). Genetic instability in colorectal cancers. Nature 386:623–627.
Lindor NM, Burgart LJ, Leontovich O, Goldberg RM, Cunningham JM, Sargent DJ, et al (2002). Immunohistochemistry versus microsatellite instability testing in phenotyping colorectal tumors. J Clin Oncol 20:1043–1048.
Lynch HT, de la Chapelle A (2003). Hereditary colorectal cancer. N Engl J Med 348:919–932.
Popat S, Hubner R, Houlston RS (2005). Systematic review of microsatellite instability and colorectal cancer prognosis. J Clin Oncol 23:609–618.
Poynter JN, Siegmund KD, Weisenberger DJ, Long TI, Thibodeau SN, Lindor N, et al (2008). Molecular characterization of MSI-H colorectal cancer by MLHI promoter methylation, immunohistochemistry, and mismatch repair germline mutation screening. Cancer Epidemiol Biomarkers Prev 17:3208–3215.
Ruschoff J, Dietmaier W, Luttges J, Seitz G, Bocker T, Zirngibl H, et al (1997). Poorly differentiated colonic adenocarcinoma, medullary type: clinical, phenotypic, and molecular characteristics. Am J Pathol 150:1815–1825.
Shashidharan M, Smyrk T, Lin KM, Ternent CA, Thorson AG, Blatchford GJ, et al (1999). Histologic comparison of hereditary nonpolyposis colorectal cancer associated with MSH2
and MLH1 and colorectal cancer from the general population. Dis Colon Rectum 42:722–726.
Shia J (2008). Immunohistochemistry versus microsatellite instability testing for screening colorectal cancer patients at risk for hereditary nonpolyposis colorectal cancer syndrome. J Mol Diagn 10:293–300.
Shia J, Ellis NA, Klimstra DS (2004). The utility of immunohistochemical detection of DNA mismatch repair gene proteins. Virchows Arch 445:431–441.
Sinicrope FA, Foster NR, Thibodeau SN, Marsoni S, Monges G, Labianca R, et al (2011). DNA mismatch repair status and colon cancer recurrence and survival in clinical trials of 5-fluorouracil-based adjuvant therapy. J Natl Cancer Inst 103:863–875.
Slattery ML, Curtin K, Anderson K, Ma KN, Ballard L, Edwards S, et al (2000). Associations between cigarette smoking, lifestyle factors, and microsatellite instability in colon tumors. J Natl Cancer Inst 92:1831–1836.
Ward RL, Turner J, Williams R, Pekarsky B, Packham D, Velickovic M, et al (2005). Routine testing for mismatch repair deficiency in sporadic colorectal cancer is justified. J Pathol 207:377–384.