Clinicopathologic Factors Identify Sporadic Mismatch Repair Defective Colon Cancers

Similar documents
COLORECTAL PATHWAY GROUP, MANCHESTER CANCER. Guidelines for the assessment of mismatch. Colorectal Cancer

COLORECTAL PATHWAY GROUP, MANCHESTER CANCER. Guidelines for the assessment of mismatch. Colorectal Cancer

Colorectal Carcinoma Reporting in 2009

Guidelines for the assessment of mismatch repair (MMR) status in Colorectal Cancer

A Review from the Genetic Counselor s Perspective

Colorectal carcinoma (CRC) was traditionally thought of

Serrated Polyps and a Classification of Colorectal Cancer

Clinicopathologic Characteristics of Left-Sided Colon Cancers with High Microsatellite Instability

Defective mismatch repair in the pathogenesis of low-grade appendiceal mucinous neoplasms and adenocarcinomas

CME/SAM. Poorly Differentiated Colorectal Cancers. Correlation of Microsatellite Instability With Clinicopathologic Features and Survival

Lynch Syndrome Screening for Endometrial Cancer: Basic Concepts 1/16/2017

High risk stage II colon cancer

CAP Laboratory Improvement Programs. Summary of Microsatellite Instability Test Results From Laboratories Participating in Proficiency Surveys

Hyperplastic polyps in hereditary nonpolyposis colorectal cancer

Colorectal cancer Chapelle, J Clin Oncol, 2010

Case Presentation Diana Lim, MBBS, FRCPA, FRCPath Senior Consultant Department of Pathology, National University Health System, Singapore Assistant Pr

GENETICS OF COLORECTAL CANCER: HEREDITARY ASPECTS By. Magnitude of the Problem. Magnitude of the Problem. Cardinal Features of Lynch Syndrome

General Session 7: Controversies in Screening and Surveillance in Colorectal Cancer

Colon Cancer and Hereditary Cancer Syndromes

Development of Carcinoma Pathways

What Pathology can tell us in the approach of localized colorectal cancer

Dall istologia alla caratterizzazione biomolecolare

TumorNext-Lynch. genetic testing for hereditary colorectal or uterine cancer

Lynch Syndrome. Angie Strang, PGY2

Hyperplastische Polyps Innocent bystanders?

T he lifetime risk of developing colorectal cancer is 5%, with. Adenoma prevalence and cancer risk in familial non-polyposis colorectal cancer CANCER

The Whys OAP Annual Meeting CCO Symposium September 20. Immunohistochemical Assessment Dr. Terence Colgan Mount Sinai Hospital, Toronto

Large Colorectal Adenomas An Approach to Pathologic Evaluation

Anatomic Molecular Pathology: An Emerging Field

Universal Screening for Lynch Syndrome

Colonic polyps and colon cancer. Andrew Macpherson Director of Gastroentology University of Bern

Disclosures. Outline. What IS tumor budding?? Tumor Budding in Colorectal Carcinoma: What, Why, and How. I have nothing to disclose

Colorectal cancer (CRC) is one of the most common

Genetic testing all you need to know

WHAT SHOULD WE DO WITH TUMOUR BUDDING IN EARLY COLORECTAL CANCER?

VENTANA MMR IHC Panel Interpretation Guide for Staining of Colorectal Tissue

Introduction. Why Do MSI/MMR Analysis?

Molecular markers in colorectal cancer. Wolfram Jochum

The pathological phenotype of colon cancer with microsatellite instability

Colonic Polyp. Najmeh Aletaha. MD

Microsatellite instability and other molecular markers: how useful are they?

Arthur Purdy Stout Society of Surgical Pathologists Companion Meeting. Microsatellite Instability and Serrated Adenomas in Common Practice

Colorectal adenocarcinoma leading cancer in developed countries In US, annual deaths due to colorectal adenocarcinoma 57,000.

Loss of Mismatch Repair Protein Expression in Epithelial Ovarian Carcinoma: A Histomorphologic Guide to Targeted Screening

Current Status of Biomarkers (including DNA Tumor Markers and Immunohistochemistry in the Laboratory Diagnosis of Tumors)

Content. Diagnostic approach and clinical management of Lynch Syndrome: guidelines. Terminology. Identification of Lynch Syndrome

Comparison of the Mismatch Repair System between Primary and Metastatic Colorectal Cancers Using Immunohistochemistry

M. Azzam Kayasseh,Dubai,UAE

Should the grading of colorectal adenocarcinoma include microsatellite instability status?

Measure Description. Denominator Statement

Colorectal Neoplasia. Dr. Smita Devani MBChB, MRCP. Consultant Physician and Gastroenterologist Aga Khan University Hospital, Nairobi

Colorectal Cancer - Working in Partnership. David Baty Genetics, Ninewells Hospital

Familial and Hereditary Colon Cancer

Colon Cancer Update Christie J. Hilton, DO

Mismatch repair status, inflammation and outcome in patients with primary operable colorectal cancer

Isolated loss of PMS2 staining is an uncommon immunophenotype

Hereditary Non Polyposis Colorectal Cancer(HNPCC) From clinic to genetics

Early colorectal cancer Quality and rules for a good pathology report Histoprognostic factors

Familial and Hereditary Colon Cancer

General Surgery Grand Grounds

Morphologic Criteria of Invasive Colonic Adenocarcinoma on Biopsy Specimens

Hereditary Gastric Cancer

Agenda 8:30 AM. Jennifer L. Hunt

Special Section Hereditary Colorectal Cancer. What s New in Hereditary Colorectal Cancer? Jeremy R. Jass, MD, DSc, FRCPath, FRCPA

Resource impact report: Molecular testing strategies for Lynch syndrome in people with colorectal cancer (DG27)

Policy Specific Section: Medical Necessity and Investigational / Experimental. October 14, 1998 March 28, 2014

B Base excision repair, in MUTYH-associated polyposis and colorectal cancer, BRAF testing, for hereditary colorectal cancer, 696

Case Study. Overview. Deleterious MLH1 mutation detected on sequencing 10/16/2014

case report Reprinted from August 2013

Systematic immunohistochemical screening for Lynch syndrome in colorectal cancer: a single centre experience of 486 patients

Immunohistochemical evaluation of mismatch repair proteins in colorectal carcinoma: the AIFEG/GIPAD proposal

Mismatch Repair Deficiency Testing for Patients with Colorectal Cancer: A Clinical and Cost-Effectiveness Evaluation

Synchronous and Subsequent Lesions of Serrated Adenomas and Tubular Adenomas of the Colorectum

ASSESSMENT OF MLH1 PROMOTER METHYLATION IN ENDOMETRIAL CANCER USING PYROSEQUENCING TECHNOLOGY OBJECTIVES

Molecular aspects of HNPCC and identification of mutation carriers Niessen, Renee

Beyond the APC era Alternative pathways to CRC. Jeremy R Jass McGill University

Immunotherapy in Colorectal cancer

Colon cancer: practical molecular diagnostics. Wade S. Samowitz, M.D. University of Utah and ARUP

FAMILIAL COLORECTAL CANCER. Lyn Schofield Manager Familial Cancer Registry

Serrated Lesions in the Bowel Cancer Screening Programme

Staging Challenges in Lower GI Cancers. Disclosure of Relevant Financial Relationships. AJCC 8 th edition and CAP protocol updates

FROM EDUCATION TO TUMOUR CHARACTERISTICS IN COLORECTAL CANCER: AN ANALYSIS OF THE PATHWAYS

Histo-prognostic factors what histopathology has to offer for clinical decision making

Pathology perspective of colonic polyposis syndromes

MSI test to distinguish between HNPCC and other predisposing syndromes of value in tailored surveillance 1

Imaging Evaluation of Polyps. CT Colonography: Sessile Adenoma. Polyps, DALMs & Megacolon Objectives

Colorectal Cancer Structured Pathology Reporting Proforma DD MM YYYY

Tumors associated with Hereditary Nonpolyposis Colorectal Cancer: Defective Mismatch Repair and Familial Risk of Cancer

Histological Type. Morphological and Molecular Typing of breast Cancer. Nottingham Tenovus Primary Breast Cancer Study. Survival (%) Ian Ellis

Ramya Thota 1, Xiang Fang 2, Shanmuga Subbiah 3. Introduction

Molecular subtyping: how useful is it?

GOBLET CELL CARCINOID. Hanlin L. Wang, MD, PhD University of California Los Angeles

GOBLET CELL CARCINOID

Arzu Ensari, MD, PhD Department of Pathology Ankara University Medical School

Mr Chris Wakeman. General Surgeon University of Otago, Christchurch. 12:15-12:40 Management of Colorectal Cancer

Guidelines on Genetic Evaluation and Management of Lynch Syndrome: A Consensus Statement by the US Multi-Society Task Force on Colorectal Cancer

Update on Colonic Serrated (and Conventional) Adenomatous Polyps

LOINC. Clinical information. RCPA code. Record if different to report header Operating surgeon name and contact details. Absent.

Molecular Subtyping of Endometrial Cancer: A ProMisE ing Change

Transcription:

Anatomic Pathology / MORPHOLOGY IN MMR-DEFECTIVE COLON CANCER Clinicopathologic Factors Identify Sporadic Mismatch Repair Defective Colon Cancers Britta Halvarsson, MD, PhD, 1 Harald Anderson, PhD, 2 Katarina Domanska, 3 Gudrun Lindmark, MD, PhD, 4 and Mef Nilbert, MD, PhD 3,5 Key Words: Mismatch repair; Microsatellite instability; Tumor-infiltrating lymphocyte; Morphology; Histopathology Abstract Identification of sporadic mismatch repair (MMR)- defective colon cancers is increasingly demanded for decisions on adjuvant therapies. We evaluated clinicopathologic factors for the identification of these prognostically favorable tumors. Histopathologic features in 238 consecutive colon cancers were linked to MMR status based on immunostaining and BRAF mutation status. MMR defects were identified in 22.7% of the tumors, with 46 classified as sporadic. When the clinical parameters of age, sex, and proximal tumor location were combined with the morphologic features with the highest relative risks (RRs), eg, mucinous differentiation (RR, 9.0), tumor-infiltrating lymphocytes (RR, 7.5), absence of necrosis (RR, 7.5), and expanding growth pattern (RR, 5.0) into a 7-factor index, the presence of at least 4 features identified the MMRdefective tumors with 92.3% sensitivity and 75.3% specificity and excluded 61.5% of the tumors from MMR testing. This clinicopathologic index thus successfully selects MMR-defective colon cancers. Chromosomal instability (CIN) and microsatellite instability (MSI) constitute the predominant tumorigenic pathways in colon cancer. 1 CIN is associated with high mutation rates in genes tightly linked to the development of colorectal cancer, eg, APC, K-ras, and TP53, and current estimates indicate that some 70 genetic alterations lead to the development of CIN tumors. 2 The MSI pathway is linked to germline mismatchrepair (MMR) gene mutations within the hereditary nonpolyposis colorectal cancer (HNPCC) syndrome, which accounts for 2% to 3% of colorectal cancer. 3,4 However, defective MMR caused by hypermethylation of the MLH1 promoter characterizes 15% to 20% of sporadic tumors, particularly within the proximal colon. 5,6 The MMR-defective tumors can be identified with high sensitivity and specificity through immunohistochemical loss of MMR protein expression, and the presence of BRAF mutations may help distinguish hereditary from sporadic tumors with loss of MLH1/PMS2 staining. 7-9 Identification of HNPCC-associated colon cancer is important to prevent additional cancer cases in family members through surveillance programs for high-risk people. 10 Identification of sporadic MMR-defective tumors is central to improved prognostication because this subset of tumors has been suggested to have a favorable prognosis and questionable benefit from adjuvant 5-fluorouracil based chemotherapy. 11-15 Several histopathologic features, including proximal tumor location, expanding growth pattern, mucinous histologic features, poor differentiation, lack of dirty necrosis, and lymphocytic reactions, ie, peritumoral lymphocytes, Crohnlike reactions, and tumor-infiltrating lymphocytes (TILs), have been linked to MMR-defective tumors. However, there are no established guidelines for the identification of these tumors. 16-21 Routine inclusion of MSI analysis and/or MMR 238 Am J Clin Pathol 2008;129:238-244 Downloaded 238 from https://academic.oup.com/ajcp/article-abstract/129/2/238/1760174

Anatomic Pathology / ORIGINAL ARTICLE protein immunostaining is, therefore, performed by many laboratories, whereas others have not taken such action for reasons of cost and partly contradictory results regarding the prognostic impact of MSI. To assess whether clinicopathologic criteria can be used to select sporadic tumors for evaluation of MMR status, we characterized the morphologic features in a consecutive, unselected series of 238 colon cancers with correlations to immunohistochemical MMR protein expression with the aim of assessing the predictive strength of combined clinical and morphologic features. Materials and Methods Cases The study was approved by the Lund University Ethics Committee, Lund, Sweden, and comprised all colon cancers diagnosed from January 1999 through December 2001 at the regional Helsingborg Hospital, Helsingborg, Sweden. The 238 tumors developed in 118 women and 109 men with a mean age of 71 years (37-92 years). Synchronous colon cancers were diagnosed in 10 patients (1 of whom had 3 tumors), and 5 patients had previously been diagnosed with colorectal cancers. Tumor location was the proximal colon in 137 tumors, the distal colon in 99 tumors, and inadequately classified in 2 cases. Rectal cancers were excluded from the study. Tumor stage was pt1 in 2.9%, pt2 in 8.0%, pt3 in 73.9%, and pt4 in 15.1% of the tumors. Tumor differentiation was moderate in 177 (74.4%) and poor (including 2 undifferentiated tumors) in 61 (25.6%) tumors, with mucinous differentiation identified in 22 tumors and signet-ring cell cancer in 1. Histopathologic Evaluation All H&E-stained routine slides were morphologically reevaluated by one investigator (B.H.) according to a standardized protocol using previously described definitions and cutoff levels Table 1. 16,18-20,22-26 The clinical information available included that provided by the surgeon referring the specimen, but clinical files or family history of cancer were not obtained. However, information about other neoplasms operated on at the same hospital was available in the Department of Pathology, Helsingborg Hospital. Tumor stage was determined according to American Joint Committee on Cancer and International Union Against Cancer criteria and grade according to the World Health Organization classification. 22 Tumors with a mucinous or signet-ring cell component identified in more than 50% of the tumor area were classified as mucinous/signet-ring cell cancers and were considered poorly differentiated. Tumors with a trabecular, solid, or medullary growth pattern or a mucinous or signet-ring cell component encompassing 10% to 50% of the tumor area were classified as heterogeneous. 18,27 The tumor growth pattern was classified as expanding or infiltrative. 24 Dirty necrosis, defined as the presence of cell detritus and inflammatory cells within glandular lumina, was scored as present or absent. 19 Lymphocytic reactions were primarily determined in infiltrating tumor components and hot-spot areas. A Crohn-like infiltrate was defined as 3 or more nodular lymphoid aggregates deep to the advancing tumor margin within a single 4 field. 18,26 Peritumoral lymphocytes were considered to be present when a cap of chronic inflammatory cells was seen in the deep invasive tumor border. 18 The presence of TILs was defined as 7 or more TILs per 10 high-power fields ( 40) on H&E-stained slides (not including intramucosal carcinomas and early invasive tumor components). 16,20 Residual adenomas were classified as tubular, tubulovillous, villous, or serrated. 22 Immunostaining and BRAF Mutation Analysis Immunohistochemical analysis was used to assess MMR defects. Fresh, 4-µm sections were immunostained using antibodies against MLH1 (clone G168-15, dilution 1:200; BD PharMingen, San Diego, CA), PMS2 (clone A16-4, dilution 1:500; BD PharMingen), MSH2 (clone FE-11, dilution 1:100; Oncogene Sciences, San Diego, CA), and for MSH6 (clone 44, dilution 1:1,000; BD Transduction Laboratories, Lexington, Table 1 Morphologic Criteria Applied in the Evaluation of Sporadic Mismatch Repair Defective Colon Cancers Factor Definition Reference Tumor localization Proximal or distal to the splenic flexure Jass et al 23 Growth pattern Pushing with even, rounded infiltration; infiltrating when Jass et al 24 invading foci identified Differentiation Well, moderate, poor, or undifferentiated Hamilton and Aaltonen 22 Mucinous differentiation Mucin in >50% of the tumor area Hamilton and Aaltonen 22 Heterogeneous differentiation Poor; mucinous and/or signet-ring cell components in <50% of the tumor area Alexander et al 25 Necrosis Dirty necrosis within glandular lumina Greenson et al 19 Crohn-like reaction 3 nodular or Crohn-like lymphoid aggregates per 10 high-power fields Young et al 18 and Graham et al 26 Peritumoral lymphocytes A cap or mantle of lymphocytes at the infiltrating tumor border Young et al 18 Tumor-infiltrating lymphocytes 7/10 high-power fields Smyrk et al 16 and Jass 20 Downloaded from https://academic.oup.com/ajcp/article-abstract/129/2/238/1760174 Am J Clin Pathol 2008;129:238-244 239 239 239

Halvarsson et al / MORPHOLOGY IN MMR-DEFECTIVE COLON CANCER KY). The EnVision detection kit (DakoCytomation, Glostrup, Denmark) was used, and the staining procedure has previously been described. 28 Immunohistochemical MMR protein expression was classified as retained when nuclear staining in the tumor cells was identified and as lost when the tumor cells showed loss of staining with retained staining in stromal, epithelial, inflammatory, or infiltrating lymphoid cells. All immunostainings were evaluated without knowledge of the results of morphologic review, and consensus was reached (without discrepant findings) by 2 investigators (B.H. and M.N.). Tumors that showed loss of MLH1/PMS2 expression were sequenced for BRAF exons 11 and 15 (for primer sequences see Davies et al 29 ) to gain support for whether the underlying cause was a germline HNPCC-associated mutation or somatic hypermethylation of the MLH1 promoter. DNA was extracted from 1-µm cores from the paraffin-embedded tumor blocks by using a protocol available at http://cc.ucsf.edu/people/waldman/protocols/paraffin.html, and sequencing was performed on an ABI Prism 3100 (Applied Biosystems, Foster City, CA). Statistical Methods Associations between defective MMR and various clinical and histopathologic factors were analyzed by means of contingency tables and the Fisher exact test. The risk of an MMR defect was calculated conditionally on each factor value, and risk ratios (RRs), including approximate 95% confidence intervals, were determined. 30 An index for the identification of sporadic MMR-defective tumors was applied, and its sensitivity and specificity is illustrated by means of a receiver operating characteristic curve. Results Overall, MMR defects, defined as loss of at least 1 of the MMR proteins, were identified in 54 (22.7%) of 238 colon cancers. These tumors were diagnosed in patients at mean age of 72 years (range, 47-87 years) and 46 (85%) of 54 were localized in the proximal colon. Among these 54 MMR-defective tumors, 7 colon cancers from 5 patients were likely to be caused by HNPCC based on loss of MSH2 and/or MSH6 staining, and 1 tumor with loss of MLH1/PMS2 staining was classified as uncertain based on a heterogeneous staining pattern. These 8 colon cancers were excluded from further analysis because other means (eg, the Bethesda guidelines) of identifying HNPCC-associated tumors have been defined and the study was aimed at identification of sporadic (non-hnpcc) MMR-defective tumors. The study was thus based on the remaining 230 colon cancers, 46 of which were considered sporadic MMR-defective tumors. The BRAF V600E mutation was identified in 34 (74%) of these, which further supports sporadic tumor development. These colon cancers developed in 115 women and 106 men, 8 of whom had synchronous colon cancers. Tumor location was within the proximal colon in 57.0%. Morphologic review Table 2 and Image 1 showed poor differentiation in 24.8%, an expanding growth pattern in 24.0%, and necrosis in 47.8%. Mucinous/signet-ring cell differentiation was identified in 10.0% of the tumors, but tumor components (10%-50% of the tumor area) with, eg, mucinous or signet-ring cell differentiation were observed in 44.3% of the tumors. Lymphocytic infiltrates appeared as Crohn-like reactions in 44.7%, peritumoral lymphocytes in 14.7%, and TILs in 16.6%. Residual adenomatous components were identified in 67 cancers, 11 of which showed serration, 22 were tubular, 31 showed villous differentiation, and 3 had tubulovillous and serrated components. The 46 sporadic MMR-defective tumors developed in patients at a mean age of 74 years (range, 60-87 years) and 35 (76%) of these tumors occurred in women. Of these tumors, 8 were synchronous MMR-defective colon cancers with concordant loss of MLH1/PMS2 and the presence of the BRAF V600E mutation in all 4 patients. Features associated with MMR-defective tumors were, as expected, overrepresented among the 46 sporadic MMR-defective colon cancers; 40 (87%) were located within the proximal colon, 27 (61%) of 44 showed an expanding growth pattern, 25 (54%) were poorly differentiated, 39 (85%) had any degree of mucinous differentiation, and 41 (89%) lacked dirty necrosis. Lymphocytic reactions were common with, eg, TILs present in 59% of the tumors. The highest RR for MMR tumor development related to mucinous differentiation in more than 10% of the tumor (RR, 9.0), TILs (RR, 7.5), lack of dirty necrosis (RR, 7.5), and an expanding growth pattern (RR, 5.0) (Table 2 and Image 1). These morphologic factors were, together with female sex, age 60 years or older, and a proximal tumor location, combined to an index equal to the number of risk factors, and the predictive power of the index to identify sporadic MMRdefective tumors is illustrated as a receiver operating characteristic curve Figure 1. The presence of at least 3 of these factors allows identification of the MMR-defective tumors with 97.7% sensitivity and 48.3% specificity and the corresponding figures for tumors with at least 4 factors were 93.2% and 75.3%, respectively Table 3. If the index was applied with a 4-factor cutoff, 61.5% of the colon cancers could be excluded from MMR analysis. Discussion Recognition of a colon cancer phenotype linked to MMR defects is beneficial for the identification of hereditary cases and for prognostic estimates in sporadic colon cancers. 240 Am J Clin Pathol 2008;129:238-244 Downloaded 240 from https://academic.oup.com/ajcp/article-abstract/129/2/238/1760174

Anatomic Pathology / ORIGINAL ARTICLE Table 2 Tumor Morphologic Features in Relation to MMR Status in Sporadic MMR-Defective Colon Cancers * Factor Whole Material Fraction MMR Defective P Risk Ratio (95% CI) Tumor location (n = 228) Proximal 131 (57.5) 40/131 (30.5) <.001 4.9 (2.2-11.2) Distal 97 (42.5) 6/97 (6.2) Growth pattern (n = 221) Expanding 53 (24.0) 27/53 (50.9) <.001 5.0 (3.0-8.5) Infiltrating 168 (76.0) 17/168 (10.1) Differentiation (n = 230) Moderate 173 (75.2) 21/173 (12.1) Poor 57 (24.8) 25/57 (43.9) <.001 3.6 (2.2-5.9) Mucinous adenocarcinoma 23 (10.0) 11/23 (47.8) Any mucinous differentiation 88 (38.3) 39/88 (44.3) <.001 9.0 (4.2-19.2) No mucinous differentiation 142 (61.7) 7/142 (4.9) Necrosis (n = 230) Absent 120 (52.2) 41/120 (34.2) <.001 7.5 (3.1-18.3) Present 110 (47.8) 5/110 (4.5) Crohn-like reaction (n = 226) Present 101 (44.7) 29/101 (28.7).002 2.4 (1.4-4.2) Absent 125 (55.3) 15/125 (12.0) Peritumoral lymphocytes (n = 225) Present 33 (14.7) 12/33 (36.4).02 2.1 (1.2-3.5) Absent 192 (85.3) 34/192 (17.7) Tumor-infiltrating lymphocytes (n = 223) Present 37 (16.6) 27/37 (73.0) <.001 7.5 (4.7-12.2) Absent 186 (83.4) 18/186 (9.7) CI, confidence interval; MMR, mismatch repair. * Data are given as number (percentage) or number of MMR defective/total (percentage). The number possible to evaluate varied between 44 and 46. Fisher exact test. Includes 2 undifferentiated tumors. Includes 1 signet-ring cell cancer. Sensitivity 1.00 0.75 0.50 0.25 0.00 0.00 0.25 0.50 1 Specificity 0.75 1.00 Figure 1 Receiver operating characteristic (ROC) curve demonstrating sensitivity and specificity for predicting sporadic mismatch repair using a 7-factor index with age ( 60 years), female sex, proximal tumor location, and the morphologic variables tumor-infiltrating lymphocytes, lack of dirty necrosis, any mucinous differentiation (>10% mucinous areas), and expanding growth. Area under the ROC curve = 0.92. HNPCC-associated tumors were excluded from this study because Bethesda guidelines have been developed for the identification of this subset and because the clinical and morphologic features differ between sporadic and HNPCC-associated MMR-defective tumors. 18 Within our series, 36 tumors fulfilled the Bethesda guidelines, and all but 2 of 7 likely Table 3 MMR Defects in Relation to Number of Factors * in 218 Sporadic MMR-Defective Colon Cancers No. MMR Deficient/ Sensitivity Specificity No. of Factors Proficient (%) (%) 0 0/3 100 0 1 0/29 100 1.7 2 1/52 100 18.4 3 2/47 97.7 48.3 4 8/25 93.2 75.3 5 5/13 75.0 89.7 6 16/4 63.4 97.1 7 12/1 27.3 99.4 MMR, mismatch repair. * The factors were female sex, age 60 years, proximal location, expanding growth, lack of necrosis, mucinous differentiation, and tumor-infiltrating lymphocytes. HNPCC-associated tumors would have been identified using these guidelines (data not shown). The main aim of our study, however, was to evaluate the predictive strength of tumor morphologic features for the identification of sporadic MMRdefective colon cancers. Epigenetic alterations in several cancer-associated genes, including MLH1, are age-dependent, and the sporadic MMR-defective colon cancers predominantly develop in females and at older age, which is in agreement with 70% of the tumors occurring in females and at a mean age of 74 years. 17,18,31 MMR-defective tumors show a predilection for Downloaded from https://academic.oup.com/ajcp/article-abstract/129/2/238/1760174 Am J Clin Pathol 2008;129:238-244 241 241 241

Halvarsson et al / MORPHOLOGY IN MMR-DEFECTIVE COLON CANCER A B C D Image 1 Tumor morphologic features that best determined mismatch repair (MMR) status in colon cancer. A, Tumor-infiltrating lymphocytes (H&E, 40). B, Necrosis (a lack associated with MMR defects) (H&E, 4). C, Expanding growth pattern (H&E, 2). D, Mucinous differentiation (H&E, 2). the proximal colon, with 75% to 85% of the MMR-defective tumors being located here, which is in line with our finding of 87% of MMR-defective tumors being proximal.17,18 Concordant MMR status has been demonstrated in synchronous tumors from the same person, suggesting an individual predilection for tumors developing along a certain pathway. However, MMR-defective synchronous tumors do not necessarily signify HNPCC but may indicate other patientspecific factors such as exposure or yet unidentified genetic causes.23 MSI is identified in 1 of 3 of synchronous, as well as metachronous colon cancers, and, among these, 80% to 90% show concordant MSI status in the tumors.32 Differences apply, however, between these groups; hypermethylation of 242 Am J Clin Pathol 2008;129:238-244 Downloaded 242 from https://academic.oup.com/ajcp/article-abstract/129/2/238/1760174 the MLH1 promoter is identified in about 80% of synchronous tumors but in only about 40% of metachronous tumors.32,33 These results are supported also by the observation that 9 of 10 patients with synchronous tumors showed concordant MMR status. Several morphologic features have been associated with the development of sporadic MMR-defective tumors, including expanding growth, poor and mucinous differentiation, and lymphocytic infiltration. The identification of an expanding growth pattern varies between 15% and 93%, and in our series it was 61% (27/44).17,19,25,34 The presence of an expanding growth pattern was associated with a relative risk of 5.0 for an MMR-defective tumor. Poor and mucinous differentiation

Anatomic Pathology / ORIGINAL ARTICLE has, in most studies, been identified in one third to one half of the MSI tumors. 17,19,35,36 Mucinous tumors represent 10% to 15% of colon cancer, and between one third and one half of these are MMR-defective. Mucinous tumor components (although not reaching the amount requested for a diagnosis of a mucinous cancer) are frequent in MMR-defective tumors. 19,23,27,35 In our series, 54% of the MMR-defective tumors were poorly differentiated, and 85% showed any degree of mucinous differentiation; these factors were associated with RRs of 3.6 and 9.0 for MMR-defective tumors, respectively. Lymphocytic tumor infiltrates have been associated with MMR-defective tumors and include Crohn-like lymphoid reactions in 34% to 69% of the tumors, peritumoral lymphocytes in 30% to 64%, and TILs in 28% to 90% of the tumors. 17,19,21,23,25 We also identified correlations between these features and MMR-defective tumors (Table 2). In accordance with the findings in previous studies, 17,19,25,27 TILs showed a particularly high RR (7.5). Combinations of histopathologic features have been reported to improve the identification of MMR-defective tumors, and assessment of TILs in combination with mucinous areas has been reported to identify MMR-defective tumors with 70% to 80% sensitivity and 80% to 95% specificity. 25 When combining TILs with any amount of mucinous differentiation and absence of dirty necrosis, the sensitivity increased to about 100%, and it has been suggested that about half of the colon cancers thereby can be excluded from analysis of MMR status. 19,20,25 We combined age of 60 years or older, female sex, and a proximal tumor location with the features with the highest RRs, ie, mucinous differentiation, TILs, lack of necrosis, and an expanding growth pattern into an index aimed at identifying MMR-defective tumors (Figure 1). We thus demonstrated that sporadic MMR-defective tumors may successfully be identified and, at the same time, a substantial fraction of the tumors can be excluded from testing; the presence of 3 or more factors identified sporadic MMRdefective tumors with 97.7% sensitivity, excluding 39.0% of the tumors from MMR assessment, and if 4 or more factors were required, 93.2% sensitivity was reached, and 61.5% of the tumors could be removed from MMR testing. Our study demonstrated strong associations between several morphologic features and MMR-defective tumors and defined an index based on the clinical factors of sex, age, and tumor location and the morphologic factors of mucinous differentiation, an expanding growth pattern, TILs, and lack of dirty necrosis that successfully identifies MMR-defective tumors. We suggest that such an approach may be used to select tumors that should undergo evaluations of MMR status and should be further evaluated in independent materials. From the Departments of 1 Pathology and Cytology and 4 Surgery, Helsingborg Hospital, Helsingborg, Sweden; 2 Cancer Epidemiology and 3 Oncology, Lund University Hospital, Lund, Sweden; and 5 Clinical Sciences, Copenhagen University and Clinical Research Centre, Hvidovre University Hospital, Copenhagen, Denmark. Financial support was granted from the Swedish Cancer Society, the Swedish Research Council, the Nilsson Cancer Fund, the Zoega Fund, and the Region Skåne Research Funds, Sweden. Address reprint requests to Dr Halvarsson; Dept of Pathology, Helsingborg Hospital, 251 87 Helsingborg, Sweden. Acknowledgments: We acknowledge Eva Rambech and Anna Karlsson for technical help and excellent immunostaining. References 1. Lengauer C, Kinzler KW, Vogelstein B. Genetic instabilities in human cancers. Nature. 1998;396:643-649. 2. Sjoblom T, Jones S, Wood LD, et al. The consensus coding sequences of human breast and colorectal cancers. Science. 2006;314:268-274. 3. Aaltonen LA, Salovaara R, Kristo P, et al. Incidence of hereditary nonpolyposis colorectal cancer and the feasibility of molecular screening for the disease. N Engl J Med. 1998;338:1481-1487. 4. Kerber RA, Neklason DW, Samowitz WS, et al. Frequency of familial colon cancer and hereditary nonpolyposis colorectal cancer (Lynch syndrome) in a large population database. Fam Cancer. 2005;4:239-244. 5. Cunningham JM, Christensen ER, Tester DJ, et al. Hypermethylation of the Hmlh1 promoter in colon cancer with microsatellite instability. Cancer Res. 1998;58:3455-3460. 6. Thibodeau SN, French AJ, Cunningham JM, et al. Microsatellite instability in colorectal cancer: different mutator phenotypes and the principal involvement of Hmlh1. Cancer Res. 1998;58:1713-1718. 7. Lindor NM, Burgart LJ, Leontovich O, et al. Immunohistochemistry versus microsatellite instability testing in phenotyping colorectal tumors. J Clin Oncol. 2002;20:1043-1048. 8. Ruszkiewicz A, Bennett G, Moore J, et al. Correlation of mismatch repair genes immunohistochemistry and microsatellite instability status in HNPCC-associated tumours. Pathology. 2002;34:541-547. 9. Domingo E, Laiho P, Ollikainen M, et al. Braf screening as a low-cost effective strategy for simplifying HNPCC genetic testing. J Med Genet. 2004;41:664-668. 10. Renkonen-Sinisalo L, Aarnio M, Mecklin JP, et al. Surveillance improves survival of colorectal cancer in patients with hereditary nonpolyposis colorectal cancer. Cancer Detect Prev. 2000;24:137-142. 11. Ribic CM, Sargent DJ, Moore MJ, et al. Tumor microsatelliteinstability status as a predictor of benefit from fluorouracilbased adjuvant chemotherapy for colon cancer. N Engl J Med. 2003;349:247-257. 12. Popat S, Hubner R, Houlston RS. Systematic review of microsatellite instability and colorectal cancer prognosis. J Clin Oncol. 2005;23:609-618. 13. Benatti P, Gafa R, Barana D, et al. Microsatellite instability and colorectal cancer prognosis. Clin Cancer Res. 2005;11:8332-8340. 14. Lanza G, Gafa R, Santini A, et al. Immunohistochemical test for Mlh1 and Msh2 expression predicts clinical outcome in stage Ii and Iii colorectal cancer patients. J Clin Oncol. 2006;24:2359-2367. Downloaded from https://academic.oup.com/ajcp/article-abstract/129/2/238/1760174 Am J Clin Pathol 2008;129:238-244 243 243 243

Halvarsson et al / MORPHOLOGY IN MMR-DEFECTIVE COLON CANCER 15. Sinicrope FA, Rego RL, Foster N, et al. Microsatellite instability accounts for tumor site related differences in clinicopathological variables and prognosis in human colon cancers. Am J Gastroenterol. 2006;101:2818-2825. 16. Smyrk TC, Watson P, Kaul K, et al. Tumor-infiltrating lymphocytes are a marker for microsatellite instability in colorectal carcinoma. Cancer. 2001;91:2417-2422. 17. Ward R, Meagher A, Tomlinson I, et al. Microsatellite instability and the clinicopathological features of sporadic colorectal cancer. Gut. 2001;48:821-829. 18. Young J, Simms LA, Biden KG, et al. Features of colorectal cancers with high-level microsatellite instability occurring in familial and sporadic settings: parallel pathways of tumorigenesis. Am J Pathol. 2001;159:2107-2116. 19. Greenson JK, Bonner JD, Ben-Yzhak O, et al. Phenotype of microsatellite unstable colorectal carcinomas: welldifferentiated and focally mucinous tumors and the absence of dirty necrosis correlate with microsatellite instability. Am J Surg Pathol. 2003;27:563-570. 20. Jass JR. Role of the pathologist in the diagnosis of hereditary non-polyposis colorectal cancer. Dis Markers. 2004;20:215-224. 21. Yearsley M, Hampel H, Lehman A, et al. Histologic features distinguish microsatellite-high from microsatellite-low and microsatellite-stable colorectal carcinomas, but do not differentiate germline mutations from methylation of the Mlh1 promoter. Hum Pathol. 2006;37:831-838. 22. Hamilton SR, Aaltonen LA, eds. Pathology and Genetics of Tumours of the Digestive System. Lyon, France: IARC Press; 2001. World Health Organization Classification of Tumours. 23. Jass JR, Do KA, Simms LA, et al. Morphology of sporadic colorectal cancer with DNA replication errors. Gut. 1998;42:673-679. 24. Jass JR, Ajioka Y, Allen JP, et al. Assessment of invasive growth pattern and lymphocytic infiltration in colorectal cancer. Histopathology. 1996;28:543-548. 25. Alexander J, Watanabe T, Wu TT, et al. Histopathological identification of colon cancer with microsatellite instability. Am J Pathol. 2001;158:527-535. 26. Graham DM, Appelman HD. Crohn s-like lymphocytic reaction and colorectal carcinoma: a potential histologic prognosticator. Mod Pathol. 1990;3:332-335. 27. Shia J, Ellis NA, Paty PB, et al. Value of histopathology in predicting microsatellite instability in hereditary nonpolyposis colorectal cancer and sporadic colorectal cancer. Am J Surg Pathol. 2003;27:1407-1417. 28. Halvarsson B, Lindblom A, Rambech E, et al. Microsatellite instability analysis and/or immunostaining for the diagnosis of hereditary nonpolyposis colorectal cancer? Virchows Arch. 2004;444:135-141. 29. Davies H, Bignell GR, Cox C, et al. Mutations of the BRAF gene in human cancer. Nature. 2002;417:949-954. 30. Rothman K. Modern Epidemiology. Boston, MA: Little, Brown; 1986. 31. Kakar S, Burgart LJ, Thibodeau SN, et al. Frequency of loss of Hmlh1 expression in colorectal carcinoma increases with advancing age. Cancer. 2003;97:1421-1427. 32. Dykes SL, Qui H, Rothenberger DA, et al. Evidence of a preferred molecular pathway in patients with synchronous colorectal cancer. Cancer. 2003;98:48-54. 33. Velayos FS, Lee SH, Qiu H, et al. The mechanism of microsatellite instability is different in synchronous and metachronous colorectal cancer. J Gastrointest Surg. 2005;9:329-335. 34. Michael-Robinson JM, Biemer-Huttmann A, Purdie DM, et al. Tumour infiltrating lymphocytes and apoptosis are independent features in colorectal cancer stratified according to microsatellite instability status. Gut. 2001;48:360-366. 35. Kakar S, Aksoy S, Burgart LJ, et al. Mucinous carcinoma of the colon: correlation of loss of mismatch repair enzymes with clinicopathologic features and survival. Mod Pathol. 2004;17:696-700. 36. Wright CM, Stewart ID. Histopathology and mismatch repair status of 458 consecutive colorectal carcinomas. Am J Surg Pathol. 2003;27:1393-1406. 244 Am J Clin Pathol 2008;129:238-244 Downloaded 244 from https://academic.oup.com/ajcp/article-abstract/129/2/238/1760174