Colon Cancer Family Registry: An International Resource for Studies of the Genetic Epidemiology of Colon Cancer

Introduction

Family studies have served as a cornerstone of genetic research on colorectal cancer (1, 2). Studies of multiple-case families led to the discovery of the best characterized colorectal cancer susceptibility syndromes: hereditary nonpolyposis colorectal cancer (HNPCC; hereafter called Lynch syndrome; refs. 3, 4) and familial adenomatous polyposis (FAP; ref. 5). The primary genetic bases of these dominant, relatively high-penetrance syndromes were identified in kindreds as mutations in the DNA mismatch repair (MMR) genes for Lynch syndrome (6) and in components of the APC or Wnt signaling pathway for FAP (7-10). Somatic (11) and epigenetic (12, 13) alterations in these same genes are involved in sporadic colorectal cancer as well. Although mutations in the MMR and APC genes are established causes of colorectal cancer (14), many important questions remain, including (a) the population frequency of these mutations; (b) the molecular and pathologic characteristics, as well as family-history patterns, most strongly associated with these mutations; (c) the penetrance of these mutations by age and sex; and (d) the genetic and environmental risk factors that modify penetrance. Recognizing the importance of further characterizing these pathways and identifying additional susceptibility genes and environmental risk factors, as well as their joint effects, in colorectal cancer, the National Cancer Institute in 1997 funded the Cooperative Family Registry for Colorectal Cancer Studies, which was competitively renewed in 2002 as the Colorectal Cancer Family Registry (Colon CFR). The Colon CFR joins the Breast Cancer Family Registry (15) as a research infrastructure for investigators interested in conducting population-based and clinic-based studies of the genetic and molecular epidemiology of these two common cancers.

The Colon CFR was developed by a multidisciplinary team of international investigators as a consortium of six field centers and an informatics support center (ISC). The participating centers are Cancer Care Ontario (CCO), Fred Hutchinson Cancer Research Center (FHCRC), Mayo Clinic, University of Hawaii (UHI), University of Southern California (USC), University of Melbourne (Australasia). The ISC developed a central database with limited data sets from all the participating centers. By the end of the second grant cycle of the Colon CFR in 2007, data and biospecimens would have been collected from nearly 10,000 probands and nearly 27,000 selected affected and unaffected relatives and controls. To facilitate the use of this resource by the scientific community, we describe the design and methods of the Colon CFR and the baseline data on colorectal cancer cases, their families, and controls, and provide a brief overview of the research and plans to date.

Materials and Methods

Overview

The Colon CFR is mandated to build and maintain a registry of high-risk colorectal cancer patients and families; collect detailed personal and family histories, including epidemiologic risk factors; obtain biological specimens; and conduct follow-up of the participants for colorectal neoplasia and other outcomes. The goal is to stimulate a broad cooperative effort based on this resource. After peer review of applications in 1997, six research institutions were selected from the United States, Canada, and Australia. Case ascertainment pools were derived from both cancer registries (population-based families) and cancer family clinics (clinic-based families). For phase I (1997-2002) center-specific ascertainment schemes further focused recruitment on families at increased risk for colorectal cancer based on age at onset and/or family history. Expansion of the Colon CFR continued in phase II (2002-2007), with new recruitment and data collection (phase II enrollees), follow-up data collection, and substantial molecular characterization of all phase I proband and family member enrollees. The infrastructure and progress are described below, and details can be found at http://epi.grants.cancer.gov/CFR/about_colon.html.

Structure

The participating centers are supported by the National Cancer Institute through the Cooperative Agreement grant mechanism. The investigators of the Colon CFR comprise a multidisciplinary team of epidemiologists, molecular biologists, medical and surgical clinicians, geneticists, pathologists, biostatisticians, and behavioral scientists. A steering committee is responsible for developing and applying Colon CFR–wide policies and procedures, setting research agendas, monitoring study progress, reviewing research proposals, addressing operational problems, and resolving intraconsortium issues. Voting members of the Steering Committee are defined as the principal investigator and one coinvestigator from each of the six research institutions and the ISC, one community representative, one elected nonaffiliated research scientist with expertise in multidisciplinary and translational colorectal cancer research, and the National Cancer Institute project officer. The Steering Committee has established several working groups that coordinate the work of the Colon CFR. These working groups include Epidemiology and Analysis; Biospecimens and Microsatellite Instability (MSI); Molecular Characterization and Pathology; Access and Tracking; Diet; Prevention Trials; Treatment and Outcomes; Behavioral and Education; and Technical. These groups have produced the protocols and procedures used at each site and collaborate closely with the ISC to assure the timely and accurate transmission of data. The Advisory Committee provides external oversight of the research agenda and reviews specific applications to use consortium resources. Additionally, the ISC manages the storage, archiving, and dissemination of deidentified data from each of the Colon CFR centers. In collaboration with the Colon CFR investigators and working groups, the ISC has developed and implemented a comprehensive system for processing, storing, and retrieving core data. Using secure FTP for transfers, the centers upload the data files according to regular schedules. The ISC has worked closely with the Colon CFR investigators and working groups to define quality control procedures. The ISC runs extensive computer edits on transmitted data and posts any resulting edit discrepancies to a secure Web site. The ISC has developed additional core software systems and assembles data sets for approved investigator requests (Fig. 1 ).

Ascertainment and Recruitment

The Colon CFR includes six study centers that have used varying designs and sampling schemes to recruit families (Table 1 ). This diversity is a primary strength, as it offers standardized data relevant to populations that differ in their underlying genetic and environmental risk profiles. The same questionnaires and biospecimen protocols have been used for all families so that data can be combined to address numerous research questions, if necessary, using a proband weighting system (described briefly below) and selection of comparison groups, if appropriate.

In general, two recruitment strategies have been used: population-based and clinic-based. Phase I recruitment began in 1997 and was completed in July 2002. Phase II recruitment continues with centers targeting certain subgroups for recruitment, including (a) families with possible MMR mutations, including those cases with incident colorectal cancer diagnosed under the age of 50 years (all centers) and/or families with multiple cases of colorectal cancer (Mayo Clinic) or (b) families from specific racial and ethnic groups. Enhanced recruitment of Japanese Americans (UHI) and African Americans (USC) is funded under a separate grant.

Phase I Population-Based Recruitment (1997-2002). All six Colon CFR centers enrolled families with incident colorectal cancer identified through population-based cancer registries in defined geographic areas. Sampling probabilities differed according to age at diagnosis, race, and/or family history of colorectal cancer, depending on center. Sampling schemes ranged from inviting all eligible cases to participate (FHCRC) to a single-stage design in which all early-onset and a fraction of older-onset cases were recruited, irrespective of family history, sex, or any other factor (Australasia), to a two-stage stratified sampling design in which race/ethnicity and/or age at diagnosis were used to select case probands to approach for family history data collection and the resulting family history information to select probands for enrollment (UHI, Mayo Clinic, CCO, USC). Table 1 describes recruitment for population-based centers. At all centers except Australasia, all case probands reporting a first-degree family history of colorectal cancer were eligible and recruited. In Australasia, sampling fractions were based on age at diagnosis, not family history. After enrollment of the case proband, his/her permission and assistance was sought to contact eligible and available relatives. The selection of family members for recruitment was center-specific, ranging from first-degree only to the inclusion of second- and third-degree relatives.

Because of these differing sampling schemes, for some analyses it may be necessary to use proband weights. These weights are defined as the inverse of the probability that a case proband was selected for recruitment. For example, for the two-stage designs, the overall sampling probability is the product of the stage 1 and stage 2 sampling probabilities. For FHCRC, where all case probands of eligible age were eligible, all case probands receive a sampling probability of 1. The UHI contributed only family history–positive cases, thus proband weights by family history are not definable and data from this center are most suitable for analyses limited to familial forms of colorectal cancer. Proband weights have been calculated for all population-based families in the Colon CFR and have been transmitted to the ISC. Weights are available to users through the ISC.

During Phase I, controls were either randomly sampled from the general population living in the relevant recruitment area using Medicare and Driver's License files (FHCRC), telephone subscribers lists (CCO), or electoral rolls (Australasia), or were ascertained through the case proband, either as the spouse (UHI, Australasia, and Mayo Clinic) or as an unaffected family member (USC).

Phase I Clinic-Based Recruitment. Three centers also recruited families with multiple or early-onset cases of colorectal cancer through clinical settings: Mayo Clinic, USC (Cleveland Clinic subcenter), and Australasia (seven family cancer clinics across Australia and New Zealand). Clinic-based probands, defined as the first family member enrolled in the Colon CFR registry, may or may not have had a personal history of colorectal cancer. Instead, eligibility was based on one or more of the following criteria: two or more relatives with a personal history of colorectal cancer or Lynch syndrome cancer, a proband diagnosed with colorectal cancer at a young age, or a proband presenting at a clinic with Lynch syndrome or Lynch-like syndrome. After proband enrollment, permission and assistance were sought to contact eligible and available relatives. Control participants were also ascertained through the proband: spouse (Mayo Clinic, Australasia) or unaffected family member (USC). Table 2 provides an overview of the sampling strategies for the clinic-based families.

Phase II Recruitment (2002-2007). To increase the sample size of MMR carriers and examine genetic and environmental risk factors for known, as well as yet unknown, genetic syndromes, recruitment was expanded to (a) more recently diagnosed population-based incident probands diagnosed under the age of 50 years and their first degree relatives and (b) more clinic-based families presenting at cancer family clinics (Table 2). The Colon CFR is also enhancing its minority component by recruiting additional Japanese Americans and African Americans, two groups in the United States whose colorectal cancer risk varies from that of the White population. Population-based controls continued to be recruited during phase II at FHCRC.

Protocols and Procedures

The Colon CFR centers all use standardized core protocols and instruments for personal and family history data, biospecimens collection, laboratory analysis, and data transmission and summaries. These materals were developed by the relevant working groups; questionnaires and protocols are described in detail at http://epi.grants.cancer.gov/CFR/index.html.

Core Questionnaire. Family history and personal exposures (except diet; see below) have been consistently collected at all sites to facilitate joint analyses. Each selected proband, population-based and spouse controls, completed the family history questionnaire either in person (USC), by telephone (FHCRC, USC, Australasia), or by mail (UHI, CCO, Mayo Clinic). The questionnaire acquired information on the number, sex, and birthdates of first-degree relatives (parents, siblings, and children), their cancer history, vital status, and, if deceased, date of death. All cancers, except for nonmelanoma skin cancers, were recorded with dates of diagnoses. Pedigrees were expanded to second- and third-degree relatives in several centers, depending on center-specific protocols and the reported family history.

Depending on the sampling schema, subjects also completed the epidemiologic questionnaire that captures information on established and suspected risk factors for colorectal cancer, including medical history and medication use, reproductive history (for female participants), physical activity, demographics, alcohol and tobacco use, race and ethnicity, and limited dietary data. The questionnaire was given to selected Colon CFR participants, including cases, controls, and affected and unaffected relatives, and defines a participant as an enrollee. Australasia and the UHI also used a short proxy version to collect limited information on selected individuals who were deceased, unavailable, or unwilling to participate. Centers customized questions for local usage, in particular for different language conventions and brand names, and were allowed to add additional questions of local interest. Detailed question-by-question manuals for administration were developed for further standardization.

Dietary Questionnaire. Diet is important in the genesis of colorectal cancer. Four Colon CFR centers collected dietary information using a self-administered diet history questionnaire from enrollees (cases, controls, and family members). The diet history questionnaire developed at UHI for use by the Multiethnic Cohort study in Hawaii and California was used at UHI, CCO, and USC (16). Australasian centers used a locally validated dietary questionnaire developed for a cohort study of >40,000 Greek, Italian, and Australian-born citizens of Melbourne (17). Both instruments collected information on the frequency of food consumption and portion size.

Biospecimens. For hypothesis-driven research to be based on this resource, extensive characterization of the families in the resource and of their cancers was required. Biospecimens (blood and paraffin-embedded tumor tissue) were obtained from cases and selected relatives and controls according to established Colon CFR protocols. Initial processing included blood sample separation and aliquoting (or tissue sectioning) and storage.

Blood or Buccal Cell Collection. A standardized protocol was used to collect and process blood samples. In general, blood was collected in EDTA and/or ACD (or heparin) vacutainer tubes. From these blood samples, whole blood spots were prepared on Schleicher and Schuel specimen collection paper (“Guthrie cards”), dried, and stored at room temperature; plasma and buffy coat (or other WBC component) were isolated, aliquoted, and stored; viable lymphocyte cells were separated by Ficoll separation procedure, aliquoted, slow-frozen, and stored in liquid nitrogen for eventual EBV-transformation to a lymphoblastoid cell line.

For subjects who did not consent to a blood draw, some centers sought a buccal cell sample. To ensure the highest possible DNA yield and quality, we used the specimen collection and processing protocol based on the mouthwash protocol of Lum and Le Marchand (18). The mailed kit contained an informed consent document, detailed instructions, and all materials needed for producing the sample and returning it to the study center. In general, the acceptability of this protocol and the quality of DNA has been high (19).

To facilitate collaborative research on colorectal cancer, a supplementary award was obtained that enabled all centers to extract germline DNA from blood or mouthwash samples for all phase I samples. Colon CFR–wide quality control studies on a sample of DNA specimens, including checking DNA concentration, absorbance ratio (260/280), DNA fragment size, and PCR amplification at six microsatellite loci were done by Coriell Institute for Medical Research.

EBV Transformation. To provide an unlimited supply of this valuable DNA, as well as RNA, supplementary funds were obtained to transform and maintain lymphoblastoid cell lines using EBV (20) on selected phase I participants. Using standard protocols, immortalization and subsequent DNA extractions were done at four laboratories: Coriell Institute for Medical Research, Mayo Clinic, University of Queensland, and Hospital for Sick Children in Toronto. The selection of samples for transformation was based on the proband's risk status: probands from families with two or more cases of colorectal cancer and probands diagnosed with colorectal cancer under the age of 50 years had the highest priority. To date, ∼4,100 specimens have been EBV-transformed.

Pathology Materials. Paraffin-embedded colorectal tumors and diagnostic pathology reports from colorectal cancer–affected probands and relatives were obtained from treating institutions, following informed consent. A standardized protocol to select and section representative blocks was developed by the Pathology working group. In general, sections were cut from each tumor and normal tissue block and stained with H&E. Stained sections were reviewed by center pathologist(s) who selected a colorectal cancer tumor block when possible, consisting of 70% to 80% tumor cells, and a block consisting of normal tissue for further sectioning. For planned immunohistochemistry and future protein-based studies, ten 4-μm or 5-μm paraffin sections from the tumor block were prepared and mounted unbaked on Probe-On Plus slides. For MSI testing and other future tumor DNA-based studies, ten 5-μm tumor sections were prepared and either mounted on glass slides or stored in microfuge tubes. When available, ten 5-μm normal tissue sections were similarly prepared and stored for DNA-based studies. DNA extractions from tumor and normal tissue for phase I recruitment was completed as part of the registry-wide MSI project, with remaining DNA stored at the local Colon CFR centers.

Confirmation of Colorectal Cancer Diagnoses. Confirmation of cancer diagnoses was sought for all reported colorectal cancers and, at some centers, for other reported invasive cancers. The level of confidence regarding a cancer diagnosis was classified into one of six categories (in decreasing order): (a) pathologist review of slides, (b) review of the pathology report; (c) cancer registry report or medical record(s) indicating treatment for the specific type of cancer; (d) report on a death certificate; (e) self-report; and (f) report by a relative. Pathology reports for all probands' qualifying colorectal cancers were reviewed, and data, including site, stage, histology, grade, distant spread, nodular involvement, etc., were abstracted onto a standardized review form developed by the Pathology Working Group. All pathology reviews are entered into a database and submitted to the ISC.

Characterization of Colorectal Cancer Tumors. One of two distinct molecular pathways leading to colorectal carcinogenesis involves the MMR pathway (21). Because it became evident that research of nearly any type on colorectal cancer would require characterizing participants' tumors according to DNA MMR competency, a competitive supplement was obtained to fund the Colon CFR–wide collection of tumors and characterization by tumor MSI, the phenotypic indicator of MMR competence (21). MSI testing was done on all available phase I Colon CFR tumors using 10 markers: four mononucleotide markers (BAT25, BAT26, BAT40, BAT34C4), four dinucleotide markers (D5S346, D17S250, D18S55, D10197), and two complex markers (ACTC and MycL; ref. 22). In general, tumors were classified as microsatellite stable (0% of unstable loci), MSI-low (>0% to <30% of unstable loci), or MSI-high (≥30% of unstable loci); unequivocal results for at least four markers were required to establish and report MSI status. MSI was completed on all affected phase I probands and affected family members with tumors meeting testing criteria and, similarly, on a selected sample of phase II affected participants (clinic-based and minority participants).

Immunohistochemical staining, a simpler and less expensive test than MSI analysis, has proved to be a valid alternative approach to identify tumors with MMR (23). Thus, in addition to the MSI testing, we did immunohistochemistry on all obtained clinic-based probands and affected relative tumors and, at a minimum, microsatellite instable population-based tumors to establish the presence or absence of protein expression for MLH1, MSH2, and MSH6 (and, at some centers, PMS2), primary protein constituents of the MMR machinery. Staining was done on 4-μm to 5-μm thick formalin-fixed, paraffin-embedded tissue sections and was conducted at three centers: Mayo Clinic, University of Queensland, and Mount Sinai Hospital. Commercially available monoclonal antibodies were used with control sections in each staining run, and assessment was standardized with protocols used for that laboratory. Although, we did not use standardized antibodies in all sites, site-specific antibodies and technical details are reported by Lindor et al. (23).

Molecular characterization testing of selected phase I colorectal cancer enrollees was completed during phase II (Fig. 2 ). Briefly, mutation testing for MLH1, MSH2, and MSH6 via dHPLC was conducted at three Colon CFR Centers: CCO, Australasia, and USC. Furthermore, the detection of large genomic rearrangements by MLPA was conducted under the direction of Dr. Stephen Thibodeau at the Mayo Clinic and the evaluation of methylation status of hMLH1 gene promoter was completed under the direction of Dr. Peter Laird at USC.

Follow-up

Follow-up activities in phase II have included active and passive follow-up of phase I enrollees (excluding unrelated controls). The active follow-up has been conducted 4 to 5 years after original enrollment and includes updating the family history, administering a follow-up epidemiologic and medical history questionnaire, and ascertaining and recruiting additional family members when indicated (e.g., new diagnosis of colorectal cancer) or when not previously feasible (e.g., were under age at time of family recruitment). Pathology materials (tumor blocks and diagnostic slides) are sought on all newly diagnosed colorectal cancers, as well as collection of pathology reports or other documents for verification, such as death certificates and medical records. Passive follow-up has allowed ascertainment of major end points (cancer diagnoses and death) by linkage to tumor registries and death indices. In addition, centers have used newsletters and other mailings to participants to maintain contact and a current address.

Results

Enrollment

From 1998 to 2002 (phase I), the six Colon CFR centers enrolled a total of 10,366 families (including controls) into the Colon CFR (Table 3 ). These represented 5,111 population-based case families, 4,474 population- and spouse-based control families, 441 clinic-based families with an affected proband, and 145 clinic-based families with an unaffected proband. Since 2002 (phase II), the centers enrolled individuals under the age of 50 years; therefore, an additional 2,672 case families, 354 population-based, and 68 spouse control families have been recruited. Table 2 describes data collection by study phase and ascertainment method; Table 3 describes the probands in terms of sex, age at onset, race, family history, and tumor characteristics. Active follow-up has been completed on >15,700 phase enrollees. Dietary questionnaires have been completed by nearly 4,000 probands, 2,600 controls, and 11,000 family members.

Proband and Family Characteristics

The Colon CFR contains subgroups of probands and families with specific characteristics representing a continuum of risk for colorectal cancer (Table 4 ). The relatively high portion of probands diagnosed at a younger age and/or with a positive first-degree family history reflects, in part, the sampling strategies used at the Colon CFR centers (excluding FHCRC wherein all cases with age of 20-74 years, regardless of family history, were enrolled during phase I). Thirty eight percent (3,706) of all probands were diagnosed under the age of 50 years. Twenty one percent (2,005) of probands had one first-degree relative with colorectal cancer, and 6% (597) reported two or more first-degree relatives with colorectal cancer. Based on these family reports, we identified 799 (8% of all probands) who met the Amsterdam I criteria (24).

Characterization of Colon CFR Tumors

MSI Analysis. The prevalence of MSI in the Colon CFR is based on the evaluation of over 4,000 phases I and II affected probands with tumors adequate for testing (Table 4). Of these participants who had their MSI evaluated, 72% were stable (microsatellite stable), 11% showed low instability (>0% to <30% unstable markers), and 17% showed high instability (≥30% unstable markers). MSI status varied by ascertainment source: clinic-based probands were twice as likely to be MSI-high as population-based probands (34% versus 17%).

Immunohistochemistry Analysis. Immunohistochemistry was used to test for the absence of protein expression for hMSH2, hMLH1, and hMSH6 (and PMS2 at Mayo Clinic, Australasia, CCO, and FHCRC). Of the 4,588 phases I and II tumors tested, 19% showed loss of expression for one or more proteins. Specifically, 499 (11%) showed loss of expression of either MLH1 or MLH1 and PMS2, 169 (4%) showed loss of expression of either MSH2 or MSH2 and MSH6, 89 (2%) showed loss of expression of MSH6 only, and 84 (2%) showed loss of expression of PMS-2 (not all centers included PMS2 testing in their immunohistochemistry protocol). Protein loss depended on sampling frame, with clinic-based proband lesions having the highest proportion of MMR protein loss.

Discussion

The Colon CFR represents a unique and valuable resource for a wide range of studies on the etiology and prevention of colorectal cancer. In an international collaborative environment, the world's largest collection of colorectal cancer families has been assembled, complete with epidemiologic data, biospecimens, and tumor tissues. A thorough search is being conducted for MMR mutations, epigenetic, environmental, and other causes for this common disease. Colon CFR participants reflect a continuum of risk that will be invaluable for the characterization of known genes, the identification of new genes, and the investigation of variants in pathways known to be involved in colorectal cancer. The results of such studies will be relevant to other cancer sites as well.

The Colon CFR has many strengths, including the extensive enrollment of affected cases, relatives, and controls which will permit a wide variety of classic and novel designs. Furthermore, the high quality biological data and existing characterization of all tumors for MSI and germline MMR gene mutations will enable the conduct of additional molecular studies. There are some limitations to the current registry design. First, differences in sampling frames and probabilities across Colon CFR centers (25) make summarizing registry data a whole challenge; varying sampling probabilities may be addressed analytically by incorporating individual proband weights (26). However, differences in population sampling (i.e., between clinic-based and population-based) may yield disparate estimates of penetrance (e.g., for major cancer susceptibility genes; ref. 27), and therefore, stratified analyses will be essential. Conversely, concordance of results will greatly strengthen inferences. Second, study centers vary with respect to control populations. The use of siblings, spouses, or unaffected population controls may yield different strengths of association with respect to genes, phenotypes, or other common exposures. These issues may be addressed by, first, comparing stratum-specific estimates based on similar control types, and second, in the absence of heterogeneity, summarizing across centers using all control types. Third, some of the study participants eligible for this study did not participate. Indeed, there is a great range of response proportions depending on the population and design. If participation depends on the presence of a gene (or genes) or phenotype, such as MSI, the prevalence of the gene in the population may have been under- or overestimated (28); furthermore, if participation depended jointly on exposure and disease status, our estimation of any association may be biased (29). It has been shown, however, that the assessment of gene-environment interactions in case-control studies is not subject to selection bias under the assumption that genotype does not influence participation conditional upon exposure and disease status (30). Fourth, not all subjects provided blood samples for genetic studies or tissue for MSI and other future studies. Because illness and family history may be factors in this loss, we may, again, have a study group that represents different genetic and phenotypic profiles than the true source population. Nevertheless, overall, the disparate designs and multiple study centers of the Colon CFR have yielded a range of risks that will support many research opportunities.

The resources collected and developed by the Colon CFR are available for use by the international scientific community. Investigators interested in pursuing collaborative research projects that use these resources are invited to submit applications for access to available data and biological materials. The first step in the application process is the submission of a “concept description,” a short form describing the proposed project and the types of resources being requested. After submitting a concept description, applicants develop a full application with the support of a scientific liaison assigned to their proposed project. The scientific liaison is an internal CFR investigator familiar with the available resources and the application process. Full applications are evaluated by the reviewing committees with careful consideration given to their scientific merit, overlap with ongoing CFR studies, and the effect of the proposed study on the inventory of biological materials. The Advisory Committee is responsible for the final approval or disapproval of applications for CFR resources. New concept descriptions are accepted through an on-line submission form hosted by the ISC. A detailed description of the application process and a link to the online submission form can be found at http://epi.grants.cancer.gov/CFR/research.html. After approval, study support is available through the ISC, as well as individual study sites, if required. Study reports should include Colon CFR authors as appropriate, based on qualifying contributions (31).

In summary, the Colon CFR is now the largest resource for multidisciplinary investigators to study colorectal cancer (Fig. 3 ). The worldwide nature and unique designs ensure that investigators will be able to study colorectal cancer as it occurs in many different populations. The Colon CFR resource is now of even greater research potential, with follow-up data on the original Colon CFR families, a sizable number of identified mutation carriers, new enrollment of minority populations, and easy access to quality biospecimens and molecular, epidemiologic, and pathology data from all centers. Colon CFR investigators have already participated in numerous projects with external researchers from throughout the world and have initiated Colon CFR projects that include new collaborators, including young investigators. This rich resource will be available to the international community of scientists interested in studying colorectal cancer for many years to come.