Hereditary nonpolyposis colorectal cancer: diagnostic strategies and their implications

Abstract

Objectives: Hereditary Nonpolyposis Colorectal Cancer (HNPCC) has been defined clinically and genetically. The disorder has traditionally been recognized in kindreds with a clustering of related cancers in association with mutations in DNA mismatch repair genes. HNPCC is associated with a substantially increased risk for several forms of malignancy but particularly colorectal and endometrial cancer. There were three main objectives of this report: (1) to assess the sensitivity, specificity, and reliability of laboratory and genetic tests commonly used in evaluating patients for HNPCC (analytic validity); (2) to summarize the accuracy of commonly used clinical and laboratory characteristics for predicting the presence of HNPCC in patients with colorectal cancer (clinical validity) and use these estimates to describe the efficiency of various strategies for identifying patients with a mismatch repair mutation; (3) to describe the benefits and harms related to screening and testing patients with colorectal cancer and their family members for HNPCC.

Data sources: Published literature identified through an electronic search (through April 2006), review of relevant bibliographies, and suggestions from technical experts.

Review methods: We evaluated studies critically and summarized the data qualitatively or by meta-analysis when studies used similar methodology and endpoints. We used decision trees to describe the efficiency of various strategies for identifying patients with HNPCC from a hypothetical population of patients with colorectal cancer.

Results: We included a total of 104 studies of which 40 addressed issues related to clinical validity, 3 to analytic validity, and 61 to benefits and harms. We identified only three studies on analytic validity and thus there exists a major gap in the published literature with regard to the accuracy and reliability of specific tests used in the evaluation of HNPCC. Among unselected patients with colorectal cancer who fulfilled the Amsterdam I criteria, 44% (95% CI: 35, 52%) carried pathogenic mismatch repair mutations (mainly in the MLH1 and MSH2 genes). The proportion was somewhat higher (51% [95% CI: 35, 66%]) among studies that performed sequencing on all available samples. The prevalence of MMR mutation carriers may be higher when genetic testing includes evaluation for large genomic deletions/rearrangements and when testing is also performed on MSH6 and PMS2. Approximately 71% (95% CI 63, 78%) of colorectal cancers from patients who fulfilled the Amsterdam I criteria demonstrated microsatellite instability while 40% (95% CI: 28, 53%) demonstrated loss of protein expression by immunohistochemistry. Of nine clinical strategies considered for detecting the presence of mismatch repair mutations in patients with colorectal cancer, the combination of three clinical predictors (age less than 50 years old at diagnosis; or a history of colorectal or endometrial cancer in a first degree family member; or the presence of multiple, synchronous or metachronous colorectal or endometrial cancers in the proband) combined with either immunohistochemistry (IHC) or MSI testing of tumor tissue identified a similar number of patients with mismatch repair mutations as other more complex strategies. There was little published information regarding potential harms associated with screening individuals with HNPCC-related cancers using clinical criteria (e.g. the Amsterdam criteria), MSI or IHC testing. Limited data suggested that testing probands for MMR mutations was not associated with severe psychological impact following formal counseling. Pre-test genetic counseling had good efficacy in improving knowledge about HNPCC and resulted in a high likelihood of proceeding with genetic testing, satisfaction in the decision to undergo genetic testing, and decreasing depression and distress levels among family members of HNPCC probands with cancer and among asymptomatic individuals from HNPCC families. Identification of HNPCC mutations was associated with an increase in the likelihood that family members of probands with CRC would undergo cancer-screening procedures. HNPCC family members who underwent cancer-screening procedures had a lower risk of developing HNPCC-related cancers and lower mortality rates than those who did not take actions. However, all of the relevant studies suggesting these benefits had important limitations. Survival was increased among asymptomatic HNPCC family members who received colonoscopy screening, regardless of their mutation status. There was limited direct evidence related to harms of the cancer-screening procedures in family members of probands with HNPCC. However, complication rates associated with these procedures in other settings are probably similar.

Conclusions: This report characterizes the accuracy of clinical and laboratory predictors of MMR mutations that can be used to identify patients with an increased risk of having MMR mutations. However, the sensitivity, specificity, and reliability of the tests used to evaluate individuals for suspected HNPCC is not known confidently. Data regarding the net benefits and harms associated with predictive genetic testing in patients with HNPCC-related cancers and their families members is incomplete but suggest that such testing improves compliance with screening procedures. At-risk family members who undergo screening colonoscopy have a reduced risk of developing HNPCC-related cancers and lower mortality. However, all studies supporting these benefits had important limitations.