The inclusion of moderate-penetrance cancer susceptibility genes in multigene panel testing poses challenges regarding the optimal management of individuals found to have pathogenic variants in these genes. A recently published counseling framework has provided evidence-based guidance for moderate-penetrance breast and ovarian cancer genes.1 However, no such framework exists for moderate-penetrance colorectal cancer (CRC) susceptibility genes, including CHEK2, APC*I1307K, and monoallelic MUTYH, which are among the most common multigene panel testing findings.2 Currently the only recommendations to address this risk are from the National Comprehensive Cancer Network, which recommends that individuals who carry pathogenic CHEK2 or APC*I1307K variants irrespective of family history, or monoallelic MUTYH with a family history of CRC, undergo earlier and more frequent CRC screening, similar to individuals with a first-degree relative with CRC.3 Whether early and increased CRC screening is truly justified for such carriers in the absence of a family history of CRC is unknown.

To better elucidate the appropriate timing of colonoscopy initiation, we calculated the cumulative lifetime risk (CLTR) of CRC as a multiple of the US Surveillance, Epidemiology, and End Results Program (SEER) estimates of ever developing CRC and the observed risk for selected genetic variants (CHEK2 1100delC, CHEK2 I157K, APC*I1307K, monoallelic MUTYH). Population age-specific incidence rates for CRCs were obtained from the 2010–2014 SEER cancer statistics for all races.4 Average relative-risk multipliers were derived from a systematic meta-analysis.5 We estimated 5-year and CLTR using previously described methods and applied the estimated odds ratio (OR) for each genetic variant to population age-specific incidence data.6

Moderate-penetrance CRC genes

CHEK2

Pathogenic CHEK2 variants are present in ~1% of Caucasians of European descent,7 with evidence for CRC risk limited to the two most common founder mutations 1100delC and I157T. Although multiple reports demonstrated a nonsignificant CRC risk,8 two meta-analyses reported a modestly increased CRC risk with ORs of 1.88 (95% confidence interval (CI), 1.29–2.73)5 and 2.11 (95% CI, 1.41–3.16)9 in CHEK2 1100delC carriers. However, given some irregularities in the analysis, the reported 2.11 OR9 appears to have been an overestimation with adjusted data suggesting an OR of 1.8 (95% CI, 1.2–2.7),10 similar to that in the report by Ma et al.5 In comparison, the relative risk of CRC in individuals with a family history of CRC is reported as 2.25 (95% CI, 2.0–2.53).11 For CHEK2 I157T, a meta-analysis demonstrated an OR of 1.61 (95% CI, 1.40–1.87) for CRC,12 similar to that in the meta-analysis by Ma et al.5 (OR 1.56; 95% CI, 1.32–1.84). CRC risk may be increased in CHEK2 carriers with a CRC family history, thereby implicating family history as a potentially important risk modifier.12 However, there is no strong evidence that CHEK2 is associated with CRC before age 50, nor an earlier age of CRC diagnosis.8

APC*I1307K

The APC*I1307K founder germ line variant, present in 7% of Ashkenazi Jews, is associated with a 1.96 (95% CI, 1.37–2.79) OR for CRC risk in Ashkenazi Jews.5,13 Age at CRC diagnosis in APC*I1307K carriers is not different from that in noncarriers.13 Association of APC*I1307K with a family history is less clear with studies among CRC patients reporting a family history of CRC ranging from 0 to 28%.13

Monoallelic MUTYH

Monoallelic MUTYH mutations are among the more common germ line abnormalities discovered via multigene panel testing, regardless of the population being tested, with a ~2% prevalence.2 CRC risk estimates associated with monoallelic MUTYH mutations are conflicting. The Colon Cancer Family Registry examined 223 monoallelic MUTYH carriers and their relatives estimating that such individuals had a cumulative 0.8% risk of CRC through age 50, regardless of CRC family history, versus 0.3% risk for noncarriers.14 However, a large meta-analysis including >25,000 MUTYH monoallelic carriers and >18,000 controls reported only a very slight increase in CRC risk (OR 1.17; 95% CI, 1.01–1.34) that was based on weak overall evidence.5

Managing CRC risk

Given uncertainties in the management of individuals with moderate-penetrance CRC gene variants, we estimated CRC risk associated with these variants to help better define possible CRC risk-reduction strategies. Estimated 5-year and CLTR of CRC for average-risk individuals and moderate-penetrance gene mutation carriers are shown in Table 1. For the average-risk individual, where current guidelines recommend initiating screening colonoscopy at age 50, a CLTR of CRC of 0.6% is reached by age 50–54.3 However, for CHEK2 1100delC, CHEK2 I157T, and APC*I1307K carriers, this same level of CLTR of CRC is reached by age 45–49. Therefore, if initiation of CRC screening is based on the average-risk individual’s CLTR of CRC, earlier initiation of screening colonoscopy in this timeframe in pathogenic CHEK2 and APC*I1307K carriers without a family history of CRC would seem reasonable (Table 2). Although our data is based on risk associated with CHEK2 1100delC and CHEK2 I157T, we would recommend considering all pathogenic CHEK2 mutation carriers equivalently given the current paucity of data on CRC risk associated with other CHEK2 variants. For monoallelic MUTYH carriers, cumulative risk of CRC is 0.7% by age 50–54, similar to individuals at average risk. Therefore, using similar reasoning, there does not appear to be an indication for earlier initiation of colonoscopy in monoallelic MUTYH carriers in the absence of a family history, consistent with current National Comprehensive Cancer Network recommendations.3

Table 1 Estimated 5-year and lifetime colon and rectal cancer risk for individuals with moderate-penetrance mutations in selected genes
Full size table
Table 2 Colonoscopy screening recommendations for individuals with moderate-penetrance mutations in selected genes
Full size table

While such recommendations apply to patients with no family history of CRC, the presence of a family history increases risk, as individuals with a first-degree relative with CRC have a 2.25 (95% CI, 2.0–2.53) relative risk of CRC.11 As such, earlier and more frequent colonoscopy is recommended for individuals with a family history of CRC based on the strength of the family history and the age of the CRC diagnoses (Table 2). Importantly, when assessing the CRC risk in moderate-penetrance CRC gene carriers, the screening age and interval should be based on the family history when a family history–based assessment results in earlier and/or more frequent screening recommendations than those dictated by consideration of the moderate-penetrance gene variant alone.

While our analysis has several limitations including risk estimates derived from a single meta-analysis, estimates based on limited studies resulting in wide confidence intervals, and the assumption that gene-associated risk is constant over lifetime, these estimates encompass a multitude of studies and are currently the most representative of actual population risk. The use of SEER estimates for absolute-risk calculations may not be applicable to other countries with different population-specific risks, or to populations such as African Americans, where absolute risk of CRC may be higher. Moreover, as SEER estimates are based on the inclusion of individuals at average risk as well as those at a higher risk of CRC, our derived risk estimates may reflect a slight overestimation of actual risk. Finally, our screening recommendations are based on the CRC CLTR threshold reached at initiation of screening colonoscopy at age 50 for the average-risk individual. Although it is conceivable that the age at which to initiate colonoscopy may be altered in the future, our risk-adapted screening strategy can be easily modified to meet such potential shifts in risk thresholds.

Cascade testing

The utility of cascade testing is another important consideration with moderate-penetrance genes. As pathogenic CHEK2 variants confer an increased risk of breast cancer7 with enhanced breast cancer screening already recommended,15 cascade testing in relatives of CHEK2 carriers is reasonable based on breast cancer risk alone, although it may not alter breast screening management in the presence of a strong breast cancer family history. Given the modest CRC risk associated with CHEK2 variants, and the resulting recommendation for more aggressive CRC screening, it also seems reasonable to offer cascade testing to relatives of CHEK2 carriers based on CRC risk alone to allow for individualized CRC risk management. For APC*I1307K carriers, there are currently no additional cancer risk-reduction strategies that would be employed other than enhanced CRC screening. Given the modest, but significantly increased, risk of CRC associated with APC*I1307K, as well as the resulting recommendations for earlier CRC screening for this population, an argument could also be made for offering cascade testing for this variant solely for CRC risk stratification. As CRC risk associated with monoallelic MUTYH mutations does not appear to merit increased CRC screening, cascade testing for this purpose alone may not be justified. However, given the high prevalence of monoallelic MUTYH carriers, cascade testing may be considered for family planning purposes, as recognition of carrier status in parents may help to identify offspring at risk for MUTYH-associated polyposis due to biallelic MUTYH mutations. Finally, as future research may provide more clarity regarding CRC and possibly other cancer risks, the decision of whether to pursue cascade testing will likely continue to evolve.

Conclusion

Increased uptake of multigene panel testing will continue to lead to the detection of germ line variants that confer a modest increase in cancer risk with questionable associated clinical significance. In hereditary breast cancer, emerging literature16 suggests that many clinicians make inappropriately aggressive management recommendations for individuals with these variants. As CRC screening is not without its own risks, we advocate that a risk-adapted screening strategy be incorporated into the management of individuals with moderate-penetrance CRC susceptibility variants. For such germ line testing to truly translate into effective and rational risk reduction, larger studies will need to be performed, ideally with the inclusion of genetic and environmental modifiers (e.g., single-nucleotide polymorphisms, aspirin, etc.), to better quantify the risk associated with moderate-penetrance variants so that management recommendations can be more optimally tailored to the actual degree of CRC risk.