In this multicenter LS cohort, several Jewish ethnic groups displayed recurring, presumably founder, DCVs in the known DNA MMR genes. Notably, DCVs in MSH2 accounted for about 50% of all Israeli Jewish cases of LS. Seven common recurring DCVs in MSH2 were detected in both AJ and non-AJ families, contributing to LS in 209/588 families.

These findings are similar to those reported in other genetically homogeneous populations. In Iceland, 2 recurrent DCVs, one in PMS2 and one in MSH6, contributed to LS in 78% of the population [19]. In Finland, one founder DCV in MLH1 was found to be responsible for 50% of Finnish carriers [20].

Only 19 DCVs were identified in PMS2 of which 2 were recurrent. This finding may reflect the low penetrance of monoallelic PMS2, as the cohort described here consisted mostly of cancer patients diagnosed with LS and family members.

A considerable number of unique DCVs (125/157), found in one or two families, accounted for LS across various ethnicities. These values raise concerns regarding the efficacy of the two-step testing process in which known founder DCVs are tested first, and comprehensive testing is only funded by the ministry of Health for highly suspicious patients. Rather, we advocate for comprehensive genetic testing to ensure rare variants are not overlooked. Rates of detection of recurring DCVs were ethnicity-dependent (Table 1). Although these rates may support the more cost-effective two-step genotyping process, several disadvantages need to be considered: the turn-around time for final results is longer, rare variants may be overlooked, especially in the lower-penetrant MSH6 and PMS2 genes, and cases with more than one DCV may be missed.

The frequency of DCVs in most minority groups is not well-documented because of their underrepresentation in global databases like gnomAD (https://gnomad.broadinstitute.org) and All of Us (https://www.researchallofus.org). However, an analysis of over 30,000 alleles from the local population has provided an estimate of founder DCVs within various ethnically matched groups. The MSH2 c.1906G > C variant, the most prevalent founder variant among AJ with LS, has a carrier frequency of 0.03% in AJ both globally in gnomAD v4 and in our local exome database. The combined frequency of the four founder DCVs in AJ detected in our exome data is approximately 1 in 900. The PMS2 c.2192T > G; p.Leu731* variant, which is considered rare worldwide, was relatively common in our cohort, with a frequency of 1:588 among individuals of Irani-Iraqi-Kurdish descent. Indeed, this variant was associated with CMMRD in the 2 affected seemingly unrelated Iranian families. We did not find evidence of a high frequency of other founder DCVs in other ethnically matched general populations. These findings should guide any regulatory decisions pertaining to population screening and spouse-testing, when applicable.

This work adds information on 21 DCVs, including 12 that have not been previously mentioned in the scientific literature and 9 that were previously classified as VUS (Tables 4 and 3). Our variant classification was based on the integration of clinical data, pathology, molecular pathology, the Insight database [16, 21], and data obtained from our internal datasets regarding the local frequency and distribution of variants. Some of the reclassifications were a result of cascade testing which enabled segregation within families.

It has been reported that VUS are most commonly reported variants among ethnic groups, such as Asian, Black, and Hispanic individuals and Sephardic Jews [10], adding to the health disparities already affecting these populations. The fact that almost half (45%) of our cohort were Sephardic Jews enabled proper diagnosis in this group and the addition of these variants to the global database. Expanding knowledge on DCVs in different subpopulations may also increase awareness among physicians and contribute to a higher rate of referrals for genetic counseling and testing.

Twenty-five genotyped individuals were double heterozygotes, carrying a DCV in a LS gene and an additional DCV in another cancer-susceptibility gene. Of them, twelve individuals had DCVs that are considered low penetrant, and 10 also had a BRCA2 DCV. The increased use of multigene panels has led to the identification of individuals harboring more than one pathogenic variant in cancer-susceptibility genes. In 2016, Whitworth and colleagues [22] published examples of patients who had, what they termed, multilocus inherited neoplasia alleles syndrome (MINAS). We previously reported a case series of double carriers of LS and hereditary breast and ovarian cancer pathogenic variants, suggesting that double MSH2/MSH6 and BRCA1/BRCA2 carriage is not associated with early disease onset or a more severe phenotype [23]. Conversely, we reported a child with “POLE-Lynch syndrome” carrying a de novo PMS2 DCV and inherited POLE DCV, manifested by an aggressive medulloblastoma with a unique genomic signature [24]. More data are needed to further our understanding of the clinical implications and consequences of double heterozygosity so that more targeted surveillance schemes can be offered.

This publication has several limitations. Data were collected from eight genetic institutes and high-risk clinics; however it is not population-based. Thus, in the absence of a national registry, it is challenging to determine the population frequency or the rate of LS in the general population, or among CRC and endometrial cancer patients. Carriers were identified via heterogenous venues, such as reflex screening, founder vriant testing, NGS based testing, cascade screening, or due to incidental findings. In addition, the lack of ongoing clinical update about carriers and family members, limits our ability to provide information on variant penetrance and expression. However, this cohort represents high-risk clinics from all over the country.