Keywords

Introduction

Advances in molecular testing have both increased recognition of heritable cancer syndromes and provided tools for their clinical diagnosis. Familial cancer syndromes that manifest in endometrial cancer include Lynch syndrome and Cowden syndrome, with very rare contributions by Cowden-like syndromes. Germline BRCA mutations have not yet been directly associated with increased endometrial cancer risk, but do appear to predispose patients to endometrial carcinogenesis indirectly through high rates of tamoxifen exposure.

An underlying cancer syndrome should always be considered in very young endometrial cancer patients, particularly in the setting of aberrant tumor morphologies or endometrioid adenocarcinoma without concomitant obesity or other evidence of estrogen excess. A personal or family history of relevant malignancies should also provoke concern. That said, some syndromic cancers manifest outside of a clinicopathologically concerning context and may warrant universal tumor screening. As the interpreter and caretaker of the tumor tissue, the pathologist is positioned to synthesize clinical, morphologic, and molecular data and suggest a work-up for an underlying germline mutation, and should therefore be well-acquainted with the features of heritable cancer in the endometrium.

Lynch Syndrome

Lynch syndrome is among the most common heritable cancer syndromes and predisposes patients to malignancies at a variety of sites, most notably the endometrium and lower gastrointestinal tract, and less commonly the ovaries, skin, renal pelvis, stomach, and brain. Endometrial carcinomas occur in 60–80% of women with Lynch syndrome and represent the sentinel malignancy in many of these patients. Between 2 and 5% of all endometrial cancers are associated with Lynch syndrome, and recognizing them as such allows for the identification and prevention of subsequent malignancies through increased surveillance and intervention programs [1,2,3,4,5].

Molecular Basis

Lynch syndrome is most often attributable to germline mutations in one of four mismatch repair genes: MLH1, PMS2, MSH2, and MSH6. These four genes encode proteins which dimerize into a MLH1–PMS2 complex and an MSH2–MSH6 complex. The two dimerized pairs form a four-protein complex that recognizes DNA mismatches and recruits repair machinery for excision and replacement of aberrant nucleotides. The prevalence and disease penetrance of endometrial cancer varies according to the implicated gene. MSH2 and MSH6 mutations are more commonly associated with endometrial carcinomas than are MLH1 and PMS2 mutations, a distribution that contrasts with Lynch syndrome-associated colorectal carcinoma. MSH6 mutations impart a particularly high risk of endometrial cancer development, with up to 71% of patients developing disease by age 70. Lifetime risk is considerably lower for PMS2 mutations at 12% by age 70, and ranges from 21 to 54% for MLH1 and MSH2 mutations [6].

In rare instances, the heritable defect lies not in one of these four mismatch repair genes, but in the related gene EPCAM. Mutations in the 3′ end of the EPCAM gene lead to hypermethylation of the MSH2 promotor region, disabling MSH2 and leading to dual loss of MSH2 and MSH6 [7,8,9]. Still more uncommon are recently described heritable mutations in MLH1 promoter mechanisms. In such patients the MLH1 gene is intact however MLH1 protein production is inhibited by hypermethlyation [10].

It is critical to emphasize that the vast majority of hypermethylated endometrial cancers are sporadic and are not associated with the exceedingly rare inheritance pattern described above. In fact, epigenetic methylation of the MLH1 promoter region is by far the most common cause of deficient mismatch repair in the uterus, underlying approximately 25% of endometrial carcinomas [11, 12].

Clinical Features

Lynch syndrome-related endometrial carcinomas, on average, develop a decade earlier in life when compared to sporadic endometrial malignancies [13,14,15,16,17]. However, these tumors are not exclusive to younger women, and a significant proportion occur in women over 50 years of age [18, 19]. Although prior and simultaneous malignancies may flag a subset of endometrial carcinomas that arise in women with Lynch syndrome, the endometrium is often the initial site of disease in these patients, and only a minority of Lynch patients identified on universal screening will have a history of colorectal or other cancers [18,19,20,21,22].

Pathologic Features

The anatomic localization of Lynch syndrome-related endometrial cancers varies. Although some studies have shown a predilection for the lower uterine segment when compared to their mismatch repair-competent counterparts, these tumors are by no means restricted to a lower uterine locale, and many arise in the fundus and surrounding uterine walls [19, 20, 23,24,25].

Reproducible histomorphologic features have been noted in a subset of Lynch syndrome-associated endometrial cancers. Perhaps most striking are the dedifferentiated and undifferentiated carcinomas; the former is characterized by areas of well-formed glands immediately juxtaposed with confluent sheets of markedly atypical tumor, while the latter contains no glandular structures (Figs. 9.1 and 9.2) [23, 26,27,28,29,30]. Such abrupt deviations in morphology make ontological sense given that tumors with incompetent DNA repair mechanisms are expected to acquire mutations rapidly. It is important to emphasize these morphologies that have been described in Lynch syndrome-related endometrial malignancies have also been recorded in both sporadically methylated and in “Lynch-like” endometrial carcinomas (e.g., cancers with mismatch repair protein patterns suggestive of Lynch syndrome, but without demonstrable mutations on germline sequencing) [31]. This suggests that these features are not an intrinsic feature of germline mutations themselves, but rather a marker of mismatch repair dysfunction at the protein level, irrespective of whether it is acquired through somatic or heritable mechanisms.

Fig. 9.1
figure 1

Undifferentiated endometrial adenocarcinoma is uncommon, but often associated with microsatellite instability due to epigenetic methylation and less commonly, to germline mutation in MLH1

Full size image
Fig. 9.2
figure 2

Undifferentiated endometrial adenocarcinoma (depicted in Fig. 9.1) exhibits intact expression of mismatch repair proteins a MHS2 and c MSH6, with loss of expression of mismatch repair proteins b MLH1 and d PMS2. In this tumor, loss of mismatch repair proteins is secondary to epigenetic methylation of the MLH1 promotor

Full size image

Not all Lynch syndrome-associated endometrial tumors exhibit remarkable morphologies. In fact the majority display a conventional, well to moderately differentiated endometrioid phenotype without notable demarcations in differentiation or distinct intratumoral morphologies [18,19,20, 22]. Pure serous, clear cell, and carcinosarcoma phenotypes are not typical of endometrial cancers arising in the setting of Lynch syndrome, but may occasionally occur.

As in the colorectum, Lynch syndrome-related endometrial cancers have been associated with increased tumor-infiltrating and peritumoral lymphocytes in some cases [19, 23, 26, 29, 32]. Although thresholds vary across the literature, most data suggest >40–42 intratumoral lymphocytes per 10 high-power fields [19, 23].

Screening and Confirmatory Testing

Because Lynch syndrome-associated endometrial carcinomas can serve as a harbinger of carcinogenesis at other sites, screening and confirmatory testing programs are of utmost clinical importance. Screening algorithms rely on mismatch repair immunohistochemistry, MLH1 promoter hypermethylation analysis, and microsatellite instability (MSI) testing in a variety of combinations. Diagnostic confirmation can be achieved through germline sequencing, with selective enlistment of somatic tumor sequencing in cases without identified germline mutations.

Mismatch Repair Protein Immunohistochemistry

Immunohistochemistry for the mismatch repair proteins MLH1, PMS2, MSH2, and MSH6 is the preferred initial screen for Lynch syndrome. This methodology has multiple benefits: firstly, immunohistochemistry is relatively inexpensive, technically simple, and readily accessible for most practicing pathologists [33, 34]. Sensitivity for the presence of MSI exceeds 90% [35]. Furthermore, the immunohistochemical loss pattern provides information as to the underlying mismatch repair defect: because MSH6 has an obligate reliance on MSH2 for expression (but the reverse does not hold), dual nuclear loss of MSH2 and MSH6 suggests an MSH2 mutation. Notably, this pattern can also be seen with 3′ EPCAM mutations due to the hypermethylation of the MSH2 promoter region (Fig. 9.3). On the other hand, isolated MSH6 loss indicates a possible MSH6 mutation.

Fig. 9.3
figure 3

Endometrial carcinomas may exhibit loss of MSH2 and MSH6 mismatch repair proteins due to germline mutation in a, c, e MSH2 or in b, d, f EPCAM. In both cases c, d MSH2 protein is not expressed, but in cases with mutations in EPCAM, expression for f EPCAM is also lost, while it is retained in tumors with e MSH2 mutations

Full size image

A similar pattern is observed with the MLH1/PMS2 pairing: because PMS2 is not expressed in the absence of MLH1 (MLH1 can be expressed in absence of PMS2), simultaneous loss of tumor nuclear expression of MLH1 and PMS2 signals a deficiency in the MLH1 protein. Importantly, this can be due to either epigenetic MLH1 methylation (Fig. 9.2) or, much less commonly, MLH1 germline mutations (Fig. 9.4). Isolated loss of PMS2 suggests a germline PMS2 mutation.

Fig. 9.4
figure 4

a, b Clear cell carcinoma of the endometrium with loss of c MLH1 and d PMS2 proteins secondary to germline mutation in MLH1

Full size image

A variety of algorithms have been proposed for the screening of endometrial carcinomas for Lynch syndrome. It is well-established that limiting screening to patients with age and history-based clinical risk as defined by the Amsterdam and Bethesda criteria misses affected patients [19, 22, 25, 31, 36]. Although screening methodologies for endometrial cancers remain a subject of debate, most experts in the field advocate some form of universal testing as is currently recommended for colorectal cancer (Fig. 9.5) [18, 19, 22, 25, 37].

Fig. 9.5
figure 5

Algorithm for evaluating endometrial cancer for possible Lynch syndrome. A more cost effective approach utilizing only 2 mismatch repair antibodies (MSH6 and PMS2) will capture most cases

Full size image

Screening approaches also differ with respect to the antibodies enlisted. Although many centers screen using a 4-antibody panel including MLH1, PMS2, MSH2, and MSH6, mismatch repair protein dimerization patterns allow for an alternative 2-antibody approach that enlists PMS2 and MSH6 as an initial screen. Current data suggests that the 2 and 4-antibody approaches show comparable efficacy in the detection of mismatch repair deficits [38, 39]. Given the relative rarity of MLH1 and PMS2 mutations in Lynch syndrome-related endometrial carcinomas, MSH6-only screening has also been proposed, although some evidence suggests that such focused panels will miss occasional Lynch syndrome patients [19, 25].

Mismatch repair immunohistochemistry interpretation is relatively straightforward, but is not without caveats. Intact expression is defined as the presence of any nuclear staining within the tumor, but sometimes staining is patchy and may be faint, particularly for the MSH6 antibody. MSH6 staining is prone to patchy, irregular staining and may pose problems on small biopsy samples. External positive and negative controls are desirable, but absence (or deficiency) of mismatch repair protein in a tumor can only be diagnosed in the presence of internal positive control staining with the antibody under evaluation. Care must be taken to specifically evaluate tumor cell nuclei as some mismatch repair deficient tumors may contain numerous intraepithelial lymphocytes that may lead to an erroneous diagnosis of intact expression. Cases that continue to present diagnostic difficulty on careful review should be classified as equivocal and subjected to second-line testing (such as MSI testing or, if clinical suspicion for heritable cancer is high, directed germline testing).

Occasionally, aberrant mismatch repair protein expression patterns may be observed. Loss of all 4 mismatch repair proteins may occur in tumors with underlying germline MSH2 mutations and concomitant MLH1 epigenetic methylation (Fig. 9.6) [40]. Also, some tumors with underlying MLH1 germline mutations may contain a nonfunctional protein that continues to be expressed on immunohistochemistry [41]. This latter aberrant expression pattern appears to be more common in colorectal cancer, in which MLH1 germline mutations are more common. Zonal loss of MLH1 and PMS2 may also be encountered [42]. This is easily recognized in the hysterectomy specimen, but may not be apparent in an endometrial sampling. It has been suggested this may reflect increased tumor aggressiveness, but that has not been our experience. Apparent isolated loss of PMS2 protein expression may be associated with MLH1 hypermethylation with heterogeneous MLH1 protein expression [43].

Fig. 9.6
figure 6

Endometrial carcinoma with loss of all 4 mismatch repair proteins: a MLH1; b MSH2; c MSH6; and d PMS2. In many cases this is due to mutation in MSH2 with epigenetic methylation of MLH1

Full size image

Microsatellite Instability Analysis

Mismatch repair defects lead to frequent replicative errors in short repetitive genomic regions known as microsatellites. The finding of MSI therefore serves as an indirect proxy for the presence of dysfunctional mismatch repair. PCR-based MSI testing measures repeat lengths of dinucleotide and mononucleotide markers (most commonly BAT25, BAT26, NR21, NR24, and NR27) and compares normal and tumoral tissue. Instability at two or more of these loci is classified as MSI-high, instability at a single locus is MSI-low, and an absence of instability is considered MS-stable [44].

Although not favored as a preliminary screen due to its inaccessibility at many centers, high cost, and inability to direct germline sequencing efforts, MSI testing can play an important role in the Lynch syndrome work-up in several situations. First, MSI testing can be enlisted in cases with equivocal immunohistochemistry results. Second, MSI has utility in resolving the differential for Lynch-like cancers as high level MSI supports the presence of a true mismatch repair defect (and argues against false immunohistochemistry results). Finally, MSI testing can be enlisted in patients with a negative MMR immunohistochemistry screen, but a strong clinical suspicion for a hereditary syndrome. Although MMR immunohistochemistry is more sensitive than MSI testing (particularly for MSH6 and PMS2 mutations, where MSI may fail to detect more than a quarter of cases), it has been reported that up to 10% of endometrial cancers with underlying MMR mutations and MSI may be missed by immunohistochemical screening [35].

MLH1 Promoter Methylation Analysis

Because immunohistochemical loss of MLH1 and PMS2 is most often attributable to sporadic methylation, PCR-based hypermethylation testing represents an important next step for endometrial cancers demonstrating this pattern, in order to prevent the perpetuation of unwarranted concern and further work-up for Lynch syndrome. In the colon and rectum BRAF testing is a reliable surrogate for the presence of MLH1 hypermethylation; however, this is not the case in the uterus [11, 12, 19, 45]. MLH1 hypermethylation demonstrates a heritable pattern in an exceedingly small minority of patients, therefore demonstration of hypermethylation effectively excludes Lynch syndrome in the absence of compelling clinical/family pedigree evidence of a familial syndrome [10].

DNA Mismatch Repair Gene Mutation Analysis

When mismatch repair protein loss and methylation data suggest a heritable syndrome, confirmatory germline sequencing is required for a diagnosis of Lynch syndrome. Because the mutations that underlie Lynch syndrome vary considerably, this requires whole genome sequencing of the suspected gene.

There is some variability in commercially available germline testing protocols and capabilities. Not all platforms have included EPCAM sequencing, although that is now performed with increasing frequency when relevant (e.g., loss of MSH2/6 without detection of mutations in either gene). Many assays are also unable to detect cryptic MSH2 gene inversions, which can account for a falsely “normal” germline result in patients with loss of MLH1/PMS2 and no evidence of MLH1 promoter hypermethylation [46, 47].

Somatic Gene Mutational Analysis

Historically, loss of mismatch repair protein expression (and the absence of MLH1 hypermethylation for those showing MLH1/PMS2 dual loss) was considered tantamount to a Lynch syndrome diagnosis. We now know that a considerable portion (up to 50%) of such immunohistochemically deficient cases will fail to show mutations on directed sequencing [48,49,50]. The possible underlying etiologies of these “Lynch like” tumors include: (1) somatic alterations (including loss of heterozygosity and biallelic somatic mutations); (2) inaccurate immunohistochemistries; and (3) undetected germline mutations. In discordant cases that prove MSI-high on MSI testing, direct tumor testing can be performed to ascertain whether somatic mutations and/or loss of heterozygosity account for the observed mismatch repair dysfunction. Demonstration of a tumor-specific mutation that is not observed on adequate germline sequencing effectively eliminates a germline cancer predisposition syndrome.

Cowden Syndrome

This autosomal dominant syndrome is extremely rare (affecting approximately 1 in 200,000) and accounts for a far smaller proportion of endometrial carcinomas than does Lynch syndrome [51,52,53,54]. Patients with Cowden syndrome are characterized by macrocephaly and a predilection for the development of multiple hamartomas involving the gastrointestinal tract (Fig. 9.7) and skin (facial trichilemmomas, acral keratoses, mucosal/cutaneous papillomatoses) [55]. In addition to endometrial carcinomas, Cowden syndrome patients are vulnerable to breast, thyroid, ovary, uterine cervix, colon, urinary bladder and renal malignancies [52, 53, 56, 57]. As with Lynch syndrome, endometrial cancers that arise in patients with Cowden syndrome present, on average, a decade prior to their mutation-negative counterparts.

Fig. 9.7
figure 7

Colon polyps in Cowden syndrome are typically small and sessile. a They exhibit an expanded and fibrotic lamina propria with b mild gland distortion and lymphoid follicles with some degree of smooth muscle proliferation and chronic inflammation. Ganglion cells, nerve fibers, and adipocytes within the lamina propria may also be seen

Full size image

Pathologic Features

Endometrial carcinomas associated with Cowden syndrome are classically of the endometrioid subtype (Fig. 9.8) [58,59,60]. However, recent evidence suggests that uterine serous carcinomas, clear cell carcinomas, mucinous carcinomas, and carcinosarcomas are also diagnosed in these patients [61].

Fig. 9.8
figure 8

a, b High grade endometrioid endometrial adenocarcinoma in patient with Cowden syndrome. Despite the high grade appearance, this tumor does not harbor a p53 mutation

Full size image

Molecular Basis

Germline mutations in the phosphatase and tensin homolog (PTEN) gene, a tumor suppressor, located on 10q23.3 underlie Cowden syndrome. However, the identification of a PTEN mutation has virtually no specificity for Cowden syndrome because between 77 and 94% of all endometrial cancers display this mutation [62]. This is true across the molecularly identified endometrial subtypes with the exception of high copy number (serous) tumors: the other three types [polymerase (ultramutated), microsatellite-unstable (hypermutated), and low copy number (endometrioid)] all show PTEN mutations in the majority of cases [62]. Futhermore, PTEN mutations can be found in a variety of hyperplastic and non-neoplastic endometria including normally cycling glands [63].

Confirmatory Testing

Combined with the extremely low prevalence of Cowden syndrome, the frequency of PTEN mutations in sporadic endometrial carcinomas and in non-neoplastic endometria obviates any utility of PTEN immunohistochemistry in Cowden syndrome screening and severely limits the utility of somatic tumor testing. Clinical screening criteria therefore play an important role in directing patients toward germline testing, with the recently released PTEN Cleveland Clinic risk assessment tool showing promise as a triage device [57]. Ultimate confirmation of a Cowden syndrome diagnosis relies on the identification of a germline mutation by sequencing.

Related Syndromes

Cowden-like syndromes have been identified in patients with mutations in succinate dehydrogenase genes (SDH-B, SDH-C, and SDHB-D) as well as killen (KLLN) genes [61, 64, 65]. In addition to endometrial carcinomas, patients with SDHB-D mutations are prone to paragangliomas, pheochromocytomas, thyroid carcinomas, renal carcinomas, gastrointestinal stromal tumors, and perhaps breast cancers [65]. Germline promoter methylation of KLLN, which shares a transcriptional start site with PTEN, has been described in patients with a clinical impression of Cowden syndrome but no identifiable PTEN mutations [64]. Testing for SDHB-D and KLLN mutations may therefore be indicated in patients with a clinical scenario highly suspicious for Cowden syndrome whose germline testing fails to identify alterations in PTEN.

Familial Breast Ovarian Cancer Syndromes (BRCA Mutations)

Germline mutations in the BRCA1 and BRCA2 genes are notoriously linked to increased risk of ovarian and breast carcinoma. There is ongoing debate, however, as to whether or not inherited BRCA mutations also increase the risk of endometrial carcinoma. Initial work has suggested that BRCA mutations carriers are at no increased risk for endometrial carcinoma, while several subsequent studies have shown that risk is increased, but appears to be commensurate with and attributable to tamoxifen exposure [66,67,68]. However, recent data suggest that although the overall risk for uterine cancer after risk reducing salpingo-oophorectomy is not increased, the risk for serous or serous-like endometrial carcinoma is increased in women with germline BRCA1 mutations [69]. When evaluating individual patients, it is important to keep in mind that increased somatic tumor testing is likely to identify a growing number of somatic BRCA mutations within endometrial carcinomas, and such results should not be interpreted as indicative of an inherited BRCA mutation in the absence of confirmatory germline testing.

Polymerase Proofreading-(POLD1) Associated Syndrome

Women with germline mutations in POLD1 exonuclease are at risk for endometrial carcinoma (57.1% of female carriers), in addition to attenuated colorectal polyposis (>60% POLD1 mutation carriers have ≥2 adenomas; on average, 16 adenomas), colorectal carcinoma (60–64% of carriers), and brain tumors (5.8%). Although the incidence is still under investigation, POLD1 exonuclease mutations appear to account for 1% of MMR-proficient familial and/or early-onset nonpolyposis colorectal carcinomas [70].

Li-Fraumeni Syndrome

Li-Fraumeni syndrome (LFS) is a cancer predisposition syndrome associated with the development of soft tissue sarcoma, osteosarcoma, pre-menopausal breast cancer, brain tumors, adrenocortical carcinoma, and leukemias [70]. A variety of other neoplasms may occur, including ovarian and endometrial cancer. Affected patients harbor a germline mutation in TP53. Intensive surveillance programs for the core cancers associated with the syndrome are instituted at an early age; affected patients should avoid exposure to radiation therapy, whenever possible, to reduce the risk of secondary radiation-induced malignancies [1, 7].