As recently reported in Nature, investigators in The Cancer Genome Atlas Research Network performed an integrated genomic, transcriptomic, and proteomic characterization of 373 endometrial carcinomas, including 307 endometrioid and 66 serous or mixed histology cases, using array- and sequencing-based techniques.¹ Their findings indicate that endometrial cancers can be classified into at least four subtypes on the basis of molecular characteristics. Moreover, they found that a significant proportion of high-grade endometrioid tumors share a molecular phenotype with uterine serous tumors, and that uterine serous tumors share molecular characteristics with ovarian serous and basal-like breast carcinomas.

Somatic Copy Number Alteration

Investigation of somatic copy number alterations in the endometrial carcinomas indicated that most of the serous (94%) and mixed histology (62%) cases clustered with a small proportion (12%) of endometrioid tumors in a group characterized by extremely high rates of somatic copy number alteration. This group was also characterized by frequent TP53 mutations (90%), low rate of microsatellite instability (6%), and a lower rate of PTEN mutations than in other endometrioid tumors (11% vs 84%). Tumors in this “serous-like” cluster were associated with significantly worse progression-free survival than tumors in the other three somatic copy number alteration clusters, which contained almost exclusively endometrioid tumors (97%).

Classification into Four Groups

With a combination of somatic nucleotide substitutions, microsatellite instability status, and somatic copy number alteration, the endometrial tumors could be classified into four groups: (1) an ultramutated group with unusually high mutation frequencies despite microsatellite stability, and a unique nucleotide change spectrum (POLE group); (2) a hypermutated group of tumors with microsatellite instability, most with MLH1 promoter methylation (microsatellite instability group); (3) a lower mutation frequency group, including most of the microsatellite-stable cancers (copy number–low group); and (4) the group consisting primarily of serous-like cancers with extensive copy number alteration and low mutation frequencies (copy number–high group).

The ultramutated group (7% of tumors) was characterized by an increased somatic transversion frequency, with all the tumors having mutations in the exonuclease domain of the POLE gene (involved in nuclear DNA replication and repair). Surprisingly, this group was associated with favorable progression-free survival, as the tumor cells possibly continued to accumulate unrepaired somatic mutations. Novel recurrent mutations were identified at codons 286 and 411 of POLE in most of these tumors. The microsatellite instability endometrioid tumors had a mutation frequency approximately 10-fold greater than microsatellite-stable endometrioid tumors and few copy number alterations. The microsatellite stable, copy number–low group had a high frequency of CTNNB1 mutations (52%), with this gene being the only one with a higher mutation frequency than in the microsatellite unstable cases.

The copy number–high group contained most of the serous cases and approximately 25% of the grade 3 endometrioid cases. In addition to increased somatic copy number alterations, frequent TP53 mutations, and infrequent PTEN mutations, this group was characterized by few DNA methylation changes and low estrogen/progesterone receptor levels. In addition to the high frequency of TP53 mutations, these tumors also had a high frequency of FBXW7 and PPP2R1A mutations, a profile previously reported as common in uterine serous but not endometrioid cancers.

Shared Features across Cancer Types

Additional analyses showed similar molecular features among uterine serous carcinomas, high-grade serous ovarian carcinomas, and basal-like breast carcinomas, including similar focal somatic copy number alteration and gene expression patterns. These three subtypes share a high frequency of TP53 mutations (84%–91%) and low frequency of PTEN mutations (1%–2%), although the uterine serous cancers were found to have higher frequencies of FBXW7, PPP2R1A, PIK3CA, and ARID1A mutations than the high-grade serous ovarian carcinomas or basal- like breast tumors.

As noted by the investigators, despite the greater frequency of mutations in uterine serous carcinomas, the shared molecular characteristics of these three tumor subtypes suggest “new opportunities for overlapping treatment paradigms.”

Overall, the investigators found that endometrial cancers have more frequent mutations in the PI3K/AKT pathway than any other tumor type they have studied thus far. Most endometrioid tumors had few copy number alterations or TP53 mutations, but frequent mutations in PTEN, CTNNB1, PIK3CA, ARID1A, KRAS, and newly described mutations in ARID5B (SWI/SNF [SWItch/Sucrose NonFermentable] chromatin remodeling complex gene).

In addition to the finding of molecular similarities among uterine serous, ovarian serous, and basal-like breast cancers, the investigators noted that the endometrioid tumors shared similarities with colorectal cancers—eg, frequent microsatellite instability, POLE alteration correlated with ­ultrahigh mutation frequency, and frequent activation of WNT/CTNNB1 signaling—although the endometrioid cancers are also characterized by KRAS and CTNNB1 mutations and a distinct mechanism of pathway activation.

For the serous-like tumors delineated on the basis of somatic copy number alterations, the investigators identified alterations that may be candidates for therapeutic targeting, including amplifications of ERBB2, FGFR1, and FGFR3 and LRP1B deletion, an alteration recently associated with resistance to liposomal doxorubicin in serous ovarian cancer.

Clinical Implications

The investigators note that early-stage type I endometrioid tumors are often treated with adjuvant radiotherapy and that similarly staged type II serous tumors are treated with chemotherapy. The findings indicate a similar molecular phenotype among roughly 25% of the high-grade endometrioid tumors and the uterine serous carcinomas, which “suggest that genomic-based classification may lead to improved management of these patients.”

In this regard, the investigators concluded, “Clinicians should carefully consider treating copy-number-altered endometrioid patients with chemotherapy rather than adjuvant radiotherapy and formally test such hypotheses in clinical trials. Furthermore, the marked molecular differences between endometrioid and serous-like tumors suggest that they warrant separate clinical trials to develop the personalized treatment paradigms that have improved outcomes in other tumor types, such as breast cancer.” ■

Disclosure: The study was supported by the National Institutes of Health.

Reference

1. The Cancer Genome Atlas Research Network: Integrated genomic characterization of endometrial carcinoma. Nature 497:67-72, 2013.