A Call for Molecular Profiling in Uterine Carcinosarcoma

Douglas Levine, MD

Uterine carcinosarcoma (UCS) is a rare, aggressive gynecologic cancer that occurs in less than 5% of patients diagnosed with uterine cancer. UCSs display features of both endometrial carcinoma and sarcoma, making it very tricky to diagnose, says Douglas Levine, MD.

In a recently published paper in Cancer Cell, Levine and other researchers charted the molecular biology of 57 patients with confirmed cases of UCS. Sixty thousand individual characteristics were uncovered, which then were narrowed down to 9,149 genetic mutations. With this data—in addition to existing information on UCS—researchers identified 14 genes that were most commonly associated with UCS.

The heterogeneity of the mutation burden suggests that each tumor may have some unique features that allows it to be targeted by novel therapeutics, says Levine.

“These results really call for molecular profiling in this rare-type of tumor. It is probably a good use of molecular profiling, because there are a lot of available targets to identify. It also provides an explanation as to why these tumors are particularly aggressive.”

OncLive: Could you provide some background information on this paper?

In an interview with OncLive, Levine, director, Division of Gynecologic Oncology, NYU Perlmutter Cancer Center, discussed the findings of the study and early steps toward developing therapeutic regimens for UCS.Levine: We had published a study from The Cancer Genome Atlas (TCGA) of endometrial cancer—the most common type of cancer of the uterus—back in 2013. At the time, we had wanted to do a substudy on UCS, which is a very aggressive type of endometrial cancer, but we were not sure if we could collect enough cases because of the relative rarity of the tumor.

The management decision at the time was to just study the more common types of endometrial or uterine cancer, but soon after that, a set of rare tumor projects were started by TCGA and UCS featured prominently as one of the rare tumor types. We scoured the country and were able to get 57 cases that met the qualifications for inclusion into the TCGA, which are quite stringent. It requires an untreated tumor specimen from the primary site of the disease—the uterus, not a metastasis—and it has to be a sufficient size and quality to perform numerous molecular analyses.

What were the findings?

This is also a very difficult tumor to diagnose pathologically. We had a group of expert pathologists to make sure that these particular cases met the strict criteria for carcinosarcoma because it is a tricky tumor to diagnose. Sometimes, when a tumor is very aggressive, there is a desire to label it as a carcinosarcoma, but it may not meet the full criteria for having a carcinoma and sarcoma that is together in the tumor. So, it was very important that we had a high quality and rigorous pathology review to know that these tumors were of the proper classification. This is also one of the main strengths of this study—this is a very high quality set of tumors.There were a couple of main findings. One is that there is a heterogeneous set of mutations found in these tumor types and those mutations are present in a number of different cancer pathways, many of which have the ability to be targeted with drugs that are either FDA approved or in clinical trials. So, one conclusion is that molecular profiling of this particular rare tumor may help to direct investigational or off-label therapy.

The second main finding is that there was a very strong epithelial-to-mesenchymal transition (EMT) signature, which means their tumors have features of classical carcinomas but have dedifferentiated into sarcoma-like tumors. And so, this is a classic example of EMT which is found in many tumor types as they are invading and metastasizing, but this EMT phenomenon likely accounts for the very aggressive nature of this disease. A third major finding was that the EMT regulation is under epigenetic control, which means it is not directly driven by cancer mutations, but its driven by alterations in micro-RNA genes, which then regulate the expression of other regular genes that drives the EMT phenomenon.

Number four, is that there has been a long debate about where these tumors actually come from, are they actually 2 separate tumors that bump into each other? Are they tumors that separate into subtypes early on or later on? And while we did not provide all of the definitive data to answer these questions, we were able to demonstrate that the tumors were clonal. Meaning, there does not seem to be a combination of different subtypes of genetic defects of genetic mutations in the tumor, they appear to be clonal, which means the whole population of tumor cells likely comes from the same precursor. This has been what is commonly believed in the literature up until now.

Are there any next steps being taken with this data?

The 2 final things we demonstrated in this paper were that the multiple molecular platforms show that the tumors have both the features of a carcinoma and a sarcoma—which is not a surprise—but we demonstrated that at the protein and transcriptome level. Finally, we found a subset of tumors that have features that look like an endometrial cancer precursor, but the majority of tumors had features that looked like a serous endometrial cancer precursor, which is the more aggressive type of endometrial cancer.A number of other groups are studying carcinosarcomas because rare tumors are very exciting. Some groups have microdissected the 2 components of the tumors and again, seem to find that there are shared mutations. Other research is affirming the fact that these come from a common precursor. This is the largest study looking for therapeutic targets with the highest quality specimens. Carcinosarcomas can also occur in the ovary and so other groups have often lumped together ovarian and UCSs, but there are differences—so this tumor was very homogenous in that regard.

We've also just completed a therapeutic trial through our cooperative group, the Gynecologic Oncology Group (GOG). The trial, GOG 261, compared 2 different chemotherapy regimens. It has not yet published results, but we are very hopeful that we can apply some of these molecular approaches to that clinical trial to try to figure out which subgroups of patients respond better to that treatment.

We have a proposal before the cooperative group to allow us to do this type of testing on specimens from that clinical trial. In fact, the group was very astute and they put out a call for translational research to be done on that clinical trial that myself and some of my colleagues responded to. We're now working together to determine what groups of patients respond better based on their molecular profiling.

Cherniack AD, Shen H, Walter V, et al. Integrated molecular characterization of uterine carcinosarcoma. Cancer Cell. 2017;31(3):411-423. doi:10.1016/j.ccell.2017.02.010.