NEW YORK (GenomeWeb News) – In a study appearing online today in Nature, members of the Cancer Genome Atlas, or TCGA, Research Network described the endometrial cancer types that they tracked down through a comprehensive, genomics-based analysis of the disease.
The group brought together information on almost 400 tumors tested with half a dozen genomic, transcriptomic, and/or proteomic approaches — from whole-exome and low-coverage genome sequencing to find somatic mutations and structural variants, respectively to experiments exploring gene expression, protein, and microRNA patterns in the tumors.
From this data, the researchers outlined four molecular sub-groups of endometrial cancer — up from two endometrial cancer sub-types described previously. The improved classification scheme highlights some of the main genetic glitches that can occur in endometrial cancer, they noted, and reveals a subset of apparently high-risk tumors within an endometrial cancer subtype that is typically subjected to less aggressive treatment.
The newly defined endometrial cancer categories seem to hold promise for predicting patient outcomes and for improving some patients' treatment options, Memorial Sloan-Kettering Cancer Center's Douglas Levine, corresponding author on the study, told GenomeWeb Daily News, though he noted that additional trials are needed to explore that possibility.
For the most part, endometrial cancers are currently classified into either endometrioid or serous sub-types based on histological information or tumor pathology. While early stage endometrioid cancers are typically treated with radiotherapy, if anything, clinicians tend to use more aggressive chemotherapy-based approaches when tackling serous tumors, even at fairly early stages.
On the genetic side, Levine noted, it's been appreciated that endometrial cancers are prone to features such as microsatellite instability and mutations involving genes such as TP53 or PIK3CA.
And there have been efforts to more fully understand genetics of the disease, including exome-sequencing studies published in the Journal of the National Cancer Institute and Nature Genetics last year. Even so, the complete suite of genetic and epigenetic processes that go awry in endometrial cancer remains largely unknown.
In an effort to take a comprehensive look at endometrial cancer, TCGA members applied a range of genomics-related to technologies to 373 tumors. Of those, they performed whole-exome sequencing on 248 tumors, along with matched normal samples. Around 100 tumor-normal sets were subjected to low-coverage whole-genome sequencing as a means of finding structural alterations.
And across most of the available tumor samples, the team used array-based approaches to look at gene expression, protein expression, miRNA expression, and DNA methylation analyses.
Together, the mutation profiles and other features in the tumor set pointed to the presence of four molecular subtypes for endometrial cancer: a group of serous and serous-like tumors that includes a subset of the histologically-defined endometrioid cancers; a collection of tumors characterized by high microsatellite instability and a high mutation rate; another sub-group with relatively infrequent copy number changes; and a set of so-called POLE tumors with a hyper-mutation profile.
The analyses revealed similarities between some endometrial cancer subtypes and other forms of cancer, too. For instance, the researchers' data supports the notion that there is some genetic overlap between tumors from the endometrioid group and colorectal carcinoma.
For tumors from the high-risk serous subtype, meanwhile, they saw genetic and genomic features that resemble those found in serous ovarian cancer and basal-like breast cancer, albeit with more frequent mutations to genes such as PIK3CA, FBXW7, PPP2R1A, and ARID1A.
These and other molecular features identified in the current study could have prognostic and treatment implications, they noted. For instance, survival patterns in the POLE subtype seem to be favorable, despite the rampant mutations present in the genomes of those tumors.
In contrast, individuals with serous or serous-like tumors appear to do much worse, Levine noted.
That, in turn, suggests that there could be a benefit to offering more aggressive treatment to individuals with endometrioid cases falling in the new serous-like category, which is characterized by a higher-than-usual burden of copy number changes coupled with frequent mutations in the TP53 gene, a common cancer culprit.
"Clinicians should carefully consider treating copy-number-altered endometrioid patients with chemotherapy rather than adjuvant radiation," Levine and his co-authors wrote, "and formally test such hypotheses in prospective clinical trials."
For their part, researchers involved in the current study recently completed the accrual process for a clinical trial that will involve a little more than 300 endometrial cancer patients being treated with various chemotherapy regimens.
By prospectively collecting and tumors and determining their molecular subtypes, Levine said, it should be possible to track outcomes in relation to the four subtypes identified in the current study and, eventually, to get a better sense of treatment response and survival patterns in each group.