Editors' ChoiceCancer

Cancer Therapeutics: Best Informed by Genes or Genomes?

See allHide authors and affiliations

Science Translational Medicine  08 May 2013:
Vol. 5, Issue 184, pp. 184ec77
DOI: 10.1126/scitranslmed.3006450

Cancer is a heterogeneous disease, even when from the same tissue of origin. By examining the molecular profiles of 373 endometrial carcinomas, researchers from the Cancer Genome Atlas Research Network have indeed identified four molecularly defined disease subtypes that may inform treatment decisions––but based on characteristics of the cancer genomes rather than on specific gene mutations.

The authors obtained 363 profiles of somatic copy number alteration, 248 exome sequencing profiles, 333 mRNA expression profiles, 293 protein expression profiles, 367 microRNA expression profiles, and 373 DNA methylation profiles. The genomic data (copy number alterations, somatic mutations, and microsatellite instability) defined four disease subtypes. Unsupervised clustering of the samples on the basis of mRNA expression and methylation profiles was consistent with the same four subtypes.

Two of the subtypes––ultramutated and serous-like––were strongly associated with good and poor prognosis, respectively. The ultramutated group had high mutation frequencies and mutations of POLE, a gene that encodes an enzyme involved in DNA replication and repair. The serous-like group had low mutation rates but extensive copy number alterations; the poor prognosis of this group is in accord with findings that genomic instability (as reflected in abnormalities such as copy number alterations) correlate with poor prognosis in many other cancers.

The authors also identified recurrently mutated genes, which corresponded to previously well-characterized cancer-related genes (with the notable exception of POLE). However, the subclasses of endometrial cancer were more clearly discriminated by their coarse-grained genomic characteristics (frequency of mutation and copy number alterations) than by (combinations of) specific mutations or activated pathways (although POLE mutations caused high overall mutation rates). In the prevailing view of cancer therapy’s future, genomic data from an individual’s tumor will reveal mutational profiles that show a small number of driver events (gene gain/loss of function events that lead to pathway activation), which in turn will determine the appropriate targeted therapeutic regimen. In contrast to this view, the results in this paper suggest that general features of endometrial tumor genomes (mutation frequency and copy number variation) are sufficient to inform choices among conventional treatment alternatives (radiotherapy versus chemotherapy).

In fact, categorizing these cancers into the new genomic subtype groups described by the authors would reclassify up to 25% of high-grade endometroid tumors as uterine serous-like carcinomas––therefore indicating a different adjuvant therapy for these patients: chemotherapy instead of radiotherapy.

The Cancer Genome Atlas Research Network, Integrated genomic characterization of endometrial carcinoma. Nature 497, 67–73 (2013). [Full Text]

Navigate This Article