In a study published in Nature Genetics, researchers showed that categorizing copy number abnormalities into distinct sets from patient samples in high-grade serous ovarian cancer (HGSOC) had prognostic value.1
Researchers characterized 7 types of copy number signatures using shallow whole-genome sequencing and targeted amplicon sequencing of TP53 across 253 samples from 132 patients. Using this information, the researchers identified the patterns of copy number change that defined each copy number signature.
Copy number signatures were assigned based on 6 features: the breakpoint count per 10 Mb of genetic material, the copy number of the segments, the difference in copy number across adjacent segments, the breakpoint count per chromosome arm, the lengths of oscillating copy number segment chains, and the overall size of the segments.
The researchers determined that the underlying mutational process for copy number signature 1 in HGSOC is likely to be breakage-fusion-bridge (BFB) events. These BFB events occurred in copy number signature 1 as a result of oncogenic RAS signaling and telomere shortening, they reasoned.
To validate their approach, the investigators correlated copy number signature 1 exposures with mutation data, SNV signatures, and other genomic information derived from deep- and exome-sequencing studies across 2 independent cohorts.
Researchers constructed the remaining copy number signature sets using the same approach they used to create the copy number signature 1 set, teasing out each set’s defining features, genomic associations, and exposures. The group also proposed distinct underlying mechanisms for each type of copy number signature set.
To explore which, if any, associations existed between copy number signature exposures and overall survival, the researchers looked at a compiled dataset of 575 diagnostic samples. They found that copy number signature exposure was significantly predictive of survival (P = .002). In addition, exposures to copy number signature 1 or copy number signature 2 significantly predicted a poor outcome (P =.05 and P =.03, respectively), while exposures to copy number signatures 3 and 7 significantly predicted a good outcome in HGSOC (P =.05 and P =.006, respectively).
Using their prediction model, the investigators analyzed 36 patient samples consisting of paired diagnostic and relapse data from the BriTROC-1 study. They determined that exposure to copy number signature 1 at diagnosis significantly predicted platinum-resistant relapse (P =.02). Treatment with chemotherapy was not associated with any change in copy number signature exposures, which, according to the investigators, suggested that the mutations that occur in HGSOC are relatively stable and the patterns of copy number change are not likely influenced by the treatment regimen administered to the patient.
The findings from the paper may help researchers determine the genomic characteristics and predictors of patient outcome in other types of cancers that are also characterized by copy number changes and genomic instability, including esophageal, non-small cell lung, and triple-negative breast cancers.
Of specific interest may be the identification of high exposure to signature 1, which was shown in HGSOC to predict resistance to chemotherapy: “…powerful intrinsic resistance mechanisms are present at the time of diagnosis and can be readily identified using copy number signature analysis,” the researchers wrote.1 Plus, they added, copy number signatures “open new avenues for clinical trial design by highlighting contributions from underlying mutational processes that depend on RAS and PI3K–AKT signaling.”1 They concluded that RAS signaling, in particular, may be an important future target for the prevention of resistance to platinum-based therapies.
Macintyre G, Goranova TE, De Silva D, et al. Copy number signatures and mutational processes in ovarian carcinoma [published online August 13, 2018]. Nat Genet. doi:10.1038/s41588-018-0179-8
This article originally appeared on Cancer Therapy Advisor