NEW YORK (GenomeWeb) – A new computational look at copy number alterations in ovarian cancers, published online in PLOS One, has uncovered patterns that appear to correspond with platinum-based chemotherapy response and patient survival time.
"We believe this is a first step toward bringing ovarian cancer into the age of precision medicine," the study's senior author Orly Alter, a bioengineering and human genetics researcher at the University of Utah, said in a statement.
With the help of a mathematical method called tensor generalized singular value decomposition, or tensor GSVD, Alter led a team from the University of Utah that modeled array-based copy number profiles for matched tumor and normal samples from hundreds of women with ovarian serous cystadenocarcinoma who were assessed through the Cancer Genome Atlas.
Using this approach, the researchers picked up three main co-occurring groups of amplifications and deletions that were associated with treatment response and, consequently, survival times.
"What made our discovery possible is our new technique for mathematical modeling," Alter noted.
"We demonstrated the tensor GSVD in comparative modeling of patient- and platform-matched but probe-independent [ovarian tumor] and normal DNA copy-number profiles from TCGA," she and her co-authors wrote. "The modeling resulted in new insights into the poorly understood relations between an [ovarian tumor]'s genome and a patient's survival phenotype."
The researchers used the tensor GSVD method to scrutinize TCGA tumor and matched normal data for 249 individuals with ovarian cancer at a range of stages, establishing so-called tensors that maintain the multidimensional nature of the data types being considered.
In particular, the team was searching for survival-related copy number changes that were exclusive to tumors and could be reliably detected using a given testing platform. From the discovery set, it saw three copy number alteration sets that seemed to offer clues to patient outcomes and/or treatment response.
The first, which included an amplification of POLD2 on chromosome 7 and deletion of the chromosome 7 gene RPA, appeared to lend more DNA stability to tumors, for example, leading to longer-than-usual survival times for ovarian cancer patients.
The team also saw somewhat longer survival times and better treatment response for individuals whose tumors contained a deletion of the PABPC5 gene and amplification of BCAP31, both on the X chromosome — features expected to boost immune responses in the cell.
In contrast, the researchers' results pointed to diminished survival times for patients with a combination of copy number changes involving parts of chromosomes 6 and 12 that appear to bolster tumor cell immortality, notably deletions affecting the CDKN1A and MAPK14 genes and amplifications involving RAD51AP1 and KRAS.
The study's author validated these findings using data for another 148 individuals with stage III or IV ovarian cancer who were profiled by TCGA, identifying gene expression patterns that coincide with each copy number-classified patient group.
Based on these findings, the team concluded that the co-occurring sets of copy number alterations and related expression profiles — either alone or in combination with tumor staging information — might offer more refined clues to survival and treatment outcomes in individuals with ovarian cancer.