A high score based on the expression of 23 genes involved in repairing DNA damage after platinum-based chemotherapy for ovarian cancer was associated with a 5-year survival of 40% versus 17% for women who had a low score, investigators reported.
The gene score outperformed clinical factors associated with ovarian cancer outcome and was the only baseline factor that had a significant association with overall survival, according to an article published online in the Journal of the National Cancer Institute.
“By selecting genes in pathways known to be involved in repair of platinum-induced DNA damage, we produced a score that is predictive of overall survival, progression-free survival, and recurrence-free survival in ovarian cancer patients following platinum-based chemotherapy,” David Kozono, MD, PhD, of Dana-Farber Cancer Institute in Boston, and co-authors wrote in their discussion.
“Moreover, we were able to validate our results in two additional data sets of advanced stage ovarian cancer patients treated with platinum-based chemotherapy and demonstrate that the score is prognostic for survival only when it is composed of genes from relevant DNA repair pathways, further strengthening the use of biologic rationale in molecular signature construction.”
To date no effective means have been developed for predicting outcomes after treatment for ovarian cancer.
Several research groups have developed gene arrays for predicting short- and long-term survival in ovarian cancer. In general, these arrays have comprised genes involved in diverse cellular activities and have yet to gain widespread acceptance or use, the authors noted in their introduction.
Use of a hypothesis-driven strategy might improve the accuracy and reproducibility of gene arrays for predicting ovarian cancer outcomes. Kozono and colleagues hypothesized that patients with favorable versus poor outcomes after platinum-based chemotherapy would have tumors with differential expression of genes associated with repair pathways for platinum-induced DNA damage.
However, they hypothesized that low expression would be associated with improved outcome.
Using The Cancer Genome Atlas database, the authors examined gene-expression data for 151 DNA repair genes identified in tumors from 511 patients with serous ovarian cystoadenocarcinoma. They selected 23 genes involved in repair of platinum chemotherapy-induced DNA damage and assigned scores for low (0 to 10) and high (11 to 20) expression.
The genes represented the ataxia telangiectasia mutated, Fanconi Anemia/homologous recombination, nucleotide excision repair, and translesion synthesis repair pathways. The gene set comprised 19 genes for which expression above the median level was associated with improved survival and four for which higher expression predicted worse survival.
The investigators compared the gene-expression scores with overall survival, progression-free survival, recurrence-free survival, and complete response. Results were validated in two independent gene-expression microarray data sets.
The study population included 464 patients with advanced disease (stage III or IV). More than 60% of the patients received a platinum-taxane combination as initial chemotherapy.
The association between high gene-expression score and better overall survival was statistically significant (P<0.001), as were results in the smaller of the validation sets (P=0.021) but not the larger (P=0.055).
The survival difference translated into a 68% reduction in the hazard ratio for death (HR 0.32, P<0.001).
A high gene-expression score also was associated with significant reductions in hazard ratios for complete response (HR 0.60, P<0.001), recurrence-free survival (HR 0.84, P<0.001), and progression-free survival (HR 0.80, P<0.001).
The authors did not have an explanation for the counterintuitive finding of increased gene expression predicting improved survival. They speculated that increased expression of DNA repair genes might reflect an attempt to compensate for defective pathways.
The unresolved issues notwithstanding, the authors concluded that the gene-expression scoring system is ready for evaluation in a clinical trial.
Authors of an accompanying editorial disagreed, describing the results as another step toward identification of useful biomarkers in ovarian cancer, but still far removed from clinical application. Kozono and co-authors did not prove their original hypothesis, the editorialists charged, and they failed to validate the counterintuitive hypothesis in the larger of the two validation data sets.
"The premature application of inadequately validated biomarkers may adversely impact the successful implementation of individualized therapies," Elizabeth M. Swisher, MD, of the University of Washington, and co-authors wrote in conclusion.
The authors of the article and the editorial had no relevant disclosures.
Primary source: Journal of the National Cancer Institute
Source reference: Kang J, et al “A DNA repair pathway-focused score for prediction of outcomes in ovarian cancer treated with platinum-based chemotherapy” J Natl Cancer Inst 2012; DOI:10.1093/jnci/djs177.
Additional source: Journal of the National Cancer Institute
Source reference: Swisher EM, et al “Molecular scores to predict ovarian cancer outcomes: A worthy goal, but not ready for prime time” J Natl Cancer Inst. 2012; DOI: 10.1093/jnci/djs203.