Women with ovarian cancer usually present with advanced-stage disease, resulting in high relapse rates with 40% surviving less than 5 years. The standard of care for high-grade ovarian cancer is resection followed by platinum-based therapy; however, around 25% of women have immediate treatment resistance. Identifying high-risk patients early on would allow physicians to consider alternative treatment plans. In 2011, Roel Verhaak and colleagues from The Cancer Genome Atlas Research Network published a prognostic ovarian cancer gene signature, which showed that the expression levels of many genes correlate with survival. Verhaak's team have now “taken this finding to the next level.”

Credit: © Thinkstock Images/Comstock/Thinkstock

The researchers compiled a dataset of more than 1,350 ovarian carcinoma gene-expression profiles matched with clinical outcome information, and separated this data into a training set and a validation set: they termed the model 'CLOVAR'. Previously published models classified patients into low-risk and high-risk groups, with median survival in the high-risk group ranging from 33 months to 41 months. Verhaak continues, “using the CLOVAR model we grouped cases into multiple risk groups, with the worst-outcome group showing a median survival of 23 months and a platinum resistance rate of 63%. We found that our model could be further enhanced by including additional risk factors such as BRCA1 and BRCA2 mutation status, age, stage, and residual disease.”

Verhaak summarizes the key findings: “we believe these findings present a major step forward in the clinical applicability of outcome prediction in ovarian cancer, and start to approach the breast cancer setting in which gene-signature-based tests such as MammaPrint® are used to predict response to therapy.” Verhaak's team plans to validate the predictive ability of their model, including CLOVAR, age, stage, BRCA1 and BRCA2 mutation status, and residual disease, in a prospective setting.