Date Published: February 3, 2009
Publisher: Public Library of Science
Author(s): Kate Lawrenson, Simon A Gayther
Abstract: Kate Lawrenson and Simon Gayther discuss two studies of ovarian cancer inPLoS Medicine, one on clinico-pathological heterogeneity and one on gene expression profiling.
Partial Text: From a clinical perspective, epithelial ovarian cancer is something of an enigma. Despite improvements in aggressive debulking surgery and the initial good response of patients to platinum-based chemotherapies, there has been little improvement in the survival rates for over three decades. About 65% of women with epithelial ovarian cancer will die within five years of their diagnosis . Early-stage ovarian cancers are often asymptomatic and the recognised signs and symptoms, even of late-stage disease, are vague. Consequently, most patients are diagnosed with advanced disease, and it seems unlikely that symptoms alone could help to improve the proportion of tumours that are diagnosed at earlier, more treatable stages.
There is continued hope that the most recent scientific advances and discoveries will have the potential to affect patient care (translational research). For example, the last decade has seen revolutionary developments in the approaches used to characterise solid tumours at the molecular level. For some cancer types, the molecular characterisation of tumours has led to better strategies for predicting disease outcome, so that treatments can be targeted more effectively, and to the development of new therapies. Many of these approaches have been tried and tested for ovarian cancer too, but they have so far failed to deliver on the anticipation of a new biomarker or gene signature that can improve our understanding of the disease. Two studies published in PLoS Medicine, one in December 2008 and one in the current issue, shed some much needed light on the clinical challenge of ovarian cancer.
For studies such as those described by Huntsman and colleagues and Crijns and colleagues, there is a critical need to validate the findings in independent sample sets. This validation can be performed as part of an intra-study design, in which a second set of cases are evaluated by the same investigators using a similar experimental methodology; or, as Crijns and colleagues have done, by using publicly available data from one or more completely independent studies. Crijns and colleagues used published data from an expression microarray analysis of 118 primary serous ovarian cancers. Even though this analysis had been performed on a different microarray platform, the investigators were able to identify a 57-gene signature, which was a sub-set of their 86-gene signature, that was able to predict which patients fell into the low- and high-risk prognostic groups they had defined.