Biological markers in ovarian cancer - Implications for clinical practice

No study has demonstrated a prognostic factor unique to ovarian cancers, and most studies have concentrated on genes analysed in more common tumours, particularly breast cancer. These tumours may share some hereditary genes and morphological features, but differ widely in natural history and response to therapy, and in the large group of patients with advanced disease and moderate to poor differentiation, no one factor has emerged as pre-eminent. Many potentially useful factors such as the growth factor receptor ERBB-2 are overexpressed in less than 10% of cases. Full assessment of the significance of the genetic changes described may require information on the functional effect of each of the different sites of mutations, and in the case of TP53 it is those mutations in the DNA binding regions of the protein which are most likely to adversely affect prognosis. There is preliminary evidence that mutations in the BRCA1 gene may also be an adverse factor, but many other genetic changes may be redundant or reflect interactions with other undiscovered genes. Such techniques are currently slow, but techniques are under development so that multiple genes may now be examined simultaneously from a few hundred cells using the polymerase chain reaction. One hypothesis is that patterns of gene abnormalities will emerge reflecting biological behaviour, rather than abnormalities in any one gene or codon. This may be independent of the functional group of the gene under study (growth factor, suppressor, repair, angiogenesis or metastasis). Analyses of multiple gene expression may lead to important but complex patterns of molecular differences between normal, dysplastic and malignant tissues, or between good and poor prognosis groups of tumours. As the alliance between biochemistry and molecular biology begins to deduce the rules governing the relationship between gene sequence and function, clinical scientists can start to make progress on the molecular basis of prognosis and the related problem of predicting of response to treatment. Improved prognostic markers in early stage patients could be used to predict which patients in the FIGO stage IA/B, IIA/B categories of well or moderately differentiated histology are likely to progress within 5 years. The principal management decision which follows from that analysis, is the rationale for adjuvant chemotherapy. A second issue in that group is the frequency of follow up and the extent of monitoring investigations required, as the majority of these patients have a minimal risk of relapse for the first 2-3 years following diagnosis. Based on present evidence, the analyses which measure proliferation, such as ploidy, have shown promise in multi-variate analyses, while more biochemically defined markers including MYC or Ki67 have yet to show additional information over and above stage and histology. In more advanced disease patients, the most widely studied is the TP53 gene, which is clearly related to prognosis, although not always independently of stage and grade in multivariate analyses. A second issue with this gene is its relation to cytotoxic agent induced apoptosis. It is likely that further data on the drug specificity of this property, and the association with DNA binding site mutation will soon emerge. Other factors such as those reflecting resistance to anticancer therapy (e.g. LRP or MDR1), may also be of more relevance, as the clinical management issues are principally those of selection of number, type and dose of cytotoxic drugs. uPA, its receptor and inhibitor seem worthy of further study. A third clinical problem is the management of residual masses of tissue after first-line therapy. This tumour tissue will reflect changes acquired as a result of cytotoxic drug exposure as well as those related to the generation of the malignant phenotype. Biopsy material will become available on tumour tissue in ovarian cancer as concepts such as second-look laparotomy are replaced by interventional debulking surgery and management of late relapse. A molecular profile, which predicted the likelihood of a further response to chemotherapy, would prove extremely useful, and assessment of TP53, MDM2, BCL2 and BAX status would appear to be valuable candidates.