Commentary

Predictive cancer genomics—what do we need?

Evaluation of the prognostic parameters of gynecologic cancer has shown their failure for classification according to the clinical behavior or the prediction of its outcome. This weakness has important implications on prognosis and treatment. The increasing understanding of the complexity of the human genome, coupled with the development of high throughput analysis techniques and bioinformatics tools, has changed our concepts on cancer biology, by shifting our targets to a global analysis of the transcriptome and the proteome, linking genes and their products into functional pathways. These approaches permit the documentation of expression patterns of thousands of genes within a cell. With the use of DNA microarray technology, it is feasible to identify signature patterns of expression in tumor samples that faithfully correlate with its biology, providing accurate prognosis for each cancer patient and thus a rational customized treatment. At this stage, there is a need for systematic studies for the validation of these novel approaches. In this review, we provide a basic background of the concept of the technology, highlight several emerging issues from their applications on gynecologic cancer, discuss a series of important themes and problems regarding their interpretation and relevance for the clinicians, and comment on future areas of research.

Many cytotoxic agents for the adjuvant treatment of breast cancer are available, but they have produced only modest results, even when the tumor burden is low. This relative lack of efficacy may be attributed, in part, to the nonspecificity of the current regimens. Additionally, there is evidence that the chemotherapy doses used in clinical practice are not optimal, which potentially compromises the outcomes when the thresholds of dose intensity are not reached. Variations in treatment underscore the need to return to the basics of chemotherapy administration: dose, schedule, concentration threshold, and therapeutic index. In patients with metastatic breast cancer a clear dose-response curve has been shown with some agents, including anthracyclines. The E-max model, which in its simplest form assumes a direct relation between the dose of a drug and its effect, may be used to improve dosing in the adjuvant treatment of breast cancer. Consistent with this model, threshold effects have been observed in treatment with both anthracyclines and paclitaxel for breast cancer. There is also evidence that using dose-dense schedules may produce better outcomes with some regimens. Maintaining chemotherapy agents at full dose on schedule is crucial to treatment success, especially in adjuvant therapy. Consequently, treatment practices should use both dose intensity and dose compression to increase the likelihood of positive outcomes in patients with breast cancer.

Mechanisms of resistance to cancer chemotherapy are poorly understood. Molecular pathways involving genes associated with inherited cancer syndromes could represent mechanisms of chemoresistance. Microarray techniques can identify simultaneous alterations in the mRNA expression of multiple genes, but identification of the exact mechanism responsible for a particular phenotype, including resistance to a specific drug, remains problematic. Genes in which mutations cause inherited cancers play vital roles in apoptosis, growth arrest and/or DNA repair, and are inactivated by somatic mutations, deletions or hypermethylation in most cancer tissues. Similarities between carcinogenic injury and cell damage created by chemotherapeutics suggests that somatic inactivation of such genes is crucially important to drug sensitivity.