Oral presentation | Open | Published:
Assessing prognosis for early breast cancer: clinical versus genetic profiles
Breast Cancer Researchvolume 9, Article number: S1 (2007)
Since the introduction of systemic adjuvant chemotherapy (ACT) and endocrine therapy in the early 1970s, the determination of risk of recurrence and death from breast cancer became a critical piece of information in the selection of the optimal postoperative treatment strategy. Classical histopathological prognostic factors included tumor size, regional lymph node metastases and number of axillary nodes involved, tumor grade, presence of lymphovascular invasion, and, more recently, estrogen receptor (ER) and progesterone receptor status, measurement of proliferative activity (S-phase fraction, mitotic index, Ki-67), and HER2 overexpression/amplification. As isolated factors, they have limited predictive ability in the case of individual patients. For that reason, prognostic indices were developed. The most successful is Adjuvant!Online, an online nomogram developed by Peter Ravdin. This nomogram incorporates tumor size, axillary nodal status, tumor grade, ER status, age and comorbidity. The nomogram will provide an assessment of recurrence and mortality rates at 10 years, including deaths due to comorbid conditions. In addition, the nomogram also calculates relative and absolute benefit from various adjuvant interventions: tamoxifen, aromatase inhibitors, and first-generation, second-generation and third-generation ACT regimens. The prognostic and predictive value of this nomogram has been externally validated, with a margin of error ≤1%. Over the past decade, high-throughput technologies have been developed based on gene expression profiling. These include between two and a couple of hundred genes, and have the ability to separate patients with excellent outcomes from those with higher risk. One of these prognostic profiles has been externally validated and is currently undergoing testing for clinical utility in a large, multicenter, prospective randomized trial (MINDACT). Another approach was based on prospectively identifying a set of genes from the literature and from the results of gene expression profiling. Mathematical modeling then led to the selection of 16 genes related to cell proliferation, ER-driven genes, HER2 and proteases, as well as five 'housekeeping' genes (OncotypeDx). This assay is based on RT-PCR, is reproducible and applicable to archival, paraffin-embedded material, and has been shown to predict prognosis in patients with lymph-node-negative, ER-positive primary breast cancer. Further testing indicated that the assay might also predict sensitivity to tamoxifen, or first-generation adjuvant chemotherapy. This assay is also under evaluation for clinical utility in a large, multicenter, prospective randomized trial (TailoRx). Whether these multigene predictors of prognosis will have greater utility than Adjuvant!Online remains to be determined. In the meantime, exploratory analyses are ongoing to identify reliable predictors of response to individual drugs and modern combination drug regimens. These are expected to lead to individualized selection of treatment, or personalized medicine.