The Human Genome Project was completed in 2003 and represented the culmination of 13 years of research coordinated by the US Department of Energy and the National Institutes of Health. Not only did this accomplishment result in the identification of the approximately 20,000 to 25,000 genes in human DNA, but the endeavor also helped foster the creation of newer, faster, and cheaper methods of gene sequencing. The result has been a rapid and expansive permeation of every specialty of medicine by genomic analysis, diagnostics, and prognostic markers. The impact on the field of surgical oncology, and specifically breast cancer, has been astounding. For most solid organ cancers, multidisciplinary treatment with curative intent is dependent on the ability to achieve complete resection, followed by adjuvant systemic therapy and/or radiotherapy. For many patients with breast cancer, this standard approach is effective and lasting. For other patients, however, recurrence and metastasis occur despite the best standard treatment. Historically, we have been ill-equipped, in most instances, to predict which patients will fare well and which will not. Because cancer is a disease so intimately associated with genetic mutation, genomic exploration should theoretically result in a better definition of the factors resulting in such disparate outcomes.
The International Cancer Genome Consortium was formed in 2008 to coordinate efforts to sequence 500 tumors from 50 cancers. As a result, full genome sequences have been reported for breast cancer. There are multiple potential ways in which this genomic information may ultimately translate to improved patient care and outcomes. Trastuzumab(Drug information on trastuzumab) (Herceptin), a monoclonal antibody that inhibits the human epidermal growth factor receptor 2 (HER2)/neu receptor, is now in widespread use in the treatment of breast cancer. It represents a success story in targeted therapy development, conferring an improved survival when used in the setting of HER2/neu-amplified breast cancer. However, one important lesson that has emerged thus far is the realization that simply knowing a cancer’s genetic sequence and identifying genetic variability do not necessarily translate into improved therapeutics. The enormous complexity and heterogeneity of cancer often contributes to the failure of targeted therapies. This variability in the success of certain targeted therapies is not surprising when one considers the number of prospective mutations that can be identified within a single cancer. For example, within the genome of the basal-like breast cancer that was sequenced, there were 27,173 point mutations, 200 of which were in protein-coding regions. The difficulty is in determining which mutations cause and accelerate carcinogenesis, as opposed to those that are simply by-products of failed DNA-repair mechanisms. A further challenge is posed by the finding that many genomic mutations interface with multiple signalling pathways in ways that vary widely among patients with the same histologic cancer—ie, with similar clinicopathologic features. Thus, therapies that target a single gene or a single signalling pathway (such as HER2/neu) may be highly effective in one patient and not at all effective in others.
One of the fastest growing areas of oncology-driven molecular tools is that of the development and implementation of molecular diagnostics for prognostication. Those of us caring for patients with breast cancer have witnessed an exponential rise in the number of available diagnostics over the past several years. Gökmen-Polar and Badve review the data on the commercially available assays, with a primary focus on the 21-gene assay (Oncotype DX) and the 70-gene signature (MammaPrint). They question whether these tests are ready for prime time—to which we would respond that, ready or not, they are already in widespread use. Most would agree that molecular profiling tests have the potential to improve clinical decision making. What remains challenging for most clinicians, however, is determining when the test should be utilized and how best to incorporate the results into the making of treatment decisions. We agree with Gökmen-Polar and Badve that these molecular tools should not replace the prognostic significance of standard clinicopathological features, but rather should be used to strengthen treatment decisions or help with decisions in patients in whom the clinicopathological findings are indeterminate. Tang et al recently published their findings regarding the use of a combinatorial tool that includes not only the recurrence score but also clinicopathological parameters (RSPC tool) and found that the RSPC tool further refined the assessment of distant recurrence risk and reduced the number of patients classified as intermediate risk. At our institution, treatment decisions are made in a multidisciplinary setting, and we selectively use the multigene classifiers. Typical cases in which we use these include patients in whom decisions regarding adjuvant systemic therapy are unclear based on routine parameters, patients who are undecided or unsure about the systemic therapy recommendations (and in whom a multigene classifier can further strengthen the decision made), and as part of clinical trial enrollment. We strongly discourage the use of these assays when the results are not expected to influence our clinical decisions.
The genomic revolution and the resulting molecular tools that continue to be developed promise enormous potential for what all of us are seeking: personalized patient care and prevention strategies. Molecular prognostication is here and, as clinicians, we need to be leaders in understanding the technologies, interpreting the results, and determining their utility in clinical practice. Future randomized studies should focus on determining which patients benefit most from the inclusion of molecular diagnostics in treatment decision making and on the development of treatment algorithms that incorporate patient factors, histologic and biologic findings, and molecular markers.
Financial Disclosure: The authors have no significant financial interest or other relationship with the manufacturers of any products or providers of any service mentioned in this article.