Twenty-one-gene assay: challenges and promises in translating personal genomics and whole-genome scans into personalized treatment of breast cancer.

TO THE EDITOR: In the September 1 issue of Journal of Clinical Oncology, Goldstein et al conclude that the 21-gene assay is a better recurrence predictor than the currently used Adjuvant!—a validated algorithm based on classical clinicopathologic features—and that it can be used for tailoring chemotherapy among individual patients with hormone receptor (HR) –positive operable breast cancer. The use of genomic markers for treatment decision making without consideration of standard conventional factors is not realistic. Personalized cancer treatment in an environment of personalized medicine is one of the major goals of clinical research in oncology. Personal genomics, pharmacogenomics, and whole-genome scans have stimulated major interest by national anticancer organizations, industry, academia, and the general public and have contributed to intensive, internationally collaborative research efforts. But despite these global efforts, only a few robust prognostic and predictive biomarkers, like HR and human epidermal growth factor receptor 2 (HER-2), have been validated and widely used in clinical practice. Although the 21-gene assay represents one, and perhaps the most promising, gene expression profiling marker to reach the level of testing in a large-scale randomized controlled trial (Trial Assigning Individual Option for Treatment [TAILORx]), the translation of genomics research into evidence-based clinical use faces multiple challenges. In addition to its strengths, the study by Goldsteinet al has several limitations: First, it is a retrospective, relatively small study of 465 patients that compared the predictive utility of the 21-gene assay with that of Adjuvant! Second, the study population, including both node-negative and node-positive patients who all received chemohormonal therapy, differs from the original population—HR-positive, node-negative patients who received tamoxifen—in whom the 21gene assay recurrence score (RS) was developed. Third, despite methodological strengths, there are several weaknesses in the techniques used to assess the predictive value of the original RS with regard to chemotherapy benefit. Fourth, and perhaps most important, neither the original RS study 8,9 nor the present study have considered the new standard targeted agents: aromatase inhibitors for HR-positive, postmenopausal patients and trastuzumab for HER-2–postitve patients. Because therapeutic strategy is changing, markers developed in tissues of patients who had not received current standard agents have limited value in current clinical practice. Therefore, the utility of RS for postmenopausal HR-positive and/or HER-2–positive patients is unknown. Two randomized trials test the efficacy of the 21-gene assay (TAILORx) and the 70-gene signature (Microarray in NodeNegative Disease Avoids Chemotherapy [MINDACT] trial). Genomics research in collected biospecimens (biobanks) will add important new information. It was once thought that simply stratifying individual patients into subgroups at high or low risk for developing breast cancer or recurrence and tailoring the most effective preventive or therapeutic intervention would revolutionize management of breast cancer. However, given the heterogeneity and complexity of breast cancer, this goal now appears elusive. Indeed, some individual primary tumors contain small cancer cell subpopulations, like breast cancer stem cells, that vary in metastatic ability and treatment response. Moreover, different deregulated signaling pathways and various biologic processes, including angiogenesis, tumor microenvironment, and dormancy, prove that individualized treatment approaches in solid cancers will require extremely hard work and sophisticated protocols. There is promise that the next-generation sequencing technology will provide newer genotyping platforms with more than 1 million single-nucleotide polymorphisms and copy number variants. Genome-wide association studies that used platforms with smaller number of single-nucleotide polymorphisms/copy number variants have discovered novel genetic risk variants. Although each risk variant confers only a small effect on cancer risk, and thus it is irrelevant, their combination may have clinical implications. Completion of breast cancer genetic mapping and functional studies to define the role of key genes and signaling pathways will enable future personalized studies. Comparison of whole-genome scans from relapsed nonresponders and relapse-free responders in largescale, prospective, unbiased studies that record both “classic” clinicopathologic features and novel genetic risk variants might result in the development of individualized therapeutic approaches preventing both locoregional and distant recurrences. The results of the two currently underway randomized trials on genomic biomarkers, TAILORx and MINDACT, will substantially contribute to evidence-based decisions in the treatment of early breast cancer.