Complimentary to the assessment of long-established markers, a large effort has been put forth by many in order to determine whether larger networks of biomarkers and pathways may provide for a better description of a specific resistance phenotype. To this end, gene expression profiling by cDNA expression microarrays has subclassified breast cancer into four distinguished molecular classes based upon similarities in gene expression characteristics [45–49].

In addition to the routine assessment of ER, PR, HER2, and Ki67, several multigene signatures are currently being utilized to better predict outcomes and treatment response. The most commonly used tests are MammaPrint and OncotypeDX, which are both cleared for clinical use and have been approved by the Food and Drug Administration (FDA). The assessment and utilization of the recurrence score (RS) with Oncotype DX is endorsed by both ASCO and NCCN treating guidelines to assess the need and potential benefit of chemotherapy in early-stage breast cancer. These tests are typically used for patients with ER-positive tumors that are node-negative and are not recommended for patients with HER2-positive and triple-negative tumors.

The Oncotype DX assay was developed using real-time reverse transcriptase polymerase chain reaction (RT-PCR) for gene expression in paraffin-embedded tumors. This assay examines the expression of 21 genes (16 breast cancer genes and 5 reference genes) to estimate the 10-year distant recurrence risk of breast cancer and the potential benefit of adjuvant chemotherapy and hormonal therapy in patients with estrogen receptor-positive breast cancer [101–106]. This test has been validated from large retrospective studies of women with early-stage, estrogen receptor-positive breast cancer without lymph node involvement, with a subgroup of patients with <4 nodes involved with metastatic disease. This test was originally designed to assign a recurrence score that reflected the 10-year risk of recurrence in women with ER-positive disease with no involved lymph nodes. Patients were treated with either tamoxifen alone or tamoxifen and chemotherapy in two large NSAPBP studies (B14, n-668 and B20, n-651) [105–107]. The results of the tests are strongly influenced by the expression of the estrogen receptor and are negatively impacted by the absence of progesterone expression and a high HER2 expression. The assay may not be clinically as relevant in those patients with HER2 overexpressing and ER-positive tumors.

The assay initially grouped tumors in three groups with low risk (recurrence score 0–18), intermediate risk (recurrence score 18–30), and high risk for recurrence (recurrence score >31). Patients with a recurrence score in the low- and intermediate-risk groups were found to have limited additional benefit from adjuvant chemotherapy compared to tamoxifen alone. A special focus of the prospective studies has been to quantify the benefits of adjuvant chemotherapy in the intermediate-risk group. While initially only evaluated for women with node-negative disease, more recent studies have been expanded for women with 1–3 or >4 lymph nodes involved with metastatic disease. It also shows that they may not benefit from adding chemotherapy to hormonal therapy, as the benefits over adjuvant chemotherapy may not warrant the excess risk in toxicity [104, 108–111].

A retrospective analysis of the recurrence score in the ATAC trial evaluated recurrence scores from 1,231 evaluable patients treated with either tamoxifen or anastrozole. This trial further supported the use of the recurrence score to assess outcomes for distant recurrence in both node-positive (n = 306) and node-negative patients (n = 872), showing that the prognostic value for aromatase inhibitors is similar to those found with tamoxifen. The distribution of risk groups was similar in both groups, with the high risk representing 15 % and 17 %, respectively, of the entire group, and 59 % and 52 %, respectively, for those with a low risk based upon their recurrence score. However, the 9-year risk of distant recurrence was considerably higher in node-positive patients compared to node-negative patients, demonstrating both its prognostic and predictive value in each setting. The distant recurrence rates were 4, 12, and 25 % in low, intermediate, and high recurrence score, node-negative patients and 17, 28, and 49 % for node-positive patients. Further studies have suggested that the RS remains to be a very strong predictor for late recurrences and may be used to determine who should receive hormonal therapy beyond 5 years [112].

The impact of the Oncotype DX breast cancer assay on clinical decision making has been quite dramatic. A recent analysis of over 4,000 women revealed that the utilization of this assay altered the recommendations for adjuvant therapy in 33 % of the patients. These studies further suggested that adjuvant chemotherapy was offered in 6 % of low-risk patients, 37 % intermediate, and 83 % with high recurrence score [104, 108–110, 113–117].

Lastly, we await the results of prospective studies, such as the TailorRx and Responder trials, which will further elucidate the value of this test for clinical decision making (Fig. 22.1).

The MammaPrint gene assay utilizes cDNA microarrays obtained from fresh samples of breast cancer (not paraffin-embedded) to subclassify tumors into five subtypes, based upon their distinct gene expression profiles. These include luminal epithelial cell phenotypes (subtypes A and B), a basal epithelial cell type phenotype, a HER2 (+) phenotype, and a group of cancers expressing a “normal-like” gene profile. Based on earlier observations that the ER levels vary considerably between luminal or basal types [46–48], Van’t Veer and colleagues designed a 70-gene microarray platform to identify a “poor-prognosis signature” involving genes associated with the cell cycle, tumor invasion, metastasis, and angiogenesis [19, 21, 118]. An initial validation of this gene assay revealed a poor-prognosis signature as well as a good-prognosis signature. Those in the former group had a significantly worse 10-year overall survival rate [19, 21, 118, 119]. While initially developed as primarily a prognostic test, it has been further developed into a test with predictive value in adjuvant and neoadjuvant trials. Retrospective studies have suggested that the patients with a good-prognosis signature are unlikely to have a significant pathological response rate to chemotherapy [119–123].

The PAM50 gene assay test was recently approved in the United States to determine the risk of recurrence (ROR) score for patients with stage I/II, node-negative or stage II, node-positive (1–3 nodes) disease that is hormone receptor positive. A clinical challenge posed by the Oncotype DX assay is the uncertainly on whether to treat patients with an intermediate-risk recurrence score. The PAM50 assay measures the gene profile of 50 genes based on the four basic tumor subtypes (luminal A, luminal B, HER2 enriched, and basal-like) with the goal of defining a better distinction between the intermediate- and high-risk groups.

In two recent studies, tumor samples were obtained from 1,017 women with hormone receptor-positive breast cancer who received either anastrozole or tamoxifen in the TransATAC trial. The recurrence score utilizing the PAM50 (ROR) assay was then compared to the recurrence score for the Oncotype DX assay, showing that the ROR may offer a more definitive discrimination between intermediate- and high-risk groups. They identified more patients that were reclassified as high risk based upon the ROR score compared to the Oncotype DX recurrence score [124–131]. Furthermore, this test may be able to determine a group of patients that benefit from paclitaxel-based chemotherapy [126]. However, further prospective studies will be needed to validate the use of this assay in order to determine its validity and ultimate clinical use in decision making.