The evolution of prognostic and predictive markers in breast cancer

Hani Katerji, Huina Zhang, David G. Hicks, Bradley M. Turner

[Abstract] Breast cancer is the most frequent cause of cancer death in women living in less developed regions of the world, and the second leading cause of cancer death in women living in more well-developed regions of the world. Breast cancer treatment options continue to evolve, with multiple treatment modalities now available; however, finding reliable predictive and prognostic biomarkers that will aid in selecting the appropriate treatment for breast cancer patients who would most likely respond to a specific therapy is still a major challenge, and has been the focus of many research groups throughout the past few decades. In this review, we attempt to provide an overview of the current prognostic and predictive markers in breast cancer including traditional, immunohistochemical, and genomic assays, and assess the potential clinical use of these markers.

[Key words] Breast cancer; Prognosis; Prediction; Biomarkers; Evolution

1 Introduction

Breast cancer is the most commonly diagnosed cancer worldwide in 2020 and the fifth leading cause of cancer mortality worldwide[1]. A growing understanding of the complex biology of breast cancer has shown that it is a diverse group of diseases that differ in histological features, clinical behavior, prognosis, transcriptional programming, and response to therapy[2].

The increased understanding of the biology of breast cancer has revolutionized the management of breast cancer, leading to more specific and personalized treatment protocols based on the biological characteristics of this disease. Still, only 2%-15% of patients who get chemotherapy based on the clinicopathologic features of the tumor benefit from it[3]. This has created the need for improved methodologies for prognosis(information on the likelihood of cancer progression in untreated patients) and prediction(information on the probability of therapy response) of breast cancer recurrence, and breast cancer response to treatment, respectively[4].

In the seminal paper by Perou et al[5], classification of breast cancer by molecular subtyping was classically described in 2000. In that study, a systematic investigation of gene expression patterns was done on thousands of human genes from different breast tumors in an effort to develop a methodology for classifying these tumors based on their gene expression patterns. That study resulted in the classification of breast cancers into five main subtypes, each of which showed distinct patterns of gene expression: 1)luminal A; 2)luminal B; 3)normal breast-like; 4)human epidermal growth factor receptor 2(HER2) enriched; and 5)basal-like subtypes. Each subtype differed in incidence, patterns of recurrence, survival, and response to therapy[6](Table 1).

Table 1 Intrinsic molecular subtypes of invasive breast cancer: biologic and clinical features(reproduced from reference[7])

In this review, we discuss the traditional pathologic markers and several multigene expression based assays, and their evolving role as predictive and prognostic markers in the treatment planning for breast cancer patients. Furthermore, we provide a practical step-wise approach to risk assessment in breast cancer patients as it applies to therapeutic decision making and treatment planning.

2 Traditional prognostic markers

2.1 Lymph node status

The single most important traditional prognostic marker for breast cancer is the lymph node status. Similar to other carcinomas, metastatic breast cancer is thought to progress through several stages. For breast cancer, these stages may include a progression from atypical hyperplasia, to intraductal carcinoma, to local invasion, followed by metastasis to the lymph nodes and/or distant sites[8-9]. A direct relationship exists between the number of involved lymph nodes and the risk of metastasis. Some studies[10-12]have shown that as the number of involved lymph nodes increased, the survival status decreased, regardless of tumor size.

Staging of the axilla has always played a central role in the treatment of breast cancer, as axillary lymph nodes have long been recognized as a route for breast cancer to spread systemically. The ACOSOG Z0011(Alliance) randomized clinical trial showed that among women with low-stage breast cancer, who 1)did not have palpable axillary adenopathy and, 2)had <3 sentinel lymph nodes containing metastases, the 10-year overall survival for patients treated with sentinel lymph node biopsy alone was non-inferior to those treated with axillary lymph node dissection[13]. As a result, targeted sentinel lymph node excision has become the standard of care, resulting in a reduction of the incidence and complications of axillary lymph node dissection.

2.2 Tumor size

The size of the primary tumor is one of the most important factors following the nodal status for decision making on adjuvant treatment. As with lymph node status, tumor size reflects the patient′s tumor burden, and has also been found to act as an independent prognostic indicator. As tumor size increases, survival decreases regardless of lymph node status. Carter et al[10]reported that with tumors <1 cm, the 5-year overall survival was close to 100%, compared to 89% for patients with a tumor size between 1 cm and 3 cm, and 86% for patients with a tumor size between 3 cm and 5 cm. It is also thought that the risk of developing metastases increases with tumor size, as the larger the cancer is at diagnosis(reflecting a higher tumor burden), the more cells are available to metastasize[14]. As a result, detection of smaller, early stage tumors can substantially reduce the likelihood of metastatic spread. Primary tumor size can be measured by imaging, or more accurately evaluated after surgical excision.

2.3 Histologic features

Histologic grade is another widely used parameter that helps in determining the prognosis in patients with invasive breast cancer[15]. Low-grade tumors have a more favorable prognosis, and high-grade tumors have a less favorable prognosis. The Nottingham grading system is a grading system for invasive breast cancer that results in a combined score based on three histologic factors: tubular formation, nuclear pleomorphism, and mitotic index[16]. Each of these histologic factors is assigned a score from 1 to 3, with 1 being consistent with a lower histologic grade, and 3 being consistent with a higher histologic grade. These scores are then added together. If the combined Nottingham tumor score is between 3 and 5, the tumor is assigned a Nottingham grade of 1, consistent with a well differentiated tumor and a more favorable prognosis. If the combined Nottingham tumor score is 8 or 9, it is assigned a Nottingham grade of 3, consistent with a poorly differentiated tumor and a less favorable prognosis. If the combined Nottingham tumor score is 6 or 7, it is assigned a Nottingham grade of 2, suggesting a prognosis between that of a Nottingham grade 1 and 3[17-18]. In 2017, the Nottingham grading system was incorporated into the American Joint Committee on Cancer(AJCC) for breast cancer staging[19]. When strict diagnostic criteria are applied, it has been shown that lower grade histologic subtypes of breast cancer and certain specific histologic subtypes, such as tubular carcinoma and mucinous carcinoma, have a more favorable prognosis when compared to higher grade histologic subtypes[20-21].

2.4 Immunohistochemical stains

The role of immunohistochemical stains(IHC) has increased dramatically in recent years and became a critical part of the routine workup for breast cancer. Different institutions use different sets of routine IHCs; however, the most characteristically used markers are estrogen receptor(ER), progesterone receptor(PR), and HER2. The positive or negative expressions of these three markers define the four different breast cancer subtypes: luminal A, luminal B, HER2 enriched, and triple negative(TN)[22]. Each breast cancer subtype has a different prognosis, with different implications for chemotherapeutic options. As such ER, PR, and HER2 provide prognostic and predictive information for the currently available anti-hormonal and anti-HER2 therapies.

Generally, ER expressing breast cancers are better differentiated, and have indolent behavior, a better prognosis, and are eligible for anti-estrogen hormonal therapy[23]. Even tumors with very low levels of ER expression have the potential to benefit from hormonal therapy. In addition to ER, PR is also a routinely tested transcription factor, and is largely regulated by ER. Both ER and PR markers are expressed in a large subset of breast cancers. Even though some studies suggest that the role of PR may not be as useful clinically as ER[24], multiple studies suggested that the loss of PR expression in ER-positive tumors is associated with a worse prognosis, a more aggressive clinical course, and decreased response to tamoxifen therapy[25-26]. The American Society of Clinical Oncology/College of American Pathologists(ASCO/CAP) guidelines recommend that ER and PR should be considered positive if ≥1% of tumor cells show nuclear staining of any intensity[27].

HER2 belongs to a family of transmembrane tyrosine kinase receptors that plays an important role in the regulation of cellular signaling that affects cell growth, differentiation, and survival[28]. HER2/neu gene amplification and/or protein overexpression has been identified in 15%-20% of invasive breast cancers[29]. Studies have shown that HER2 enriched breast cancer is associated with an aggressive clinical course and poor outcome[30]. The role of HER2 and its biology has been extensively studied, leading to the development of the drug trastuzumab, a humanized monoclonal antibody that directly targets the HER2 receptor. Trastuzumab has become an important therapeutic option for patients with HER2 positive breast cancer[31]. The success and evidence based clinical benefit of trastuzumab was followed by expanding the horizon for other HER2 targeted agents, including the humanized monoclonal antibody pertuzumab[32], tyrosine kinase inhibitors(lapatinib, neratinib, tucatinib)[33]and the antibody-drug conjugated ado-trastuzumab emtansine, which had shown improved outcomes in patients with HER2-positive early breast cancer who had residual invasive disease after completion of neoadjuvant therapy[34].

HER2 protein expression can be detected by IHC, whileHER2 gene amplification can be identified byin-situhybridization(ISH), most commonly fluorescent in-situ hybridization(FISH). Initial evaluation of HER2 by IHC has been widely-used due to its availability and low cost, and the equivocal case will be automatically reflexed to ISH testing. The evaluation of HER2 expression by immunohistochemistry in breast cancer is semi-quantitative rather than qualitative, since a background level of up to 20,000 HER2 receptor molecules is expressed in all breast epithelial cells[30]. Based on the intensity of the stain and the membranous staining pattern, tumors are classified as negative(scored as 0 or 1+), equivocal(2+) or positive(3+)[35].

In addition to the aforementioned markers, Ki-67 is another immunohistochemical stain that has been shown to be helpful for distinguishing breast cancers at a higher risk for recurrence[36]. Several studies have shown that Ki-67 has an independent predictive and prognostic value in terms of response to endocrine therapy and chemotherapy, as well as overall survival in breast cancer patients[37]. In a study that included 1951 cases of primary breast cancers, it was found that triple negative breast cancers showed the highest Ki-67 index[mean(50.9±23.7)%], followed by HER2/neu[mean(42.6±21.6)%] and luminal B cancers[mean(34.9±20.05)%], while luminal A cancers showed lowest Ki-67 index[mean(23.6±19.7)%][38]. However, the use of Ki-67 in routine clinical practice in breast cancer remains controversial due to the lack of a standardized procedure for Ki-67 assessment, and the persistence of several issues of debate with regards to the Ki-67 assay interpretation and the lack of a validated cut-off point[39].

2.5 Genomic tests

The introduction of genomic testing into the clinical practice has changed the approach of how breast cancer patients are evaluated for risk of recurrence and treatment alternatives. In addition to the conventional pathologic factors and biomarkers used for assessing tumor behavior, several multigene expression-based assays that offer prognostic and predictive information for risk of recurrence and treatment efficacy have become established and commercially available over the past several years.

Multigene expression-based assays differ in the technological platforms used to measure gene expression, the number and identity of assessed genes, and the patient populations used for clinical validation. Here we discuss the more commonly used multigene expression-based assays(Table 2).

Table 2 Commercially available molecular profiling tests for prognostication in ER positive breast cancer(reproduced from reference[40])

2.6 Oncotype DX

Oncotype DX(ODX) is a quantitative real-time reverse transcriptase polymerase chain reaction(qRT-PCR) mRNA-based multigene assay developed through collaborations between Genomic Health Incorporated and the National Surgical Adjuvant Breast and Bowel Project(NSABP). It is performed on RNA extracted from formalin-fixed paraffin-embedded(FFPE) breast cancer tissue samples that are ER-positive and lymph node-negative[41]. The development of the ODX assay depended mainly on cohort studies with long-term follow-up of patients with early breast cancer[41-42]. The assay evaluates mRNA from 21 genes, among which 16 are cancer-related genes, and 5 reference genes used to check RNA integrity and to normalize expression levels. The evaluated 16 cancer-related genes are heavily influenced by proliferation-related genes, hormone receptor-related genes, and HER2-related genes[41]. An algorithm is used to calculate an ODX recurrence score(ODXRS) that ranges from 0 to 100, reported as either low risk(<18), intermediate risk(18-30), or high risk(>30). The ODX assay is validated for predicting the risk of breast cancer recurrence at 10 years[41], and for predicting benefit from adjuvant chemotherapy in ER positive lymph node negative breast cancer[41,43-45]. The TAILORx results suggest that an ODXRS of less than 11 has an extremely low risk of developing recurrence, and that women with early breast cancer who are older than 50 years with an ODXRS <26 have a similar risk to patients in the lower risk ODX category[46]. The ODXRS provides a more accurate and reproducible measure of breast cancer risk of recurrence and therapeutic effectiveness than the standard evaluation of histologic variables alone.

2.7 MammaPrint and BluePrint

MammaPrint(MP) is a genomic assay that uses microarray technology to measure the expression of 70 genes that are involved in cell cycle and proliferation, invasion and metastasis, angiogenesis, and signal transduction[47]. This assay is marketed by Agendia, BV(Amsterdam, Netherlands). Its main use is as prognostic marker for risk of distant recurrence at 5 years. MammaPrint classifies patients into two risk groups: low risk, with a 1.3% risk of distant recurrence at 5 years, and high risk, with an 11.7% risk of distant recurrence at 5 years. Test validation of MP was performed in a study that included 295 breast cancers with available banked, fresh-frozen tissue from Netherlands Cancer Institute, including patients with and without lymph node involvement[48]. Results showed significantly different outcomes for patients with low-risk signatures(14.8% 10-year distant recurrence) vs high-risk signatures(50% 10-year distant recurrence). Interestingly, poor prognosis correlated with the traditional high-risk pathologic factors such as increased tumor size and high histologic grade.

The clinical benefit of MP was recently validated in a prospective randomized trial, the MINDACT trial[49]. This was a randomized phase Ⅲ study, which enrolled more than 6,600 patients with breast cancer who were ER positive and HER2 negative, ER positive and HER2 positive, ER negative and HER2 positive, or triple negative. Patients had MP performed to determine genomic risk, while clinical risk was determined by using Adjuvant! Online[50-51](the modified version 8.0 with HER2 status). The study included only patients with discordant genomic and clinical risk assessments. Patients were randomized to chemotherapy vs no chemotherapy. Results showed that patients classified as low risk by MP but at high risk based on clinical risk had an excellent outcome, with a 5-year distant metastasis-free survival of 94.7%. In these clinically high risk patients, the use of MP rather than the traditional criteria resulted in a 46% reduction in the use of adjuvant chemotherapy. Over all, when MP was used rather than the traditional criteria, there was a 14% reduction in the use of adjuvant chemotherapy. This study resulted in guideline recommendations by ASCO, suggesting that if a breast cancer patient is clinically high risk with a low risk MP, then MP may be used to inform decision to withhold adjuvant chemotherapy[52]. ASCO does not recommend using MP in breast cancer patients who are clinically low risk[52].

BluePrint(BP), an 80-gene breast cancer molecular subtyping microarray-based test also marketed by Agendia that was developed using an IHC-based clinical subtype as a guide[53]. The BP assay measures the expression of 80 genes that assess functional pathways which determine the intrinsic breast cancer molecular subtypes by measuring the similarity of the tested tumor to a representative profile luminal-type(58 genes), basal-type(28 genes), and HER2-type(4 genes)[54]. For each tested tumor, the similarity to all three representative profiles is calculated and the subtype with the greatest magnitude is determined to be the breast cancer molecular subtypes[54]. When BP is combined with MP, breast cancers can be classified into a luminal A signature(low risk), a luminal B(high risk), a HER2 signature, and a basal-type signature. Studies have shown that tumors of different BP signatures exhibit differences in long term survival and response to neoadjuvant therapy. Tumors with luminal-type BP signature have more favorable distant metastasis(DM) free survival but less pathological complete response to neoadjuvant therapy, whereas tumors with basal-type BP signature and HER-type BP signature tumors have less favorable DM free survival but are more sensitive to chemotherapy[53,55].

2.8 Prosigna

Prosigna is a second-generation multigene expression assay, comprised of 50 discriminator genes and 8 controls, popularly known as the Prediction Analysis of Microarray 50(PAM50) gene signature[56]. This test has been cleared by the United States Food and Drug Administration(US-FDA) for postmenopausal patients with ER positive cancer, who have undergone treatment, and only receive adjuvant endocrine therapy[5]. The assay is marketed by Prosigna(NanoString Technologies, Seattle,WA). Originally it comprised of 50 discriminator genes and 8 controls; however, it was found that four of those genes did not add any prognostic value, so a 46-gene expression profile is used to assign cancers to four PAM50 molecular subtypes: luminal A, luminal B, HER2 enriched, and basal-like. An 18-gene subset of the 46-gene panel is used to calculate a proliferation score, which in combination with the molecular profile and the pathologic tumor size determines the 10-year Prosigna risk of recurrence(ROR) score[57]. For lymph node negative patients, the ROR is classified as low risk(<5%, scores 0-40), intermediate risk(about 10%, scores 41-60), and high risk(>15%, scores 61-100).For lymph node positive patients, the ROR is classified as low risk(about 5%, scores 0-40) and high risk(about 25%, scores 41-100)[58]. Including the tumor size, the ROR risk strengthens the Prosigna test as a prognostic assay; however, the predictive value of this assay has not been widely tested.

3 Using IHC as a prognostic and predictive tool

Genomic testing is expensive, and not available to certain patients, particularly patients in less developed countries around the world. The list price reported in the Genomic Health 2017 Annual Report for the invasive breast carcinoma ODX test was ＄4,620[59-60]. Many of the previously discussed genomic assays measure similar variables assessed in the traditional pathologic evaluation of breast carcinoma, including tumor proliferation, tumor grade, and IHC for ER, PR, and HER2. As such, an alternative approach for risk-stratifying breast cancer is to use IHC-based markers for clinical testing, and it has been suggested that IHC can be used independently for subtyping breast cancer, as well as risk stratifying certain types of breast cancer patients[61-63].

Immunohistochemical markers such as ER,PR,HER2, and Ki-67 have showed great promise for risk-stratifying breast cancer patients into groups similar to subtypes that have been defined by gene expression studies[64-66]. The use of certain immunohistochemistry antibody panels for routine pathologic evaluation of newly diagnosed breast cancer patients might provide useful information for guiding clinical decisions about adjuvant therapies rapidly and cost-effectively; however, antibody panels must be validated rigorously using multiple patient cohorts from multiple institutions to define their practicality and clinical utility before entering routine clinical practice[64], as lack of reproducibility and reliability due to poor assay standardization is a concern[67].

Several national quality assurance programs have created guidelines to include standardization of pre-analytical, analytical, and post-analytical testing factors, as well as mandatory proficiency testing, resulting in improvement in the quality, reliability, and inter-laboratory agreement for breast cancer assays[68-72]. This has made the use of antibody panels results more feasible for predicting breast cancer risk. Below we discuss three of the more commonly known immunohistochemistry antibody panel assays and techniques used for predicting breast cancer risk of recurrence.

3.1 Mammostrat? immunohistochemistry-based multiplex assay

Mammostrat?(GE/Clarient, Aliso Viejo, CA), is an IHC assay that measures the expression of five proteins(p53, HTF9C, CEACAM5, nDRG1, and SLC7A5) in ER positive, lymph node negative, HER2 negative breast cancers. These five proteins are believed to be associated with risk of breast cancer recurrence[73-74]. Mammostrat? was developed in a predominantly post-menopausal cohort, and has been suggested for use in older patients with newly diagnosed, early stage breast cancer[75]. The expression levels of the five proteins are translated into a risk index classifying patients into low, moderate, or high risk for recurrence over 10 years. Mammostrat? has been validated using three independent institutional cohorts of patients[74], as well as using archival tissue samples from the NSABP B14 and B20 clinical trials. Bartlett et al[73,76]have also confirmed the efficacy and prognostic significance of Mammostrat? in a validation study of 3,837 cases from tamoxifen or exemestane treated node-positive or node-negative patients who were enrolled in the Tamoxifen Exemestane Adjuvant Multinational(TEAM) trial. Acs et al[77]have suggested that Mammostrat? may provide a better estimation of tumor behavior in certain subgroups of low-grade breast carcinomas. Murray et al[78]have suggested that this benefit is time-dependent, and most prognostic over the first five years of follow-up, with an ability to potentially predict early recurrence during this time period. Mislick et al[79]have suggested that Mammostrat? costs savings for women with early-stage breast cancer compared to ODX.

Mammostrat?, is currently considered investigational and/or experimental and is not generally accepted as standard of care in the evaluation or management of breast cancer patients.

3.2 IHC4 score

The IHC4 score was developed in 2011 at the Wolfson Institute of Preventative Medicine, London, UK on patients included in the translational arm of the retrospective cohort from the Arimidex, Tamoxifen, Alone or in Combination(ATAC) trial of ER-positive breast cancer patients[80]. The IHC4 assay is performed on FFPE tissue and uses the immunohistochemical assessment of ER, PR, Ki-67, and HER2 to calculate a risk score using weighting factors and an algorithm. The assay was validated and found to be prognostic in a cohort of 786 breast cancer patients, both pre- menopausal and post-menopausal status, none of whom received adjuvant therapy[81-82], and has been reported to have prognostic utility similar to the ODX recurrence score[80]. A recent study suggested that the IHC4 score was predictive of pathologic response following neo-adjuvant chemotherapy[83-84]. Investigators combined the IHC4 score with clinical and pathologic variables including the involvement of regional lymph nodes, size and extent of the primary tumor, histologic grade, and patient age to create an IHC4+ C score[85], which improved the prognostic accuracy of the IHC4 score. Recent studies have further validated the use of the IHC4+ C score for identifying patients at low risk who potentially can avoid adjuvant radiotherapy[86].

It is unclear how the IHC4 and IHC4+ C scores might perform with decentralized testing, given the interobserver and interlaboratory variability of semi-quantitative IHC testing. As such, the IHC4 and IHC4+ C scores should be considered investigational until additional prospective studies are done to further evaluate their interobserver and interlaboratory variability.

3.3Magee equations

The recent literature[59,64,87-92]suggests that algorithms and models that use ranges of ER, PR, HER2 and Ki-67 results can provide information similar to that from multigene assays with significant cost savings[65]. In 2008, Flanagan et al[87]published a linear equation(the original Magee equation) using different combinations of standard histopathologic variables, including the mitotic index, ER H-score, PR H-score, HER2, Ki-67, and tumor size, to calculate a recurrence score, which was shown to correlate well with the ODX recurrence score. In 2013, Klein et al[90]revised the original Magee equation, publishing three new linear equations(the new Magee equations), using the Nottingham score, ER H-score, PR H-score, HER2, Ki-67, and tumor size. These new Magee equations also calculated a recurrence score, which also was shown to correlate well with the ODX recurrence score. In 2015, Turner et al[91]published a modification of these new Magee equations, using a modification of the H-score, making it easier to use the equations(the modified Magee equations). These modified Magee equations were shown to correlate well with the ODX recurrence score, and were published with an algorithmic approach using an average modified Magee score supporting a “stepwise” risk stratification approach, with the omission of ODX testing in certain breast cancer patients, and the reflex of ODX testing in others. An algorithmic approach using the Magee equations published by Bhargava et al[93-94]also supports the potential for this “stepwise” risk stratification approach. Several other studies have reinforced that the use of the Magee equations can be helpful in predicting patients likely to have either a low or high ODX score, with the suggestion that it may be reasonable to omit multigene assays in certain situations, particularly when the cost is a consideration[95-99]. Turner et al[59]subsequently validated the 2015 modified Magee algorithmic approach, publishing the Rochester Modified Magee algorithm(RoMMa) in a multi-institutional study, with outcome data showing that only 2.0% of patients classified as low risk by the RoMMa had a breast cancer recurrence over 5-10 years of follow-up.

The Magee equations have also been suggested to be predictive for pathologic response to neoadjuvant chemotherapy[100-101]and to be predictive for distal metastatic risk in male breast cancer patients[102].

It is unclear how the Magee score might perform with decentralized testing, given the interobserver and interlaboratory variability of semi-quantitative IHC testing. As such, the Magee score should be considered investigational until additional prospective studies are done to further evaluate their interobserver and interlaboratory variability.

4 Summary

The new paradigms of pathologic diagnostic testing for breast cancer patients continue to evolve with the increased importance of personalized patient care. As multigene assay testing continues to offer improvements on current practice, the traditional clinical-pathologic paradigm provides a complement to this next-generation of testing, in an effort towards providing cost-effective, and cost-efficient medical care for breast cancer patients. A practical “stepwise” approach when risk-stratifying breast cancer patients might be to use information from the traditional clinical-pathologic paradigm to help identify patients with clinical and pathological metrics that will likely elicit information that is similar to multigene assay testing. In these cases, multigene assay testing may not provide any additional significant clinical utility, and would likely not be cost-effective or cost-efficient. Multigene assay testing could then be limited to cases where the assay results would potentially provide clinical utility beyond the available clinical and pathologic metrics. The potential cost savings to the health-care system would be significant. The successful integration of the traditional clinical-pathologic paradigm into these new paradigms of pathologic diagnostic testing will depend on the reproducibility of interobserver and interlaboratory results in both centralized and decentralized laboratory settings.

As new prognostic and predictive models emerge over the next several years, both the traditional clinical-pathologic paradigm and the next generation of multigene assay development will continue to evolve. New prognostic and predictive models include the use of nontraditional biomarkers, miRNA-based signatures, predictive algorithms, predictive nomograms, and predictive models based on the microenvironment of the tumor[64]. As prognostic and predictive models arrive on the scene, attention must be paid to the validity of prognostic and predictive modeling in diverse ethnic populations, because it is not clear how differences in ethnicity affect the outcomes predicted by the currently used prognostic and predictive models[64]. With these things in mind, the future of prognostic and predictive testing for breast cancer patients seems bright.