Adjuvant Systemic Therapy

Minal Shah

Claudine J. D. Isaacs

Minetta C. Liu

A significant number of women with invasive breast cancer lack any clinical evidence of metastatic disease at initial diagnosis but subsequently develop distant spread to such sites as the liver, lung, or bone. The aim of adjuvant systemic therapy, defined as the administration of cytotoxic chemotherapy, biologic therapy, and/or endocrine manipulation after the completion of definitive surgery, is to eradicate clinically inapparent micrometastatic disease. Randomized clinical trials and the overviews published by the Early Breast Cancer Trialists’ Collaborative Group (EBCTCG) (1,2,3) have shown that adjuvant therapy for early-stage breast cancer is effective in reducing the rate of recurrence and mortality from this disease. In this chapter, we review the current approach to adjuvant systemic treatment, with a focus on the factors involved in patient selection, the available treatment modalities, and recommendations for the management of specific subgroups of patients.

Prognostic Factors

Overview

Determining the appropriate adjuvant systemic therapy regimen for an individual patient requires an estimate of her underlying risk of disease recurrence and a determination of the therapy to which she is most likely to respond. The former is based primarily on tumor-related prognostic factors that reflect tumor biology irrespective of the treatment administered. This is in contrast to predictive factors, which are indicative of responsiveness to specific therapies and are discussed later in association with general treatment recommendations. It is important to note, however, that individual factors may be prognostic, predictive, or both.

Prognostic markers are patient- or tumor-related factors that allow for a determination of clinical outcome in the absence of systemic therapy. The use of prognostic factors provides the clinician with a means to estimate the benefits of adjuvant therapy and thus forms the basis for adjuvant treatment recommendations. Prognostic factors tend to reflect the underlying biology of the tumor, such as its ability to proliferate, invade, and induce angiogenesis. Tumor size, hormone receptor status, Her2/neu (i.e., c-erbB2) status, histopathologic features, including tumor type and grade, and the number of involved lymph nodes are all markers that have documented prognostic value. Other indicators, such as proliferative rate, the presence of lymphatic or vascular space invasion, tumor angiogenesis, and the presence of micrometastases in lymph nodes or bone marrow, have varied significance in different studies. Gene expression profiles generated by high-throughput, RNA-based methodologies are also under active investigation. Data regarding potential new prognostic factors are constantly emerging, but the studies are frequently challenging to interpret because they are based on retrospective analyses of a relatively small number of patients, include a mix of treated and untreated women, and often do not control for other known prognostic factors.

Specific Prognostic Factors

Lymph Node Status

Axillary lymph node status is recognized as the single most important predictor of breast cancer recurrence. A Surveillance Epidemiology and End Results (SEER) analysis of 24,740 breast cancer cases demonstrated a clear association between the number of involved lymph nodes and survival (4). Five-year overall survival (OS) was 92% for women with node-negative disease, 81% for those with one to three positive axillary lymph nodes, and 57% for those with four or more involved nodes. An analysis based on approximately 10,000 breast cancer patients included in the San Antonio Database confirmed these findings with a striking correlation between disease-free survival (DFS) and OS and axillary lymph node status (5).

Tumor Size

A SEER study with more than 13,000 node-negative patients demonstrated a significant correlation between tumor size and patient outcomes (4). Five-year relative OS of close to 99% was seen in patients with tumors measuring <1 cm, compared with 89% for those with tumors 1 to 3 cm and 86% for those with tumors 3 to 5 cm. In addition, a retrospective analysis of 767 women with node-negative cancer conducted by Rosen and colleagues reported a statistically significant association between tumor size and long-term outcome, with an estimated 20-year recurrence-free survival (RFS) of 88% for women with tumors measuring ≤1 cm, 72% for those with 1.1- to 3.0-cm tumors, and 59% for those with 3.1- to 5.0-cm tumors (6).

Histopathologic Features

Certain histopathologic subtypes, including pure tubular, papillary, and medullary carcinomas, are associated with a particularly good prognosis and long-term recurrence rates of less than 10% (7). In addition, histologic grade has been shown to be an important predictor of clinical outcome. One study of 1,262 women found a statistically significant correlation between histologic grade and 5-year DFS for both node-negative and node-positive patients (8). Nonetheless, the application of histologic grade in determining prognosis has been hampered by interobserver variability. Guidelines such as those proposed by Elston and Ellis have therefore been developed in an attempt to improve the reproducibility of tumor grading (9). These grading systems provide a standardized approach by which to score tumors that incorporates values for nuclear grade, mitotic activity, and architectural grade as measured by the degree of tubule formation. A number of studies using this grading system demonstrate that tumor grade provides statistically significant, independent prognostic information on multivariate analysis (10,11,12).

Hormone Receptor Status

Women with estrogen receptor (ER)-positive tumors have a small but statistically significant survival advantage of 9% to 10% in comparison to those with ER-negative tumors. In the National Surgical Adjuvant Breast and Bowel Project (NSABP) B-06 study, ER positivity was associated with a 5-year DFS of 74% and a 5-year OS of 92% (13), compared with 66% and 82%, respectively, for patients with ER-negative tumors. Similar results were obtained by the University of Texas Health Sciences Center, where studies with more than 2,000 women demonstrated a 5-year OS of 84% in patients with ER-positive disease and 75% in patients with ER-negative disease (14,15). In addition, more recently published results from a large retrospective study of untreated patients indicate that progesterone receptor status adds modest independent prognostic information over and above that provided by ER status alone (16).

Lymphatic and Vascular Space Invasion

Several groups have demonstrated that peritumoral lymphatic and blood vessel invasion predict for both local and distant recurrence (17,18,19,20,21). The Ludwig Breast Cancer Study Group, for example, conducted a study of women with node-negative breast cancer who were randomized to receive treatment with perioperative chemotherapy; the presence of peritumoral vascular invasion was associated with an approximately 15% increase in the risk of recurrence at 5 years (17). Others have demonstrated that the DFS and OS of women with tumors <2 cm was adversely impacted by the presence of lymphatic or vascular space invasion, irrespective of lymph node positivity or negativity (21).

Proliferative Rate

A number of methods are available for determining the rate of tumor cell proliferation, including DNA flow cytometry for ploidy and S phase fraction (SPF), as well as determinations for the thymidine labeling index, mitotic index, and extent of bromodeoxyuridine incorporation. While most evidence suggests that ploidy is not a significant prognostic factor, several studies support the notion that tumors with a high SPF are associated with inferior outcomes when compared to those with a low SPF. In one large study analyzing the outcome of 28,000 women followed in the San Antonio Database, high SPF was found to be an independent predictor of mortality (relative risk = 1.29; p <0.0001) (22). Several studies employing a variety of other methodologies to determine proliferative rate also suggest that rapidly proliferating tumors are associated with a poorer outcome (23,24,25,26). There are several drawbacks to the routine use of this information, however, particularly in relation to methodologic differences and a lack of standardization in estimating the proliferative rate.

Her2/Neu Status

The Her2/neu (c-erbB2) protooncogene was first identified in 1983. It is a member of the epidermal growth factor receptor family, and it is overexpressed in 20% to 30% of invasive breast cancers (27). Her2/neu positivity is defined by protein overexpression (assessed by immunohistochemistry) and/or gene amplification (assessed by fluorescence in situ hybridization or chromogenic in situ hybridization) (28). Most studies report that Her2/neu overexpression or amplification in both node-positive and node-negative patients is associated with poorer DFS and OS (25,29,30,31,32,33,34,35,36). The interpretation of these studies may have been hampered by small study size, the inclusion of systemically treated and untreated patients, and variable methods of Her2/neu analysis. Nonetheless, it appears that Her2/neu is an independent prognostic factor in addition to its role as a strong predictor of response to such Her2/neu-directed therapies as trastuzumab and lapatinib and to its questionable role as a predictor of response to anthracycline-based chemotherapy (37,38).

Upa and Pai-1

High levels of uPA and/or its inhibitor, PAI-1, appear to be promising and significant prognostic indicators (39,40,41,42). In a large retrospective study employing the European Organization for Research and Treatment of Cancer’s data sets (42), high levels of uPA and/or PAI-1 were strong independent predictors of survival in node-negative patients who received no adjuvant systemic therapy; the absolute differences in 10-year RFS and OS were 35% and 28%, respectively, between those with the highest levels of these factors compared with those with the lowest levels. Results from a prospective randomized clinical trial confirm that these factors offer important prognostic information (39,40,41). The final 10-year analysis of systemically untreated, node-negative participants in this trial demonstrated a DFS rate of 87% in patients whose tumors had low levels of uPA and PAI-1, in comparison to 77% in patients with a high level of either of these factors (p = 0.01). In addition, uPA/PAI-1 levels and tumor grade were the only independent predictors of outcome on multivariate analysis. The routine use of this factor in clinical practice is hampered by the need for fresh, non-paraffin-embedded tissue. The ongoing Node Negative Breast Cancer-3 trial will further evaluate the utility of this assay.

Adjuvant!

Adjuvant! (http://www.adjuvantonline.com) is a Web-based program that utilizes traditional prognostic factors including age, comorbidities, nodal status, tumor size, tumor grade, and hormone receptor status to compute and project 10-year rates of DFS and OS without and with selected adjuvant systemic treatment regimens (43). A population-based, retrospective validation study of Adjuvant! revealed that the observed and predicted outcomes were within 1% to 2% for OS, event-free survival (EFS), and breast cancer-specific survival (BCSS) in the absence of systemic therapy and with systemic therapy, respectively (44). This tool has a number of limitations, however, and adjustments are required to account for such additional factors as lymphovascular space invasion, younger age, and Her2/neu status. In addition, further validation is needed to evaluate the accuracy of predicted outcomes by Adjuvant! in the setting of more modern chemotherapy and endocrine therapy regimens.

Gene Expression Signatures

The use of gene expression profiling is a promising technique for identifying novel prognostic and predictive factors. Two platforms are commercially available for use in the clinical setting, both of which are undergoing large-scale prospective validation.
The 21-gene recurrence score (Oncotype DX; Genomic Health, Inc., Redwood City, CA) uses a real-time reverse transcriptase-polymerase chain reaction-based method and requires formalin fixed, paraffin-embedded tumor tissue. The assay includes genes related to cellular proliferation, estrogen receptor signaling, the Her2/neu pathway, and tumor invasive potential in addition to five reference genes that serve as an internal control to generate scores of 0 to 100 (45,46,47). Three risk categories have been defined on the basis of recurrence scores (RS) from banked tumor tissue and known clinical outcomes from postmenopausal patients with node-negative, hormone receptor–positive breast cancer treated with tamoxifen versus placebo in the absence of chemotherapy on the NSABP B-14 clinical trial: low risk (RS <18), intermediate risk (RS 18 to 31), and high risk (RS >31), with estimated rates of distant recurrence at 10 years of 6.8%, 14.3%, and 30.5%, respectively (47). Another retrospective study based on similar patients treated on the NSABP B-20 clinical trial of adjuvant chemotherapy and tamoxifen versus tamoxifen alone supports the prognostic utility of the recurrence score and demonstrates the potential to predict for benefit with chemotherapy, seen clearly in the high-risk group and questionably in the intermediate-risk group (48). Evidence-based guidelines, including those put forth by the National Comprehensive Cancer Network (http://www.nccn.org) and the American Society of Clinical Oncology (30), endorse consideration of the 21-gene recurrence score in making treatment decisions for patients with hormone receptor-positive, node-negative, early-stage breast cancer. Limited retrospective data suggest that this finding may be applicable to patients with hormone receptor-positive, node-positive breast cancer as well (49). The ongoing multicenter TAILORx trial in the United States will prospectively validate the 21-gene recurrence score in hormone receptor-positive, node-negative breast cancer by following patients with scores <11 on endocrine therapy alone, randomizing patients with scores of 11 to 25 to chemotherapy and endocrine therapy, and following patients with scores >25 on chemotherapy and endocrine therapy.

The 70-gene profile (MammoPrint; Agendia, Huntingdon Beach, CA), developed by investigators in the Netherlands, is available for clinical use on either fresh-frozen tumor tissue or fresh tissue collected and shipped at room temperature in RNAlater (50,51). Large-scale retrospective validation of this profile was performed on frozen breast tumor specimens collected from 326 node-negative patients through a collaborative effort of the TRANSBIG consortium (52). The risk of recurrence was estimated for each patient using both the Adjuvant! program and the 70-gene profile for comparison to observed clinical outcomes. There was a high degree of concordance, although the 70-gene profile provided additional risk stratification independent of the clinical variables used in Adjuvant!. The ongoing multicenter MINDACT Trial in Europe is prospectively evaluating the utility of 70-gene profile versus standard clinical-pathologic criteria in appropriately selecting breast cancer patients with zero to three involved lymph nodes for adjuvant chemotherapy.

Occult Axillary Lymph Node or Bone Marrow Metastases

Although more detailed analysis of the axillary lymph nodes by serial sectioning and/or immunohistochemical analysis increases the likelihood of detecting a positive node, the prognostic significance of this finding is uncertain (53,54,55,56). This is particularly true for sentinel lymph nodes. Ongoing prospective clinical trials, such as the American College of Surgeons Oncology Group (ACOSOG) Z0010 study and NSABP B-32, will address the prognostic relevance of occult-involved sentinel nodes. Until the results of these studies are available, these findings must be considered investigational and should not be incorporated into patient management (57). In addition, several retrospective studies have examined the independent prognostic significance of occult bone marrow involvement, with conflicting results (58,59,60,61,62), although one prospective study did demonstrate that the presence of cytokeratin-positive micrometastatic cells in the bone marrow independently predicted for a poorer outcome at 3 years of follow-up (58,59). The ACOSOG Z0010 study is also examining the significance of occult tumor involvement in the bone marrow, and the results of this study should shed further light on the independent significance of this finding.

Summary

Despite the number of recognized and proposed prognostic factors, the determination of an individual patient’s risk of recurrence remains challenging. Axillary lymph node status remains the most important determinant of patient outcome. In addition, tumor size, hormone receptor status, and histopathologic features play a role in predicting the risk of breast cancer recurrence, particularly in lymph node-negative patients. A number of gene signatures are also under active investigation as potential prognostic indicators. Their role has yet to be clearly defined, but ongoing research and validation should assist in realizing the goal of objectively individualizing therapy to minimize toxicity from nonbeneficial therapies.

Principles of Adjuvant Chemotherapy

Overview

Randomized clinical trials have demonstrated that adjuvant therapy reduces the risk of recurrent breast cancer and improves the survival of patients with both lymph node-negative and lymph node–positive disease. The results of the EBCTCG overview (1,2,3) indicate that the percentage reduction in the risk of recurrence and mortality is identical for both node-negative and node-positive patients. Therefore, the higher risk of relapse in node-positive patients suggests they will derive a greater absolute benefit from adjuvant therapy than those with node-negative disease. For example, a woman with node-positive breast cancer who has a 10-year risk of recurrence of 70% and a 10-year risk of mortality of 50% in the absence of systemic therapy would expect to obtain a 25% absolute improvement in her DFS and a 15% improvement in her OS when treated with optimal adjuvant therapy. However, a node-negative patient with a 10-year risk of recurrence of about 40% and 10-year risk of mortality of 20% in the absence of systemic therapy would have about a 13% absolute improvement in her risk of recurrence and a 5% absolute improvement in her risk of death when treated with adjuvant therapy (1,2,3,43). Although overall these represent only modest improvements in the rates of
survival, these differences translate into thousands of lives saved yearly in the United States alone because of the high incidence of breast cancer. The identification of the patients most likely to benefit from therapy and the use of optimal adjuvant regimens are therefore key to maximizing the benefits of systemic treatment.

Benefits of Therapy

Randomized trials of adjuvant chemotherapy have been ongoing since the mid-1970s, and the long-term effectiveness of this therapy has been established. Initial chemotherapy trials were conducted in node-positive patients, demonstrating that chemotherapy administration is associated with significant reductions in the rates of recurrence and death. The Milan trial randomized women with node-positive breast cancer to receive no adjuvant systemic therapy or to receive the combination of cyclophosphamide, methotrexate, and 5-fluorouracil (CMF) over 12 months (63). At 20 years’ follow-up, the RFS rate was 36% for those receiving CMF and 27% in the control patients, with an OS rate of 34% in those treated with chemotherapy and 24% in patients who had mastectomy alone. Of note, a second early study conducted by the NSABP confirmed the benefits of chemotherapy in this patient group (64).

Similarly, significant improvements in DFS with the use of adjuvant chemotherapy have been demonstrated in node-negative women (65). The NSABP B-13 trial randomized patients with node-negative, hormone receptor–negative breast cancer to receive either 12 months of methotrexate and 5-fluorouracil (MF) or no adjuvant therapy (66). With 8 years of follow-up, a significant increase in DFS was seen: 74% in the MF group compared with 59% in the control population. An intergroup study confirmed that chemotherapy was associated with a significant decrease in the rate of recurrence, with a 5-year DFS of 83% in the treated group versus 61% in the control population (67).

The meta-analysis from the EBCTCG provides a comprehensive overview of the benefits of adjuvant chemotherapy and endocrine therapy (1). Data regarding 15-year survival for patients who received prolonged (>2 months) combination chemotherapy versus no chemotherapy demonstrated clear benefits in favor of adjuvant therapy. Overall, combination chemotherapy resulted in a 12.4% and 4.2% absolute reduction in the risk of recurrence for women age <50 and 50 to 69 years, respectively, and a 10.0% and 3.0% absolute reduction in breast cancer–related mortality for women age <50 and 50 to 69 years, respectively. These reductions in risk were statistically significant (2p < 0.00001) irrespective of nodal status, although the benefits were greatest in the setting of node-positive disease.

Although the overall efficacy of adjuvant chemotherapy is firmly established, the optimal regimen for women with early-stage breast cancer continues to be refined for various subsets of patients. In particular, questions have arisen regarding the role of anthracyclines, taxanes, and dose-dense chemotherapy regimens. In terms of anthracycline-based chemotherapy, there are data to suggest that four cycles of doxorubicin and cyclophosphamide (AC) over 12 weeks are equivalent to six cycles of classical CMF over 24 weeks (68,69). In contrast, the most recently published EBCTCG overview demonstrated that about 6 months of anthracycline-based chemotherapy is associated with a 38% reduction in annual breast cancer-related deaths for women age <50 years and a 20% reduction in annual breast cancer–related deaths for women between ages 50 and 69 years irrespective of nodal status (1). Other randomized clinical trials have also demonstrated superiority of six cycles of an anthracycline-containing regimen over CMF in high-risk patients. One study, led by the National Cancer Institute of Canada, randomized premenopausal and perimenopausal women with node-positive disease to receive six cycles of classical CMF or six cycles of cyclophosphamide, epirubicin, and 5-fluorouracil (CEF), and DFS and OS were significantly prolonged in women receiving CEF (57,58); the 10-year DFS rates were 52% versus 45% (hazard ratio [HR] 1.31; p = 0.005), and the 10-year OS rates were 62% versus 58% (HR 1.18; p = 0.05) (70). The Intergroup INT-0102 study, on the other hand, focused on node-negative patients and compared six cycles of classical CMF to six cycles of cyclophosphamide, doxorubicin, and 5-fluorouracil (CAF), demonstrating equivalence in 10-year DFS (77% vs. 75%; HR 1.09; p = 0.13; 2p = 0.26) and only marginal superiority in 10-year OS (85% vs. 82%; HR 1.19; p = 0.03; 2p = 0.06) (71).

The role of taxanes in the adjuvant treatment of patients with node-positive breast cancer has also been examined. Initial studies compared four cycles of standard 3-weekly AC to four cycles of standard 3-weekly AC followed by four 3-weekly cycles of paclitaxel, and more favorable clinical outcomes were obtained with the addition of paclitaxel (72,73). Furthermore, additional benefits may be achieved with increased dose density, as demonstrated by Cancer and Leukemia Group B (CALGB) protocol 9741; administration of AC and paclitaxel at the same doses in 2-weekly treatment cycles with granulocyte growth factor support (as opposed to the standard 3-weekly cycles) led to incremental but statistically significant improvements in DFS (7% absolute improvement; HR 0.74; p = 0.01) and OS (2% absolute benefit; HR 0.69; p = 0.01) with 3 years’ median follow-up (74).

Other randomized clinical trials confirm the DFS and OS benefits at 5 years with sequential or concurrent anthracycline and taxane-based chemotherapy in patients with operable, node-positive invasive breast cancer (Table 18.1). These benefits have been extrapolated to include patients with high-risk, node-negative disease. In the absence of data to confirm the superiority of one regimen over another, acceptable treatment regimens include the following: (a) four cycles of 2-weekly AC followed by four cycles of 2-weekly paclitaxel (74), (b) four cycles of 3-weekly AC followed by 12 weeks of weekly paclitaxel (75), (c) four cycles of 3-weekly AC followed by four cycles of 3-weekly docetaxel (75,76), (d) six cycles of 3-weekly docetaxel, doxorubicin, and cyclophosphamide (77), and (e) three cycles of 3-weekly 5-fluorouracil, epirubicin, and cyclophosphamide followed by three cycles 3-weekly of docetaxel (78). More than 20,000 women have been enrolled in a variety of studies investigating the combination of an anthracycline and a taxane (79,80,81).

Given the efficacy of taxane-based chemotherapy regimens and the small risk of anthracycline-associated cardiotoxicity, there is interest in the potential to eliminate anthracycline use in the adjuvant treatment of selected patient populations. The most recent randomized trial is US Oncology Protocol 9735, in which patients with stages I to III operable breast cancer were randomized to AC versus docetaxel and cyclophosphamide (TC). DFS and OS with a median follow-up of 6.9 years favor
TC with absolute differences of 6% (HR 0.69; p = 0.018) and 4% (HR 0.73; p = 0.045), respectively (82,83). Additional studies are needed to reproduce these findings.

Table 18.1 Taxane-Containing Chemotherapy Regimens for High-Risk, Early-Stage Breast Cancer

Trial	Nodal Status	Schema	N	Disease-free Survival	Overall Survival
CALGB 9344 (72)	Node-positive	AC × 4q3wk	3,121	65% at 5 yr	77% at 5 yr
		AC × 4q3wk → P × 4q3wk		70% at 5 yr*	80% at 5 yr*
NSABP B-28 (73)	Node-positive	AC × 4q3wk	3,060	72% at 5 yr	85% at 5 yr
		AC × 4q3wk → P × 4q3wk		76% at 5 yr*	85% at 5 yr
BCIRG 001 (77)	Node-positive	FAC × 6q3wk	1,491	68% at 5 yr	81% at 5 yr
		TAC × 6q3wk		75% at 5 yr*	87% at 5 yr*
CALGB 9741 (74)	Node-positive	A × 4 → P × 4 → C × 4q3wk}	2,005	75% at 4 yr	90% at 3 yr
		AC × 4 → P × 4q3wk
		A × 4 → P × 4q3wk}	82% at 4 yr*		92% at 3 yr*
		AC × 4 → P × 4q2wk
PACS01 (78)	Node-positive	FEC × 6q3wk	1,999	73.2% at 5 yr	86.7% at 5 yr
		FEC × 3q3wk → T × 3q3wk		78.4% at 5 yr	90.7% at 5 yr
USO 9735 (82)	0–3 nodes	AC × 4q3wk	1,016	75% at 7 yr	82% at 7 yr
	positive	TC × 4q3wk		81% at 7 yr	87% at 7 yr
ECOG 1199 (75)	Node-positive	AC × 4 → P × 4q3wk	4,950	76.9% at 5 yr	86.5% at 5 yr
	OR high-risk	AC × 4 → P × 12q1wk		81.5% at 5 yr^†	89.7% at 5 yr^†
	node-negative	AC × 4 → T × 4q3wk		81.2% at 5 yr^†	87.3% at 5 yr
		AC × 4 → T × 12q1wk		77.6% at 5 yr	86.2% at 5 yr
A, doxorubicin; C, cyclophosphamide; E, epirubicin; F, 5-fluorouracil; P, paclitaxel; T, docetaxel. *p < 0.05. ^†p < 0.05 in comparison to AC → P × 4q3wk.