NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Seidenfeld J, Samson DJ, Rothenberg BM, et al. HER2 Testing to Manage Patients With Breast Cancer or Other Solid Tumors. Rockville (MD): Agency for Healthcare Research and Quality (US); 2008 Nov. (Evidence Reports/Technology Assessments, No. 172.)

  • This publication is provided for historical reference only and the information may be out of date.

This publication is provided for historical reference only and the information may be out of date.

Cover of HER2 Testing to Manage Patients With Breast Cancer or Other Solid Tumors

HER2 Testing to Manage Patients With Breast Cancer or Other Solid Tumors.

Show details

4Discussion and Future Research

The human epidermal growth factor (EGF) receptor-2 (HER2; also referred to as HER2/neu and as ERBB2) gene is amplified and the HER2 protein overexpressed in approximately 18–20 percent of breast cancer cases. Evidence from multiple randomized trials demonstrates that adding trastuzumab, a therapeutic monoclonal antibody that targets HER2, to adjuvant chemotherapy regimens for HER2-positive breast cancer improves patient outcomes. HER2 also is overexpressed in varying proportions of other epithelial malignancies such as ovarian, thyroid, lung, salivary gland/head and neck, stomach, colon, and prostate cancers. This evidence report is a systematic review of the evidence on novel applications of HER2 testing to the management of cancer patients including: potential for response to trastuzumab among breast cancer patients who have negative, equivocal, or discordant HER2 assay results; use of HER2 assay results to guide selection of breast cancer treatments other than trastuzumab (i.e., chemotherapy regimen or hormonal therapy regimen); the use of serum HER2 to monitor treatment response or disease progression in breast cancer patients; and use of HER2 testing to manage patients with ovarian, lung, prostate, or head and neck tumors.

HER2 assay results are influenced by multiple biologic, technical and performance factors. Since many aspects of HER2 assays have not been standardized until very recently (Wolff, Hammond, Schwartz, et al., 2007), the effects of these disparate influences cannot be isolated in the existing literature that compares results of different methods. Discordances between IHC and FISH results might arise in one of three ways. They may be artifacts of one accurate or two inaccurate tests. Alternatively, they may reflect a threshold issue, either related to changes in threshold definitions over time, or an inherent problem of using a continuous measure to classify patients dichotomously. Finally, discordant test results might accurately reflect a small number of different patients with respect to the biologic mechanisms that can increase membrane levels of the HER2 protein. This clearly affects the interpretation of evidence on the use of “HER2 status” to predict treatment or disease outcomes, which presumes accurate classification by tissue assays. Future studies reporting outcomes as a function of HER2 status should report separately on patients with concordant, equivocal, and discordant assay results.

To assess the quality of the available evidence on using HER2 status to guide treatment decisions, we took a two-fold approach. First, we applied a hierarchical framework to evaluate how informative various designs and analytic strategies would be to predict outcomes according to HER2 status. The most informative would be a trial in which randomized assignment to treatment groups was stratified by HER2 status or patients were randomized to receive treatment guided by HER2 results or not. Prespecified subgroup analyses guard against the problems of data dredging. In contrast, post-hoc subgroup analyses may generate hypotheses, but do not support strong inferences about effectiveness. The least-informative situation would be a single-arm study that presents univariate comparisons of HER2 groups. To further assess the quality of predictive studies, we adapted the “Reporting Recommendations for Tumor Marker Prognostic Studies” (REMARK) statement (McShane, Altman, Sauerbrei, et al., 2005). Good quality characteristics of predictive studies include: prospective design; prespecified hypotheses about relation of marker to outcome; large, well-defined, representative study population; marker assay methods well-described; blinded assessment of marker in relation to outcome; homogeneous treatment(s), either randomized or rule-based selection; low rate of missing data (≤15 percent); and a well-described, well-conducted multivariate analysis of outcomes.

Overall, few trials included in this evidence report were intended or designed to investigate the key questions of the systematic review. With exception of two trials (Seidman, Berry, Cirrincione, et al., 2004; Martin, Pienkowski, Mackey, et al., 2005), evidence for this review consisted mostly of post-hoc analyses on subgroups not directly randomized, selected, or stratified by HER2 status. Nearly all were secondary or correlative analysis on patient subgroups with archived tissue samples available for HER2 testing. Direct comparison of baseline and prognostic factors for HER2-positive and HER2-negative subgroups were infrequently reported, so it is uncertain whether these subgroups were well balanced in such studies.

Going forward, cancer therapy trial protocols should incorporate elements to facilitate robust analyses of the potential of HER2 to improve treatment management by providing predictive information on disease progression and response to treatment. These elements include:

  • Detailed reporting of how HER2 status was ascertained, including assay methods, thresholds, validation, and quality assurance measures.
  • Since HER2 status is now routinely ascertained for all newly diagnosed breast cancer patients, relevant data should be recorded for all participants, and accessible to other researchers, to permit subgroup analyses of outcomes by HER2 status. Investigators of ongoing and completed trials should similarly contribute data or tissue samples to large international collaborations for patient-level meta-analysis.
  • Perform stratified randomization by HER2 status or prospectively specify HER2 subgroup analysis of outcomes.
  • Report on and, where sample size permits, prospectively specify subgroup analyses on participants with equivocal or discordant HER2-positive assay results (e.g., IHC 2+ and FISH positive). Of particular importance is recording, tracking and making accessible to researchers assay results and outcomes data on patients with equivocal results of initial assays, as defined in the ASCO/CAP guideline.
  • Future studies should report more completely and statistically compare subgroups from each treatment arm by HER2 status for known prognostic factors and baseline characteristics of patients and their tumors.

The case for these measures is strongest for breast cancer therapy trials, as the relation of HER2 status to outcomes of treatments other than trastuzumab has been hypothesized, but not confirmed; and as many therapeutic agents, classes, and regimens have been and will be tested. The case for incorporating these measures into therapeutic trials for other cancers is less compelling in that the relationship of HER2 status to outcome of any therapy has not been established. However, the argument in favor of doing so is that it would provide an efficient approach to screening for such relationships. This approach can be generalized to promising biomarkers other than HER2. Because existing evidence is scant on serum HER2 (sHER2) results to predict disease progression or treatment outcome, serial collection of serum samples at standard intervals to be assayed subsequently for sHER2 levels in trials planned for other purposes offers an opportunity to examine potential utility.

For Key Question 2, potential for response to trastuzumab among breast cancer patients who have equivocal, discordant, or negative HER2 assay results, evidence is scant, but intriguing. Future research should address whether there are other markers that might predict for these subgroups response to therapy that targets HER2. For example, studies will address patients with triple-negative or “basal-like” breast cancer and not merely patients who are negative on HER2 tests. In addition, patients with HER2/CEP17 ratios either ≥1.8 but <2.0 or ≥2.0 but <2.2 deserve attention. Post-hoc analysis from the only trial found to address this question in HER2-negative patients reported an association of benefit from trastuzumab with presence of polysomy 17. It is likely most efficient to evaluate markers with the potential to identify trastuzumab-responsive patients in samples and patients pooled across the large adjuvant trastuzumab trials that have already been completed.

For Key Question 3a, use of HER2 results to guide selection of breast cancer treatments other than trastuzumab (i.e., chemotherapy regimen or hormonal therapy regimen), there are suggestions that HER2 status predicts clinical benefit from certain regimens. Future trials that compare adjuvant chemotherapy regimens with versus without an anthracycline, or with versus without a taxane, could determine HER2 status at the time of diagnosis, and stratify randomization by HER2 assay results. This approach might provide more definitive tests for hypotheses about whether an anthracycline or a taxane improves outcomes of HER2-negative patients. For emerging targeted therapies in the adjuvant setting, and for all therapies in the neoadjuvant and advanced disease settings, future trials should prospectively collect data on HER2 status and prospectively define hypotheses they will test on treatment outcomes in HER2 subgroups.

The most attractive and pragmatic approach currently available is work that could be done using the individual patient data of the Early Breast Cancer Trialists Collaborative Group (EBCTCG).

  • EBCTCG has published an individual patient-level meta-analysis (17 trials, N=14,000; minimum 5 years followup) of randomized trials that compared CMF versus an anthracycline-based regimen for adjuvant chemotherapy of breast cancer.
  • The absolute difference in rates of relapse or death favored anthracyclines by 3 percent at 5 years and 4 percent (SE: 1) at 10 years.
  • The opportunity is to access as many archived tumor specimens as possible of participants in these trials and determine HER2 status using IHC and ISH with current scoring thresholds.
  • If sufficient tumor samples can be obtained and tested, this would permit a rigorous assessment of the benefit from anthracyclines to HER2-positive and HER2-negative patients.

In the ASCO 2007 update of recommendations for the use of tumor markers in breast cancer (Harris, Fritsche, Mennel, et al., 2007), the expert panel continued to recommend that HER2 status should not be used to withhold endocrine therapy from HER2-positive patients, nor should it be used to select a specific endocrine therapy. They summarized conflicting results on these issues, especially data addressing the hypothesis that aromatase inhibitors may be more effective than tamoxifen in HER2-positive patients. In our evidence review, the same data supplemented by recent studies still does not support conclusions about how well HER2 status predicts relative outcomes of different endocrine therapies in patients with hormone-receptor-positive breast cancer. Research is ongoing to compare tamoxifen and newer hormonal agents (i.e., aromatase inhibitors or selective estrogen receptor modulators) in hormone-receptor-positive patients. Implications of HER2 status should be prospectively investigated in ongoing and future trials that compare hormonal therapies. In addition, retrospective analyses of HER2 status should be conducted for additional completed trials comparing hormonal therapies, an attractive approach because long-term followup of outcomes is already available. Of particular importance, given the inverse relationship between hormone-receptor and HER2 status, is accumulating a large enough sample of patients who are both hormone-receptor positive and HER2 positive to evaluate the use of hormonal therapy in HER2-positive patients.

For Key Question 4, current evidence does not support conclusions on sHER2 as a predictor of outcomes after treatment by any regimens in any setting of breast cancer treatment. Evidence primarily focused on first-line or second- and subsequent-line treatment of metastatic disease using variety of regimens. Furthermore, these studies used different thresholds for a positive sHER2 result and varied on whether patient selection required positive tissue HER2 status. There were only three randomized trials and only one used multivariate analysis, while two single-arm studies performed multivariate analysis. The quality of reporting on multivariate analyses was poor. Univariate analyses provide very limited information value, suggesting candidate variables for future multivariate analyses. These studies do not support clear conclusions for whether sHER2 predicts disease progression, treatment response, or outcomes of any specific treatment regimen. A potentially useful approach to filling the evidence gap is to identify one or more completed randomized, controlled trials with banked serial serum specimens and either known tissue HER2 status or banked tissue from all, or nearly all, randomized patients. Once identified, an appropriate multivariate analysis could assess the relation between sHER2 changes and treatment outcomes. In addition, future trials should prospectively collect serial sHER2 samples.

For Key Question 5, this systematic review did not find evidence to support conclusions on serum or tissue HER2 testing to predict treatment outcomes for malignancies in any of these sites: lung, ovary, head and neck, or prostate. Overall, the evidence is weak and heterogeneous with respect to treatment regimens and thresholds for positive HER2 test results. Of 22 studies addressed for the four types of malignancies, there were no randomized trials that could have analyzed HER2 by treatment effect interactions. Six multivariate analyses in single-arm designs were performed, all of which were poorly described, so it is unclear if they were well conducted. Data from these exploratory analyses did not consistently find that HER2 status predicts treatment results. Univariate analyses provide very limited information value, at best suggesting candidate variables for future multivariate analyses. Future studies of nonbreast malignancies should prospectively collect and test serum and tissue specimens for HER2 status and perform planned multivariate analyses, preferably in randomized trials.

Given the human and financial cost of cancer therapy trials, the limited resources available, and the long duration of followup needed to assess outcomes particularly for early stage or slowly growing cancers, it is imperative that tumor tissue blocks be collected, optimally fixed, saved, and made available for correlative tumor marker studies from all randomized patients. Agreement to share blocks with investigators should be made a condition for institutions seeking to participate in cooperative group trials.


  • PubReader
  • Print View
  • Cite this Page

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...