© 2001 by Oxford University Press
Journal of the National Cancer Institute Monographs, No. 30, 85-87,
2001
© 2001 Oxford University Press
Is HER-2/neu a Predictor of Anthracycline Utility? No.
Correspondence to: George W. Sledge, Jr., M.D., Indiana Cancer Pavilion, RT-473, 575 Barnhill Dr., Indianapolis, IN 46202 (e-mail: gsledge{at}iupui.edu).
 |
ABSTRACT |
|---|
 Top Abstract IntroductionProblems With Positive TrialsProblems With ConsistencyBiologic ProblemsProblems With ExtrapolationConclusionReferences |
|---|
HER-2 has a well-established role as a prognostic indicator in breast cancer and as a predictor for response to trastuzumab. Recent studies have also suggested that it may serve as a predictor of response to anthracycline-based therapies. This article argues that the data are insufficient to accept this hypothesis as scientifically established. The argument is developed along several lines: first, that the trials used to support a predictive role for HER-2 have real flaws with regard to this hypothesis; second, that HER-2 is a remarkably inconsistent predictor of anthracycline response when examined in a broader context that includes preoperative and metastatic disease; third, that preclinical data fail to support the hypothesis; and finally, that even if accepted, the hypothesis is difficult to extrapolate to the everyday world of breast cancer.
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionProblems With Positive TrialsProblems With ConsistencyBiologic ProblemsProblems With ExtrapolationConclusionReferences |
|---|
HER-2 is clearly important in the natural history of breast cancer. This importance derives from its role in tumor growth, invasion, and metastasis (the clinical summation of which is its role as a prognostic factor in early-stage breast cancer), as well as its role as a predictor of response to trastuzumab. These roles for HER-2 seem firmly established.
Recent U.S. cooperative group trials have suggested that patients receiving anthracycline-based chemotherapy are most likely to benefit if their tumors overexpress the HER-2 protein. This role of HER-2, as a predictor of anthracycline utility, seems promising and, if confirmed, clearly would be of importance. But to what extent should we accept the data presented to date?
I would like to suggest that what might be called the HER-2 hypothesis suffers from several potentially fatal flaws: 1) The data used to support the hypothesis are inherently flawed, as is the analysis of these data; 2) a broader view of the available data suggests that there are problems with consistency across the spectrum of breast cancer; 3) there is no solid biologic basis for the hypothesis; and 4) even if we were to accept the general hypothesis, practical extrapolation is difficult.
|
PROBLEMS WITH POSITIVE TRIALS |
|---|
|
TopAbstractIntroductionProblems With Positive TrialsProblems With ConsistencyBiologic ProblemsProblems With ExtrapolationConclusionReferences |
|---|
Several trials have reported positive results (1–3). Individually and collectively, these trials suffer from potential and actual flaws. All of the data presented to date (with the exception of a subset analysis of Cancer and Leukemia Group B [CALGB] 8541) (4) use immunohistochemical techniques, and such techniques are associated with well-known concerns. Immunohistochemical accuracy depends on proper tissue preservation, antigen retrieval, reagent specificity, type of technique used, protocol used for grading positivity, and observer experience. The ability of the antibodies used in immunohistochemical assays to detect HER-2 is highly variable, and the correlation of immunohistochemistry with the "gold standard" of fluorescence in situ hybridization is often relatively low. Similarly, immunohistochemical techniques often correlate poorly with each other (5,6).
All of these problems potentially come into play in the studies used to support the HER-2 hypothesis. In each study, tissue collection was retrospective. In one of the studies (2), the antibody used was switched mid-study, and the correlation coefficient for the two antibodies used was 91%—good but not perfect. The same study evaluated interobserver reproducibility in the hands of two expert breast pathologists, with an overall R2 = 76%, a statistically significant result that is less than reassuring when one is dealing with potentially life and death decisions. In another, immunohistochemistry was done, in many cases, on old slides rather than off paraffin-embedded tissue (1). No two studies used the same scoring technique: In one study as few as 1% positive cells was used as a cutoff for positivity, whereas in another 50% positivity was used. One might interpret consistently positive results obtained from such different technical approaches as evidence of the strength of the overall hypothesis, yet it is equally reasonable to consider such studies as being essentially uninterpretable for comparative purposes.
Leaving aside the potential technical flaws inherent in immunohistochemical techniques, analysis of the reported trials leaves something to be desired. All involved retrospective, post hoc analyses and should, therefore, properly be viewed as hypothesis-generating rather than as proof-of-principal studies. In the CALGB trial, an initial analysis reported positive results, a second analysis with a somewhat larger set of patients failed to reach statistical significance, and a third analysis combining the two was once again positive. Because of an unbalanced randomization, the Southwest Oncology Group report involves analyses of truly small numbers of patients in the subset of HER-2-positive, estrogen receptor-positive patients. Tests for interaction did not reach statistical significance in any of the three trials with regard to overall survival, and for disease-free survival, only in the National Surgical Adjuvant Breast and Bowel Project (NSABP) trial. As discussed by Ravdin in another article in this issue (7), the number of patients required to answer this question with any degree of scientific rigor is large, and it is larger than any current study affords.
Finally, what are we to make of the available trials themselves? It is tempting to view these trials through the lens of the HER-2 hypothesis, in effect to assume that each represents an attempt to test the HER-2 hypothesis. In reality, each trial tested a treatment hypothesis unrelated to the (post hoc) HER-2 hypothesis. As such, the best we can do (or should do) is to rephrase the HER-2 hypothesis in terms of the actual hypothesis tested in each protocol, yet to do so gives us some sense of the problems inherent in using these trials to support the HER-2 hypothesis.
For instance:
CALGB 8541. An increased dose intensity of combination chemotherapy with cyclophosphamide, doxorubicin, and 5-fluorouracil (CAF) (please note, not doxorubicin alone; all three agents were dose escalated) is beneficial in HER-2-positive patients; moderate-dose intensity CAF given for a longer duration (but with the same total dose as the high-dose arm) and low-dose intensity CAF are not beneficial in HER-2-positive patients.
NSABP B-11. Addition of low-dose, chronically administered doxorubicin to melphalan and 5-fluorouracil is beneficial in HER-2-positive but not in HER-2-negative patients.
Southwest Oncology Group 8814. CAF (please note, not doxorubicin alone) chemotherapy is beneficial in HER-2-positive, estrogen receptor-positive patients but has no benefit in HER-2-negative, estrogen receptor-positive patients.
It is important to note that only one of these trials (NSABP B-11) specifically evaluates the role of doxorubicin (as opposed to a doxorubicin-containing combination). Within the context of two of the trials above, it would be just as reasonable to assume that 5-fluorouracil or cyclophosphamide efficacy was modulated by HER-2 as was doxorubicin's efficacy. Those who support the HER-2 hypothesis have suggested that what is common to all three of these trials is that "more" doxorubicin always appears to work better than "less" doxorubicin in HER-2-positive patients, yet the low-dose-intensity arm of the CALGB trial uses essentially the same dose intensity as the doxorubicin-containing arm of NSABP B-11. How can the results of low-dose-intensity doxorubicin be strikingly ineffective in one trial, and yet so strikingly impressive in another?
One might argue that such cross-trial comparisons are inherently problematic, so that it is unfair to compare the results of these two similar low-dose-intensity arms. This is exactly the point. Those who lump together studies with such disparate primary hypotheses, drugs, schedules, and dose intensities for the purpose of supporting a post hoc hypothesis are guilty of exactly this practice. One would not accept the results of (or even contemplate performing) a meta-analysis of these very different trials in terms of their original hypotheses, yet, in effect, we are asked to perform an informal meta-analysis, a kind of science by gestalt, in a statistically less rigorous fashion in support of the HER-2 hypothesis.
|
PROBLEMS WITH CONSISTENCY |
|---|
|
TopAbstractIntroductionProblems With Positive TrialsProblems With ConsistencyBiologic ProblemsProblems With ExtrapolationConclusionReferences |
|---|
Although most doxorubicin-based adjuvant trials have been reported to be positive, at least one epirubicin-based analysis has been reported to show no differential benefit for HER-2 positivity. Untch et al. (8) compared a standard-dose epirubicin/cyclophosphamide combination with a dose-intense epirubicin/cyclophosphamide regimen. Only in the HER-2-negative subgroup was there a benefit for the dose-intense regimen, a result that contrasts strikingly with the CALGB results. Given that epirubicin is at least as beneficial in the adjuvant setting as doxorubicin, how could one explain the absence of HER-2-related benefit?
Similarly, Colozza et al. (9) evaluated the effect of HER-2 status in a trial comparing epirubicin with combination adjuvant chemotherapy with cyclophosphamide, methotrexate, and 5-fluorouracil in patients with American Joint Committee on Cancer stage I or II breast cancer. With a median follow-up of 5.6 years and 266 patients evaluable for HER-2 status, epirubicin treatment had no statistically significant impact on the outcome of patients with HER-2-positive tumors.
One would expect a consistency of effect across all stages of breast cancer if the HER-2 relationship were real. We would not expect a predictive factor to have a benefit in stage II breast cancer but not in stage III or IV breast cancer, yet this is what we are asked to accept for the HER-2/doxorubicin interaction. Numerous trials (10–19) have examined the effect of HER-2 overexpression on response to preoperative or metastatic anthracycline-based regimens. These trials (summarized in Table 1
) are inconsistent in their results, but generally they fail to demonstrate a positive relation. While the number of patients entered in these trials is small relative to the adjuvant trials, the number of events is large, since virtually every patient entered is evaluable for response.
|
|
BIOLOGIC PROBLEMS |
|---|
|
TopAbstractIntroductionProblems With Positive TrialsProblems With ConsistencyBiologic ProblemsProblems With ExtrapolationConclusionReferences |
|---|
One would expect that, if HER-2 overexpression conferred sensitivity to doxorubicin in the clinic, a similar effect might be seen in the laboratory. Pegram et al. (20) tested this hypothesis by transfecting four breast cancer cell lines with HER-2 and then exposing them to doxorubicin in vitro. No alteration in chemosensitivity was observed in any of the transfected breast cancer cell lines compared with the parent cell lines or in a related in vivo nude mouse xenograft model. These observations argue against a direct role for HER-2 amplification in anthracycline sensitivity.
It has been suggested that HER-2's proposed relationship to anthracycline sensitivity might, in fact, be as a surrogate for topoisomerase II-
. Anthracyclines are, of course, topoisomerase inhibitors. Topoisomerase II- is located close to HER-2 on chromosome 17, and recent evidence (14) suggests strongly that its amplification may be associated with sensitivity to anthracyclines in the metastatic setting. While this proposal seems reasonable, two observations may be made regarding it: 1) Jarvinen et al. (21) have examined the relation of HER-2 and topoisomerase II- in patients with metastatic breast cancer. As it turns out, the two are not on the same amplicon, and when HER-2 is overexpressed, topoisomerase II- is as often deleted as amplified. 2) Assuming that the explanation is correct, what benefit would be gained using a surrogate marker for topoisomerase II-, when one could presumably measure the "real thing" with greater correlative power?
|
PROBLEMS WITH EXTRAPOLATION |
|---|
|
TopAbstractIntroductionProblems With Positive TrialsProblems With ConsistencyBiologic ProblemsProblems With ExtrapolationConclusionReferences |
|---|
The problems associated with accepting the results seen with the available adjuvant trials mirror the problems seen with HER-2 testing in general. Immunohistochemical analysis of HER-2 is only imperfectly correlated with fluorescence in situ hybridization and is subject to considerable interobserver variation, problems with technique, and problems with tissue preservation. None of the current trials have even used the same immunohistochemical techniques. How then may we extrapolate the currently available results to clinically available testing kits?
Similarly, proponents of the HER-2 hypothesis should understand its logical practical conclusions. It is safe to guess that virtually all lymph node-positive, HER-2-positive patients currently receive doxorubicin-based therapy. Two thirds to three quarters of breast cancer patients are HER-2 negative. The same studies suggesting a benefit for doxorubicin in HER-2-positive patients suggest a relative lack of added benefit in HER-2-negative patients. Given the Oxford overview (22) demonstration of benefit for anthracycline-based regimens in the adjuvant setting, do we have sufficient confidence in the HER-2 hypothesis that we are willing to omit doxorubicin in HER-2-negative patients?
|
CONCLUSION |
|---|
|
TopAbstractIntroductionProblems With Positive TrialsProblems With ConsistencyBiologic ProblemsProblems With ExtrapolationConclusionReferences |
|---|
HER-2 testing has many real benefits. These benefits should not blind us to the real concerns surrounding the use of HER-2 as a therapeutic predictor for anthracyclines. We currently lack a solid biologic basis for the proposed linkage. The available positive studies have real uncertainties. There are studies contradicting the linkage in the adjuvant, neoadjuvant, and metastatic settings. Extrapolation from current data to the clinic is problematic. The cumulative weight of these concerns calls the HER-2 hypothesis into question. Until more solid data emerge, HER-2 positivity should not be accepted as a predictor of doxorubicin sensitivity.
|
REFERENCES |
|---|
|
TopAbstractIntroductionProblems With Positive TrialsProblems With ConsistencyBiologic ProblemsProblems With ExtrapolationConclusionReferences |
|---|
1
Paik S, Bryant J, Park C, Fisher B, Tan-Chiu E, Hyams D, et al. erbB-2 and response to doxorubicin in patients with axillary lymph node-positive, hormone receptor-negative breast cancer. J Natl Cancer Inst 1998;90:1361–70.
2
Thor AD, Berry DA, Budman DR, Muss HB, Kute T, Henderson IC, et al. erbB-2, p53, and efficacy of adjuvant therapy in lymph node-positive breast cancer. J Natl Cancer Inst 1998;90:1346–60.
3 Ravdin P, Green S, Albain K, Boucher V, Ingles J, Pritchard K, et al. Initial report of the SWOG biological correlative study of c-erbB-2 expression as a predictor of outcome in a trial comparing adjuvant CAF T with tamoxifen (T) alone [abstract]. Proc ASCO 1998;17:97a.
4 Dressler L, Berry D, Liu E, Cowan D, Broadwater G, Tse J, et al. Amplification of erbB2 by fluorescent in situ hybridization (FISH): an alternate method to predict outcome following dose-escalated CAF in stage II, node positive breast cancer patients. A Cancer and Leukemia Group B (CALGB) study [abstract]. Proc ASCO 1999;18:75a.
5
Press MF, Hung G, Godolphin W, Slamon DJ. Sensitivity of HER-2/neu antibodies in archival tissue samples: potential source of error in immunohistochemical studies of oncogene expression. Cancer Res 1994;54:2771–7.
6 Mitchell MS, Press MF. The role of immunohistochemistry and fluorescence in situ hybridization for HER2/neu in assessing the prognosis of breast cancer. Semin Oncol 1999;26 (4 Suppl 12):108–16.
7 Ravdin PM. Is Her2 of value in identifying patients who particularly benefit from anthracyclines during adjuvant therapy? A qualified yes. J Natl Cancer Inst Monogr 2001;30:80–4.
8 Untch M, Konecny G, Lebeau A, Thomssen C, Kahlert S, Sattler D, et al. Dose-intensification (DI) of anthracycline in the adjuvant treatment of high risk breast cancer (HRBC) and c-erbB-2 overexpression [abstract]. Proc ASCO 1998;17:103.
9 Colozza M, Gori S, Mosconi A. Cerb-B2 expression as a predictor of outcome in a randomized trial comparing adjuvant CMF vs single agent epirubicin in Stage I–II breast cancer (BC) patients [abstract]. Proc ASCO 1999;18:70.
10 Jarvinen TA, Holli K, Kuukasjarvi T, Isola JJ. Predictive value of topoisomerase IIalpha and other prognostic factors for epirubicin chemotherapy in advanced breast cancer. Br J Cancer 1998;77:2267–73.[ISI][Medline]
11 Sjostrom J, Krajewski S, Franssila K, Niskanen E, Wasenius VM, Nordling S, et al. A multivariate analysis of tumour biological factors predicting response to cytotoxic treatment in advanced breast cancer. Br J Cancer 1998;78:812–5.[ISI][Medline]
12 Vincent-Salomon A, Carton M, Freneaux P, Palangie T, Beuzeboc P, Mouret E, et al. ERBB2 overexpression in breast carcinomas: no positive correlation with complete pathological response to preoperative high-dose anthracycline-based chemotherapy. Eur J Cancer 2000;36:586–91.
13 Stender M, Neuberg D, Wood W, Sledge G. Correlation of circulating c-erb B-2 extracellular domain (Her-2) with clinical outcome in patients with metastatic breast cancer [abstract]. Proc ASCO 1997;16:154a.
14 Isola J, Tanner M, Holli K, Joensuu H. Amplification of topoisomerase II alpha is a strong predictor of response to epirubicin-based chemotherapy in HER-2/neu-positive metastatic breast cancer [abstract]. Breast Cancer Res Treat 2000;64:31.
15 MacGrogan G, Mauriac L, Durand M, Bonichon F, Trojani M, de Mascarel I, et al. Primary chemotherapy in breast invasive carcinoma: predictive value of the immunohistochemical detection of hormonal receptors, p53, c-erbB-2, MiB1, pS2 and GST pi. Br J Cancer 1996;74:1458–65.[ISI][Medline]
16 Kling KM, Ostrzega N, Schmit P. Breast conservation after induction chemotherapy for locally advanced breast cancer. Am Surg 1997;63:861–4.[Medline]
17 Niskanen E, Blomqvist C, Franssila K, Hietanen P, Wasenius VM. Predictive value of c-erbB-2, p53, cathepsin-D and histology of the primary tumour in metastatic breast cancer. Br J Cancer 1997;76:917–22.[Medline]
18 Gregory RK, Powles TJ, Salter J, Chang JC, Ashley S, Dowsett M. Prognostic relevance of cerbB2 expression following neoadjuvant chemotherapy in patients in a randomised trial of neoadjuvant versus adjuvant chemoendocrine therapy. Breast Cancer Res Treat 2000;59:171–5.[CrossRef][ISI][Medline]
19 Zapf M, Flope A, Storer B, Gown AM, Ellis GR, Livingston RB, et al. Breast cancer tumor associated proteins as predictors of response to neoadjuvant doxorubicin-based chemotherapy [abstract]. Proc ASCO 1998;17:104a.
20 Pegram MD, Finn RS, Arzoo K, Beryt M, Pietras RJ, Slamon DJ. The effect of HER-2/neu overexpression on chemotherapeutic drug sensitivity in human breast and ovarian cancers cells. Oncogene 1997;15:537–47.[CrossRef][ISI][Medline]
21 Jarvinen TA, Tanner M, Barlund M, Borg A, Isola J. Characterization of topoisomerase II alpha gene amplification and deletion in breast cancer. Genes Chromosomes Cancer 1999;26:142–50.[CrossRef][ISI][Medline]
22 Polychemotherapy for early breast cancer: an overview of the randomised trials. Early Breast Cancer Trialists Collaborative Group. Lancet 1998;352:930–42.[CrossRef][ISI][Medline]
23 Petit T, Ghnassia J, Rodier J, Borel C, Escande A, Mers R, et al. Relationship between erbB-2 status and neoadjuvant chemotherapy response is dependent on anthracycline dose intensity. Proc ASCO 2000;19:96a.
24 Rozan S, Vincent-Salomon A, Zafrani B, Validire P, De Cremoux P, Bernoux A, et al. No significant predictive value of c-erbB-2 or p53 expression regarding sensitivity to primary chemotherapy or radiotherapy in breast cancer. Int J Cancer 1998;79:27–33.[CrossRef][ISI][Medline]
25 Steger GG, Wenzel C, Shmidinger MP, Rudas M, Taucher S, Gnant MF. HER2 status predicts complete pathologic response (pCR) in primary, operable breast cancer treated with neoadjuvant epirubicin/docetaxel + G-CSF. Breast Cancer Res Treat 2000;64:31.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||