Home: Educational Supplement: Appendix

Is Her-2/neu a Predictor of Anthracycline Utility
in Adjuvant Therapy? A Qualified Yes.

Peter M. Ravdin, M.D., Ph.D.

The use of Her-2/neu overexpression to select patients who might benefit from adjuvant anthracyclines was first suggested in 1994. Now, 6 years later, the use of Her-2 to guide adjuvant chemotherapy still remains a matter of controversy. Nonetheless, there is some evidence that Her-2 can and should be used for this purpose. It is therefore worthwhile to examine the reasons why this evidence has fallen short of being convincing, and why Her-2 overexpression has not yet become a standard method to help select adjuvant therapy.

The principal shortcoming of the evidence is that none of the adjuvant trials whose results have been used to examine this question were designed for that purpose. They were designed to answer treatment questions only. Moreover, a two-arm trial becomes a four-arm trial when stratified by a biomarker, and Her-2 is a particularly nettlesome biomarker because it is overexpressed in only about 20 to 30 percent of cases. As a result, the overexpressing subset is generally small. That weakens the power of the studies.

How big would an ideal trial be? Estimating the sample size as a two-arm trial in node-positive patients with adjuvant anthracycline-based therapy being twice as effective in Her-2 overexpressing patients, and about 20 to 30 percent of the patients overexpressing Her-2, the number of patients required for the trial would be 2,000 to 3,000. Given that the largest reported studies dealing with Her-2 in adjuvant therapy involved about 1,000 patients, it is clear that most of what we know is based on underpowered studies, many of them with a power of 0.50 or less.

This would seem to be an ideal area for a meta-analysis, but that would be somewhat problematic because of another feature of the studies pertaining to Her-2. Unlike treatment trials with well-defined standardized endpoints (disease-free survival [DFS] and overall survival [OS]), the endpoints (stratification) for using Her-2 are not well defined. There was no uniformity in the trials in how Her-2 testing was done; a number of different systems (based on percentage of cells staining, stain intensity, and stain cellular location) were used to interpret whether Her-2 was overexpressed by a given tumor. Thus, some of the variation in the results of the studies may be due to simple methodological differences.

It has been argued that the mechanism by which Her-2 overexpression leads to a sensitivity to adjuvant anthracyclines is unclear. The mechanism is indeed rather obscure, with some preclinical systems suggesting that it exists, and some finding just the opposite—that Her-2 overexpression may confer resistance to anthracyclines. Adding to the confusion is the fact that Her-2 overexpression in patients with metastatic disease is not unequivocally associated with response to anthracyclines, and in some instances seems to have been associated with resistance (Vincent-Salomon, Carton, Freneaux, et al. 2000; Sjostrom, Krajewski, Franssila, et al. 1998; Jarvinen, Holli, Kuukasjarvi, et al. 1998). These studies of metastatic disease were generally small, however, and may be confounded by the biology of Her-2, which in general is associated with unfavorable features (such as faster cell proliferation) that may mask or counterbalance the benefits of higher efficacy.

What is clear is that arguments for the use of Her-2 in selecting adjuvant chemotherapy are not as simple as the arguments for use of the estrogen receptor (ER) in selecting adjuvant endocrine therapy. In the case of the estrogen receptor, there are very clear molecular correlates, preclinical system evidence, and supportive evidence in metastatic disease for use of the ER. In addition, the ratio of response rates in metastatic disease based on the estrogen receptor’s presence or absence is 10:1. It is clear from the data that the usefulness of Her-2 in assessing the benefit of anthracyclines will be more modest.

Having conceded these points, it is important to review how this HER-2 hypothesis was developed and the consistency of the clinical trial evidence that supports it. The initial report was from the CALGB, which analyzed a trial in which node-positive breast cancer patients were randomized between one of three different dose intensities of cyclophosphamide and Adriamycin. Only in the Her-2 overexpressing patients did the dose make a difference, and thus it appeared that Her-2 overexpression identified a subset of patients who particularly benefited from anthracyclines (Muss, Thor, Berry, et al., 1994).

The CALGB’s own attempt to replicate those results has been controversial. The second CALGB study used patients from the second half of the trial, but the benefit was not as clear and did not reach statistical significance. But by using elaborate statistical techniques and combining data from the two halves of the study, the investigators were again able to demonstrate that Her-2 overexpression identified a subset of patients with anthracycline sensitivity (Thor, Berry, Budman, et al., 1998). Those studies have inspired a number of other tests of whether Her-2 overexpression identified a subset of patients who were particularly likely to benefit from adjuvant therapy.

The first of those studies used results from NSABP B-11. In that trial, patients were randomized between melphalan and 5-fluorouracil (PF) or the same drugs plus doxorubicin (PAF). Patients overexpressing Her-2 were the only subset for whom PAF resulted in a statistically significant risk reduction, despite the fact that the subset was only about half the size of the subset with low (normal) Her-2 expression. The test for a statistically significant difference in risk reduction between the subsets reached statistical significance for DFS and showed a strong trend as a predictor of OS (Paik, Bryant, Park, et al., 2000).

A second NSABP study extended these observations. In this study, based on material from NSABP B-15, tumor samples from the arms receiving CMF or CA were analyzed. Again, the hazards of relapse and death were more strongly reduced in the patient subset overexpressing Her-2 and receiving CA. The test for interaction did not reach statistical significance, but a trend toward greater benefit was seen (Paik, Bryant, Tan-Chiu, et al., 1998).

The same trends were seen in S8814, where postmenopausal ER+ patients were randomized between tamoxifen and CAF plus tamoxifen (a 3 to 10 randomization). Again, chemoendocrine therapy was superior only in the Her-2 overexpressing patients, but once again the difference did not reach statistical significance (Ravdin, Green, Albain, et al., 1998).

The last large study to provide data for the present review was EORTC 10854, in which node-negative premenopausal women were randomized to one perioperative (immediately postoperative) cycle of CAF or to no further adjuvant therapy. At 4 years of followup, a reduction in the hazard of relapse was seen in both the Her-2 positive and Her-2 negative subsets, but appeared larger in the Her-2 positive subset. Here too, the difference did not reach statistical significance (Clahsen, van de Velde, Duval, et al., 1998).

In summary, a pattern is emerging in adjuvant studies showing that the Her-2 positive (overexpressing) subset derives greater benefit from adjuvant therapy than the Her-2 negative subset. The low statistical power of the four studies makes this effect difficult to demonstrate statistically, but the studies do suggest that such an interaction exists.

References

Clahsen PC, van de Velde CJ, Duval C, Pallud C, Mandard AM, Delobelle-Deroide A, et al. p53 protein accumulation and response to adjuvant chemotherapy in premenopausal women with node-negative early breast cancer. J Clin Oncol. 1998; 16:470-9. Abstract.

DiLeo A, Larsimon D, Beauduni M, et al. CMF or anthracycline-based adjuvant chemotherapy for node-positive (N+) breast cancer (BC) patients (PTS): 4 year results of a Belgian randomised clinical trial with predictive markers analysis. Proc Am Soc Clin Oncol 1998; 18:258a. Abstract.

Jarvinen TA, Holli K, Kuukasjarvi T, Isola JJ. Predictive value of topoisomerase IIalpha and other prognostic factors for epirubicin chemotherapy in advanced breast cancer. Bri Cancer 1998; 77:2267-73. Abstract.

Muss HB, Thor AD, Berry DA, Kute T, Liu ET, Koerner F, et al. c-erbB-2 expression and S-phase activity predict response to adjuvant therapy in women with node-positive early breast cancer. Eng Medicine 1994; 330:1260-6. Abstract.

Paik S, Bryant J, Park C, Fisher B, Tan-Chiu E, Hyams D, et al. HER2 and choice of adjuvant chemotherapy for invasive breast cancer. NSABP Protocol B-15. Submitted 2000. Abstract.

Paik S, Bryant J, Park C, Fisher B, Tan-Chiu E, Huams D, et al. erbB-2 and response to doxorubicin in patients with axillary lymph node-positive, hormone receptor-negative breast cancer. J Nat Cancer Inst 1998; 90:1361-70. Abstract.

Ravdin P, Green S, Albain 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 alone. Proc Am Soc Clin Onco 1998; 17:97a. Abstract.

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. Bri Cancer 1998; 78:812-5. Abstract.

Thor AD, Berry DA, Budman DR, Muss HB, Kute T, Henderson IC, et al. erbB2, p53, and efficacy of adjuvant therapy interactions in node-positive breast cancer. Nat Cancer Inst 1998; 90: 1346-1360. Abstract.

Vera R, Albanell J, Lirola JL, et al. HER2 overexpression as a predictor of survival in a trial comparing adjuvant FAC and CMF in breast cancer. Proc Am Soc Clin Oncol 1998; 18:265a. 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 Cancer 2000; 36:586-91. Abstract.

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 Cancer 2000; 36:586-91. Abstract.