HER-2/neu Overexpression As a Poor Prognostic Factor for Patients with Metastatic Breast Cancer undergoing High-dose Chemotherapy with Autologous Stem Cell Transplantation

The HER-2/neu receptor has attracted great attention as a potential prognostic factor in breast cancer. The significance of HER-2/neu in breast cancer patients receiving standard-dose chemotherapy is relatively well established. Many studies have found HER-2/neu overexpression to be correlated with shorter PFS³ and OS, as well as with resistance to hormone therapy and higher relapse rate after conservative surgery, standard-dose chemotherapy with or without radiotherapy alone (1, 2, 3, 4, 5, 6, 7, 8, 9, 10). HDCT produces a high tumor response rate for patients with metastatic breast cancer and has 20% long-term PFS. Only few studies addressed the prognostic significance of HER-2/neu overexpression in breast cancer patients undergoing HDCT (11, 12, 13, 14), especially patients with distant metastases. Therefore, we hypothesized that HER-2/neu overexpression could serve as an unfavorable predictive factor independently from other prognostic factors for OS and PFS in metastatic breast cancer patients who underwent HDCT.

Between August 1991 and April 1998, 236 women underwent HDCT for metastatic breast cancer at The University of Texas M. D. Anderson Cancer Center (15). Patients (63) had biopsy samples of primary tumor collected at the time of diagnosis that were available for immunohistochemical analysis of HER-2/neu overexpression. Patient characteristics are listed in Table 1. In addition to traditional evaluation criteria, we used a scoring system developed by Swenerton to assess the burden of metastatic disease at the time of progression and after salvage therapy as described previously (16).

All patients were given standard-dose induction chemotherapy with the intent of maximizing tumor response before HDCT was begun. Hematopoietic stem cells were collected after treatment with hematopoietic colony-stimulating factors alone or in combination with cyclophosphamide 1.5 g/m² IV, etoposide 250 mg/m² IV, and cisplatin 40 mg/m² on days 1–3. After stem cell collection, all patients were given HDCT consisting of cyclophosphamide, 2 g/m² IV (total dose 6 g/m²); thiotepa, 240 mg/m² (total dose 750 mg/m²); and carmustine, 150 mg/m² IV, all on days −6 to −4 before stem cell infusion (day 0).

HER-2/neu expression was assessed in formalin-fixed, paraffin-embedded archived tumor specimens. For immunohistochemical staining, 4 μm-thick sections were cut from a representative paraffin block of invasive carcinoma. The sections were mounted on charged slides, deparaffinized in xylene, and rehydrated in descending grades of ethanol. Sections were then subjected to heat-induced epitope retrieval by immersion in a 0.001 m concentration of citrate buffer preheated to >90°C and heated in an electric steamer. Endogenous peroxidase activity was blocked by a 5-min treatment with 3% hydrogen peroxide. The slides were incubated with anti-HER-2/neu e2-4001 monoclonal antibody (1:100 dilution) on a Dako Autostainer. The staining was performed by using LSAB-2 peroxidase kit (Dako). 3-3-diaminobenzedine was used as chromogen. The slides were counterstained with Mayer hematoxylin. The immunohistochemical preparations were interpreted independently by two pathologists (Y. S. K. and A. A. S.), both of whom were unaware of clinical outcome. The intensity of HER-2/neu overexpression was graded by 0, 1+, 2+, or 3+. Overexpression of HER-2/neu was defined as positive 3+ membranous staining in >10% of the invasive carcinoma cells.

The information regarding ER and PR status of the primary tumors was retrieved from patients’ medical records. Tumor specimens were considered negative for ER and PR if staining for both receptors was negative; specimens were considered positive for ER/PR if staining for either or both receptors was positive. Patients whose ER/PR status was unknown were grouped with those who were ER/PR positive because of the common clinical practice of giving antiestrogen drugs to patients whose ER/PR status is unknown.

PFS was defined as the time between the date of the stem cell transplant and the date of documented disease progression or death from any cause. OS was defined as the time from the date of the transplant to the date of death from any cause. Those patients who were free of progression and those who were still alive at last follow-up were counted as censored observations in the respective PFS and OS curves; the most recent follow-up date (up to July 31, 2000) was used. The Kaplan-Meier method (17) was used to estimate survival distributions, accounting for censoring, in the total group and in specified subsets; comparisons among survival distributions were based on the Log-rank test. The χ² test was used to assess possible associations between other factors and HER-2/neu expression.

In addition to directly comparing survival outcomes with respect to HER-2/neu expression, we also adjusted for potential differences in other pretreatment patient characteristics. In earlier analyses of factors associated with survival outcomes in the total group of 236 patients, we identified the combination of ER/PR expression, number of disease sites, length of disease-free interval, and change in Swenerton score as influencing outcome in these patients.⁴ Hence, we used a multivariate regression technique (18) that included these four factors, and we tested whether adding a term for HER-2/neu status could account for possible variability in PFS and OS outcome.

To confirm that the subgroup of patients for whom tumor samples were available (63 patients) was a random sample of the total group (236 patients) who underwent HDCT for metastatic breast cancer, we compared patient characteristics between those for whom tumors were available and those for whom they were not. There was no statistically significant difference in ER/PR status, time from diagnosis to appearance of metastasis, location and number of metastatic sites, difference in Swenerton score, disease status before HDCT, and conversion from PR to CR after HDCT (Table 1). Median progression-free and OS times were also similar between the two groups (Table 1).

Of the 63 tumor samples available, in 22 samples, there was a positive membranous staining in >10% of the invasive carcinoma cells. Therefore, 22 patients (35%) were considered to have HER-2/neu overexpression.

Table 2 summarizes patient characteristics for the 22 patients with HER-2/neu-overexpressing tumors and for the remaining 41 patients with low HER-2/neu-overexpressing tumors. Patients with HER-2/neu-overexpressing tumors were more likely to have been presented with metastatic disease, and a smaller proportion of them achieved >50% reduction in disease involvement, according to the Swenerton score, during induction therapy. HER-2/neu-overexpressing tumors (12 of 19) were negative for ER/PR, and 29 of 40 HER-2/neu-negative tumors were positive for ER/PR. Thus, 41 of 59 cases (69%) had correspondence between HER-2neu overexpression and absence of ER/PR expression (κ statistic = 0.34). Patients with HER-2/neu-overexpressing positive tumors tended to have worse prognostic features overall.

To assess the correlation between HER-2/neu overexpression and response to HDCT, we evaluated the conversion from PR to CR after HDCT. A group of 33 patients were in PR before HDCT. A group of 11 patients achieved CR after HDCT. The conversion rate was lower among patients with HER-2/neu overexpression (3 of 14, 21%) compared with patients without HER-2/neu overexpression (8 of 19, 42%), P = 0.28.

Among the 63 patients, the estimated median OS was 36 months, and the estimated median PFS was 11 months. A group of 50 patients progressed, including 20 patients with HER-2/neu overexpression. A group of 38 patients died, including 17 patients with HER-2/neu overexpression. As shown in Figs. 1 and 2, patients whose tumors overexpressed HER-2/neu tended to have shorter statistically significant progression-free intervals and survival. In the multivariate analysis of PFS, HER-2/neu overexpression remained an independent prognostic factor (hazard ratio: 1.99, P = 0.04). In the multivariate analysis of OS, HER-2/neu overexpression did not approach statistical significance (hazard ratio: 1.60, P = 0.23). Results are summarized in Table 3.

Metastatic breast cancer is a heterogeneous disease, and similar treatment modalities may offer different levels of benefit to different subsets of patients. The benefit of HDCT for metastatic breast cancer patients as a group is still undetermined. However, it is important to identify patients who would have a poor PFS or OS from HDCT. HER-2/neu overexpression might be such an identifier.

HER-2/neu plays an important role in cell proliferation and differentiation, and its overexpression seems to correlate with an increase in the proliferative activity of breast cancer cells (6, 8). The association between HER-2/neu overexpression and poor survival has been documented in male as well as female breast cancer patients, suggesting that this phenomenon is universal across the sexes (9). Some authors reported that the tumors of patients with HER-2/neu overexpression can render the sensitivity to doxorubicin and paclitaxel (19, 20, 21); however, these studies did not include controls for proliferation rate, and, therefore, their findings could reflect an increase in the chemosensitivity of cells that had higher proliferative activity.

Fluorescence in situ hybridization and HercepTest immunohistochemical staining are the most commonly used methods for HER-2/neu status assessment. However, our group has already reported the high concordance rate of results obtained by anti-HER-2/neu e2-4001 monoclonal antibodies to results of fluorescence in situ hybridization and HercepTest assays (22). Our laboratory has been using anti-HER-2/neu e2-4001 monoclonal antibodies for years, and changing to HercepTest would result in a major increase in cost to the patients.

Many others have found, as we did, that HER-2/neu overexpression correlated with absence of ER expression (3, 6, 7, 8, 21). The exact molecular mechanism of this phenomenon remains to be elucidated, but the finding of estrogen-mediated down-regulation of HER-2/neu in breast cancer cell lines implies the existence of reciprocal control at the level of gene regulation (23). Considering the well-established adverse prognostic effect of the absence of ER expression and its common coexistence with HER-2/neu overexpression in breast cancer, some investigators have expressed doubt as to whether the HER-2/neu overexpression has independent prognostic significance. Nevertheless, the independent prognostic significance of HER-2/neu overexpression in breast cancer has been established in patients treated with standard-dose chemotherapy and now, from our results, for patients with metastatic breast cancer patients treated with HDCT.

The conversion rate from PR to CR after HDCT may become an important factor that may determine the long-term outcome of those who undergo HDCT. Most Phase II studies show 20–40% PR to CR conversion rate, including our study showing 31% (15). In contrast, Philadelphia study shows only 8% conversion rate (24). In this current study, the low conversion rate among HER-2/neu-overexpressing tumors might reflect the sensitivity to alkylator used in the HDCT setting; however, the small number of observations precludes drawing a definitive conclusion. The lower conversion rate from PR to CR after HDCT found in our study might reflect the sensitivity to alkylator used in HDCT setting; however, the small number of observations precludes drawing a definitive conclusion. Therefore, this observation needs to be confirmed with a large number of patients.

In univariate analyses, we found that HER-2/neu overexpression correlated with shorter OS and PFS, a finding that agrees with those of other studies of involving standard or high-dose chemotherapy for high-risk primary or metastatic breast cancer (11, 12, 13, 14, 25). In our multivariate analyses, HER-2/neu overexpression remained a statistically significant predictor for PFS but not for OS. This finding could well have reflected the small number of patients and low statistical power of our study. The fact that ER/PR status also was not significant for OS in multivariate analyses tends to support this explanation.

The introduction of trastuzumab to clinical practice has provided a powerful modality for enhancing sensitivity to chemotherapeutic agents and suppressing tumor growth (26, 27). In cell line model HER-2/neu antibodies also rendered breast cancer cells more sensitive to radiation (28). Two trials underway at M. D. Anderson Cancer Center and at Diana Faber Institution are incorporating trastuzumab in maintenance therapy after HDCT.⁵ This approach may allow using trastuzumab to eradicate minimal residual disease after high-dose chemotherapy and improve the long-term survival.

In conclusion, the correlation between HER-2/neu overexpression and shorter OS and PFS in patients with metastatic breast cancer given HDCT with autologous stem cell transplantation may indicate that HER-2/neu overexpression is an important prognostic factor. Despite the correlation between HER-2/neu overexpression and the absence of ER/PR expression and other unfavorable disease factors, simultaneous consideration of these factors in a modeling procedure suggested that HER-2/neu overexpression correlated with a shorter PFS independently of other factors.

We thank Christine Wogan of the M. D. Anderson Cancer Center’s Department of Scientific Publications for critically reviewing the manuscript.