|Home | About | Journals | Submit | Contact Us | Français|
To determine whether (1) immunohistochemical (IHC) HER2 status (ie, 2+ or 3+), (2) degree of fluorescence in situ hybridization (FISH) amplification according to (2a) HER2/CEP17 ratio or (2b) HER2 gene copy number, or (3) polysomy significantly influenced clinical outcome for patients with human epidermal growth factor receptor 2 (HER2) –positive breast cancer enrolled in the Herceptin Adjuvant trial of trastuzumab versus no trastuzumab administered after completion of chemotherapy.
IHC and/or FISH analyses were performed locally and required central confirmation as indicating HER2 positivity for trial entry. FISH data from the central HER2 analysis on patients in the 1-year trastuzumab and no trastuzumab arms were assessed in relation to disease-free survival (DFS) after a median 2 years of follow-up.
Central FISH results were available for 2,071 (61%) of the 3,401 patients randomized to the 2 arms. Among patients with FISH-positive disease, (1) the hazard ratios for trastuzumab versus no trastuzumab were 0.56 (95% CI, 0.32 to 0.99) for locally IHC2+ cases (n = 340) and 0.80 (95% CI, 0.40 to 1.61) for centrally IHC2+ cases (n = 299). There was no significant prognostic relationship between (2a) HER2 FISH ratio, (2b) HER2 copy number, or (3) polysomy and DFS in the control arm or predictive relationship defining differential benefit from trastuzumab.
There was no evidence for reduced benefit of trastuzumab in HER2 IHC2+FISH+ cases. The degree of HER2 amplification does not influence prognosis or benefit from adjuvant trastuzumab in patients treated with prior adjuvant chemotherapy.
The human epidermal growth factor receptor 2 (HER2) is a 185-kd glycoprotein with tyrosine kinase activity. Its overexpression in breast cancer is a key feature of the pathobiology of the disease and is associated with poorer prognosis.1,2 Amplification is the primary mechanism of HER2 overexpression.3 Approximately 15% of all newly diagnosed invasive breast carcinomas are HER2 positive (HER2+) but the proportion is higher among tumors with higher grade and patients with positive nodal status. As a result approximately 25% of patients considered appropriate for adjuvant chemotherapy are HER2 positive.4
Trastuzumab is a humanized monoclonal antibody directed to the external domain of HER2. Four large and one small randomized trial of trastuzumab (Herceptin; Genentech, South San Francisco, CA) in early breast cancer have shown that this treatment can significantly improve the patient outcome when trastuzumab is applied alongside and/or subsequent to adjuvant chemotherapy.5–8 Decreases in recurrence and mortality of approximately one half and one third, respectively, have been demonstrated. As a result trastuzumab has become standard of care for the treatment of HER2+ early breast cancer.
The diagnosis of HER2 overexpression is based on two analytic approaches that are most commonly used in conjunction: immunohistochemistry (IHC) and in situ hybridization (ISH).9 IHC reveals overexpression of HER2 on the cell membrane. The degree of staining is most frequently described on a scale of 0 to 3, with 3+ being considered unequivocally positive (> 10% of cells with intense circumferential membrane staining according to US Food and Drug Administration, or > 30% immunostained tumor cells as per the recent American Society of Clinical Oncology/College of American Pathologists (ASCO/CAP) recommendations9), 2+ equivocal and 0/1+ negative. ISH reveals the number of HER2 gene copies per cell and has been most commonly conducted with a fluorescence in situ hybridization (FISH) probe. A second probe that hybridizes to the centromeric region of chromosome 17 (CEP17) can be used and this allows the ratio of HER2 gene copies per chromosome to be calculated. A ratio of 2.0 or greater is regarded as positive according to the US Food and Drug Administration recommendations, though a cutoff of 2.2 or greater has been proposed more recently in the ASCO/CAP guidelines.9
The aims of the current study were to determine whether within the control and 1-year treatment arms of HERceptin Adjuvant (HERA), the largest of the adjuvant trials, there were significant differences in clinical outcome according to IHC status (ie, 2+ or 3+) in FISH-amplified cases or according to the degree of FISH amplification. We also assessed the impact of increased HER2 gene copy number as a result of polysomy on prognosis and treatment benefit and the impact of polysomy itself in an HER2-positive population.
The HERA (Breast International Group [BIG] 01-01) trial is an international, intergroup, open-label, phase III randomized trial involving women with HER2-positive (overexpressed or amplified) early stage invasive breast cancer who completed locoregional therapy (surgery plus radiotherapy where indicated) and a minimum of 4 courses of chemotherapy. The study design, eligibility criteria, treatment schedules, monitoring and statistical analysis plan have been described in detail elsewhere.5 Eligibility criteria included node-positive disease or node-negative disease if the pathological tumor size was greater than 1 cm. The institutional review board at each of the participating institutions approved the study protocol. All patients gave written informed consent.
After local regional therapy patients were randomly assigned to one of three groups: observation only (no trastuzumab), 8 mg/kg trastuzumab intravenously as a loading dose followed by 6 mg/kg every 3 weeks for 1 year, or the same schedule of trastuzumab for 2 years: outcome data is not yet available for the 2-year arm. Randomisation took place within 7 weeks from the last chemotherapy cycle or 6 weeks from the end of radiotherapy or definitive surgery, whichever was last.
The primary end point of the trial was disease-free survival (DFS), and this is the only end point reported here. The data presented here are from the analysis conducted after a median 2-years of follow-up.10 DFS was defined as time from randomisation to the first occurrence of any of the following events: recurrence of breast cancer at any site; the development of ipsilateral or contralateral breast cancer including ductal carcinoma in situ; second nonbreast malignant disease other than basal cell or squamous cell carcinoma in situ of the cervix; or death from any cause without documentation of a cancer-related event.
Central confirmation (TARGOS, Kassel, Germany) of the HER2-positive status of tumors was required in all cases before randomisation as follows. A result on IHC (DAKO, HercepTest) at the central lab of 3+ (> 10% stained cells) was required for confirmation of the status of tumors assessed by the participating institution as 3+, and a positive result from FISH (PathVysion, Vysis; HER2:CEP17 ratio ≥ 2.0) for HER2 amplification was required for tumors that were assessed in the participating institution as IHC2+ or FISH-positive. FISH was not performed before randomisation on cases that were locally deemed IHC3+ and confirmed IHC3+ centrally. The results from both local and central analyses were recorded in the case report forms.
For the purposes of this study additional FISH analyses were performed retrospectively for patients not having a central FISH analysis before randomisation. Tissue sections of 4 microns were used from one of two resources: 1 sections cut at the time of central analysis and stored at TARGOS; 2 sections taken in TransHERA, a parallel project to HERA in which formalin fixed paraffin embedded (FFPE) tumor excision blocks have been collected retrospectively for translational research studies. These retrospective FISH analyses were conducted either at TARGOS or in the European Institute of Oncology (Milan, Italy) after an initial pilot study to ensure common standardized procedures.
The database closure for this analysis was February 14, 2006. Between December 2001 and June 2005, there were 1,698 patients randomized to observation (no trastuzumab) only and 1,703 patients randomly assigned to 1-year trastuzumab. We calculated Kaplan-Meier estimates for 3 year DFS rates according to subgroups and estimated SEs and 95% CIs according to the Greenwood formula.11,12 We calculated Kaplan-Meier estimates for DFS curves for observation patients according to defined subgroups. Tests for interaction between treatment, respective pathological parameter and DFS were carried out using Cox proportional hazard models.
Central FISH results were available on 2,071 (61%) of the 3,401 patients randomized to the 1-year trastuzumab or no trastuzumab arms of HERA. Of these 1,131 results were derived during the eligibility screening and 940 were derived from material banked either as spare sections at TARGOS or as sections taken for TransHERA. DFS was similar for the subgroup of 2,071 patients at 3 years to that for the overall set of 3,401 patients both for the treated (80.6% and 81.2%, respectively) and untreated arms (74.3% and 75.1%, respectively). The HR for trastuzumab versus no trastuzumab in the subgroup was 0.69 (95% CI, 0.55 to 0.86) versus 0.64 (95% CI, 0.54 to 0.76) in the overall population.
To determine whether patients with IHC2+FISH+ breast cancer derived significant benefit from trastuzumab, analyses were performed in two partially overlapping populations: (i) in the 340 cases that were locally IHC2+ and centrally FISH+ the hazard ratio was 0.56 (95% CI, 0.32 to 0.99); and (ii) in the 299 cases that were centrally IHC2+ and centrally FISH+ the hazard ratio was 0.80 (95% CI, 0.40 to 1.61). The overlap between these populations was only 53 patients, mainly because the majority of local IHC2+ central FISH+ cases did not have a central IHC result and the majority of the central IHC2+ central FISH+ cases were local IHC3+.
We examined the ER status for patients in the 2 IHC2+FISH+ subgroups described above because it is known that ER positivity and HER-2 protein levels are inversely correlated.13 In the locally IHC2+FISH+ subgroup, 60.3% of cases were ER+ compared with 45.6% in the locally IHC3+FISH+ (P < .0001). Similarly, 63.2% of the centrally IHC2+FISH+ cases were ER+ compared with 42.2% of the centrally IHC3+FISH+ (P < .0001).
The distribution and range of FISH ratios and HER2 copy numbers are shown in Figures 1A and and1B,1B, respectively. Analyses of FISH ratio and HER2 copy number with clinical outcome were conducted by comparison of approximate quartiles of the respective data sets: exact quartiles were calculated and then the closest integer cutoff was applied to define the 4 approximate quartiles (shown in Fig 1A and and1B).1B). There was no significant relationship between FISH ratio levels or HER2 copy number on prognosis in the untreated arm (Fig 2A and 2B). Although there was an apparent trend towards decreasing effectiveness of trastuzumab with increasing FISH ratio (Fig 3A; Table 1) this was not statistically significant (P = .29) and the trend was less evident when HER-2 was quantified by copy number only (P = .38) (Fig 3B; Table 2). FISH ratio was found to correlate inversely with ER-positivity (based on locally performed ER analyses): the proportion ER+ was 64%, 46%, 46% and 39% for the categories of more than 2.0 to 4.0, more than 4.0 to 6.0, more than 6.0 to 8.0 and more than 8.0, respectively (P < .0001).
There were only 41 patients with tumors that by central FISH analysis had HER2 ratios less than 2.0. We therefore performed no evaluation of possible benefit from trastuzumab in this subgroup.
The distribution of chr17 copy numbers/cell is shown in Figure 1C. Two sets of comparative analyses were conducted: (i) between three groups defined by copy numbers/cell ≤ 2.0, 2.0 to 3.0 and more than/= 3.0; and (ii) between groups defined as polysomic or not where polysomy was considered as at least 30% of nuclei with 3 or more CEP17 signals. For both (i) and (ii) there was no significant difference in prognosis (Fig 2C and and2D)2D) and no significant difference in the effectiveness of trastuzumab (Fig 3C and and3D;3D; Tables 3 and and4)4) according to chr17 copy number.
Large randomized trials have established trastuzumab as the standard of care in patients with early HER2-positive breast cancer also suitable for treatment with chemotherapy.5–7 Although there has been much debate about the appropriate method of diagnosing HER-2 status widely accepted guidelines have recently been published that confirm the combined use of IHC and/or FISH as an appropriate means of establishing positivity.9 It is clear however that the degree of HER-2 overexpression/amplification varies markedly in patients considered positive: the wide range and the continuum of this variability is well demonstrated by the analyses shown in this study (Fig 1).
The relatively low level of HER-2 protein expression in the subgroup of patients with IHC2+/FISH+ disease might conceivably lead to a lower degree of benefit in such patients. The design of the central testing algorithm in HERA resulted in there being two sets of samples available to address this issue. In both the locally IHC2+/FISH+ and centrally IHC2+/FISH+ cohorts, ER was positive in a greater proportion than in the corresponding IHC3+ cohort. Given that there is a well-described inverse relationship between HER-2 protein levels and ER positivity in breast cancers,13 these data provide supporting evidence for lower levels of HER-2 protein in both of the IHC2+/FISH+ groups compared with IHC3+ cases. We found a statistically significant benefit in the 340 patients (with 49 events) that were diagnosed locally with IHC 2+ and centrally with FISH+ disease. For the group of 299 patients that was diagnosed centrally with IHC2+ and FISH+ disease, there was no significant benefit found but the 95% CI for the hazard ratio included the overall hazard ratio (0.64) and there was no statistically significant difference from the overall result by the test of heterogeneity. Data from the NSABP B31 trial also suggested benefit in the IHC2+/FISH+ group while the NCCTG N9831 trial did not. In both cases, however, the number of DFS events was very low14,15 and a meta-analysis is needed to provide useful information.
The HERA trialists have recently shown that the proportional benefit from trastuzumab does not vary substantially across subgroups of HER2-positive patients that differ markedly in prognosis.16 By the current analyses we have now established that there are no marked differences in benefit according to the degree of amplification whether this is expressed in terms of the HER2/CEP17 ratio or the absolute number of HER2 copies; the latter varies from the former because of the presence of polyploidy of chr17 that can be assessed from the use of the 2-probe FISH test used in this study. It is also notable that prognosis did not differ in the nontrastuzumab arm according to the level of amplification. All patients on HERA received adjuvant chemotherapy5; therefore this result reflects prognosis after chemotherapy and the clinical outcome would in part reflect benefit from that chemotherapy. The amplicon which harbors HER2 contains several other biologically significant genes that may or may not be coamplified with HER217 and influence the benefit from certain chemotherapy, such as the apparent interaction between TOP2A and benefit from anthracyclins.18,19 Such interactions may have influenced our ability to identify any relationship of HER2 amplification and prognosis per se.
Our data indicate that differences in HER2 expression resulting from different FISH ratios, HER2 copy number or ploidy have no significant relationship with benefit from trastuzumab. This implies that there is a strong threshold effect whereby any degree of amplification above the cutoff ratio of 2.0 is of equal clinical significance. A consequence is that misdiagnosis of patients as having HER2-negative tumors, which inevitably is most likely to occur at amplification levels close to the threshold,20 would deny a patient the potential of deriving full benefit of trastuzumab. This is in marked contrast to the strong direct relationship between expression levels for the other major target of breast cancer therapy, ER-alpha, and benefit from tamoxifen21 This latter relationship indicates that while misdiagnosis of low-positive ER patients is highly undesirable the loss of potential benefit from tamoxifen will be small for most patients.
The precise level of the threshold of HER-2 amplification/overexpression in relation to benefit from trastuzumab has recently been brought into doubt by data from the NSABP-B31 trial that indicate that the subgroup of 174 patients conventionally considered as having HER2-negative disease (centrally tested IHC and FISH negative) appeared to benefit from trastuzumab (relative risk for DFS, 0.34; 95% CI, 0.14 to 0.80).22 A similar analysis in 103 patients in the NCCTG trial also showed a trend towards benefit of trastuzumab but this was not statistically significant (HR 0.51; 95% CI, 0.21 to 1.2).15 Similarly the HERA trial had only 41 patients that could contribute to this analysis and we deemed it inappropriate to analyze this small group but hope to contribute these to a more detailed assessment together with samples from the other trials to clarify this issue.
Chromosome 17 polysomy (defined as at least one more copy of chromosome 17 than normal) in the absence of HER2 amplification is sometimes associated with 2+ IHC staining but rarely with 3+ indicating that such polysomy has little influence on HER2 gene expression in breast carcinoma23 and no apparent influence on the clinical presentation or behavior in HER2-negative tumors.24 In the present study we assessed the importance of polsomy in the HER2-positive population. There is no agreed definition of cut-offs for polysomy. We applied an analysis which used a cutoff of a mean 3 CEP17 signals per cell as mentioned in the recent ASCO/CAP guidelines9 and also a cutoff of at least 3 copies per cell in at least 30% of nuclei as used for analysis of the NCCTG trial.25 The latter gave a substantially larger population that was considered polysomic but in neither case was there a statistically significant association with clinical outcome. There were insufficient patients to assess the possible importance of polysomy according to different levels of gene amplification.
The strengths of the current work include the analysis being conducted within a randomized trial and the FISH analysis being conducted centrally in just two well-controlled laboratories that had conducted a pilot study to ensure comparability between them. The study population amounted to only 61% of the HERA population in the 1-year and no trastuzumab groups and as such there must be a caveat that results may have been different in the whole population. However, this seems unlikely since the clinical outcome for the subpopulation was very similar to that in the overall trial. The subset of samples amounted to over 1,000 cases in each arm giving a high probability of finding any substantial effect on prognosis. These same numbers are, however, modest to determine whether there is a significant interaction with treatment, such that only a major interaction is incompatible with the overall data from the HERA trial; collaborative overview analysis with other trialists is envisaged to provide greater statistical power for analyses of interaction with treatment.
In conclusion, the current analysis indicates that in HER2-positive early breast cancer there is no significant effect of HER2 FISH ratio, HER2 copy number or polysomy of chr17 on either prognosis or benefit from trastuzumab in a chemotherapy-treated population.
We are indebted to the women who participated in the study; to the Breast European Adjuvant Study Team (BrEAST) data center; to the Frontier Science Team; to the Breast International Group (BIG) Secretariat; the HERA steering committee; the independent data monitoring committee; the Pathology Laboratory at TARGOS, Kassel, Germany; the Istituto Europeo di Oncologia, Milan, Italy; Hoffman-La Roche; and all of the doctors who had actively participated to the success of the HERA trial.
Supported in part by NHS (National Health Service) funding to the Royal Marsden National Institute for Health Research Biomedical Research Centre (M.D.).
Presented in part at the 30th Annual San Antonio Breast Cancer Symposium, December 13-16, 2007, San Antonio, TX.
Terms in blue are defined in the glossary, found at the end of this article and online at www.jco.org.
Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
Clinical Trials repository link available on JCO.org.
Clinical trial information can be found for the following: NCT00045032.
Although all authors completed the disclosure declaration, the following author(s) indicated a financial or other interest that is relevant to the subject matter under consideration in this article. Certain relationships marked with a “U” are those for which no compensation was received; those relationships marked with a “C” were compensated. For a detailed description of the disclosure categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors.
Employment or Leadership Position: Bruce Jordan, Roche Professonal Diagnostics (C) Consultant or Advisory Role: Mitch Dowsett, Roche (C), DAKO (C), GlaxoSmithKline (C); Josef Rueschoff, Roche (C); Martine Piccart-Gebhart, Roche (C), GlaxoSmithKline (C); Brian Leyland-Jones, Roche (C) Stock Ownership: None Honoraria: Mitch Dowsett, Roche; Martine Piccart-Gebhart, Roche, GlaxoSmithKline; Brian Leyland-Jones, Roche Research Funding: Mitch Dowsett, Roche, GlaxoSmithKline; Josef Rueschoff, Roche; Oliver Stoss, Roche; Giuseppe Viale, Roche; Richard D. Gelber, Roche; Brian Leyland-Jones, Roche Expert Testimony: None Other Remuneration: None
Conception and design: Mitch Dowsett, Worta McCaskill-Stevens, Josef Rueschoff, Giuseppe Viale, Richard D. Gelber, Martine Piccart-Gebhart, Brian Leyland-Jones
Financial support: Bruce Jordan
Administrative support: Mark Abramovitz
Provision of study materials or patients: Josef Rueschoff, Oliver Stoss, Giuseppe Viale
Collection and assembly of data: Mitch Dowsett, Martine Piccart-Gebhart, Evandro de Azambuja, Urania Dafni, Josef Rueschoff, Stella Dolci, Oliver Stoss, Giuseppe Viale, Richard D. Gelber
Data analysis and interpretation: Mitch Dowsett, Martine Piccart-Gebhart, Worta McCaskill-Stevens, Evandro de Azambuja, Urania Dafni, Josef Rueschoff, Bruce Jordan, Stella Dolci, Oliver Stoss, Giuseppe Viale, Richard D. Gelber, Martine Piccart-Gebhart, Brian Leyland-Jones
Manuscript writing: Mitch Dowsett, Richard D. Gelber
Final approval of manuscript: Mitch Dowsett, Martine Piccart-Gebhart, Worta McCaskill-Stevens, Evandro de Azambuja, Urania Dafni, Josef Rueschoff, Bruce Jordan, Stella Dolci, Mark Abramovitz, Oliver Stoss, Giuseppe Viale, Richard D. Gelber, Martine Piccart-Gebhart, Brian Leyland-Jones