The following is a summary of recently presented information about the possible use of trastuzumab (Herceptin) in the adjuvant therapy of breast cancer. Adjuvant! does not yet make estimates of benefit of adjuvant trastuzumab because the follow-up of these trials is so short. 

At a special session of the May 2005 ASCO meeting in Orlando the results of 3 randomized trials testing trastuzumab (Herceptin) in the adjuvant therapy of Her2 positive breast cancer were presented.  This information has subsequently been published [1-3].  Two additional trials were presented at the 2005 San Antonio Breast Cancer Symposium: BCIRG 006 and the FinHer trial [5-6]. 

 All of these trials contained control arms in which patients did not receive trastuzumab, and all patients in the trials received adjuvant (or in the case of some patients in the HERA trial neoadjuvant) chemotherapy. In the HERA trial randomization occurred after adjuvant chemotherapy. In B31, N9831, BCIRG 006, and FinHer trials the patients were randomized prior to getting chemotherapy.  In the HERA trial and one arm of N9831 trastuzumab was given only after all chemotherapy was completed, in the other trials it was given concurrently with at least some of the adjuvant chemotherapy. 

The following table provides an overview of the trastuzumab trials as of December 2005. It illustrates how powerful trastuzumab is in reducing the risk of early relapse, how short the follow-up is for these trials, and how some questions remain still open. For example, why does transtuzumab appear relatively ineffective when given after chemotherapy in N9831, but strongly effective when given after chemotherapy in HERA? 

Proportional Risk Reductions

Combined Analysis of NSABP B31 and N9831

Because of the many similarities in eligibility and treatment in the B31 and N9831 trials, the decision was made to do a combined analysis [1]. This combined analysis included results from the control arms of these two trials, and the arms of the trials in which patients received trastuzumab starting concurrently with paclitaxel and continuing for a total of 1 year. An arm of N9831 in which patients received trastuzumab after all the chemotherapy was completed was not included in the analysis 

Blue boxes represent cycles of AC, red circles paclitaxel either every 3 weeks or every 1 week, yellow circles represent weekly trastuzumab. 

All patients received 4 cycles of doxorubicin and cyclophosphamide (AC * 4), followed by 12 weeks of paclitaxel. A slight difference between the trials was that in the NSABP trial the paclitaxel was given every 3 weeks, and in the Intergroup trial it was given on a weekly basis. Overall paclitaxel was given every 3 weeks to 48% of the patients, and on a weekly schedule to 52% of the patients.  Trastuzumab was given weekly starting with the first paclitaxel treatment and continuing for a total of 52 weeks. 

Median follow-up was short, only 2.0 years (2.4 years in the B31 and 1.5 years in N9831). 

The eligibility criteria shown in red are of special importance to this trial. The N9831 trial had been amended to include high risk node negative patients, so there were a few (6% overall) patients who were high risk node negative. (Defined as tumor size > 1cm if ER and PgR negative, and tumor size > 2 cm if ER and/PgR positive.)

Because of the concern about possible cardiac toxicity, patients with a history of cardiac history were not eligible. Patients also must have had a normal left ventricular fraction (LVEF). [[ No past or active cardiac disease included: history of myocardial infarction, history of congestive heart failure, angina pectoris requiring medication, arrhythmia requiring medication, clinically significant valvular disease, uncontrolled hypertension, LVH, and cardiomegaly on CXR.]]

Because of concern about the quality of Her2 determinations done at some sites, a special effort was made to standardize Her2 determinations. To be eligible patients had to have Her2 determined by immunohistochemistry as +++ or to have Her2 gene amplification as determined by FISH.  The B31 trial required that the Her2 determination be done at an approved laboratory. The N9831 trial required that the determinations be done at a central laboratory

Patient and tumor characteristics were balanced between treatment arms

.  

The primary endpoint of the analysis was disease free survival (DFS).  There was both a strong statistically significant absolute advantage for trastuzumab (8% at 3 years), and also relative advantage (hazard ratio 0.48). This is a 52% reduction in recurrence and mortality events. There were statistically significant early improvements in DDFS and OS as well. Subset analysis was done for DFS and showed benefit irrespective of age, tumor size, number of positive nodes (although patients who were node negative were included in the analysis), ER status, or which trial they participated in. There was also a trend to less contralateral breast cancer in the trastuzumab treated group (4 cases vs 6 cases) and also fewer second primary cancers in the trastuzumab treated cases (5 cases vs 20 cases; p = 0.002).

Of particular concern in this trial was cardiac safety. For the B31 trial it was reported that cumulative incidence at 3 years of Grade 3 or 4 New York Heart Association congestive heart failure was in 0.8 and 4.1% of the patients in the control and trastuzumab groups respectively. Of the 31 patients developing CHF, 27 have been followed for at least 6 months, and only 1 reported persistent symptomatic problems. In the N9831 trial it was reported that cumulative incidence at 3 years of Grade 3 or 4 New York Heart Association congestive heart failure was in 0.0 and 2.9% of the patients in the control and trastuzumab groups respectively. One of the trastuzumab treated patients died of cardiomyopathy.  .

Very careful cardiac monitoring was an important element in this trial. Not only was a normal LVEF required at entry, but LVEF monitoring was required after the completion of the 4 cycles of AC * 4, and then 3, 6, and 12 months after starting trastuzumab. Approximately 20% of the patients had to stop trastuzumab because of decreases in LVEF. Even with this careful monitoring some patients developed CHF.  In fact for patients older that 50 and with on-study LVEF's of 50-54% the risk of developing CHF was ~20!  Although much of this CHF was reported to be reversible (4), the long term consequences of this are unknown. 

Preliminary Analysis of All Three Arms of N9831

An initial analysis of N9831 was reported at the ASCO meeting. This study only had a median follow-up of 1.5 years. The HR for no trastuzumab versus trastuzumab starting together with paclitaxel was 0.55 (p < 0.001). The HR for no trastuzumab versus trastuzumab starting after paclitaxel was 0.87 (p < 0.29). This would suggest that trastuzumab must be given concurrently with paclitaxel to have a strong beneficial effect in adjuvant therapy. However, as reviewed below, the HERA trial suggests that trastuzumab may have strong beneficial effects even if given alone without concurrent chemotherapy, a result which is in apparent conflict with the inferences that might be drawn from the sequencing comparisons made from the results of N9831. 

Preliminary Analysis of the HERA Trial

The third adjuvant trastuzumab trial presented at ASCO was the HERA trial and these results have subsequently been published.  This trial had a different design than the B31 and N9831 trials. In this trial all patients were randomized and started trastuzumab after completing adjuvant chemotherapy. The chemotherapy regimen was not as standardized as in the North American trials, but had to be one of the approved regimens. 

The major questions this trial was designed to answer was whether 1 year of trastuzumab improved outcome and whether 2 years of trastuzumab would be even better. This analysis compared the early results of Arm A (no trastuzumab) with Arm B (1 year of trastuzumab). This trial enrolled 3,307 patients. It was reported at only 1 year of follow-up. 

The eligibility criteria for this trial were far broader that for the North American Trials. Patients must have received chemotherapy for their early breast cancer, but they could have received it either as adjuvant of neoadjuvant therapy. The chemotherapy received must have been at least 4 cycles of an approved regimen. Node negative patients could have been included, but they must have had a tumor T1c > 10 mm in size. Patients must have had a LVEF > 55%.  Her2 must have been confirmed to be +++ by IHC or FISH + by the central lab. 

The median age was approximately 50.  One third of the patients participating in this trial were node negative. About 95% of the patients participating in this trial had received an anthracycline as part of their therapy.  About one quarter had received a taxane. About 50% of the patients were ER positive. Most of these patients received some form of hormonal therapy.

The HERA trial like the B31 and N9831 trials showed a striking improvement in DFS, reducing the risk of relapse and mortality events by about 46%.  This result is statistically significant. The difference in outcome is also large in absolute terms with about an 8% advantage at 2 years of follow-up. The advantage was evident irrespective of nodal status, hormone receptor status, or age.  Improvement in DDFS was also seen (reduction in risk by 51% p < 0.001) . There was a non-significant trend for an improvement in OS. It is still too early for an analysis of OS but there appears to be a trend developing favoring trastuzumab (29 deaths in the trastuzumab group and 37 deaths in the control group, reduction of 24 % p = ns).  It must, however, be cautioned again that this analysis is based on only 1 year of follow-up.

In the HERA trial there was more cardiac toxicity in the group that got trastuzumab . This was true both when cardiac toxicity was assessed by the measurement of LVEF to < 50% and a fall of > 10% (2% of the patients in the control group and 7% of the patients in the trastuzumab group), or severe CHF in 0.6% of the trastuzumab treated patients and none of the controls. This was a lower incidence than in the B31 patients, but whether this difference was due to a difference in the entry criteria (a higher LVEF of 55% was required in the HERA trial), or the fact that the trastuzumab was not given concurrently with chemotherapy in HERA, or that the chemotherapy may have be different, or some other factor is unclear. 

This trial is of special significance because it shows that trastuzumab can have a make effect even if not given concurrently with chemotherapy. 

BCIRG 006

This trial enrolled 3,222 patients and has been reported with 23 months of follow-up. 

Both node negative (29%) and node positive patients (71%) were enrolled. Only 54% of the patients were ER positive. 52% of the patients were younger than 50. 

There were 3 treatment arms. The standard arm was AC * 4 every 3 weeks followed by docetaxel  * 4 every 3 weeks. The second arm was the same therapy but with trastuzumab started concurrently with the docetaxel. There was a 3rd arm that had the unique feature that it did not include an anthracycline. This arm used a combination of docetaxel and a platinum (nearly always carboplatin) with the rationale that such an arm would hopefully be efficacious, but would allow trastuzumab to be given without significant increases in cardiotoxicity.  When given concurrently with chemotherapy the trastuzumab was given on a weekly basis. Afterward the trastuzumab was given every 3 weeks. Total duration of trastuzumab treatment was 1 year. 

Both the trastuzumab arms were better than the standard arm in terms of DFS. For the AC to docetaxel with trastuzumab arm the hazard ratio was 0.49 (p-value <0.001. For the platinum docetaxel trastuzumab are the hazard ratio was 0.61 (p-value < 0.001).  At 2 yrs of follow-up there the DFS was 86%, 93%, and 91% respectively for the control, control plus trastuzumab, and the carboplatin/docetaxel arms respectively. 

There was no statistically significant difference between the hazard ratios of the two trastuzumab containing arms, but the modest number of events made this comparison one of low statistical power. 

Grade 3 / 4 Toxicities

The trial did find that  symptomatic cardiac events were increased in the AC to docetaxel trastuzumab arm compared to the standard arm, but between the non anthracycline containing arm vs control arm. Control: vs Control plus trastuzumab: 1.0% vs 2.3%, p-value=0.02.  Control vs docetaxel/carboplatin trastuzumab: 1.0% vs 1.3% p value=0.11. Absolute LVEF decline of >10% occurred in 9% of the patients in the control arm, 17% in the control plus trastuzumab arm., and 8% in the docetaxel/carboplatin trastuzumab arm. 

This trial shows is consistent with other trials in that it shows that trastuzumab can improve the efficacy of anthracycline / taxane regimens. It also shows that by omitting the anthracycline, the cardiac toxicity of trastuzumab can be largely avoided. The carboplatin docetaxel trastuzumab arm was clearly more efficacious that the control arm, but whether it is slightly less efficacious than the control arm plus trastuzumab arm cannot be addressed at this time with any statistical certainty.  One might be tempted to speculate that the carboplatin docetaxel trastuzumab regimen might be particularly relevant to the treatment of Her2 amplified patients with pre-existing cardiac issues, but it should be kept in mind that these patients were for the most part excluded from this trial and therefore the cardiac safety in this patient population is unknown. 

The FinHer Trial

This trial enrolled 1,010 patients and has been reported with 3.2 years of follow-up. Patients were either node positive, or node negative with tumors > 2 cm in size and PgR -. All ER positive patients received 5 years of tamoxifen.  All patients were randomized between:

3 cycles of docetaxel (100 mg/m2 every 3 weeks) followed by 3 cycles of FEC (600/60/600) mg/m2 every 3 weeks

or

8 weekly doses of vinorelbine (25 mg/m2) followed by 3 cycles of FEC (600/60/600) mg/m2 every 3 weeks

Patients who had Her2 amplification (n = 231) as detected by CISH (chromagen in situ hybridization) were additionally randomized between 9 weekly doses of trastuzumab or control. The first dose of trastuzumab was given at 4 mg/kg, with the other doses given at 2 mg/m2.  

There was a slight statistically sigificant imbalance between treatment arms in tumor size with T1ab tumors being found in 6% of the patients getting doctaxel and 10% of the patients getting vinorelbine, but this imbalance was minor, and other patient characterists were well balanced with 11% of the patients being node negative, 72% of the patients being ER positive, and 23% of the patients having Her2 amplification. Patients with severe cardiac disease, severe hypertension, or CHF of any degree were excluded. 

The following slide shows the basic design of the trial. 

Purple boxes represent cycles of docetaxel. Red boxes represent weekly vinorelbine. Blue boxes represent cycles of FEC. Yellow circles represent weekly trastuzumab. 

The results of this trial showed a striking advantage for docetaxel over vinorelbine. 

The results of this trial also show a striking advantage for trastuzumab over control. These results have important implications because they show a similar improvement in outcome with only 9 weeks of trastuzumab, much shorter course that in any other trial. Although the effects on OS do not reach statistical significance the expected trend is there, and the trial is underpowered at this time to evaluate this endpoint. 

The docetaxel arm had considerably more hematologic toxicity. When given at 100mg/m2 the rate of febrile neutropenia was 41% so the initial dose was decreased to 80mg/m2 and even at this level the rate  of febrile neutropenia was still 15%. The average docetaxel related rate of febrile neutropenia was 24% as compared to the 3% rate in the vinorelbine arm. In terms of risk of CHF these were no reported instances of CHF in the trastuzumab arm. Risk of a decrease in LVEF of > 15% was actually lower in the trastuzumab arm (3% vs 6%).  It might surprise some that there was no CHF in the trastuzumab treated patients, particularly given that the serum half life of trastuzumab is long and it would certainly have been present during the start of the anthracycline treatment. It may be present in this trial but simply missed because of the small size of this trial.  It might be noted that in a neoadjuvant study of trastuzumab given concurrently with paclitaxel followed by FEC there was only 1 patient out of 45 who developed grade 3 CHF so perhaps such regimens can be given with acceptable cardiac safety.  More data bearing in this question would be helpfu. 

This fascinating study opens the door to the exploration of short courses of trastuzumab. 

Summary: 

The Her2/trastuzumab target / targeted agent pairing seems to have over the short term strongly beneficial effects in the predominantly node positive patient population it was tested in. 

The balance of risk versus benefit over the long term is more uncertain because we have only short term follow-up.  As with most agents used in adjuvant therapy, the value of adding trastuzumab is most uncertain in low risk (stage 1) patients. For these patients the balance of benefit and risk is particularly uncertain, and can only be estimated with any degree of accuracy as more detailed safety and efficacy information becomes available.  Results (anticipated in 12/2005, for both for efficacy and cardiac safety) of BCIRG-006 a Phase III adjuvant trial evaluating the non-anthracycline containing regimen of docetaxel, carboplatin, and trastuzumab are awaited with interest. 

1) Romond EH, Perez EA, Bryant J, et al. Trastuzumab plus Adjuvant Chemotherapy for Operable HER2-Positive Breast Cancer. N Engl J Med 353:1673-84, 2005.

2) Perez EA, Suman VJ, Davidson N, et al.

NCCTG N9831 May 2005 ASCO Update

3) Piccart-Gebhart MJ, Procter M, Brian Leyland-Jones, B. Trastuzumab after Adjuvant Chemotherapy in HER2-Positive Breast Cancer. N Engl J Med 2005;353:1659-72.

The HERA Study Team. Trastuzumab (H: HerceptinŽ) following adjuvant chemotherapy (CT) significantly improves disease-free survival (DFS) in early breast cancer (BC) with HER2 overexpression: the HERA Trial. December 2005 San Antonio Breast Cancer Symposium. Abstract 11. 

4) Geyer, Jr. CE, Bryant J, Romond E, et al.  Cardiac safety analysis of the first stage of NSABP B-31, a randomized trial comparing the safety and efficacy of Adriamycin and cyclophosphamide (AC) followed by Taxol to that of AC followed by Taxol plus Herceptin in patients (Pts) with operable, node-positive (N+), HER-2 overexpressing breast cancer (HER2+BC).  Abstract 23, SABCS 2003. 

5) Slamon D, Eiermann W, Robert N, et at. Phase III randomized trial comparing doxorubicin and cyclophosphamide followed by docetaxel (AC®T) with doxorubicin and cyclophosphamide followed by docetaxel and trastuzumab (AC®TH) with docetaxel, carboplatin and trastuzumab (TCH) in HER2 positive early breast cancer patients: BCIRG 006 study. December 2005 San Antonio Breast Cancer Symposium. Abstract 1. 

6) Joensuu H, Kellokumpu-Lehtinen P-L, Bono P,  et al. Trastuzumab in combination with docetaxel or vinorelbine as adjuvant treatment of breast cancer: the FinHer Trial. December 2005 San Antonio Breast Cancer Symposium. Abstract #2. 

7) Buzdar AU, Valero V, Ibrahim N, et al. Prospective data of additional patients treated with neoadjuvant therapy with paclitaxel followed by FEC chemotherapy with trastuzumab in HER-2 positive operable breast cancer, and an update of initial study population. December 2005 San Antonio Breast Cancer Symposium. Abstract #5049. 
 

11/05