This Phase III study of trastuzumab/paclitaxel with or without carboplatin in advanced breast cancer was spawned by the results of the pivotal trial by Slamon and colleagues, in which the combination of paclitaxel/trastuzumab improved the response rate to the 40 percent range and the time to progression to 6.9 months compared to paclitaxel alone.

We couldn’t add doxorubicin to the paclitaxel/trastuzumab combination because in the pivotal trial, 28 percent of patients in the group given an anthracycline, cyclophosphamide and trastuzumab had cardiotoxicity. We knew of preclinical synergy between the taxanes and carboplatin, as well as three first-line therapy trials showing response rates between 52 percent and 62 percent with the combination of paclitaxel and carboplatin. Therefore, adding carboplatin seemed an obvious next step in evaluating the paclitaxel/trastuzumab combination.

We recruited 196 patients with Stage IV, HER2-positive breast cancer, of whom 191 were eligible and 186 were evaluable for response. As in the trial by Slamon and colleagues, we enrolled patients with IHC 2+ and 3+ disease, but as the data became available, we found that only 30 percent of the patients with IHC 2+ had FISH-positive disease. Therefore, we changed our eligibility requirements so that patients with IHC 2+ disease also had to have FISH-positive disease. Patients had to have measurable disease and a normal left ventricular ejection fraction. They were ineligible if they received adjuvant taxanes or more than 360 mg/m2 of doxorubicin.

Patients were randomized to receive trastuzumab/paclitaxel, the successful arm of the pivotal trial, or the combination plus carboplatin. Paclitaxel was administered at 175 mg/m2 over three hours every 21 days, trastuzumab was administered at a standard loading dose of 4 mg/kg followed by weekly 2 mg/kg, and carboplatin was administered at an AUC of six every 21 days. As in the pivotal trial, physicians had to give six cycles of chemotherapy, but could discontinue chemotherapy and continue the trastuzumab after that.

The addition of carboplatin improved both the response rate and time to progression. The primary endpoint was the response rate, which improved from 36 percent with the two-drug regimen to 52 percent with the addition of carboplatin, with a P value of 0.04.

Time to progression was a secondary endpoint in the trial. The time to progression in the trastuzumab/paclitaxel control arm was similar to what was seen in the pivotal trial by Slamon and colleagues. The addition of carboplatin increased the time to progression from 6.9 months to 11.2 months.

We looked at survival, although it was early to do so as over 120 patients are still alive. The preliminary analysis shows a trend for improvement with the three-drug regimen. In the patients with IHC 3+ disease, we saw an improvement in survival, with a P value of 0.06, approaching 0.05, and the population with FISH-positive disease showed a similar trend. It will be important to see if the survival advantage persists.

The trastuzumab/paclitaxel/carboplatin regimen was well-tolerated. The only significant difference in toxicity was increased myelosuppression, which we expected to see from the addition of carboplatin. However, there were no significant differences in terms of serious complications, such as infectious complications, significant neutropenia or fever. Other toxicities, such as neuropathy, allergic responses, nausea and arthralgias, were comparable in both arms.

It is important to note that we did not use prophylactic growth factors or attempt a dosedense trial. We utilized dose reduction or dose delay when needed. In responding patients, only about 25 percent continued treatment beyond six cycles, so there are a number of important caveats when administering this regimen in order to get the benefits and avoid unacceptable toxicities. One of the questions our trial evoked was: Could we achieve the same results by treating patients with paclitaxel/trastuzumab and switching to carboplatin and trastuzumab when they progress? Historically, carboplatin is not a very effective agent when given outside the first-line setting, with response rates in the range of 10 percent; but it’s possible that in combination with trastuzumab it’s a different drug. This may be a strategy to consider in future clinical trials.

Nicholas Robert, MD