Preliminary data shows an anthracycline/taxane-based chemotherapy regimen (TaxAC) could be superior to docetaxal plus cyclophosphamide in patients with breast cancer.
Joanne L. Blum, MD, PhD
Treatment with a taxane plus anthracycline (TaxAC) showed superior disease-free survival (DFS) compared with an anthracycline-free regimen of docetaxel plus cyclophosphamide (TC) for patients with high-risk, HER2-negative breast cancer, according to findings from the phase III ABC trials. Moreover, the benefit seen with TaxAC over TC was more pronounced in patients with hormone receptor-positive tumors with more than 4 positive nodes, according to lead investigator Joanne L. Blum, MD, PhD.
The phase III ABC trials included USOR 06-090, NSABP B-46I/USOR 07132, and NSABP B-49. The studies randomized patients to a TaxAC regimen of doxorubicin or cyclophosphamide plus docetaxel or paclitaxel (n = 2052) or TC (n = 2078). The primary endpoint was invasive DFS.
In an interview withTargeted Oncology, the lead investigator of the ABC trials, Blum, oncologist and hematologist at Texas Oncology, discussed the design and results from the analysis she presented during the 2016 ASCO Annual Meeting.
TARGETED ONCOLOGY:Can you give us a recap of the ABC phase II trials?
Blum:The ABC trials were three trials that were done in sequence. These three trials examined TaxAC regimens versus TC times 6. The initial 06-090 trial, which has the longest median follow-up of all the trials of about 6.3 years, was then followed by B46I07132. I should preface that by saying that this was a collaboration with US Oncology Research, as well as the NSABP. This second trial has a median follow-up of 4.8 years.
The B46I trial also had a third arm, which was TC bevacizumab, and those data are not analyzed in the ABC trial's presentation. The third trial was B49. This trial rapidly completed accrual in 2013 and has the shortest medium follow-up of all of them at about 2.2 years.
In aggregate, we had over 4200 patients enrolled in the trial. We have over 4100 analyzed currently. The trial schema included both high-risk node-negative patients, as well as node-positive patients. The trial pre-specified stratification variables included the number of positive nodes, 0, 1-3, 4-9, or 10 more, ER or PR positive, or ER/PGR negative. Also, another stratification variable was which trial the patients were derived from.
The original two trials, 06-090 and B46I07132, allowed only for the tax arm every three weeks times six versus TC. Arm one was tax. In B49, four regimens were allowed. These included the tax regimen every three weeks times six, or AC every two weeks followed by weekly paclitaxel times 12, or AC every three weeks followed by paclitaxel times 12, or AC dose dense followed by paclitaxel dose dense. The 06-090 and B46I07132 trials just had the tax arm and B49 allowed all four arms.
Median follow-up was 6.3 years for 06-090, 4.8 for B46I, and 2.2 years for the B49 trial. The overall median follow-up is 3.2 years. The number of patients that were node negative is 41%, 44% had 1-3 positive nodes, had 11% had 4-9 positive nodes, and 4% had 10 or more positive nodes. The data was as of October 31, 2015, and at that time we had 399 events which exceeded the pre-specified 334 events. The initial analysis was done for the primary endpoint, which was invasive disease-free survival. At that time, the hazard ratio was 1.202, which exceeded the pre-specified hazard ratio of 1.18 for futility of TC versus TaxAC. Therefore, the trial was deemed ready for the initial presentation.
The results from the trial that the overall hazard ratio for the 334 patients was over 1.2, so it was presented. The data we presented was from all 399 patients, and what we saw was that there were more events in the TC arm relative to the TaxAC arm. The overall hazard ratio was 1.23. The curves begin to diverge at 2 to 3 years, the hazard ratio for the primary endpoint was statistically significant with a P value of .04 favoring TC. The other thing we analyzed was the forest plot for the pre-specified stratification variables. What we saw was that the hazard ratios favored TaxAC for both 06-090 and B46I with hazard ratios of 1.31 and 1.34 respectively. However, for B49, which has the shortest follow-up with 2.2 years, the hazard ratio is 1.
For ER/PR negative patients, the hazard ratio was strongly supportive of TaxAC. It was also positive for the hormone receptor positive group, but less so. Node-negative patients had a hazard ratio of 1, but 1-3 positive nodes, 4-9, and 10, the hazard ratio is progressively moved further to the right.
In addition, we performed a forest plot looking at hormonal and nodal receptor statuses. What was clear from that analysis was that for all the subsets, node negative 1 to 3 positive nodes, and 4 or more positive nodes, the hazard ratios were all in excess of 1.3 for the node negative patients, all favoring TaxAC. In contrast, the hazard ratio for the ER/PR positive subset, node negative actually favored TC with a hazard ratio of 0.69. While it was positive for 1 to 3 positive nodes, but particularly so for 4 or more positive nodes.
When we looked at this in even more detail, the KaplanMeier curves really separate for all of the subsets. For all the ER/PR negative group, whereas for the 4 or more positive nodes do they separate in the hormone receptor positive subset.
We presented the data in detailed granular fashion, and it's really quite interesting because this is clear disparity between the hormone receptor negative and the hormone receptor positive groups. This is an exploratory analysis, but the data are fairly compelling. The overall survivals for the two arms is comparable so far, with a median follow-up of about 4 years. We will need longer follow-up to look at overall survival. The other thing that we saw was that there were more distant events in the TC arm, as well as locoregional, compared with TaxAC, and there were 5 acute leukemias seen in the TaxAC arm when compared to the TC arm.
Our conclusions demonstrate, based on the interim analysis, that TaxAC is superior to TC in terms of our primary endpoint of invasive disease free survival. Overall survival is not affected so far, and is quite high so far, at 95%, which is encouraging. The largest benefit, the largest deltas that we saw were for the ER positive, 4 or more positive subset, as well as node-positive ER negative patients, with deltas that range from 5.8% to 11%. Quite striking.
We look forward to further follow-up, of course, and we look forward to correlative studies to determine which subset, and to define those patients who clearly benefit from an anthrocycline.
It was a large series of trials. We were fortunate that we were able to complete the non-inferiority aspect of the trial, and that we're able to provide some important information for clinicians as they decide to use an anthrocycline for their patients or not.
TARGETED ONCOLOGY:What would you say are some thing community oncologists can take away from this?
Blum:In my practice, I plan to use these data to support using an anthrocycline-based regimen in hormone-receptor negative patients. The data are compelling enough for the triple-negative subset. For node-positive patients who are hormone-receptor positive, clearly there is a benefit for four or more positive nodes, and I would use an anthrocycline in selected 1-3 positive nodes that are hormone receptor positive.