Lisa A. Carey, MD
Dual HER2-targeted neoadjuvant therapy for early breast cancer did not significantly improve pathologic complete response (pCR) rate as compared with a single anti-HER2 agent plus chemotherapy, according to results from the phase III CALGB 40601 trial.1
Further, exploratory analyses of response by tumor intrinsic subtype revealed considerable heterogeneity, including in response to therapy. The data were presented at the 2013 ASCO Annual Meeting.
In CALGB 40601, the combination of paclitaxel, trastuzumab (Herceptin), and lapatinib (Tykerb) led to pCR in 56% of patients versus 46% with paclitaxel and trastuzumab. The 10-point absolute difference did not meet the prespecified improvement required for statistical significance, primarily because the paclitaxel/trastuzumab combination achieved a higher-than-expected pCR rate.
“The addition of lapatinib to 16 weeks of paclitaxel/ trastuzumab did not meet predefined criteria for a significant increase in pCR,” said Lisa A. Carey, MD, of the University of North Carolina at Chapel Hill (UNC), who presented the data at the ASCO meeting. “Higher pCR rates were noted among hormone receptor-negative than hormone-positive tumors, as has been seen in other studies.
“Exploratory analyses by intrinsic subtype showed that HER2-positive breast cancer is molecularly heterogeneous. Intrinsic subtypes appear to differ in sensitivity to HER2-targeting agents, with the numerically highest pCR rates among the HER2- enriched subtype.”
The rationale for CALGB 40601 came from earlier trials showing an improved pCR rate in patients with metastatic HER2-positive breast cancer previously treated with two anti-HER2 agents. Carey said the trial was designed to provide additional quantitative information about pCR following neoadjuvant therapy with two versus one HER2-targeted therapies.
Investigators randomized 305 patients with stage II-III HER2-positive breast cancer to weekly paclitaxel plus trastuzumab, lapatinib, or both. Neoadjuvant therapy continued for 16 weeks, and the primary objective was pCR. The trial design had the statistical power to demonstrate an improvement in pCR from 30% with paclitaxel/trastuzumab to 50% with the three-drug regimen.
Accrual to the paclitaxel/lapatinib arm ended prematurely when a preliminary analysis of a similar ongoing trial revealed unfavorable efficacy and toxicity data, said Carey.
The lapatinib-containing regimens had more grade ≥3 toxicity, including neutropenia, rash, and diarrhea.
The response data showed pCR rates of 56% with the three-drug regimen versus 46% with paclitaxel plus trastuzumab. The paclitaxel/lapatinib arm had a pCR rate of 37% among women who had completed therapy prior to discontinuation of the arm.
About 60% of patients had hormone receptor– positive breast cancer, and a preplanned analysis showed that all three regimens resulted in higher pCR rates among women with hormone receptor– negative cancers—77% with the three drugs, 55% with paclitaxel/trastuzumab, and 37% with paclitaxel/ lapatinib versus 42%, 39%, and 31%, respectively, in hormone receptor–positive breast cancer.
Putting the results into context, Carey noted that CALGB 40601 and three other randomized trials of neoadjuvant therapy in HER2-positive breast cancer had consistently demonstrated numerically higher pCR rates with dual anti-HER2 therapy. In two of the trials (NeoSPHERE and NeoALTTO), the difference in favor of dual anti-HER2 therapy reached statistical significance.
Though consistent, the primary results of the four trials showed substantial heterogeneity in terms of pCR rates. The trials employed different chemotherapy regimens, and the length of neoadjuvant treatment varied somewhat. Even so, underlying differences in the patient populations might have played a role in the results, said Carey.
In an embedded correlational study, CALGB 40601, investigators compared pCR rates by breast cancer intrinsic subtypes. Pretreatment analysis of 160 patients showed that a third of the tumors were HER2-enriched, another third were luminal A, and more than 20% were luminal B. Normal-like (8%) and basal-like (7%) tumors accounted for the remainder.
Posttreatment analysis of 63 patients showed residual disease consisted of <5% in luminal B and basal-like tumors, 6% in HER2-enriched, 29% in luminal A, and 57% in normal-like tumors.
Analysis of instrinsic subtypes of hormone receptor status also showed considerable variability. HER2-enriched subtype predominated in hormone receptor-negative tumors (52%), whereas luminal A accounted for 43% of hormone-receptor positive tumors, luminal B for 34%, and HER2-enriched for 18%.