TNBC Subtype Responds Favorably to Neoadjuvant Bevacizumab


The addition of bevacizumab (Avastin) to standard neoadjuvant chemotherapy significantly improved pathologic complete response (pCR) rates in women with basal-like breast cancer compared with non-basal-like subtypes.

William M. Sikov, MD

The addition of bevacizumab (Avastin) to standard neoadjuvant chemotherapy significantly improved pathologic complete response (pCR) rates in women with basal-like breast cancer compared with non-basal-like subtypes, according to an analysis of the CALGB 40603 trial presented at the 2014 San Antonio Breast Cancer Symposium (SABCS).1

In an earlier analysis of the study, the addition of carboplatin or bevacizumab to standard neoadjuvant chemotherapy was shown to significantly increase pCR rates in women with triple-negative breast cancer (TNBC). In the updated analysis presented at SABCS, a distinct interaction between subtype and pCR was established, with basal-like subtypes responding favorably to bevacizumab. Moreover, expression of mRNA gene signatures of high proliferation, low estrogen expression, or high TP53 expression was associated with higher pCR rates with neoadjuvant bevacizumab.

"The addition of either carboplatin or bevacizumab to standard neoadjuvant therapy was associated with significantly higher pCR breast and pCR breast and axilla rates in patients with basal-like cancers," said lead investigator William M. Sikov, MD, associate professor of medicine at the Alpert Medical School of Brown University in Providence, Rhode Island. "There was an interaction between subtype–that is basal-like and non-basal-like–and the size of pCR increment only for bevacizumab."

In the CALGB 40603 study, 443 patients with stage II-III TNBC were treated with paclitaxel plus either bevacizumab, carboplatin, or the combination of the two. Weekly paclitaxel was administered for 12 cycles at 80 mg/m2. Patients received bevacizumab at 10 mg/kg every 2 weeks for 9 cycles and carboplatin at AUC 6 every 3 weeks for 4 cycles. All patients received didecyldimethylammonium chloride for 4 cycles.

Following neoadjuvant therapy, patients went on to receive surgery or radiation treatment. Adjuvant systemic therapy was not planned.

Three hundred sixty-seven biopsy samples were collected for intrinsic subtype analysis prior to the initiation of therapy. Evaluable samples (n = 360) were examined using paired-end mRNA sequencing, with results normalized to PAM50 and claudin-low subtypes. The primary endpoint of the study was to assess pCR, defined as the absence of residual invasive cancer in the breast (ypT0/is).

Across all samples, 87% were identified as basal-like (n = 314), 7% were normal-like, 4% were HER2 expressing, 2% were claudin-low, and <1% were luminal A. There were a total of 46 samples with non-basal-like histology. Subtype distribution was similar across all treatment arms (P = .88). The pCR rate across all treatment arms did not differ by subtype, at 54% in basal-like and 52% in non-basal-like.

"Eighty-seven percent, a higher number than we anticipated, proved to be basal-like in our analysis, with a much smaller percentage having other subtypes," explained Sikov. "The overall pathologic complete response rate in the breast was similar between patients with basal-like cancers and non-basal-like cancers."

In basal-like patients, the addition of bevacizumab improved pCR in the breast from 45% with paclitaxel alone to 64% with bevacizumab (OR = 2.15; P = .0009). In the breast and axilla combined, the pCR with bevacizumab was 57% versus 43% with paclitaxel alone (OR = 1.76; P = .014).

A clear benefit was not seen for adding bevacizumab to paclitaxel in the non-basal-like subtype. With bevacizumab plus paclitaxel, those with non-basal-like tumors experienced a pCR of just 43% compared with 64% for basal-like tumors (P = .024). 

In the basal-like subtype specifically, the addition of carboplatin to paclitaxel demonstrated a pCR rate in the breast of 61% versus 47% with paclitaxel alone (odds ratio [OR] = 1.76; P = .014). In the breast and axilla, the pCR rate with carboplatin was 56% versus 43% with paclitaxel alone (OR = 1.72; P = .018). These outcomes were similar in patients with non-basal-like subtypes, with a pCR in the breast of 58% with carboplatin.

"For carboplatin, the benefit of its addition on pathologic complete response was equivalent in the basal-like and non-basal-like patients," Sikov said. "There was no significant interaction between subtype and the benefit of adding carboplatin. This doesn't mean that carboplatin was of no benefit in the basal-like subtype, there was simply no accentuation of that benefit in the basal-like subtype compared with non-basal-like cancer."

High expression of HER2 was not prognostic for pCR in the full study population but was associated with reduced benefit from carboplatin (P = .025). Other signatures demonstrated added pCR benefit with the combination of bevacizumab and paclitaxel, including high tumor proliferation (P = .031), high TP53 mutation signature (P = .0017), and low estrogen signaling signatures (P = .0002).

"None of these signatures were predictive of greater pCR benefit for carboplatin in all patients or in the basal-like subset," Sikov elaborated.

Five immune signatures were analyzed by mRNA gene expression that reflected the extent and characterization of tumor-infiltrating lymphocytes. These signatures included: B cell, CD8, T cell, ICG, and general immune cells. All signatures analyzed were associated with pCR benefit across all subtypes; however, this benefit was not statistically significant.

"There was no association between these signatures and a greater pCR benefit with the addition of carboplatin or bevacizumab, with all P values for interaction greater than 0.2," Sikov said. "This is in contrast to findings from GeparSixto, which reported an interaction between immune signatures and a benefit with carboplatin."

In the GeparSixto trial, 11 immune signatures were linked with increased pCR by multivariate analysis in patients treated with neoadjuvant carboplatin-based chemotherapy.2 Moreover, findings from the study suggested these immune signatures could be used in addition to tumor infiltrating lymphocytes to predict benefit to neoadjuvant carboplatin.

"I would say that the jury is still out on the role of carboplatin in the neoadjuvant setting for triple-negative breast cancer. It depends on your interpretation of results presented over the last year and half from our trial and GeparSixto," Sikov said. "I don't think the results presented today will really change anyone's mind. What we haven't done is identify a subgroup in whom the benefit is much greater but on the other hand we haven't identified a subgroup that doesn't benefit from carboplatin."


  1. Sikov WM, Barry WT, Hoadley KA, et al. Impact of intrinsic subtype by PAM50 and other gene signatures on pathologic complete response (pCR) rates in triple - negative breast cancer (TNBC) after neoadjuvant chemotherapy (NACT) +/ - carboplatin (Cb) or bevacizumab (Bev): CALGB 40603/150709 (Alliance). Presented at: 2014 San Antonio Breast Cancer Symposium; December 9-13, 2014; San Antonio, TX. Abstract S4-05
  2. Denkert C, von Minckwitz G, Brase JC, et al. Expression of immunologic genes in triple-negative and HER2-positive breast cancer in the neoadjuvant GEPARSIXTO trial: Prediction of response to carboplatin-based chemotherapy. J Clin Oncol. 2014;32:5s (suppl; abstr 510).  


View more from the 2014 San Antonio Breast Cancer Symposium

Related Videos
A panel of 3 experts on breast cancer
Paolo Tarantino, MD
A panel of 3 experts on breast cancer
Sujith Samarasinghe, MD
Suzanne Trudel, MSc, MD
Consuelo Bertossi, MD
Michael R. Grunwald, MD, FACP
Kevin Kalinsky, MD, MS,
Manali Kamdar, MD
Ibrahim Aldoss, MD