SAN ANTONIO – Using Myriad Genetics' next-generation sequencing based homologous recombination deficiency (HRD) Test, researchers have identified a subset of early-stage triple-negative breast cancer (TNBC) patients who responded to platinum-based therapies, according to data presented at the San Antonio Breast Cancer Symposium.
Additionally, in a separate study presented here, this time in advanced TNBC patients, researchers reported that BRCA mutation status gauged by Myriad's BRACAnalysis test was predictive of carboplatin benefit over docetaxel.
However, another group of researchers looking at basal-like subtypes and gene expression signatures in TNBC didn't have much luck identifying best responders to the addition of carboplatin to chemotherapy; but could predict responders when Genetech's Avastin (bevacizumab) was added.
TNBC patients don't have many treatment options beyond chemotherapy since unlike other kinds of breast cancers, these tumors don't overexpress targetable markers such as estrogen receptor (ER), progesterone receptor (PR), and HER2. TNBC tumors are heterogeneous, with an estimated 15 percent of TNBC patients harboring BRCA mutations. Like breast cancer patients with BRCA mutations, TNBC patients tend to have faulty DNA repair mechanisms. As such, past trials have shown that early-stage and advanced TNBC patients benefit from DNA-damaging platinum-based therapies.
In a study involving 62 early-stage patients from two previously conducted trials on cisplatin, researchers led by Andrea Richardson from the Dana-Farber Cancer Institute analyzed archival formalin-fixed paraffin-embedded biopsy tumor samples using Myriad's myChoice HRD Test. Researchers used the diagnostic, which gauges germline and somatic BRCA1/2 gene mutations, as well as a number of other markers involved in DNA repair, to analyze patients for homologous recombination (HR) deficiency and associated the test results with their pathologic complete response (pCR). The endpoint pCR is a measure of the absence of residual invasive or in situ cancer in the resected tumor and in lymph nodes after neoadjuvant treatment.
Study investigators deemed patients to be HR deficient when they had HRD test scores of 42 or greater or BRCA1/2 mutations. Richardson's group reported that 52 percent of early-stage TNBC patients deemed to be HR deficient by the test responded to platinum-based treatment compared to only 10 percent of patients with intact HRD. More specifically, 28 percent of HR-deficient patients had a pCR, while none of the patients had a pCR in the intact HRD group.
Myriad said it is planning to offer the HRD test to TNBC patients through an early access program next year.
Meanwhile, in the metastatic TNBC setting, Andrew Tutt of Kings College London enrolled 376 patients who were classically ER, PR, and HER2-negative, or had BRCA mutations but were positive for ER or HER2. Researchers randomized these patients to either carboplatin or docetaxel, and then analyzed them for their TNBC subtype via NanoString's Prosigna PAM50 assay, for their BRCA status by BRACAnalysis, or for their HR deficiency using the HRD Test.
Upon enrollment, 29 patients were known BRCA mutation carriers, among whom 13 didn't have TNBC, but had ER-positive, HER2-negative disease. Upon analysis with BRACAnalysis, 43 patients were identified as having BRCA mutations. There was no significant difference in objective response in the overall, molecularly undifferentiated population between the carboplatin and docetaxel arms; there was also not much difference in terms of progression-free survival and overall survival.
However, in the BRCA mutation carrier population the objective response was 68 percent in the carboplatin arm compared to 33 percent for those receiving docetaxel, a significant difference. Response to these two agents in the BRCA wild-type status wasn't substantially different. In the carboplatin arm, median progression-free survival was 6.8 months in BRCA mutated patients and 3.1 months in wild-type patients. Median progression-free survival for patients on docetaxel didn't differ significantly based on BRCA status.
While there was no evidence that in the unselected TNBC population carboplatin is better than docetaxel, patients with BRCA mutations experienced "significantly greater" response and progression-free survival with carboplatin than doctaxel, Tutt said. "The TNT trial supports BRCA1/2 genotyping to inform therapy choice in metastatic TNBC and familial breast cancer," he noted.
Then, researchers gauged responses based on HRD test scores, analyzing 220 samples (81 had high HRD scores and 114 had low scores). The test didn't yield a result in 25 samples. Those with germline or somatic BRCA mutations tended to have high HRD scores.
"In the high HRD group there was no significant difference between the performance of these two agents," Tutt said at SABCS. "In the low HRD score group the performance of the two drugs was also similar." However, patients with BRCA1/2 mutated tumors and high HRD scores did have high platinum responses.
One factor impacting results may be that researchers analyzed primary tumor DNA even though the study was focused on metastatic disease, Tutt said. He suggested that further development of the HRD test is needed for the non-BRCA mutated cancers, especially in the metastatic setting. Non-BRCA mutated patients "perhaps more frequently reverse their underlying cause for homologous recombination deficiency than do those with germline or somatic mutations," he offered.
Researchers also looked at response according to basal-like subtypes, but the performance of the two agents was similar in this group. In the non-basal-like group 74 percent of those on docetaxel responded, while only 17 percent responded to carboplatin. However, Tutt cautioned not to read too much into this finding in the non-basal-like subset, given the small number of patients. Generally, in the study, carboplatin was better tolerated than docetaxel.
While HR deficiency and BRCA status tracked with response to platinum-based drugs in early and advanced TNBC patients, respectively, another group of investigators looking at the predictive ability of basal subtypes and gene expression signatures in neoadjuvant (pre-surgery) patients didn't have as much luck identifying predictive markers for carboplatin.
In the so-called CALGB 40603, a team led by William Sikov of Women and Infants Hospital in Rhode Island tried to identify subsets of TNBC who garnered a better of worse benefit with the addition of carboplatin and Genentech's Avastin (bevacizumab) to standard chemotherapy for neoadjuvant treatment. Using the PAM50 breast cancer subtyping assay they found that pathologic complete response (pCR) rates didn't vary between basal-like and non-basal-like subtypes with the addition of carboplatin, but pCR rates did significantly vary when Avastin was added.
Then, Sikov and colleagues analyzed samples from 360 patients by Illumina's HiSeq, classified tumors according to 17 previously published gene signatures, and assessed if these signatures could differentiate between pCR rates. A high proliferation signature, low estrogen signature, and high TP53 mutation signature were all predictive of pCR and also predictive of benefit with Avastin therapy. None of these signatures predicted response for the addition of carboplatin, however.
Although pCR rates didn't differ with the addition of carboplatin between basal-like and non-basal like types, Sikov noted that this didn't mean that carboplatin is of no benefit in basal-like TNBC patients. "There was simply no accentuation of that benefit in the basal-like subtype compared to non-basal-like cancers," he said. "That is, we did not identify a subtype of TNBC that did not benefit from the addition of carboplatin."
The latest study is an extension of an earlier analysis, which the CALGB/Alliance researchers published in the Journal of Clinical Oncology earlier this year. In that paper, researchers reported that TNBC patients' pCR rates increased when Avastin or carboplatin was combined with standard chemotherapy.
Although these results are encouraging, they probably won't change the standard of care, according to Sikov, particularly since it is unknown whether an improvement in pCR will necessarily translate to a survival advantage. "For carboplatin, this [study] doesn't give us much guidance" since researchers didn't identify a subgroup that had substantial improvements in pCTR, Sikov told GenomeWeb. "We're certainly hoping that other signatures we're looking at in other correlative studies will."
A more complicated picture
With regard to Avastin, the question of assessing a differential benefit for TNBC patients is complicated, Sikov said, by the fact that past studies have shown only a borderline or non-significant long term impact on outcomes from the drug. He suggested that based on this study, Genentech could reanalyze those studies "excluding the 20 percent of patients who had non-basal like cancers or 20 percent of patients who had low proliferation cancers" to see if there is "a much larger [benefit] both statistically and clinically with the addition of bevacizumab in those patients."
In 2011, the US Food and Drug Administration revoked its accelerated approval for Avastin in metastatic breast cancer saying that the drug did not improve overall survival and only marginally approved progression-free survival, but caused serious adverse events. Genentech had asked FDA to keep the drug on the market for patients who were responding to it, and to give the company more time to identify a best-responder population. During public hearings hosted by the FDA to analyze Avastin's efficacy and safety data, many patients, particularly those with TNBC, said they had experienced robust responses to Avastin treatment and pleaded with the agency to keep the drug on the market.
Genentech conducted a large study to gauge benefit of Avastin treatment in TNBC patients, but the results weren't positive. In the more-than-2,500-patient Phase III BEATRICE study, Avastin plus chemotherapy in the adjuvant setting didn't improve invasive disease-free survival or overall survival in early-stage TNBC patients compared to just chemo. In that study, Genentech also looked for biomarkers of response to early-stage TNBC patients, but failed to report definitive leads.
A Genentech spokesperson told GenomeWeb that the the data reported to date suggest that Avastin is primarily beneficial in the metastatic breast cancer setting, and that's where Genentech has focused its predictive biomarker discovery efforts. Genentech will likely not change its plans for Avastin based on CALGB 40603, and the spokesperson further pointed out that the study size was much smaller than BEATRICE.
Still, the CALGB 40603 study "raises questions as to the importance of subsets especially for bevacizumab," Sikov said, noting that he believes it is still worth pursing research to identify breast cancer subsets that benefit from Avastin treatment.
Meanwhile, Genentech is conducting a Phase III study called MERIDIAN, where it is investigating whether baseline VEGF-A levels will predict which metastatic breast cancer patients will respond to Avastin plus paclitaxel. The company hopes to report results from this trial next year.