A new study using gene expression profiling to divide patients from an adjuvant trastuzumab (Herceptin) clinical trial into subgroups based on their response to the drug has shown that adjuvant treatment in addition to chemotherapy might be effective even in HER2-negative patients.
The results, which were published in the Journal of the National Cancer Institute earlier this month, confirm an earlier finding using data from the National Surgical Adjuvant Breast and Bowel Project's B-31 trial, which compared standard chemotherapy with chemotherapy plus trastuzumab in the adjuvant setting.
In an earlier study of B-31 patients, published in the New England Journal of Medicine the NSABP researchers found, surprisingly, that HER2 status itself did not appear to be predictive of trastuzumab benefit in the trial, and that HER2-negative patients also appeared to have responded to the drug.
Among the 1,787 patients with follow-up data, 174 had breast cancers that were found to be central HER2-negative, yet these patients also appeared to benefit from trastuzumab, the group reported.
Soonmyung Paik, the new study's senior author, told PGx Reporter this week that this initial result prompted the team to go back to the samples from the trial to try to develop an alternative genomic predictor of response, and to see more precisely where the HER2-negative subjects fit into the response spectrum.
The resulting genomic classifier divided the trial cohort into low-responding, intermediate-responding, and high-responding subsets and showed clearly that almost all the HER2-negative patients in the trial fit into the intermediate group with only a few in the low-response group.
Current clinical guidelines recommend that only HER2-positive patients be treated with trastuzumab, but Paik said that his team's finding provides support for further testing of trastuzumab in HER2-negative patients, which the NSABP is now doing in another trial called B-47.
"We have already started this trial, B-47, where we are testing the efficacy of Herceptin in 3,700 HER2-negative patients, and we have recruited 2,000 so far," Paik said.
"We haven’t used the test [we developed in this study] as an upfront stratifier, but certainly we will look at it when the trial is over," he added.
In the NSABP team's recent effort to define genomic subsets with differing responses in the B-31 cohort, Paik and his colleagues used NanoString's nCounter platform to narrow a field of 462 candidate genes — including genes from earlier microarray profiling, candidate genes from other studies, and genes in the PAM 50 and Oncotype DX breast cancer assays — down to an eight-gene stratifier using two sets of original trial participants.
"In B-31, about 10 percent of the cases were essentially accidentally enrolled into the trial and were HER2-neg," Paik explained." Surprisingly, they also got the benefit, which means that HER2 is not the best predictor of Herceptin benefit, so we wanted to discover a good predictor."
The group built the model in a 588-patient discovery set made up of 516 HER2-positive and 72 HER2-negative patients, and then tested and confirmed in a separate group of 991 total — 884 HER2-positive and 107 HER2-negative — trial patients.
Applying the eight-gene prediction model in the confirmation cohort, the researchers found that the model "readily identified" three subsets of response, one with no benefit from trastuzumab, one subset with moderate benefit, and one with a large benefit—with hazard ratios of 1.58, 0.60, and 0.28 respectively.
When the group compared their stratification with central HER2 assay results from the trial, they found that the majority of HER2-negative cases were located in the moderate-benefit subgroup, confirming their earlier finding that HER2-negative patients in B-31 were also deriving some benefit.
Overall, a 40 percent reduction in recurrences would be expected from the addition of trastuzumab to adjuvant chemotherapy for patients whose gene-expression results place them in the moderate-response subset, the authors reported.
About 90 percent of HER2-negative patients from the B-31 cohort fell into the moderate group, while approximately 10 percent of HER2-positive patients were stratified into the no-benefit group.
The team also found that the no-benefit group was characterized by some other unexpected molecular alterations. Because trastuzumab targets HER2, this subset would be expected to express the lowest levels of ERBB2. However, Paik said, the no-benefit group actually had intermediate levels of ERBB2 overexpression, and high levels of ESR1 in both the discovery and confirmation cohorts.
These patients might have dominated the no-benefit group because their baseline prognosis with endocrine therapy alone was so great, Paik explained. Having seen a maximum benefit from anti-estrogen therapy, these patients would derive minimal additional benefit from trastuzumab.
Another possibility, according to the study authors, is that these high ESR1-expressing tumors might be biologically resistant to trastuzumab.
According to Paik, the molecular signature his team developed needs to be validated in additional cohorts before it could potentially be used clinically. He said other groups are currently applying the stratifier to other retrospective sample sets, including the international Herceptin Adjuvant (HERA) trial, which compared two years of treatment with trastuzumab with one year of treatment.