pCR counts
However, patients who did not have pathologic complete responses at the end of first-line de-escalated therapy had worse outcomes, with a 5-year iDFS rate of 82%, compared with 98% for patients who had achieved a pCR. This translated into a hazard ratio for invasive disease in patients with pCRs of 0.14 (P = .011).
This difference occurred despite the study requirement that all patients who did not have pCR after 12 weeks of initial therapy would receive additional chemotherapy.
Looking at the tumor subtype among patients in the paclitaxel-free arm to see whether they could identify predictors of early response, the researchers found a pCR rate of 36.5% among 85 patients with nonbasal tumors, but 0% among 7 patients with basal tumors.
The investigators identified a population of patients whose tumors could be considered nonsensitive to dual HER2 blockade alone: Those with basal tumors, those tumors with low immunohistochemical HER2 expression, and those without an early response to therapy on biopsy 3 weeks into initial therapy. Among 31 of the 92 patients in the dual HER2 arm who met this description, 2 had pCRs, Dr. Harbeck noted.
The 5-year iDFS rate among patients in the dual blockade–only arm with nonsensitive tumors was 79%, compared with 93% for patients with treatment-responsive types, although there were only 13 invasive events total in this arm.
“If we look at the whole trial population, the negative prognostic impact of what we termed nonsensitive tumors was even significant regarding dDFS, with a hazard ratio of about 5,” she said.
‘A consistent theme’
Invited discussant Lisa A. Carey, MD, ScM, of the University of North Carolina Lineberger Comprehensive Cancer Center in Chapel Hill, noted that the trial was underpowered for outcomes, but that results nonetheless suggest that patients with strongly HER2-driven tumors might get comparable benefits from less chemotherapy.
“This trial included only hormone receptor–negative, HER2-positive tumors, and these we know are likely to be HER2-enriched in terms of subtype, about three-quarters of them,”she said.
The previously reported pCR rate of 90% in the paclitaxel-containing arm, with 80% of patients requiring no further chemotherapy, resulted in the excellent 5-year iDFS and dDFS in this group, despite the relatively highly clinical stage, with about 60% of patients having clinical stage 2 or higher tumors, and more than 40% being node positive.
The idea that pCR itself can predict which patients could be spared from more intensive chemotherapy “is starting to look like a consistent theme,” she said.
Dr. Carey pointed out that in the KRISTINE trial comparing the combination of trastuzumab emtansine (T-DM1) and pertuzumab with standard chemotherapy in patients with HER2-positive stage I-III breast cancer, although the experimental combination was associated with lower pCR rates and worse event-free survival, rates of iDFS/dDFS were virtually identical for patients in both arms who achieved a pCR.
“So the question is can pCR mean that we can either eliminate additional therapy,” she said, noting that the question is currently being addressed prospectively in two clinical trials, COMPASS-pCR and DECRESCENDO.
ADAPT HER2+/HR- is sponsored by F. Hoffman-La Roche. Dr. Harbeck disclosed institutional research funding from Roche/Genentech, as well as honoraria and consulting/advising for multiple companies. Dr. Carey disclosed institutional research funding and other relationships with various companies.