For the meta-analysis, Dr. Spring and her coinvestigators searched for published studies of localized breast cancer that had 25 patients or more, featured neoadjuvant chemotherapy, and reported pCR using definitions allowed by the FDA (ypT0 ypN0 or ypT0/is ypN0), as well as recurrence and/or survival based on pathologic outcome. They excluded studies reporting only local recurrence and those using neoadjuvant endocrine therapy or neoadjuvant radiation.
Results showed that the pCR rate averaged 21.1% for the entire study population, according to Dr. Spring. But there was wide variation by tumor subtype, as expected, with a rate of less than 10% for hormone receptor–positive/HER2-negative breast cancer, to rates in the mid-30% range for triple-negative breast cancer and HER2-positive breast cancer.
Women who had pCR after neoadjuvant chemotherapy had a significantly lower risk of EFS events than peers who had residual disease (hazard ratio, 0.31; 95% probability interval, 0.24-0.39). The corresponding 5-year EFS rates were 88% and 67%.
Similarly, women who had pCR had a significantly lower risk of death (HR, 0.22; 95% probability interval, 0.15-0.30). The corresponding 5-year overall survival rates were 94% and 75%.
The EFS benefit of pCR versus residual disease was consistently seen across subgroups with triple-negative breast cancer (90% vs. 57%), HER2-positive breast cancer (86% vs. 63%), and hormone receptor–positive/HER2-negative breast cancer (97% vs. 88%). Findings were essentially the same for overall survival, according to Dr. Spring.
Among patients attaining pCR, the 5-year EFS rate was 86% for those who went on to receive adjuvant chemotherapy (HR, 0.36; 95% probability interval, 0.19-0.67) and 88% for those who did not (HR, 0.36; 95% probability interval, 0.27-0.54). The difference in hazard ratios between groups was not significant (P = .60).
Finally, the investigators conducted modeling to assess how the change in pCR rate corresponded with the change in EFS benefit. “This approach could be helpful in the design of neoadjuvant studies,” Dr. Spring explained.
“Assuming pCR is a valid surrogate endpoint, this is, it mediates all treatment effects, and that the average pCR is 50%, the magnitude of pCR change is predictive of treatment effects on EFS within a certain amount of uncertainty, based on the model,” she reported. For example, a change in pCR of 0.3 had a corresponding HR of 0.72, with a 95% probability interval of 0.68-0.77.
Dr. Spring disclosed that she has a consulting or advisory role with Novartis and that she receives institutional research funding from Tesaro. The study was supported by grants from the National Cancer Institute and Susan G. Komen.
SOURCE: Spring LM et al. SABCS 2018, Abstract GS2-03.