I-SPY2 is a multicenter phase 2 trial in high risk stage II/III breast cancer (BC) using adaptive randomization within biomarker subtypes to evaluate novel treatment agents added to standard neoadjuvant chemotherapy (NAC) in different phenotypic subsets of breast cancer. Residual cancer burden (RCB) quantifies the extent of residual disease (RD) for patients who did not achieve pathologic complete response (pCR = RCB-0).
Local site pathologists reported RCB in the I-SPY2 trial. We performed a pooled analysis of 678 patients in I-SPY2 with RCB data and known follow-up (median 2.5 years). Cox models for event-free survival (EFS) were evaluated for RCB index (continuous) and RCB classes (hazard ratio; 95% CI) in all patients and in subtypes defined by hormone receptor (HR) and HER2 status. We separately compared experimental and control arms (Wilcoxon rank sum test) in a pooled analysis of RCB index (498 patients in total) from the first six treatment comparisons that “graduated” a therapy based on ≥85% predicted probability of increasing pCR rate over control therapy in a future 300-patient phase 3 trial.
RCB index was prognostic overall (hazard ratio; 95% CI: 1.86; 1.62-2.14) and in each subtype: TNBC (2.09; 1.70-2.57, N = 224), HR-/HER2+ (2.91; 1.79-4.73, N = 69), HR+/HER2+ (1.41; 1.00-1.99, N = 134), and HR+/HER2- (2.08; 1.54-2.81, N = 251). Overall, estimates of 3-year EFS for RCB classes were: pCR 94%, RCB-I 87%, RCB-II 80%, RCB-III 62%. The distribution of RCB index decreased with graduating treatments, relative to control therapy, in TNBC (p < 0.001) and HER2+ (p = 0.03), but not in HR+/HER2- (p = 0.21). In those with RD (excluding pCR), there was a trend for decreased RCB index with graduating treatments, relative to control therapy, in TNBC (P = 0.08), but not in HER2+ (p = 0.43) or HR+/HER2- cancers (p = 0.94).
RCB determined by local site pathologists was prognostic in all subtypes of breast cancer. Observed differences in RCB index distribution between randomized treatments suggested different patterns, possibly by class of experimental treatment and phenotype of disease. Clinical trial information: NCT01042379