Predictive performance of breast cancer index (BCI) and clinical treatment score post-5 years (CTS5) in the IDEAL study.

Abstract

545 Background: The Breast Cancer Index (BCI) is a gene expression–based signature that stratifies patients based on the risk of overall (0-10y) and late (post-5y) distant recurrence (DR) and predicted the likelihood of benefit from extended endocrine therapy in MA.17, Trans-aTTom, IDEAL, and B-42 trials. The Clinical Treatment Score post-5 years (CTS5) is an algorithm incorporating four clinicopathologic variables (nodes, age, tumor size and grade), which has been shown to be prognostic for late DR. Previous analysis of CTS5 in IDEAL showed that CTS5 was not predictive of extended endocrine therapy benefit. The current analysis compared the predictive performance of BCI (H/I) and CTS5 for the benefit of extended endocrine therapy in the same subset of patients from the IDEAL study. Methods: 818 patients from the IDEAL trial who remained recurrence free 2.5y after randomization and had both BCI (H/I) and CTS5 results available were included in this study. The primary endpoint was recurrence-free interval (RFI). Absolute benefit was measured as the difference in 10-year risk of recurrences between the two treatment arms. The likelihood ratio test from Cox regression models was used to test for interaction. Results: 818 IDEAL patients (68% ≥50y at surgery; 48% T2; 47% G2; 73% N+) were included in the analysis. CTS5 stratified 167, 346 and 305 patients into Low-, Intermediate-, and High-risk groups. The analysis revealed that no group derived significant benefit with CTS5-Low showing 6.4% absolute benefit (HR = 0.61, 95% CI: 0.20-1.86, P = 0.377), CTS5-Intermediate showing 6.7% absolute benefit (HR = 0.55, 95% CI: 0.27-1.09, P = 0.080), and CTS5-High showing 2.1% absolute benefit (HR = 0.80, 95% CI: 0.42-1.51, P = 0.482). BCI (H/I) stratified 428 and 390 patients into BCI (H/I)-Low and BCI (H/I)-High groups. Only BCI (H/I)-High patients derived a significant absolute benefit of 10.5% (HR = 0.38, 95% CI: 0.18-0.79, P = 0.007), while BCI (H/I)-Low patients did not show any absolute benefit (HR = 0.94, 95% CI: 0.55-1.60, P = 0.817). The treatment by biomarker interaction was significant for BCI (H/I) (P = 0.045), but not for CTS5 (P = 0.731). When re-stratifying CTS5 categories by BCI (H/I) or vice versa, only BCI (H/I)-High patients showed consistent absolute benefit regardless of CTS5 category (19.4%, 8.1% and 8.8% in CTS5-Low, -Intermediate and -High, respectively). In contrast, CTS5-High patients did not show any benefit (-4.4%) in the BCI (H/I)-Low group. Conclusions: These results demonstrate that CTS5 does not provide predictive information to support extended endocrine therapy decision-making. Only BCI (H/I) was a predictive biomarker of benefit from extended endocrine therapy. This study further highlights the clinical utility of BCI as an endocrine response biomarker and emphasizes that prognostic information does not equate to predictive information in guiding duration of endocrine therapy. Clinical trial information: NTR3077; BOOG 2006-05; Eudra-CT 2006-003958-16.