| Validation of a prognostic algorithm based upon a five monoclonal antibody immunohistochemistry test in Tamoxifen-treated, node negative breast cancer: NSABP B14 and B20 studies.
Poster at San Antonio Breast Cancer Symposium 2006
Ross DT1, Kim C2, Tang G2, Mejia OM2, Ring BZ1, Seitz RS1, and Wolmark N.2
1Applied Genomics Inc., Sunnyvale CA, USA; 2National Surgical Adjuvant Breast and Bowel Project, Pittsburgh PA;
Abstract
Background: A strong association between disease progression and a novel five antibody immunohistochemistry (IHC) test for ER+ breast cancer was previously demonstrated in three independent institutional cohorts (Ring et al., JCO, July 2006, Ring et al, ASCO 2006). As an additional trial of this IHC test we performed a prospectively designed blinded retrospective study using tissue arrays constructed using available paraffin blocks from the combined Tamoxifen-treated arms of the NSABP B14 and B20 ER+ N- trials.
Subjects and Methods: Tissue arrays were constructed in triplicate and stained using five monoclonal antibodies targeting p53, NDRG1, SLC7A5, CEACAM5 and HTF9C using standard IHC protocols. Pre-defined scoring rules, protocols for deriving consensus scores from triplicate stains, algorithm for combining scores, and cut-points for low, moderate and high risk patient strata were applied to the 711 scored patients (550-B14 and 161-B20).
Results: Three of five markers included in the algorithm were significantly associated with recurrence free interval in univariate Cox models. In a univariate Cox model, the distribution of recurrence events between low, moderate and high risk strata for the ten year study period were significantly distinct (HR=1.304, 95%:CI 1.08-1.573, p=0.006). In a multivariate model the test contributed information independent of age, tumor size and menopausal status (p=0.007). The Kaplan-Meier (KM) estimates of the Recurrence-free-interval (RFI) were 72.6% for the high risk group, 85.6% for the moderate risk group, and 84.6% for the low risk group (p=0.001). The KM estimates of Breast- cancer-specific-death were 22.9 % for the high risk, 9.9% for the moderate risk and 9.1% for the low risk group (P<0.0001). Exploratory analysis revealed that the test may have the greatest clinical relevance in patients >60 years old with KM estimates of RFI of 78.5% for the high risk group, 89% for the moderate risk group and 92% for the low risk group.
Conclusion: IHC staining patterns using five selected monoclonal antibodies combined using a pre-defined algorithm assign patients to risk strata significantly associated with outcome in ER+ N- populations. Exploratory analysis suggests that this test will be most useful in clinical decision making for post-menopausal patients.
|
