2431

Purpose: More accurate prognosis of disease outcome with primary invasive breast cancer would support better clinical management of this disease. OncoPlan uses immunohistochemistry (IHC) to measure Shc proteins, which play key roles in aggressive cancers. Here, we sought to validate our earlier study (Davol, Can Res 2003), in which levels of PY-Shc and p66 Shc were strong, independent predictors of tumor aggressiveness.

Methods: Tumor specimens in TMA format were available from 845 women with primary invasive breast cancer (201 distant relapses and 167 disease-specific deaths (DSD), avg follow-up of 11.7 yrs), who received no systemic adjuvant therapies prior to first relapse. IHC staining of the Shc proteins was independently scored on a 0-5 scale by two pathologists, blinded to patient information. Staining intensities were related to relapse-free survival (RFS) and DSD by univariate and multivariate analyses.

Results: By univariate Cox proportional hazards analysis, p66 Shc levels as a continuous variable strongly protected against distant recurrence (HR=0.34, 0.19-0.61 95%CI, P=0.0002) and DSD (HR=0.29, 0.15-0.56 95%CI, P=0.0001). In univariate log-rank analyses, poor RFS and disease-specific survival (DSS) were observed for patients with low p66 Shc scores (P=0.002 and P=0.007, respectively), high PY-Shc scores that trended towards significance (P=0.14 and P=0.09, respectively), and for a p66 Shc_PY-Shc interaction variable (P=0.0005 and P=0.0001, respectively). By multivariate Cox proportional hazards modeling (in which p66 Shc and PY-Shc were adjusted for nodal, estrogen receptor (ER) and Her2 status, tumor size and grade, and age), p66 Shc and PY-Shc were very strong predictors of both relapse (HR=4.0, 1.8-9.4 95%CI, P=0.0005), and DSD (HR=6.2, 2.4-16 95%CI, P=0.0001). Among patients identified as low risk by p66 Shc and PY-Shc scores, only 3% relapsed, and none died from their disease. Conversely, among patients identified as high risk by p66 Shc and PY-Shc, 30% relapsed and 27% died of their disease. In node negative, ER-positive patients, and considering Her2 status, the prognostic ability of the Shc proteins (DSS HR=13, 3.2-52 95%CI, P=0.0002) compared favorably to that reported for the 21-gene assay (DSS RR=2.4, 1.1-5.2 95%CI, P=0.025). Interaction analyses provided no evidence of association of the Shc predictors with other covariates.

Conclusions: OncoPlan had a very strong ability to predict disease outcome in patients who did not receive systemic adjuvant therapy. This ability was independent of traditional clinical indicators of risk. The ability of the Shc proteins to stratify patients over a very broad range of risk should be a valuable tool in the clinical management of invasive breast cancer.

This was supported in part by the Susan G. Komen Breast Cancer Foundation

98th AACR Annual Meeting-- Apr 14-18, 2007; Los Angeles, CA