Validation of an immunohistochemical signature predictive of 8-year outcome for patients with breast carcinoma.

We recently reported that standardized quantitative immunohistochemical (IHC) assays allowed prediction of an adverse outcome among 572 node negative (N-) patients with breast carcinoma (BrCa). To further validate our prior findings, we repeated the IHC stains including a second series of BrCa diagnosed at Yale University. Tissue microarrays (TMAs) of two cohorts of patients with BrCa (418 Marseille University and 303 Yale University) were respectively investigated for IHC expression of 15 markers (HIF-1α, PI3K, pAKT, pmTOR, moesin, P21, 4(E) BP-1, P27, Ker5-6, pMAPKAPK-2, SHARP2, claudin-1, ALDH, AF6 and CD24). Quantitative measurements of immunoprecipitates densitometry assessed with an image analyzer were correlated with 8-year patients' outcome and compared in the two cohorts. The best predictive signature consisted of a combination of five markers that included HIF-1α, PI3K, claudin-1, AF6 and pAKT in N- BrCa. This combination permitted an accurate prediction of outcome in 92.34% (386/418) of N- patients in the first set (Marseille) and 89.8% (158/176) in the second set (Yale). The close results in both cohorts confirmed the validity of this original IHC signature predictive of prognosis in node negative BrCa. This validation suggests that in clinical practice, it would be possible with standardized kits (i) to identify patients with poor prognosis at diagnosis time, particularly in the N- BrCa subset, who would require more aggressive adjuvant therapy and (ii) to avoid useless expensive therapies and their side effects in N- patients with favorable prognosis.