The aim of this study was to identify a prognostic immunohistochemical signature indicative of risk of early metastasis in node-negative breast carcinomas that would also be relevant to the development of new tailored therapy. Quantitative measurements of the immunohistochemical expression of 64 markers (selected from literature data) using high-throughput densitometry (as a continuous variable) of digitised microscopic micro-array images were correlated with clinical outcome in 667 node-negative breast carcinomas (mean follow-up 102 months). Multivariable fractional polynomials model of logistic regression allowed the selection of the best combination of markers (in terms of sensitivity and specificity) to predict patient outcome without any categorisation using predefined cut-points for individual marker measurements. A highly predictive ten-marker (out of 64) signature was identified comprising PI3K, pmTOR, pMAPKAPK-2, SHARP-2, P21, HIF-1α, Moesin, p4EBP-1, pAKT and P27 that well classified 91.4% of node-negative patients (specificity 90.9%, sensitivity 93.7%, area under ROC curve 0.958) independently of estrogen receptors (ER), and progesterone receptors (PR) and HER-2 status (91.6% well classified patients when ER, PR, HER-2 excluded). It is concluded that quantitative immunoprofiling of node-negative breast carcinomas is helpful in selecting patients who should not receive aggressive adjuvant chemotherapy and provides data for the development of tailored therapy.

•  Abstract
•  Download PDF