Molecular breast cancer prognosis in node negative early breast cancer.

Le résumé fourni par la source

Abstract Abstract #1071 Background: A number of molecular signatures have been published to aid breast cancer prognosis and therapy response prediction. However, most signatures rely on measurements of dozens of genes on an array platform. Here we report the identification and validation of a prognostic algorithm based on a small number of expressed genes reflecting dominant prognostic motifs like steroid hormone receptor expression, proliferation, matrix degradation and B-cell infiltration. The test is designed for mRNA extracted from formalin fixed paraffin embedded tumour tissue (FFPE). Materials and Methods: 200 fresh frozen tumours from node negative patients, who did not receive systemic therapy after surgery, were profiled on HG-U133a arrays. Primer and probes were designed for genes identified as prognostic relevant and used to profile the corresponding paired FFPE tumour tissue of our identification cohort by kinetic PCR. A genomic algorithm was constructed and tested in FFPE tumor tissue of an independent validation cohort of 240 node negative patients that remained untreated in the adjuvant setting. Results: Based on the Prognostic-score of our algorithm we classified the patients of the validation cohort into three classes: low risk (25% of patients), intermediate risk (25%) and high risk (50%). The probability to die after having a recurrence within 10 years was 4.7% for the low risk group, 11.9% for the intermediate risk group and 33.1% for the high risk group. Kaplan Meier analysis revealed that the trend for survival (Logrank test for trend P = 0.0002) was highly significant. Adjuvant! Online calculated probabilities to die of cancer were also used to classify patients into three risk groups of comparable size (25/25/50 percent of patients). The observed risk was 9.1% in the low risk, 26.7% in the intermediate and 23.6% in the high risk group. The area under the ROC curve for Adjuvant! Online was 0.619 (95% CI, 0.52 to 0.718; P = 0.031) while for our algorithm it was 0.755 (95% CI, 0.6628 to 0.8473; P < 0.0001). Discussion: We identified a prognostic algorithm for node negative breast cancer based on kinetic PCR analysis of FFPE tumor tissue. Validation in an independent cohort showed the ability of the algorithm to classify patients into clinically relevant risk classes. Preliminary analysis shows a better prognostic performance than current clinical guidelines. Citation Information: Cancer Res 2009;69(2 Suppl):Abstract nr 1071.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.