Abstract P4-11-15: Risk stratification within luminal B breast cancer using a second generation prognostic RNA signature

Abstract

Abstract Background: Indication to adjuvant chemotherapy in early stage ER-positive breast cancer is usually based on clinico-pathological parameters, and has been recently improved by the introduction of prognostic expression profiles. Clinical or molecular features are efficient in identifying low-risk (pT1, histological grade 1, pN0, low RS, Luminal A subtype) and high-risk patients (>pT1, grade 3, N2, high RS). However, a large group of patients with intermediate clinical or molecular characteristic (grade 2, intermediate RS, luminal B) fall into the category between distinctly low and distinctly high risk and receive no treatment guidance from current decision tools. Methods: From a population of 1929 chemo-naïve, hormone treated, luminal B (Her-2 negative), pT1-pT3, pN0-N1a breast cancer patients diagnosed and treated at the European Institute of Oncology from 1997 to 2005, we selected a random subcohort of 555 cases, in a case-cohort design. All the patients with local or distant metastasis which were not already included were added to the subcohort, leading to a total of 704 patients (208 with local or distant recurrences and 496 random controls). Luminal B status was determined by the immunohistochemical analysis of ER, PgR, HER2 and Ki-67, according to the 2011 St. Gallen criteria. FFPE sections of the primary tumor were analyzed for the mRNA expression of 43 genes by multiplex quantitative PCR. A molecular score (MS) was calculated from the average expression of 23 cell cycle progression genes, the average expression of seven lymphocyte specific genes and the expression of PR and ABCC5, based on a model derived from an independent training cohort. A combined score of MS and the clinical variables of tumor size, grade and node status was modeled in the training cohort and applied to the Luminal B set. The association between MS and the risk of distant metastasis was evaluated in a weighted multivariable Cox regression model, adjusted for traditional clinical factors and Ki-67 labeling index (LI). Results: 640 samples, including 102 distant metastasis, had full clinical and expression data. In the 500 samples from the subcohort, median Ki67 LI was 21% (IQR=11%, Q1=16%, Q3=27%). Either one unit increase of Ki-67 LI (HR 1.06, 95%CI (1.04-1.08), p<0.0001) and of MS (HR 3.4, 95%CI (2.5-4.6), p<0.0001) were highly significant predictors of distant recurrence in univariable analysis. In multivariable analysis, the MS provided independent significant prognostic information after adjustment for Ki-67 LI, tumor size, grade and node status (HR 4.3, 95%CI (2.5-7.3), p=<0.0001). Using the combined score of MS and clinical variables, 383 patients or 77% of the subcohort had an estimated 10 year risk of distant recurrence of ≤10%. Similar results were obtained when samples were re-defined according to 2013 St. Gallen guidelines. Conclusions: The MS provides important prognostic discrimination beyond traditional clinico-pathological characteristics, including Ki-67 LI, in Luminal B breast cancer, and contributes in identifying a subset of patients which may be successfully treated with endocrine therapy only. Citation Format: Giancarlo Pruneri, Vincenzo Bagnardi, Davide Disalvatore, Giuseppe Curigliano, Nicole Rotmensz, Carmen Criscitiello, Darl D Flake II, Susanne Wagner, Alexander Gutin, Jerry Lanchbury, Massimo Barberis, Francesca Lombardi, Giuseppe Viale. Risk stratification within luminal B breast cancer using a second generation prognostic RNA signature [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr P4-11-15.