An RT-PCR-based multi-gene prognostic signature predicts distant metastasis of node negative, ER positive breast cancer from FFPE sections

Abstract

Listen

Translate article

506 Background: Gene expression profiles have been shown to predict distant metastasis risk in breast cancer patients. For routine medical practice, molecular prognostic tests need to be able to quantify metastasis risk with readily available samples and optimally employ regulatory-approved platforms. Methods: 142 early stage, N-, ER+, untreated breast cancer patients were chosen. Mean age was 61.8 (31 - 89) yrs with 72% > 55 yrs. RNA was isolated from 4×10 μm FFPE sections. Using real-time RT-PCR, 200 genes selected from van ‘t Veer et al, Dai et al, Paik et al and additional candidate genes were profiled. Primary endpoint was distant metastasis free survival. A prognostic signature was identified using semi-supervised principal component method. Metastasis score (MS) was calculated from gene expression levels. Results: We identified a unique 15-gene prognostic signature and 3 normalization genes. Probability of metastasis for any time period can be calculated from MS and was evaluated for 2, 5, and 10 yrs. Using median of MS to stratify patients, the high-risk group had a hazard ratio (HR) of 4.1 (95%CI 1.77–9.54, p = 0.001) vs. the low-risk group. 5-yr and 10-yr metastasis free survival rates (std. error) were 0.76 (0.053) and 0.62 (0.064) in the high-risk group and 0.96 (0.025) and 0.90 (0.038) in the low-risk group. Multivariate Cox regression analyses indicated that molecular signature had independent prognostic value with HR of 2.97 (95%CI 1.17 - 7.52, p = 0.022) after adjusting for age, tumor size and grade. Molecular signature was more prognostic in the low-grade subgroup (HR = 6.85, 95%CI 1.81–25.88) than an unstratified analysis. AUC at 5 yr was 0.65 and 0.60 for molecular signature and Adjuvant Online (AO) respectively. The prognosticator combining both had AUC of 0.67. 10-yr metastasis probabilities from MS and AO have low correlation (R2=0.17). Conclusions: A prognostic signature based upon mRNA expression of 15 genes has been identified using FFPE sections with RT-PCR for early stage, N-, ER+ patients. Metastasis score that quantifies distant metastasis risk for any time period, not confounded with any treatment effect, can help clinicians and patients to choose among different therapeutic options. (Lau and Wang contribute equally) [Table: see text]