OBJECTIVES Risk stratification based on results provided by a 21-gene assay (Oncotype DX ®) in early stage breast cancer can help to optimize hormone therapy (HT) and chemotherapy (CT) decisions, by stratifying women according to recurrence risk. Previous reviews addressing the issue were outdated or did not provide data on both overall and net decision impact regarding use of CT in this group of women. We aimed to perform a systematic review and meta-analysis on global decision impact and net change in CT use by risk category before and after assay results. METHODS We updated a broader systematic review (Carlson, search  up to 03/12). Inclusion criteria were further restricted to studies with prospective data collection. Global decision impact was estimated as the proportion of patients whose management was changed after assay results were available. Net change was estimated as the difference between the proportion of all patients who received chemotherapy before and after the test.  RESULTS Ten studies (N=1218) met our inclusion criteria; 49.75, 38.18 and 12.07% of patients were low, intermediate and high risk. Due to significant heterogeneity among studies, a random effects model is presented. Treatment decision was changed in 30.83% (95% CI 26.75 to 35.07) of all patients. From all patients, net change of CT use was 16% (95%CI 12-20) in the low risk group (CT+HT to HT); 0% (-4 to 3) in the intermediate risk; and -2% (-3 to -1) in the high risk group, in this latter case increasing CT use (HT to CT+HT). CONCLUSIONS Using a 21-gene assay showed a significant impact on treatment decisions (from 100 women, 31 will have their treatment optimized, and 14 less net CT overall). Its main effects consist of sparing chemotherapy in low risk patients, and increasing it in the high risk category.