BACKGROUND: There are five multi-gene expression based prognostic tests for breast cancer offered as reference lab tests - Mammaprint, MapQuant Dx, OncotypeDx, PAM50 Breast Cancer Intrinsic Subtype Classifier, and Theros Breast Cancer Index. Each claims to provide additional prognostic information beyond conventional prognostic markers and to aid in determining who should receive systemic therapy. Evidence for their clinical utility was reviewed to determine whether any of them should be considered as routine clinical test. METHODS: Peer reviewed publications, meeting abstracts, and information provided by company web sites have been reviewed to compile information regarding their clinical utility according to the following criteria; (1) Analytical validity and regulatory approval of the reference lab test. (2) Level of evidence for clinical utility. (3) Whether published evidences support prognostic and/or predictive claim. RESULTS: While published evidences for clinical claims for OncotypeDx and Mammaprint used reference lab tests, and the supporting evidences for other tests come from academic assays before being converted to reference lab tests, results from two large randomized clinical trials testing the clinical utility of OncotypeDx and Mammaprint are still several years away and until that time none of the markers would reach level I evidence by Marker Utility Grading System. However Oncotype Dx has reached a level IB evidence according to Simon modification to Marker Utility Grading System. Therefore OncotypeDx may be considered for routine clinical use as an adjunct to clinical and pathological information and has been incorporated into clinical guidelines in USA. While Mammaprint, MapQuantDx, and PAM50 have been repeatedly demonstrated to provide robust prognostic information, evidence for its worth as a predictive marker for chemotherapy benefit is yet to come from randomize clinical trials and therefore its utility is limited to prognostication. Meta-analysis of publicly available microarray based gene expression studies demonstrated that gene expression assays provide similar information and the most important information they provide is the proliferation activity. In untreated population, the prognostic impact of proliferation genes is limited to ER+HER2- subset since HER2+ or ER-HER2- subsets are associated with high proliferation activity. Therefore the clinical utility of these gene expression based tests is mainly for ER+HER2- subset. Since they are usually treated with adjuvant anti-estrogen therapies, for their clinical utility, demonstration of the interaction between the gene expression markers and chemotherapy in anti-estrogen treated cohort in a randomized clinical trial would be required. While OncoytpeDx is the only test supported by studies in a randomized clinical trial for adjuvant chemotherapy, other gene expression based tests are expected to provide similar information. Gene expression profiling assays as more reproducible and precise surrogates for tumor grade (MapQauntDx and Theros Breast Cancer Index) are very promising assays. However, absence of definitive predefined cut-off for defining the subset that benefit from chemotherapy validated in cohorts from randomized trials limit their clinical application.