Human breast cancer was solely classified based on clinical and immunohistochemical (IHC) findings in the past. A growing body of evidence suggests that these categorisations are rendered more precisely by intrinsic subtyping with the aim of an introduction of personalised medicine. Especially in breast cancer with the uncertain potential of disease spread, such as T1-2, Grade 2 and oestrogen receptor-positive (ER+ve) tumours, the value of chemotherapy applied to every patient has been questioned and the need for additional information on the tumour´s specific risk of recurrence is overt. It is estimated that the average risk for recurrence is 15% at 10 years in hormone-receptor-positive breast cancer. Thus, a relatively small proportion of these patients would need chemotherapy, and the main task is to stratify which patients of this cohort are at high-risk and will benefit from cytotoxic agents. Ki67, as a proliferation marker classifying high-risk tumours, has been demonstrated as a continuous marker, but not as a clear cut risk-defining instrument in recent publications. Thus, the difficulties are perceived especially at the threshold of the low to high-risk area of this marker. Reproducibility of Ki67 is to some extent uncertain considering there is inter and intra-institutional variability of up to 30% of the results. Several multigene arrays, such as MammaPrint®, Oncotype DX®, Endopredict®, and PAM50 have demonstrated clinical utility and experienced validation. The aim of this review is the description of the implementation of genomic testing in international guidelines (North American and European), with regard to incorporation of multigene arrays into the decision-making process in different clinical settings (including tumour size and IHC status). Data cut-off was 1st October, 2013. It seems that North America and some European countries have initiated a shift towards a personalised medicine with multigene arrays based on RT-PCR or microarrays.