Greater awareness and use of screening has led to an increase in the proportion of ‘early’ breast cancers. Conventional prognostic factors such as tumour size and nodal status are of limited use with these tumours, because most of them are node negative and small. There is a need for factors that are not only prognostic, but also predict response to therapy, in the selection of appropriate therapy for these low-risk patients, especially to avoid the administration of toxic therapies to patients who are unlikely to gain significant benefit. This has led to the emergence of newer molecular prognostic factors, gene-expression signatures being the latest. Although not many of these newer prognostic factors have proved clinical utility, recent studies report remarkable results with the use of gene-expression signatures as prognostic and predictive factors. These results, though promising, have been compared only with simple parameters such as tumour grade and with broad practice guidelines. Additional studies documenting superiority of gene signatures over existing prognostic algorithms such as the Nottingham prognostic index and Adjuvant! Online are necessary before their widespread routine use. Currently, gene signatures are best used as a part of randomized clinical trials such as MINDACT and TAILORx. In this review, we discuss the biological basis, scientific evidence and clinical application of conventional and molecular prognostic factors, including gene-expression signatures.