Personalising cancer treatment to optimise therapeutic efficacy while minimising exposure to the toxicities of ineffective drugs is the holy grail of medical oncology. Clinical parameters and conventional histopathological characterisations of cancers are no longer adequate to guide the practising oncologists in treatment planning. The explosion of knowledge in cancer molecular biology has led to the availability of tumour-specific molecules that serve as predictive and prognostic markers. In breast cancer, HER-2 positivity is a good predictor for success of anti-HER-2 trastuzumab monoclonal antibody therapy. K-ras mutational status predicts the likelihood of response to anti-EGFR monoclonal antibodies in advanced colorectal cancers. Similarly, EGFR mutational status in pulmonary adenocarcinoma is highly predictive for responses or otherwise to tyrosine kinase inhibitors. Notwithstanding our deeper understanding of tumour biology and the availability of predictive and prognostic laboratory tools, we are still far from achieving our dream of the perfect personalised cancer treatment, as each tumour in a particular patient is unique to itself. A much coveted, real-time, anti-tumour drug sensitivity testing in the future may one day pave the way for truly treating the right tumour with the right drug in the right patient.