Advances in our understanding of tumour biology have encouraged reassessment of tumour classification by the site of origin in favour of molecular characteristics and/or oncogenic drivers amenable to treatment. The identification of EML4-anaplastic lymphoma kinase (ALK) as an oncogenic driver in non-small cell lung cancer (NSCLC) early in the clinical development of crizotinib and the observation of promising clinical responses in patients with NSCLC harbouring ALK translocations accelerated its clinical development in ALK-positive NSCLC. Phase I and II trials of crizotinib in patients with ALK-positive advanced NSCLC reported notably high response rates that tended to be rapid and of prolonged duration. Crizotinib was well tolerated; treatment-related adverse events were typically gastrointestinal (grade 1/2) and visual disorders (almost exclusively grade 1). Crizotinib provided NSCLC symptom relief and maintained quality of life. Based on the phase I and II trial data, the US Food and Drug Administration granted approval of crizotinib in August 2011. The consistency of the crizotinib data to date suggests accurate selection of the target population for crizotinib treatment. The ability to molecularly select patients likely to respond to an investigational agent argues that future clinical development of targeted agents should be re-evaluated. Updated trial designs incorporating molecular testing, early use of enrichment biomarkers and intermediary endpoints may accelerate and optimise clinical evaluation of targeted agents. Such trial designs should allow rapid clinical evaluation, minimise exposure of patients to therapies unlikely to be of benefit and, potentially, allow accelerated drug approval in molecularly specified populations.