Background: Imatinib dose escalation is advocated for gastrointestinal stromal tumour (GIST) treatment, but its effectiveness compared with sunitinib and best supportive care (BSC) after failure at the 400 mg/day dose is unknown. Objectives: To assess the effectiveness and cost-effectiveness of imatinib at escalated doses of 600 or 800 mg/day for patients with unresectable and/or metastatic GISTs whose disease had progressed on 400 mg/day. Data sources: Electronic databases, including MEDLlNE, MEDLINE In-Process, EMBASE, BIOSIS, Science Citation Index, Health Management Information Consortium and the Cochrane Controlled Trials Register, were searched until September 2009. Review methods: A systematic review of the literature was carried out according to standard methods. An economic model was constructed to assess the costeffectiveness of seven alternative pathways for treating patients with unresectable and/or metastatic GISTs. Results: Five primary studies involving 669 people were included for clinical effectiveness; four reported imatinib and one reported sunitinib. The data were essentially observational as none of the studies was designed to specifically assess treatment of patients whose disease had progressed on 400 mg/day imatinib. For 600 mg/day imatinib, between 26% and 42% of patients showed either a partial response (PR) or stable disease (SD). Median time to progression was 1.7 months (range 0.7-24.9 months). For 800 mg/day imatinib, between 29% and 33% of patients showed either a PR or SD. Median overall survival (OS) was 19 months [95% confidence interval (CI) 13 to 23 months]. Progression-free survival ranged from 81 days to 5 months (95% CI 2 to 10 months). Median duration of response was 153 days (range 37-574 days). Treatment progression led to 88% discontinuations but between 16% and 31% of patients required a dose reduction, and 23% required a dose delay. There was a statistically significant increase in the severity of fatigue (p < 0.001) and anaemia (p = 0.015) following dose escalation. For sunitinib, median OS was 90 weeks (95% CI 73 to 106 weeks). For the cost-effectiveness review, only one full-text study and one abstract were identified, comparing imatinib at an escalated dose, sunitinib and BSC, although neither was based on a UK context. The definition of BSC was not consistent across the studies, and the pattern of resources (including drugs for treatment) and measures of effectiveness also varied. Within the model, BSC (assumed to include continuing medication to prevent tumour flare) was the least costly and least effective. It would be the care pathway most likely to be cost-effective when the cost per quality-adjusted life-year threshold was < £25,000. Imatinib at 600 mg/day was most likely to be cost-effective at a threshold between £25,000 and £45,000. Imatinib at 600 mg/day followed by further escalation followed by sunitinib was most likely to be cost-effective at a threshold > £45,000. Limitations: The evidence base was sparse, data were non-randomised and potentially biased. The economic model results are surrounded by a considerable degree of uncertainty and open to biases of unknown magnitude and direction. Conclusions: Around one-third of patients with unresectable and/or metastatic GIST, who fail on 400 mg/day of imatinib, may show response or SD with escalated doses. Between a threshold of £25,000 and £45,000, provision of an escalated dose of imatinib would be most likely to be cost-effective. However, these results should be interpreted with caution owing to the limited evidence available on outcomes following imatinib dose escalation or sunitinib for this group of patients. Funding: The National Institute for Health Research Health Technology Assessment programme.