Imatinib for the treatment of patients with unresectable and/or metastatic gastrointestinal stromal tumours: systematic review and economic evaluation.

OBJECTIVES

To assess the clinical and cost-effectiveness of imatinib in the treatment of unresectable and/or metastatic, KIT-positive, gastrointestinal stromal tumours (GISTs), relative to current standard treatments.

DATA SOURCES

Electronic databases.

REVIEW METHODS

As there were no randomised trials that have directly compared imatinib with the current standard treatment in patients with advanced GIST, this review included non-randomised controlled studies, cohort studies, and case series that reported effectiveness results of treatment with imatinib and/or other interventions in patients with advanced GIST. The effectiveness assessment was based on the comparison of results from imatinib trials and results from studies of historical control patients. Economic evaluation was mainly based on an assessment and modification (when judged necessary) of a model submitted by Novartis.

RESULTS

Evidence from published uncontrolled trials involving 187 patients, and from abstracts reporting similar uncontrolled trials involving 1700 patients, indicates that approximately 50% of imatinib-treated individuals with advanced GIST experience a dramatic clinical response in terms of at least a 50% reduction in tumour mass. At present, although useful data are accumulating, it is not possible to predict which patients may respond in this way. Fifteen studies where possible GIST patients had been treated with therapies other than imatinib or best supportive care were also identified. All imatinib-treated patients experienced adverse effects, although they were relatively mild. Overall, imatinib was reported to be well tolerated. The most common serious events included unspecified haemorrhage and neutropenia. Skin rash, oedema and periorbital oedema were the common adverse events observed. Patients on the highest dose regimen (1000 mg per day in one trial) may experience dose-limiting drug toxicity. A structured assessment was carried out of the Novartis economic evaluation of imatinib for unresectable and/or metastatic GIST. The model was clearly presented and well written, its structure and input data were transparent, and the level of simplification was reasonable in terms of the objectives and data availability. However, the original Novartis model overestimated the cost-effectiveness of imatinib because of disproportion of survival and time-to-treatment failure in the imatinib arm, and the use of a possibly biased survival curve for patients in the control arm. The original Novartis model was modified to correct these two important shortcomings, which made it less sensitive to the choice of the survival curve for the control patients. According to the modified Novartis model, the estimated cost per quality-adjusted life-year (QALY) was 85,224 UK pounds (range 51,515--98,889 UK pounds) after 2 years, 41,219 UK pounds (27,331--44,236 UK pounds) after 5 years and 29,789 UK pounds (21,404--33,976 UK pounds) after 10 years. The results from a new Birmingham model were also within the range of estimates from the modified Novartis model.

CONCLUSIONS

Evidence from uncontrolled studies indicates that the treatment with imatinib brings about clinically significant shrinkage of tumour mass in about half of patients with unresectable and/or metastatic, KIT-positive GIST. Results of modelling based on data from uncontrolled studies suggest that imatinib treatment improves survival in patients with unresectable and/or metastatic GIST. The economic evaluation modelling suggests that the cost per QALY gained ranges from 51,515 to 98,889 UK pounds after 2 years, from 27,331 to 44,236 UK pounds after 5 years, and from 21,404 to 33,976 UK pounds after 10 years. Further research is needed into quality of life within trials involving patients with advanced malignancy, and long-term follow-up of adverse events is needed. Subgroup analysis of which, if any, patient types have a better or worse response to imatinib is also required. Analysis of individual patient data may be a good way of exploring these issues. There are many uncertainties surrounding imatinib prescription, such as the length of time patients should be on imatinib, the dose, drug resistance and the optimum time-point in the disease course at which to give the drug. Secondary research such as an update of this systematic review and a reassessment of the model is highly recommended when ongoing trials reach completion.