This paper examines the challenge of conducting economic evaluations to support patient access to cancer therapies when the cost-effectiveness estimation is hampered by crossover trial design. To demonstrate these limitations, we present the submission to the Canadian Drug Review (CDR) of a cost-effectiveness evaluation of sunitinib versus best supportive care (BSC) for the treatment of gastrointestinal stromal tumour in patients intolerant or resistant to imatinib. The economic model generated an incremental cost-effectiveness ratio for sunitinib versus BSC of dollars 79,884/quality-adjusted life-year gained. Eight months after initial submission, CDR granted a final recommendation to fund sunitinib following the manufacturer's appeal against their first recommendation. Although cost-effectiveness is an important consideration in reimbursement decisions, there is a need for improved decision-making processes for cancer drugs, as well as a better understanding of the limitations of clinical trial design.