Randomized trials comparing the aromatase inhibitors, anastrozole and letrozole, to megestrol acetate (MA) in postmenopausal women with advanced breast cancer demonstrated that both agents are better tolerated than MA with comparable efficacy. In addition, one trial revealed that tumor response and time to treatment failure were significantly better with letrozole. Since oncologists are faced with a choice between three agents with at least comparable efficacy but different toxicity profiles and cost, a cost-utility analysis was conducted to quantify these differences and to determine if the new agents are more cost-effective than MA. In the absence of a randomized three-arm trial, a decision model was developed to simulate the most common therapeutic outcomes. The clinical data were obtained from an overview analysis of randomized trials. Total hospital resource consumption was collected from 87 patients with advanced disease that had failed second-line hormonal therapy. Utility estimates were obtained from interviewing a random sample of 25 women from the general public and 25 female health care professionals using the Time Trade-Off technique. The model suggested a similar duration of quality-adjusted progression-free survival between drugs (letrozole 150 days, anastrozole 153 days and MA 146 days). Letrozole had an overall cost of Can$2949 per patient which was comparable to MA at Can$2966 per patient. In contrast, anastrozole was slightly more costly than MA at $Can3149 per patient, respectively. The analysis revealed that letrozole has comparable overall costs relative to MA while providing at least equivalent quality-adjusted progression-free survival. These outcomes were largely related to its higher tumor response rate, which translated to a lower proportion of patients requiring chemotherapy. Anastrozole was slightly more costly than MA and did not demonstrate superiority in quality-adjusted progression-free survival in this palliative setting.