It has been shown that the combination of irinotecan with fluorouracil is superior to fluorouracil alone in patients with advanced colorectal cancer as first-line therapy. In a recent randomised trial (Douillard et al., Lancet, 2000), median survival was 17.4 months for patients treated with irinotecan combined with fluorouracil as compared to 14.1 months in patients who received fluorouracil alone. OBJECTIVE: The objective of the study is to relate this statistically significant difference in overall survival to differences in costs of each alternative. METHODS: Medical care consumption data were collected prospectively as part of the trial that included 387 patients. A further retrospective data collection was designed to inform on further chemotherapy and disease cost after study treatment failure. The perspective for the calculation of cost was the French National Health System. Drug costs were assumed to be equal to public prices. Hospital costs were derived from French public DRG database. Consultation costs were abstracted from the public tariff database. Censoring of medical care consumption data was handled using the method described by Lin et al. (Biometrics, 1997). Uncertainty will be explored through bootstrap analyses. RESULTS: The total costs including drug cost, cost of treatment administration, cost of the management of toxicities and cost of disease progression amounted to 182,000 FF (USD 26,500) per patient treated with irinotecan and 123,000 FF (USD 18,000) per patient treated with fluorouracil alone. When the difference in cost is related to the clinical benefit of irinotecan, the cost per life year saved amounted to 214,000 FF (USD 31,100). Uncertainly analyses are ongoing and results will be presented. CONCLUSION: First-line therapy with irinotecan extends significantly survival, at a cost that remains within the limits currently accepted for new chemotherapeutic agents. Robustness of this conclusion to uncertainty is currently under investigation.