2568 Background: For investigators, the selection of patients to be considered for phase I clinical trials is difficult, because of the lack of objective criteria for a rational decision-making process. From October 1997 to October 2002, we retrospectively assessed prognostic factors for cancer patients considered for Phase 1 trials. Methods: 148 consecutive patients who had been screened for inclusion in 6 different phase I trials were included in the present study. 70 out of them actually received the phase I treatment. Univariate (Log-Rank test) and multivariate analysis (Cox proportional hazard ratio model) were performed to determine the prognostic factors related to overall survival (OS) after screening. Results: The study comprised 63 men and 85 women, with a median age of 54 (range 23–79). The most frequent primary cancer sites were: breast (38 cases), head and neck (28 cases), lung (18 cases) and colorectal (17 cases). 91 out of them had a performance status PS = 0. The median OS of the 148 patients was 5.7 months (173 days, range 1–2,421). Univariate analysis identified PS = 1, Body Mass Index < 20, liver and visceral metastasis, serum albumin < 38 g/L, lymphocytes count < 0.7 x 109/L and granulocytes count > 7.5 x 109/L as poor prognostic factors. The Cox model identified serum albumin < 38 g/L (HR 2.51 [1.51–4.18], p=0.0001) and lymphocyte count < 0.7 x 109/L (HR 2.27 [1.13–4.62], p=0.024) as independent prognostic variables for OS. All patients presenting with both prognostic factors died within 90 days. Conclusion: We propose a simple model, easily obtained at the patient bedside, which can discriminate patients who have a life expectancy of over 3 months and thus could be enrolled in phase-I anti-cancer trials. No significant financial relationships to disclose.