Development and validation of a model that predicts early death among cancer patients participating in phase I clinical trials investigating cytotoxics

Abstract

Objective: Selecting patients for phase 1 studies remains challenging. Given the lack of clear and reliable guidance for the estimation of life expectancy, we retrospectively assessed predictive factors of early death (within 90 days following inclusion) among these patients.

Methods: Two hundred fifty-seven consecutive cancer patients enrolled in phase I studies investigating cytotoxics at Oscar Lambret Cancer Center and Institut Claudius Regaud were included in the development database. Univariate and multivariate analyses (logistic regression model) were undertaken to determine the prognostic factors. A probability tree described the rate of early death in the different prognostic subgroups. This prognostic model was then evaluated on a second independent cohort of 128 patients treated at Léon Bérard Cancer Center.

Results: The median overall survival was 8.4 months in the dataset population, and the rate of early death was 15%. In multivariate analysis, the two prognostic factors for early death were albumin <38 g/l (OR = 5.21) and lymphocytes <700/mm(3) (OR = 3.88). According to these two parameters, three prognostic subgroups were defined with early death rates of, respectively, 8/121 (6%), 19/119 (16%) and 13/17 (76%). In the validation dataset, the rates of early death according to three prognostic groups were 13/68 (19%), 20/57 (35%) and 3/3 (100%), respectively.

Conclusion: We do not recommend the enrolment of patients with albumin level below 38g/l and lymphocytes count below 700/mm(3), in phase 1 trial investigating cytotoxics. Our model is helpful to discriminate "patients with reasonable life expectancy" as defined in most phase 1 protocols.