PURPOSE: The evaluation of treatment efficacy with RECIST criteria does not take into account tumour growth dynamics. We notably investigated the impact of the pre-treatment tumour growth rate (GR) on the evaluation of treatment response. PATIENTS AND METHODS: Seventy-six patients included in phase I clinical trials had scanographic evaluations before and after starting an experimental treatment. The GR was calculated for the pre-treatment period and for the experimental period (i.e. during the new treatment). Tumour response was evaluated per protocol at week 12 and at week 24 of the experimental period according to RECIST criteria. We studied the relation between pre-treatment and experimental GRs and RECIST tumour response. RESULTS: On average the tumour GR was decreased by 40% during the experimental period; compared to the pretreatment period (p=0.03). An increased growth rate (acceleration of GR during experimental treatment compared to pretreatment) was observed in 20 (38%) of the 53 patients considered as non-progressive at week 12 according to RECIST. Conversely a decreased GR was observed in 12 out of 23 (53%) patients classified as progressive according to RECIST. The variation in the GR between the pre-treatment and experimental period was not significantly correlated with response evaluated according to RECIST at week 12 or at week 24 (p=0.45 and 0.44, respectively). CONCLUSIONS: RECIST evaluation of tumour response depends on the natural history of the tumours and poorly measures the impact of treatment on the kinetics of tumour growth. Integrating pre-treatment GR evaluations could substantially improve the assessment of treatment efficacy in drug development.