9555 Background: The European Medicine Agency (EMEA) stated that the degree of pre-treatments can modify the patient's tolerance to new treatments in pediatrics oncology, and then suggested that exploration of this potential difference should be considered. It is current practice to split a phase I trial in two groups to identify the Maximum Tolerated Dose (MTD) in each group. Yet how relevant such approach is, has not been investigated. In particular, role of baseline characteristics has not been explored. Methods: We reanalyze a large phase I trial of CPT11 in which 80 children had been included (32 heavily pre-treated pts and 48 less heavily pre-treated) to determine if baseline characteristics were associated with the heaviness of pre-treatment and could predict serious adverse events. To investigate robustness of conclusions in the context of small sample sizes, we performed an extended simulation study. We explored dose recommendations in scenarios with group difference, as measured by Odds Ratio (OR), ranging from 1 (absence of difference) to 8 (large difference) and sample size increasing from 2 * 20 to 2*40. Results: We did not find any relation between the degree and pre- treatment and neither the risk of DLT nor the baseline characteristics in this pediatric trial. However robustness appeared to be critical with high risk of misidentification of MTD in each of the two groups, whatever the prognostic value. With a group difference corresponding to OR=8 and balanced sample sizes (2*20 pts), the same MTD was identified in 10% of the simulations. Even with larger sample sizes (2*40pts), this figure reached 24% for OR=3. There is also a very high risk of identifying two different MTD although the risk is similar in the two groups (52% for 40 × 2 patients). Conclusions: Two independent sample designs in pediatric phase I trials should be avoided or reserved to limited situations when there is a strong rationale possibly based on adult data. No significant financial relationships to disclose.