Retrospective analysis of data from a phase I trial with a new adaptive method

Abstract

1067 Background: The phase I trial CAPEV CGFL 01 [ESMO, Vienne, Austria: 2004.V 52–53] aimed to find the Maximum Tolerated Dose (MTD) of i.v. Vinorelbine when associated with oral Capecitabine for three cycles of 21 days in metastatic breast cancer (MBC). Method: Vinorelbine was administered at d1 and d15 and Capecitabine b.i.d. from d1 to d14 (fixed at 2,000 mg/m2/day). Trial design was based on “3+3” scheme with predetermined dose levels of Vinorelbine starting at 25 mg/m2/day. Ten patients (age 46–67 years) with MBC were enrolled in this trial. A new phase I adaptive method [Statistical Methods in Biopharmacy. Statistical Innovations in Clinical trials, Paris September 2005] under validation was used to analyze data available from the CAPEV trial. This method performs precise definition of Dose Limiting Toxicity (DLT) and modeling of all the grades of toxicity. The method allows sequential real time analysis of data by alternating estimation and design steps. Results: Most frequent toxicities were hematological (anemia and neutropenia) and cytolysis. Three DLTs occurred for the first 6 patients included at the starting level of Vinorelbine. De-escalation to 20 mg/m2 was done at the 7th patient but a new DLT occurred for the 7th patient and lead to trial ending at the 10th patient. Firstly, by analyzing all available data, Vinorelbine MTD was calculated at 20.2±4.04 mg/m2/day. Secondly, data was processed according to the inclusion time of patients; the adaptive method calculated a Vinorelbine MTD equal to 19.5±8.34 mg/m2/day as early as the 4th patient. Conclusion: The administration scheme of i.v. Vinorelbine and oral Capecitabine tried in this study cannot be retained for MBC treatment. However, early convergence of adaptive method clearly shows its high performance in the de-escalation scheme and Vinorelbine MTD determination. With this method Vinorelbine MTD could have been determined with the 4th patient data without needing the inclusion of 6 additional patients (4th and not 7th patient). This method should be used in prospective phase I trials and in parallel with classical and clinical method of MTD determination. No significant financial relationships to disclose.