P163 - Meta-analysis of phase I dose-finding studies: application for protein kinase inhibitors developed in oncology - 20/04/23
Bienvenue sur EM-consulte, la référence des professionnels de santé.
Article gratuit.
Connectez-vous pour en bénéficier!
Résumé |
Introduction |
Recommended phase 2 doses (RP2D) are determined following dose-finding studies that include limited number of patients, resulting in a potential lack of precision. Phase I meta-analysis methods have been developed to improve RP2D selection; this study aimed to assess the feasibility of applying such meta-analysis to protein kinase inhibitors (PKI) developed in oncology.
Methods |
As part of a large meta-research project, this ancillary study required data from a systematic review performed to identify PKI and dose-finding studies. PKI selected for meta-analyses had to have at least 5 dose-finding studies (i) completed, (ii) with available results, (iii) including cancer patients, and (iv) RP2D defined based on toxicity assessment. To account for heterogeneity caused by multiple schedules and/or conditions of administration some studies were split into sub-studies. For each PKI, a Bayesian random-effects meta-analysis [1] was conducted to estimate the toxicity probability of recommended doses (marketed dose, or in the absence, RP2D).
Results |
Among the 402 PKI identified, 21 met eligibility criterion with a median number of phases I of 6 [Min-Max: 5 – 16]. After selecting similar sub-studies and removing missing data (4 sub-studies and 41 dose levels), meta-analyses were performed for 20 PKI including 115 sub-studies. Twenty-four (83%) of medians of toxicity probabilities estimated for 29 recommended doses were below the most commonly used minimum toxicity threshold of 0.20 and three (10%) medians were above the maximum toxicity threshold of 0.33. The 95% credibility intervals are rather wide.
Conclusion |
Meta-analysis can be applied to estimate toxicity probabilities of a dose evaluated in multiple trials. For PKI, toxicity probabilities of recommended doses are generally below the minimum toxicity threshold. Uncertainty around these probabilities, reflected by 95% credibility intervals, may derive from a high between-trial heterogeneity and may also be attributed to the missing data related to reporting quality. Reference 1. Ursino M. et al. Ann Appl Stat. 2021;15
Mots clés |
Meta-analysis , Random-effects , Recommended dose , Phase I clinical trials , Protein kinase inhibitors
Déclaration de liens d'intérêts |
Les auteurs n'ont pas précisé leurs éventuels liens d'intérêts.
Le texte complet de cet article est disponible en PDF.
Vol 71 - N° S2
Article 101816- mai 2023 Retour au numéro
Article précédent
- P162 - La consommation de substances psychoactives chez les jeunes NEET (« Not in Employment, Education or Training ») et ses facteurs de risques associés: revue systématique de la littérature
- C. Eyraud, C. Collin, C. Alberti, A. Dumas, P. Martin, E. Le Roux
- P164 - Quality assessment of randomized controlled clinical trials of pharmacological treatments in patients with rare diseases
- L. Paraz, B. Kassai
Bienvenue sur EM-consulte, la référence des professionnels de santé.