TY - JOUR
T1 - The performance of model-based versus rule-based phase I clinical trials in oncology
T2 - A quantitative comparison of the performance of model-based versus rule-based phase I trials with molecularly targeted anticancer drugs over the last 2 years
AU - van Brummelen, E. M.J.
AU - Huitema, A. D.R.
AU - van Werkhoven, E.
AU - Beijnen, J. H.
AU - Schellens, J. H.M.
N1 - Publisher Copyright:
© 2016, Springer Science+Business Media New York.
PY - 2016/6/1
Y1 - 2016/6/1
N2 - Phase I studies with anticancer drugs are used to evaluate safety and tolerability and to choose a recommended phase II dose (RP2D). Traditionally, phase I trial designs are rule-based, but for several years there is a trend towards model-based designs. Simulations have shown that model-based designs perform better, faster and are safer to establish the RP2D than rule-based designs. However, the superiority of model-based designs has never been confirmed based on true trial performance in practice. To aid evidence-based decisions for designing phase I trials, we compared publications of model-based and rule-based phase I trials in oncology. We reviewed 172 trials that have been published in the last 2 years and assessed the following operating characteristics: efficiency (trial duration, population size, dose-levels), patient safety (dose-limiting toxicities (DLTs)) and treatment optimality (percentage of patients treated below and at or above the recommended phase 2 dose). Our results showed a non-significant but clinically relevant difference in trial duration. Model-based trials needed 10 months less than rule-based trials (26 versus 36 months; p = 0.25). Additionally, fewer patients were treated at dose-levels below the RP2D (31 % versus 40 %; p = 0.73) while safety was preserved (13 % DLTs versus 14 % DLTs). In this review, we provide evidence to encourage the use of model-based designs for future phase I studies, based on a median of 10 months of time gain, acceptable toxicity rates and minimization of suboptimal treatment.
AB - Phase I studies with anticancer drugs are used to evaluate safety and tolerability and to choose a recommended phase II dose (RP2D). Traditionally, phase I trial designs are rule-based, but for several years there is a trend towards model-based designs. Simulations have shown that model-based designs perform better, faster and are safer to establish the RP2D than rule-based designs. However, the superiority of model-based designs has never been confirmed based on true trial performance in practice. To aid evidence-based decisions for designing phase I trials, we compared publications of model-based and rule-based phase I trials in oncology. We reviewed 172 trials that have been published in the last 2 years and assessed the following operating characteristics: efficiency (trial duration, population size, dose-levels), patient safety (dose-limiting toxicities (DLTs)) and treatment optimality (percentage of patients treated below and at or above the recommended phase 2 dose). Our results showed a non-significant but clinically relevant difference in trial duration. Model-based trials needed 10 months less than rule-based trials (26 versus 36 months; p = 0.25). Additionally, fewer patients were treated at dose-levels below the RP2D (31 % versus 40 %; p = 0.73) while safety was preserved (13 % DLTs versus 14 % DLTs). In this review, we provide evidence to encourage the use of model-based designs for future phase I studies, based on a median of 10 months of time gain, acceptable toxicity rates and minimization of suboptimal treatment.
KW - Clinical trial design
KW - Dose-escalation trials
KW - Model-based trials
KW - Rule-based trials
KW - Trial duration
KW - Trial performance
UR - http://www.scopus.com/inward/record.url?scp=84960431820&partnerID=8YFLogxK
U2 - 10.1007/s10928-016-9466-0
DO - 10.1007/s10928-016-9466-0
M3 - Review article
C2 - 26960536
AN - SCOPUS:84960431820
SN - 1567-567X
VL - 43
SP - 235
EP - 242
JO - Journal of Pharmacokinetics and Pharmacodynamics
JF - Journal of Pharmacokinetics and Pharmacodynamics
IS - 3
ER -