455P - Evolution of clinical trial design in early drug development: the use of expansion cohorts (ECs) in phase I cancer trials (PITs)

Date 01 October 2012
Event ESMO Congress 2012
Session Poster presentation III
Topics Drug Development
Presenter Arif Manji
Authors A. Manji1, I. Brana Garcia2, E. Amir3, I.F. Tannock4, P. Bedard5, A.M. Oza5, L. Siu5, A.R.A. Razak5
  • 1Division Of Paediatric Haematology/oncology (room 8222), Princess Margaret Hospital, M5T 2M9 - Toronto/CA
  • 2Division Of Medical Oncology & Hematology / Drug Development Program, Princess Margaret Hospital, M5T 2M9 - Toronto/CA
  • 3Medical Oncology, Princess Margaret Hospital, Toronto/CA
  • 4Dept Of Medical Oncology, Princess Margaret Hospital, Toronto/CA
  • 5Drug Development Program, Division Of Medical Oncology And Hematology, Princess Margaret Hospital, M5T 2M9 - Toronto/CA



ECs are frequently used to optimize the yield of PITs. However, their rationale and value have yet to be evaluated. Here, we explore the prevalence, characteristics and objectives of ECs in single-agent PITs.


We conducted a systematic review using MEDLINE and EMBASE to identify all adult single-agent PITs published after 2006. Eligibility assessment and data extraction were performed by 2 reviewers. The primary endpoint was the proportion of PITs with ≥ 1 EC. Additional endpoints included factors associated with EC inclusion and whether the EC objectives were stated and achieved.


4,557 articles were reviewed and 591 trials met eligibility criteria. 139 (24%) included ≥ 1 EC. Use of ECs in single-agent PITs increased between 2006 and 2011 (12.2% to 35.7%, Spearman's rho 0.20, p < 0.001). In PITs with ECs, a median of 22 and 17 subjects were enrolled in the dose-escalation cohorts (DECs) and ECs respectively. In unadjusted analysis, PITs were more likely to include an EC if they were multi-center (OR 2.41, 95% CI 1.52-3.82, p < 0.001), industry-sponsored (OR 1.80, 95% CI 1.12-2.91, p = 0.02), and evaluating non-cytotoxic agents (OR 2.12, 95% CI 1.30-3.47, p = 0.003). In multivariable analysis, these factors retained statistical significance except for industry sponsorship. Geographical location of study and tumor type were not significant. EC objectives were reported in 74% of trials and included safety (83%), efficacy (45%), pharmacokinetics (28%), pharmacodynamics (23%), and patient enrichment (14%). Among ECs assessing safety, the MTD was modified in 14% and new toxicities defined in 53%. Among ECs designed to assess efficacy, only 24 of 46 (52%) reported EC efficacy separately; of these, only 3 (13%) reported tumor responses in EC subjects not previously observed in the DEC.


The utilization of ECs in PITs has increased with time and is more common for multi-center trials of non-cytotoxic agents. Safety and efficacy are common objectives but 26% failed to report explicit aims. Although the majority of ECs supplement PITs with meaningful safety data, their role in assessing preliminary efficacy requires better definition.


All authors have declared no conflicts of interest.