Abstract 78P
Background
When investigating multiple agents, pairs of combination doses may initially have unclear toxicity probability orderings. The partial order continual reassessment method (PO-CRM) design was initially proposed to evaluate combination doses with unknown ordering. It was later extended to cope with late-onset toxicities by incorporating both full and partial dose-limiting toxicity (DLT) follow-up, thereby avoiding impractically long trial durations. This work shares insights from applying a novel tailored PO-TITE-CRM design in a phase I ovarian cancer combination trial to identify the maximum tolerated dose (MTD) for two DNA damage response inhibitors.
Methods
To implement the PO-TITE-CRM design, statistical parameters were calibrated considering expected DLT occurrence distribution, whilst minimising the trial duration. With up to 18 DLT evaluable patients, the model selects the toxicity ordering with the highest posterior probability among the available treatment combinations, and then adaptively allocates the dose with estimated DLT rate closest to the 25% target rate to the next patient. Extensive simulations assessed model performance across varied toxicity profiles, executed by modifying codes from R packages pocrm and dfcrm.
Results
We highlight our collaborative development of a customised PO-TITE-CRM, strategically incorporating the nuanced expected tolerability profile of the combination therapy and operational factors. Simulation results demonstrate that trial duration can be substantially reduced compared to PO-CRM design, by enabling semi-continuous accrual, without compromising on patient safety and the accuracy of MTD recommendations.
Conclusions
These findings underscore the practical utility of the PO-TITE-CRM, showing its adaptability to individual trial needs. The approach holds promise in accelerating dose-finding combination trials despite the challenge of unknown ordering. It highlights the vital role of close collaboration with the trial teams in designing and delivering innovative dose-finding trials, essential for promoting the adoption of these methodologies.
Clinical trial identification
Trial in set up.
Editorial acknowledgement
Legal entity responsible for the study
The authors.
Funding
Has not received any funding.
Disclosure
X. Hu: Financial Interests, Personal, Invited Speaker: Boehringer Ingelheim; Financial Interests, Personal, Full or part-time Employment: The Institute of Cancer Research. U. Banerji: Financial Interests, Personal, Advisory Board, Advisory Board Member: Pegasy, Carrick Therapeutics; Financial Interests, Personal, Advisory Board: Pharmenable, Ellipses Pharma; Financial Interests, Personal, Full or part-time Employment, Employed by the Institute of Cancer Research: The Institute of Cancer Research; Financial Interests, Institutional, Research Grant, Investigator Initiated Trial: Carrick Therapeutics, Chugai, Verastem Oncology; Financial Interests, Institutional, Research Grant, Pre-clinical grant: Verastem Oncology; Financial Interests, Institutional, Research Grant, Preclinical Research Grant: Avacta; Non-Financial Interests, Personal, Principal Investigator: Various Pharma Companies. C. Yap: Financial Interests, Personal, Advisory Board, Statistical Consultant: Faron Pharmaceutical; Financial Interests, Personal, Invited Speaker, Speaker at a Training Workshop: Bayer; Financial Interests, Institutional, Research Grant: AstraZeneca, Celgene, Novartis; Financial Interests, Personal and Institutional, Research Grant: Faron Pharmaceuticals.