Introduction of accelerated BOIN design and facilitation of its application
Purpose: During discussions at the Data Science Roundtable meeting in Japan, there were instances where the adoption of the BOIN design was declined, attributed to the extension of study duration and increased sample size in comparison to the 3+3 design. We introduce an accelerated BOIN design aimed at completing a clinical phase I trial at a pace comparable to the 3+3 design. Additionally, we introduce how we could have applied the BOIN design within our company, which predominantly utilized the 3+3 design for most of its clinical oncology dose escalation trials. Methods: The accelerated BOIN design is adaptable by using efficiently designated stopping criterion for the existing BOIN framework. Our approach is to terminate the dose escalation study if the number of evaluable patients treated at the current dose reaches 6 and the decision is to stay at the current dose for the next cohort of patients. In addition, for lower dosage levels, considering a cohort size smaller than 3 may be feasible when there are no safety concerns from non-clinical studies. We demonstrate the accelerated BOIN design using a case study and subsequently evaluate the performance of our proposed design through a simulation study. Results: In the simulation study, the average difference in the percentage of correct MTD selection between the accelerated BOIN design and the standard BOIN design was -2.43%, the average study duration and the average sample size of the accelerated BOIN design was reduced by 14.8 months and 9.22 months, respectively, compared with the standard BOIN design. Conclusion: We conclude that our proposed accelerated BOIN design not only provides superior operating characteristics but also enables the study to be completed as fast as the 3+3 design.
Code (0)
등록된 구현이 없습니다.
Similar Papers 제목 키워드 기반
RoboInter1.5: A Holistic Intermediate Representation Suite for Embodied World Modeling and Robotic Manipulation
Existing robot datasets remain expensive to curate, embodiment-specific, and insufficiently annotated with the fine-grained structure required for generalizable reasoning, execution, or long-horizon environment dynamics …
RoboInter: A Holistic Intermediate Representation Suite Towards Robotic Manipulation
Advances in large vision-language models (VLMs) have stimulated growing interest in vision-language-action (VLA) systems for robot manipulation. However, existing manipulation datasets remain costly to curate, highly emb…
Robot ManipulationA deep learning based steganography integration framework for ad-hoc cloud computing data security augmentation using the V-BOINC system
In the early days of digital transformation, the automation, scalability, and availability of cloud computing made a big difference for business. Nonetheless, significant concerns have been raised regarding the security …
Cloud ComputingRoboInspector: Unveiling the Unreliability of Policy Code for LLM-enabled Robotic Manipulation
Large language models (LLMs) demonstrate remarkable capabilities in reasoning and code generation, enabling robotic manipulation to be initiated with just a single instruction. The LLM carries out various tasks by genera…
Code GenerationScaling characteristics of sequential multitask learning: Networks naturally learn to learn
We explore the behavior of a standard convolutional neural net in a setting that introduces classification tasks sequentially and requires the net to master new tasks while preserving mastery of previously learned tasks.…