search for


A comparative study on the accuracy and safety of Bayesian CRM-Based MTD estimation method
Journal of the Korean Data & Information Science Society 2021;32:337-50
Published online March 31, 2021;
© 2021 Korean Data and Information Science Society.

Si-Jin Park1 · Yeji Kim2 · Sangbum Choi3

1Department of Statistical Data Science, Korea University
23Department of Statistics, Korea University
Correspondence to: This research was supported by Korea University Research Fund (K2008341) and Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (2019R1F1A1052239, 2019R1A4A1028134).
1Graduate student, Department of Statistical Data Science, Korea University, Seoul 02841, Korea.
2Graduate student, Department of Statistics, Korea University, Seoul 02841, Korea.
3Corresponding author: Associate professor, Department of Statistics, Korea University, 145 Anam-ro, Seongbuk-gu, Seoul 02841, Korea. E-mail:
Received February 8, 2021; Revised March 15, 2021; Accepted March 23, 2021.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License ( which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Phase I clinical trial is an early stage of clinical trials targeting humans based on pharmacological data obtained from preclinical trials. The objective is to verify the safety and dosage of an investigation product to be used in the phase II clinical trials that confirm the effectiveness of the drug. Maximum tolerated dose (MTD) is the biggest interest in phase I clinical trials, which has a maximum effect of the drug on a person’s tolerated toxicity level. In this paper, we conduct extensive simulation studies to compare modern Bayesian continual reassessment methods (CRM) for MTD estimation, such as BMA-CRM or BMS-CRM, with standard algorithms like SM3 under various toxicity probability scenarios. We confirm that Bayesian CRM performs very stably and safely even when pre-specified skeletons do not exactly match true toxicity probability. Our study may prove the utility of the Bayesian CRM approach for routine dose-finding analysis.
Keywords : Bayesian, continual reassessment method, maximum tolerated dose, phase I clinical trial.