Adjudication Rates between Readers in Blinded Independent Central Review of Oncology StudiesFord RR1,2*, O’ Neal M2, Moskowitz SC3 and Fraunberger J2
- Corresponding Author:
- Ford RR Clinical Trials Imaging Consulting, LLC, Belle Mead, NJ, USA Tel: 609-651-6887 Fax: 908-431-5940 E-mail: [email protected]
Received Date: September 15, 2016; Accepted Date: October 18, 2016; Published Date: October 28, 2016
Citation: Ford RR, O’ Neal M, Moskowitz SC, Fraunberger J (2016) Adjudication Rates between Readers in Blinded Independent Central Review of Oncology Studies. J Clin Trials 6:289. doi: 10.4172/2167-0870.1000289
Copyright: © 2016 Ford RR, et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Purpose: Blinded independent central review (BICR) is advocated by regulatory authorities as a means of minimizing bias and independently verifying endpoints based on medical imaging when the data is intended to support pivotal trials. However, discordance between reviewers at the BICR raises concern with regulators. There are few published metrics related to discordance rates at the BICR. Methods: We analyzed BICR data from 79 oncology clinical trials including interpretations by 23 different radiologist reviewers of 23,476 subject cases. Results: The proportion of cases requiring adjudication across all trials was 42% (95% CI: 41-42%). There is variation based on the indication. There is a significant tendency for the Adjudication Fraction (AF) to increase as the number of adjudication variables increases (p<0.001). There is also a relationship between the average number of target lesions and the AF. In trials for which there were at least 2 targets lesions per patient, the AF decreases as the number of target lesions increases (p=0.020). The data suggests a pattern whereby the AF increases as the number of assessment time points for a subject increase until approximately 7 time points and then decreases (p=0.001). The AF is independent of the response criteria. Conclusion: The AF has multiple dependencies and can be predicted based on modeling of those factors.