Abstract
Cancer drug discovery and development is being throttled by high attrition rates and skyrocketing clinical trial costs as trials get increasingly more complex. The status quo is not sustainable, and new approaches to preclinical and clinical experimentation are urgently needed as we continue our quest to combat cancer. We have gained significant knowledge in cancer molecular biology in past decades, which has led to the discovery of new biomarkers and novel targeted therapies, yet the traditional clinical trial designs are lengthy and challenged by low patient enrollment due to the inability to identify patients eligible for trials. New paradigms for trial design and conduct must be developed in order to address unmet needs. In recent years, the use of adaptive design methods has gained support in clinical trials to improve study success rate, such as the I-SPY 2 trial. Adaptive clinical trial designs enable a study to address multiple questions concurrently such as appropriate patient population, dose, dosing regimen, and effective drug combinations. This adaptive approach uses accumulating data to modify the ongoing trial to improve the power of the study and treat more patients with more effective treatments. In parallel, computational, or in silico modeling, has also become increasingly useful in identifying patient populations who are phenotypically and genotypically eligible for target-based clinical trials. One of the major challenges we are facing now is integrating the vast amount of data often required to improve clinical trial design and matching patients to clinical trials. Here we present a framework that enables combining both approaches to enhance the discovery, development, and delivery of new therapies. This undertaking is made possible through a data-sharing network, the Oncology Research Information Exchange Network (ORIEN). ORIEN is comprised of numerous cancer centers that have agreed to use the same IRB-approved protocol and consent (Total Cancer Care Protocol) to follow patients throughout their lifetime. Importantly, consented patients also agree to be recontacted for future research and clinical trials. The ORIEN clinical data collection is harmonized through data standards and common data dictionaries among all sites. The acquisition, processing, storage, and release of human tissue are also uniform across all sites, as they follow the same policies and procedures. Collectively, this powerful platform enables prospective, current, and retrospective data analyses and in silico modeling that can be iteratively enhanced with the addition of raw and derived data to identify specific study cohort(s), termed “in silico communit(ies).” Significant effort is needed for updating data and models to keep pace with patient outcomes. Like all adaptive designs, the development and refinement of a robust statistical methodology is also critical for an unbiased evaluation of the data. The key to success for this approach requires devoted collaboration between patients, cancer research community, pharmaceutical industry, and the FDA.
Citation Format: Judy Barkal, Ming Poi, William Dalton. An innovative approach to improve clinical trials using adaptive in silico design [abstract]. In: Proceedings of the AACR Special Conference on Modernizing Population Sciences in the Digital Age; 2019 Feb 19-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2020;29(9 Suppl):Abstract nr IA27.