Elimination of Instrumental Bias for Quantitative Diffusion Imaging in Clinical Oncology Trials

-
Investigator: David Newitt, PhD
Sponsor: University of Michigan

Location(s): United States

Description

To aid clinical decision making and improve oncological patient management, current multi-center/multi- platform clinical trials evaluate quantitative diffusion imaging for tissue response and quantitative lesion characterization for translational oncology applications. While providing data analysis and quality control for ongoing ECOG-ACRIN clinical oncology trials, the academic partners of the proposed AIP have identified the major source of spatial variability in quantitative diffusion measurements across scanner systems related to platform-specific nonuniformity bias in diffusion weighting. This systematic bias caused by platform-dependent gradient designs confounds quantitative diffusion metrics for characterization of tissue pathology leading to inconclusive findings and increasing the requisite subject numbers and trial costs. Through active involvement with national and international quantitative imaging initiatives (NCI-QIN, RSNA-QIBA, ISMRM, QuIC- ConCePT) as well as collaboration between several academic centers and three major vendors the consensus has been established that robust and timely solution of the technical hurdle for quantitative DWI trials requires merging expertise among commercial scientists/engineers and academic researchers to implement practical correction of spatial bias across diverse clinical MRI platforms. The goal of the proposed AIP is to design, evaluate and implement practical DW bias correction tools for quantitative diffusion imaging applications in clinical cancer trails acrss three major MRI vendors. These tools will eliminate a dominant source of scanner- specific bias manifesting as cross-platform, cross-exam variability and thereby advance longitudinal and multi- institutional translational cancer research that utilizes quantitative diffusion imaging to improve management of oncology patients. The goal will be achieved through Aim1: designing and testing practical bias correction procedures to improve quantitative diffusion imaging across representative clinical platforms and through Aim2: evaluation and implementation of practical correction procedures to enhance precision of tissue diffusivity metrics generated in clinical oncology trials. Academic members of the proposed partnership are leading experts in clinical MRI and diffusion imaging and have active collaboration with three dominant clinical MRI manufactures. The PI institution has developed a widely-utilized phantom for diffusion imaging quality control in numerous translational oncology clinical trials. Partnership members have pioneered investigations of practical diffusion bias characterization and correction procedures. Accomplishment of the project aims will eliminate significant instrumental bias that confounds current multi-center/multi-platform clinical trials that employ quantitative diffusion imaging.