University of Michigan Quantitative Co-Clinical Imaging Research Resource

点击翻译按钮获取中文摘要

PROJECT SUMMARY/ABSTRACT Quantitative imaging methods to detect cancer and assess response to therapy are cornerstones of pre-clinical drug development, clinical trials, and patient care. Success of quantitative imaging in oncology relies on standardization of protocols for image acquisition and analysis to ensure reproducibility within a single site over time and across institutions for multi-site clinical trials. Work by our group and others in the Quantitative Imaging Network and Quantitative Imaging Biomarkers Alliance continues to advance standardization procedures for clinical imaging. However, similar rigor has not been applied to pre-clinical imaging studies of cancer therapy in mice. The disconnect between standardization and validation methods incorporated into imaging studies for humans versus mice contributes to ongoing challenges with reproducibility in drug development and successful translation of new drugs to clinical medicine. To ensure direct, quantitative comparisons between pre-clinical and clinical imaging, we will establish a resource for quantitative MRI of bone marrow composition and architecture in myelofibrosis (MF), a chronic hematologic cancer marked by progressive fibrosis and destruction of bone marrow. This resource will extend quantitative imaging into hematologic cancers, a group of malignancies understudied and underserved by imaging because current methods generate largely qualitative data that cannot be used reliably as biomarkers for response to therapy. We will analyze key metrics of bone marrow disease using FDA-approved MRI sequences included in standard software packages for pre-clinical 7T and clinical 3T scanners: 1) bone marrow composition and cellularity (quantitative Dixon technique for fat/water); 2) replacement of normal bone marrow cells and bone trabecula (mobility of water (diffusion, DWI)); and 3) extent and severity of fibrosis (magnetization transfer (MT)). As part of standardization procedures for both mouse and human imaging, we will measure repeatability of imaging data using phantoms for each MRI sequence and test/retest imaging procedures for mouse and human subjects to establish confidence intervals. We also will standardize workflow for quantifying bone marrow MRI data with parametric response mapping (PRM), a voxel-wise image processing method we devised to capture spatial and temporal heterogeneity of imaging data during treatment. After establishing standard operating procedures for quantitative bone marrow MRI (Aim 1), we will apply these methods to co-clinical trials with standard-of-care and investigational therapies for MF, matching driver mutations for MF present in our patient population with our mouse model (Aim 2). To disseminate these methods to the imaging community, we will post standard operating procedures for MRI protocols, mouse models of MF, and PRM of bone marrow MRI data (Aim 3). We also will deposit curated imaging data in the TCIA, enabling other investigators to mine these data and test new hypotheses. Overall, this resource will reduce variability in quantitative bone marrow MRI in both mice and humans, improving the ability to reliably implement these imaging biomarkers to advance co-clinical trials and drug development MF and likely other hematologic malignancies.