Hyperpolarized MRI for GIST on Imatinib

Conventional imaging techniques, such as CT and MRI, have limitations in assessing GIST response to imatinib. They primarily focus on tumor size, which may not reflect early therapeutic effects, as GISTs can show necrosis or cystic changes that don't correlate with viable tumor tissue reduction. These methods also fail to capture metabolic and molecular alterations within the tumor. As a result, relying solely on size changes can lead to inaccurate or delayed treatment response evaluations, underscoring the need for more advanced imaging techniques. This project is novel because it aims to investigate GISTs undergoing imatinib treatment from a metabolic perspective by integrating hyperpolarized (HP) 13C MRI, metabolomics, and radiomics. HP 13C-MRI is a cutting-edge, non-invasive, real-time dynamic imaging technique used to monitor metabolic flux in vivo. Dynamic nuclear polarization (DNP) enhances the signal of 13C-labeled probes by up to 50,000 times. With DNP, \[1-13C\]pyruvate can be utilized to study various metabolic pathways, including its conversion to lactate (anaerobic glycolysis), alanine (transamination), and bicarbonate (an indirect marker for the TCA cycle). In our preliminary GIST cell models, significant changes in pyruvate metabolism were observed following imatinib treatment depending on their imatinib resistance. Additionally, NMR-based metabolomics analysis of tumor tissue and blood samples will provide a comprehensive view of systemic metabolic changes. These metabolic alterations will then be linked to MRI radiomics features for a more holistic understanding of treatment response. In this single-center, cross-sectional study, a prospective trial integrates advanced MRI techniques to address clinical challenges and improve patient care in GIST treatment. This 3-year project follows a non-randomized, two-group observational cohort design with 30 GIST patients undergoing imatinib therapy at Chang Gung Memorial Hospital at Linkou (CGMH). Pre-treatment assessments include standard-of-care CT, HP 13C-MRI, 1H-NMR metabolomics, and biopsy of the tumor tissue. HP 13C-MRI will be used to assess metabolic activity in the tumors, while 1H-NMR metabolomics will provide additional insights into systemic metabolic changes. Early response assessment one month after starting imatinib will involve HP 13C-MRI and metabolomics analysis, followed by routine response evaluation at four months using standard CT. Patients will be categorized into responders and non-responders based on treatment efficacy, and data from HP 13C-MRI, metabolomics, and radiomics will be analyzed. Machine learning techniques will integrate multi-omics data for more precise response prediction. We hypothesize that metabolic changes detected through HP 13C-MRI can predict GIST response to imatinib therapy. The extent of metabolic alterations will be correlated with tumor response, providing a new way to evaluate treatment effectiveness. By combining metabolic data with radiomic tissue characteristics, we aim to track dynamic changes in GIST metabolism with reproducible, quantifiable metrics. Ultimately, the integration of HP 13C-MRI and metabolomics will offer a deeper understanding of GIST pathophysiology from a metabolic perspective and could lead to new therapeutic strategies.