Lymph Node Quantification System for Multisite Clinical Trials

Brigham And Women'S Hospital, Boston MA

Investigators

Linked publications & trials

Abstract

Abstract Measuring the extent of lymph node disease burden (LNDB) is crucial for diagnosing and predicting cancer outcomes. Yet, accurately quantifying LNDB remains a challenge in clinical practice, hindering both clinical trials and patient care. This Academic-Industry Partnership proposes to address this unmet need by developing accurate and time-efficient LNDB quantification tools specifically for oncology clinical trials. Our multidisciplinary team includes extensive complementary expertise in tumor metrics, machine learning, translational technology, cancer imaging, and clinical trials informatics. The members of the team include leading academic researchers from Dana-Farber Cancer Institute, Brigham and Women's and Massachusetts General Hospitals with industry leaders from Yunu, Inc., a commercial clinical trials imaging informatics platform vendor used by 18 Cancer Centers, including 12 NCI-designated facilities. Current software and algorithms for quantifying LNDB fall short, demanding excessive manual work that impedes the use of valuable prognostic factors like total metabolic tumor volume (TMTV). To address this unmet need, we leverage unique resources: a vast collection of expertly annotated contrast CT and PET/CT datasets, a proven platform for integrating AI tools into clinical trials workflows, and an experienced team with a successful track record in deploying AI imaging solutions. A key aspect of our work involves continuously improving these quantification tools through ongoing data collection from clinical trials, ensuring their real-world effectiveness. Our project focuses on three main aims: (1) Optimize semi-automated 3D lymph node annotation, segmentation, and analysis, including enhancing technologies for data access, optimized image annotation, scalable training, and efficient deployment and integration to the Yunu system, (2) Create tools for extracting clinically relevant information from both CT and PET images to provide an integrated PET/CT clinical review mode, and (3) Integrate a data analytics platform to extract new insights within Yunu's clinical trial informatics environment. Completing these aims will directly address the critical need for quantitative LNDB assessment in precision cancer treatment. Moreover, it will pave the way for similar segmentation and quantification tools for solid tumors and other diseases. Our team has unique access to many thousands of annotated images and possesses the combined academic and commercial expertise to turn these innovative goals into reality, ultimately providing a unique set of tools for assessing LNDB in patients enrolled in clinical trials.