Info
Talk Summary:
The talk will cover the why and how of developing the Preclinical Imaging XNAT-enabled Informatics (PIXI) platform. PIXI is an open-source informatics ecosystem built to address the fragmented state of preclinical cancer imaging. Preclinical modalities parallel the clinic but suffer from vendor-specific file formats, inconsistent metadata, and legacy databases ill-suited to modern big-data and precision-medicine demands. PIXI (pixi.org) provides a unified environment to capture preclinical imaging workflows, associated metadata, and linked-biology experiments in support of NCI’s Precision Medicine Initiative and NIH’s broader Data Management and Sharing Policy. The next phase of PIXI will establish PIXI Center, a domain-specific, federated data management and sharing (DMS) repository for preclinical imaging datasets and linked-biology experiments. Beyond storage and discovery, PIXI Center will enable Centralized Learning by letting users create new projects from search results and leverage integrated Jupyter Notebooks and Docker containers for AI/ML model development on curated preclinical imaging datasets. The presentation will also put PIXI in the context of broader unmet needs to support next generation quantitative imaging research. Finally, because PIXI is a community platform, the presentation will highlight several deliberate practices for building and sustaining a mature, community-oriented computational product.

Presenter Biography:
Dr. Kooresh Shoghi is a Professor of Radiology, Biomedical Engineering, Computational Systems Biology, and Cancer Biology at Washington University in Saint Louis.

Author
Kooresh Isaac Shoghi
Washington University School of Medicine