PIXI: A preclinical imaging XNAT-enabled informatics (PIXI) platform for co-clinical imaging research

Introduction: Preclinical imaging workflows have been growing in complexity resulting in generation of large datasets. In this work, we report on efforts to develop an open-source preclinical imaging XNAT-enabled informatics (PIXI) platform to manage the workflows of preclinical image data acquisition, capture imaging-associated experiments including metadata and annotations, and to implement analytic pipelines in a unified environment.

Methods: PIXI is based on the Extensible Neuroimaging Archive Toolkit (XNAT) as the underlying informatics architecture. XNAT is used by over 200 academic institutions and industry entities as the backbone for data management across a wide range of imaging applications in clinical research, and thus offers a robust platform for the development and deployment of PIXI. Native features include support for DICOM image session data, clinical experiment data, and subject demographics, as well as user access control, integrated search, and reports. The PIXI platform includes: the PIXI Server, which provides core database, user management, visualization, and workflow functionality; PIXI Point-of-Service (PoS) interface for data entry and a PIXI Notebook for data exploration and analysis; and PIXI Apps to enable automated image processing pipelines through Docker container environment. The database is organized by “project” at the top hierarchy followed by “subject” and “encounter”. The XNAT subject model has been extended to support non-human subject metadata.

Results: In the current development phase of PIXI, positron emission tomography (PET), magnetic resonance (MR) and computed tomography (CT) DICOM images are pushed to the PIXI server for workflow management. Basic metadata is captured through the DICOM image files for search and reporting. Multi-mouse images (mouse hotel images) are supported by a custom RESTful API that automatically partitions mouse hotel images into individual image datasets for incorporation into the PIXI database. Imaging workflow information can be entered or edited through the PoS interface including animal modeling information/metadata (such as patient-derived tumor xenografts and cell-line xenografts), descriptive data (weight, caliper measurements for tracking tumor volumes, metabolic panels, etc.), and drug therapies. Additional data types are being developed to capture image-derived measurements, non-DICOM images (such as BLI), pathology, OMICS, image annotations and radiation therapies/treatments. To enable open science, PIXI instances will interconnect through a federated model to allow site-to-site data and application sharing with configurable access control. Importantly, PIXI will also support interfaces to The Cancer Image Archive (TCIA) and the planned NCI Imaging Data Commons (IDC) to enable import and export of data to facilitate open science research.

Conclusions: Overall, the development of the preclinical imaging informatics platform is expected to have a profound impact on the management of preclinical imaging datasets which will ultimately support translational precision medicine. Additional information is available on the PIXI development website at https://www.PIXI.org/ with an opportunity to become an early adopter.