Sonador IO

Sonador IO is the data management interface of Sonador used to securely store medical images so that they can be queried, retrieved, and transferred. It provides a streamlined way to share medical image data with other institutions across the Internet or other applications within the hospital.

Platform Components

Sonador IO is comprised of a rich ecosystem of libraries, tools, and programs to support medical imaging development. It enables healthcare organizations and researchers to own their data while providing integrations with clinical tools and cloud platforms.

Sonador's cloud platform is built on open source technologies including:

The Sonador web application runs within a hospital, institutional network, or a trusted cloud environment and provides security and integration APIs allowing for the systems used in Sonador to work in concert.

Applications run in containers and communicate across standard protocols such as DICOM and DICOMweb, exchanging data in standard formats such as DICOM, DICOM-SR, and DICOM-SEG. Python Client libraries provide tools to translate meta, image, and annotation data to formats which can be used by frameworks such as OpenCV, PyTorch, TensorFlow, OpenCascade, PyVista, and others for custom application development. A web based viewer (OHIF) enables clinical staff to review imaging studies, segmentations, models, and other imaging data to verify and validate findings, plans, or designs.

Medical Imaging Database

At the center of Sonador is the highly scalable imaging database, Orthanc, and set of libraries that can be deployed on-premise or in the public cloud, greatly simplifying medical data exchange. Orthanc securely stores medical images in their standard DICOM format, allowing the system to integrate with other tools in the healthcare IT infrastructure while the libraries can assist users in retrieving and moving data from one system to another.

Security and Integration

Security in Sonador IO is provided by the Sonador web application. It provides user and group management, role based access control, authentication/authorization workflows, service discovery, centralized configuration, and a set of integration APIs that allow for Orthanc, OHIF, and modalities to work together in concert.

DICOM Viewer

Sonador IO includes an embedded copy of the OHIF DICOM viewer with support for a wide range of modalities (MRI, CT, ultrasound, X-ray, and more); 2D/3D visualization; multi-planar reconstruction (MPR); segmentation (DICOM-SEG) and structured reporting (DICOM-SR); and whole-slide imaging. This allows for radiologists, researchers, and others to review and comment imaging studies without requiring them to install additional software.

Sonador Viewer: Multi-modality MRI/CT/SEG — OHIF provides provides a web-based DICOM viewer with support for a wide range of modalities including MRI, CT, ultrasound, and PET.

Sonador Viewer: Multi-planar reconstruction (MPR with segmentations) — OHIF can be expanded to include advanced image visualization and processing capabilities, such as multi-planar reconstruction, through the use of plugins. Sonador implements support for OpenID Connect (OIDC) login workflows, enabling highly customized versions of the viewer to be deployed alongside a standard build.

Image Segmentation, Registration, and 3D Analysis

Sonador IO integrates with programs such as 3D Slicer which can be used for advanced biomedical image analysis. Using Slicer, it is possible to create segmentations for surgical planning, generating training datasets for AI, or for advanced simulation.

Once data has been processed, results can be exported to DICOM segmentations; registration objects; structured reports (DICOM-SR); parametric maps; or RT structure sets, plans, images, and dose maps. Slicer supports the DICOMweb and DICOM network protocols, which makes it easy to send results back to Orthanc for verification and validation.

Sonador 3D: Knee Models in 3D Slicer — Slicer can be used to identify regions of interest in an image (segmentations) and export the data to a 3D model for use in surgical planning, mathematical modeling, or custom implant design.

Sonador 3D: Knee Models in FreeCAD — 3D models from Slicer are compatible with Computer Aided Drafting (CAD) and other tools.

Sonador Imaging Platform: Lung Segmentation Step 2 — Through its integration with Python, Slicer can be used as a platform to create interactive workflows to aid with the segmentation, review, and quantification of DICOM images.

Sonador Imaging Platform: Automated Lung Analysis — Slicer has been used for building software to help segment CT scans of COVID-19 patients automatically to quantify how patients are recovering from the disease and investigate the impact that that lung scarring may have on long-term health.

Python

Sonador IO includes a Python software library for the Orthanc REST API which allows developers to work with imaging resources using an object-relational mapping (ORM) interface similar to those provided by the popular web development framework Django and SQLAlchemy. This greatly simplifies the implementation of complex systems and allows for data stored in Orthanc to be integrated with popular Python libraries for Machine Learning, Artificial Intelligence, and 3D processing.

The client library interfaces with the security controls provided by the Sonador web application, ensuring that only authorized users are able to access resources.

Sonador 3D: Knee Models in JupyterLab — Python client libraries in Sonador allow for data to be retrieved from Orthanc for visualization or exploratory data analysis in visual development environments such as JupyterLab.

Sonador AI (Workflow Step 2): Extracting image masks from contours obtained from a 3D model — The Python library opens the door for data to be analyzed by any tool in the Python Data Science ecoystem such as SimpleITK or OpenCV. It is also possible to combine pixel data with other sources, such as point clouds or meshes.

Learn More

Security Overview

This article describes how OpenID connect and oAuth2 are used in Sonador to manage authentication of users and authorization of API requests.