This post has been republished via RSS; it originally appeared at: Healthcare and Life Sciences Blog articles.
Today Microsoft is releasing the Medical Imaging Server for DICOM, new Open Source Software (OSS) that provides developers with a powerful tool to migrate medical imaging data to the cloud and integrate imaging metadata with clinical data in FHIR using DICOM Cast technology.
DICOM + FHIR, better together
Health systems around the globe are rapidly working to bring data into the cloud to improve patient diagnosis, the quality of care, and transformative learning about the way we deliver healthcare. And that makes sense, because cloud technology can help improve data security, lower the cost of data management, and enable machine learning. But as health data transitions to the next era of cloud computing, leading health systems have quickly realized that how they bring data to the cloud matters. Getting the most from cloud computing isn’t just about moving data in existing custom formats, it’s about using the right tools to enrich and normalize your data on the way in so you can do more with it once it’s there. With health data, the open data frameworks of FHIR® (Fast Healthcare Interoperability Resource) have become a key tool for this transformation in the cloud. FHIR normalizes the standard for clinical data exchange to the cloud, helping to reduce the cost of downstream development, accelerating machine learning, and expanding opportunities for innovation.
But the use of FHIR for interoperability has only focused on structured clinical data. Until now.
Alongside clinical data, medical imaging plays a critical role in patient diagnosis and care. DICOM® (Digital Imaging and Communications in Medicine) is the international standard to transmit, store, retrieve, print, process, and display medical imaging information. DICOM is important because it ensures that medical images meet quality standards. It’s used in most health organizations for diagnosis alongside clinical inputs, and yet DICOM data is primarily maintained and managed independently of clinical data, creating barriers for access, reducing efficiency of use, and slowing the ability to research with imaging data. Microsoft developers wanted to push the boundaries of FHIR and bridge the gap. We’re releasing new tools to bring DICOM together with FHIR in the cloud so the health industry can do more with their data.
The Medical Imaging Server for DICOM
The Medical Imaging Server for DICOM streamlines the process of ingesting medical imaging data into the cloud with a simple click to deploy. Developers can deploy the open source software in minutes and setup an Azure Resource Group to enable cloud management of imaging data, including:
- REST API endpoint that enables direct ingest and persistence of imaging data in the cloud
- Low cost and scalable data store for raw imaging files
- Security through Azure Key Vault, allowing developers to import and centrally manage keys, secrets and passwords to control data access
- Connectivity to Azure Application Insights to view operational telemetry and manage performance and costs through a centralized platform
Elevating interoperability, the Medical Imaging Server for DICOM is the first cloud technology to bring together DICOM data standard and FHIR using the DICOM Cast feature. Concurrent with data upload through the Medical Imaging Server for DICOM, structured metadata from imaging files is seamlessly linked to clinical health data stored in FHIR:
- Rapid deployment of Azure Container Instance for FHIR integration
- Identification and extraction of metadata from DICOM images and push of that data through secure exchange to a FHIR endpoint
- Direct connection and deployment of the Azure API for FHIR (PaaS) or FHIR Server for Azure (OSS) to persist and manage DICOM metadata alongside other clinical data
- Downstream access to a portfolio of FHIR based tools via FHIR APIs: complex queries across clinical and imaging data sets, visualization and analytics through the Power BI and anonymization rules for de-identification.
Improving data agility, research and machine learning in the cloud
By using the Medical Imaging Server for DICOM alongside the Azure API for FHIR or other FHIR services, data reference are created between imaging data and clinical data in FHIR, setting the stage for multiple scenarios which are difficult and expensive to execute in today’s on-premises systems:
- Creating cohorts for research. Queries for patients that match data in both clinical and imaging systems can be streamlined when metadata is normalized in FHIR. Requests such as: “Give me the medications prescribed with all the CT Scan documents and their associated radiology reports for any patient older than 45 that has had a diagnosis of osteosarcoma over the last 2 years” can be executed in minutes through a FHIR service without lengthy or complex processing.
- Providing a longitudinal view of a patient during diagnosis and closing the feedback loop. Radiology is increasingly introducing AI/ML models to improve diagnosis. Machine learned models perform better when real-world feedback can be used to improve their models, but after providing a diagnosis, the silos of imaging and clinical data create limited access to clinical results or outcomes. Closing this feedback loop is essential for improved learning and future diagnosis. Through DICOM and FHIR integration, development of rapid feedback loops is possible, even when working with radiologists or ML solutions outside of the organization’s local network.
- Offsite archival and backup to leverage the cloud for lower cost storage. Many modern PACS and VNA systems can already connect directly to servers communicating with the DICOMweb standard REST APIs. For those existing solutions, the Medical Imaging Server for DICOM is a solution that can rapidly be deployed for cloud-based, off-site storage needs.
Microsoft customers are already beginning to harness the power of normalized data in exciting ways. As part of their early work with the Medical Imaging Server for DICOM, IMS, an innovative medical imaging software company based in Canada, saw a growing gap with the increased use of teleradiology and their ability to access data in real time. This need was heightened by COVID-19, which forced many radiologists to work remotely. IMS envisioned an end-to-end solution to provide complete business continuity so radiologists can effectively work remotely in cases of work-from-home orders, server failures, or security issues that caused a delay in data access. Their IMS CloudVue solution connects directly with the Medical Imaging Server for DICOM and enables users to upload, view, annotate, share, and delete images directly in Azure. The viewer displays full fidelity medical images directly in the browser. As a next step IMS plans to add support for IMS CloudSync which will provide a complete solution for uploading and synchronizing image data in Azure with the on-premises hospital data.
“We are excited to use Microsoft’s Medical Imaging Server for DICOM with IMS CloudVue and are impressed with the speed with which the Microsoft team has enabled our FDA approved viewer. IMS CloudVue leveraging the Medical Imaging Server for DICOM enables hospitals and healthcare providers a safe transition to the cloud. This platform is excellent for image sharing and machine learning, areas in which we continue to focus on innovations.” – Vittorio Accomazzi, CTO, International Medical Solutions (IMS).
Getting Started with the Medical Imaging Server for DICOM
The easiest way to get started with the Medical Imaging Server for DICOM is by visiting the GitHub page located at https://github.com/microsoft/dicom-server and clicking the “Deploy to Azure” button. After making a few quick choices, your deployment will be on its way and in less than 5 minutes, you'll have DICOM medical images in Azure.
As part of the installation process, most the following components will be deployed inside of an existing or new subscription:
- SQL Server: Indexes a subset of the metadata to support queries
- Azure Storage: Blob storage persists all Medical Imaging Server for DICOM data and metadata
- App Service Plan: Hosts the Medical Imaging Server for DICOM
- Azure Key Vault: Stores critical security information
- Application Insights: Monitors performance of Medical Imaging Server for DICOM
- Azure Container Instance: Hosts the DICOM Cast service for FHIR integration
- Azure API for FHIR: Persists the DICOM metadata alongside other clinical data
As open source technology, the Medical Imaging Server for DICOM can be deployed to any FHIR endpoint but works seamlessly with the Azure API for FHIR. Customers can easily deploy the API for FHIR by following the steps located at https://docs.microsoft.com/en-us/azure/healthcare-apis/fhir-paas-portal-quickstart.
With OSS technology, we know developers want to explore. The Medical Imaging Server for DICOM is easy to deploy in low-cost tiers to support dev/test and proof-of-concept environments. When you’re ready to scale it for production use, it can ramp up to ingest tens of thousands of images per minute in production. Performance matters, so we’ve designed it to configure autoscaling that allows your implementation to respond to changing ingestion, retrieval and query patterns over time, and with performance dashboards to quickly identify ways you can streamline data flows for efficiency.
Get started today
We’re excited about the launch of the Medical Imaging Server for DICOM! We can’t wait to see the innovative solutions our customers build. Join us in the next horizon of health data transformation and do more in the cloud with your data.
Learn more about the Medical Imaging Server for DICOM on GitHub (https://github.com/microsoft/dicom-server)
DICOM® is a registered Trademark of National Electrical Manufacturers Association
FHIR® is a registered trademark of HL7 and has been used with their permission.