Cloud-Based Medical Imaging Platform for Radiology Diagnostics

Andersen was invited to address these challenges leveraging our expertise in medical imaging. The customer set key project goals, and together we outlined the solution:

• Building a fast, modern SaaS platform that makes imaging operations smoother and speeds up radiology service delivery;
• Phasing out the outdated legacy system and putting a stronger, more reliable solution in its place;
• Ensuring business continuity by carefully moving historical data from the legacy platform into the new one;
• Scaling capacity to process up to 10,000 orders a day, over three times the old limit;
• Building a headless transfer agent to streamline and speed up DICOM image transfers;
• Working to integrate smoothly with outside systems, such as PACS and fax services, so everything stayed connected;
• Adding a rule engine to manage orders with speed and clarity across the RAD workflow.

During the Radiology Information System build, we faced a number of technical hurdles and tasks, and successfully worked through all of them:

• Lack of API compatibility. With third parties updating APIs unpredictably and without documentation, we had to validate constantly and design flexible workarounds;
• HL7 communication. The customer’s HL7 messages sometimes included inconsistent data types. Together, we anticipated this and resolved it during the project’s early stages;
• GUI-less transfer agent. Together, we tackled the challenge of integrating with clinics by creating a background Windows service to handle DICOM transfers securely;
• Integration orchestration. In collaboration with the customer, we ensured all integrations worked in sync, since downtime in a single third-party service could stop orders and delay reports;
• Historical data migration and synchronization. We migrated all historical data from the legacy platform and kept both systems in sync during the transition. Clients could finish their RAD flow on the old platform while starting on the new one, with orders updated in real time.

On the technical side, the medical imaging platform was built using:

• Amazon API Gateway;
• Amazon ECS Delivery;
• AWS Config Delivery;
• Amazon EKS Delivery;
• AWS Lambda Delivery;
• Amazon MSK Delivery;
• AWS WAF Delivery;
• Amazon RDS Delivery (Amazon RDS for PostgreSQL);
• Amazon CloudFront Delivery;
• Amazon Route 53;
• Amazon DynamoDB Delivery;
• Amazon CloudTrail, among others.
• AWS Systems Manager Delivery;

The resulting compliant clinical SaaS for radiology includes the following key features:

• Data ingestion. A local transfer agent receives DICOM files, compresses them, and generates JPEG thumbnails.
• Workflow automation. A Camunda BPMN engine drives a status model that ensures every order progresses consistently.
• Comprehensive patient records. The system merges duplicate records using Master Data Management (MDM) rules and submits them to the Assigning Authority (AA).
• Reporting and document handling. The system generates radiology reports and manages documents securely while ensuring compliance requirements are met.
• Dashboards and analytics. The system displays key metrics on configurable dashboards, allowing users to quickly visualize and analyze operational and clinical data.

While building the medical imaging platform, Andersen achieved the following results:

• Complex platform delivery. The system supports 21,000 users and processes over 10,000 orders per day.
• Meeting project requirements 100%. The system fully matched all functional, non-functional, and performance specifications.
• Data migration. We successfully moved all historical data – orders, users, files, and more – from a non-relational database to a relational one for integrity and continuity.
• Regulatory compliance. The solution complies with ISO/IEC 27001 and HIPAA standards, guaranteeing secure handling of sensitive data.
• Extensive monitoring. The solution runs on 40+ microservices and 400+ transfer agents, all tracked through robust monitoring tools (Amazon CloudWatch, AWS Systems Manager Incident Manager, Prometheus, Grafana, Loki) for health and performance insights.
• Turn-key approach. This project's delivery model gave the customer total control over budget and scope.