RaySearch Containerizes Cancer Treatment Solution on Amazon ECS

Stockholm-based RaySearch is a leading provider of software to advance cancer treatments. The company works closely with oncological clinics around the world, with more than 700 customers in 40 countries. Clinics rely on RayStation, the company’s on-premises treatment planning solution, to generate and evaluate radiation therapy treatment plans. After releasing RayIntelligence, RaySearch's oncology analytics system on Amazon Web Services (AWS), the next step was to integrate its RayStation system and make it easier for clinics to derive insights from their patient and treatment data. Windows containers on Amazon Elastic Container Service (Amazon ECS) was the answer.

RayStation is a treatment planning system typically deployed on premises to help clinicians and oncologists create and evaluate a patient’s radiation therapy treatment plans. Anonymized data from previous patients provide insights about the efficacy of different treatment levels. Using RayStation, clinics extracted this data, but the speed and scalability of their analysis depended on their on-premises IT infrastructure, which varied significantly. With RayIntelligence on AWS, RaySearch was able to leverage the speed and scale of the cloud to help clinics seamlessly turn data into insights. The challenge was bringing RayStation along, ensuring the Windows application was integrated with the new cloud-based analytics solution.

RaySearch containerized the data warehousing functions of RayStation with Windows containers on Amazon ECS, creating a scalable, lightweight agent that sits on the clinical side and extracts oncological data, then loads it into RayIntelligence. Once ingested, the data is converted to JSON or CSV files, making it easier and faster for clinicians to obtain insights from the data generated in RayStation. Now, clinics don’t need on-premises computing power to perform high-intensity tasks like processing oncological data from multiple patients. With an Amazon ECS architecture, they can ingest multiple data streams in parallel.

It was clear for RaySearch that Amazon ECS was the best path forward, where other options would have required at least a year of upfront development, cutting into go-to-market timelines. And because Amazon ECS is a fully managed service, RaySearch was able to reallocate resources from infrastructure management to product development. This shift has helped RaySearch accelerate the frequency of software RayIntelligence updates from semi-annually to quarterly. “Containerizing RayStation on Amazon ECS has been a game changer in terms of our ability to move forward with an existing solution, rather than building from the ground up,” said Dror Asaf, software architect and team lead at RaySearch.

More than 700 clinics around the world use RayStation, with every clinic analyzing data on a different schedule—some on a continual basis, some daily, and some weekly. Running RayStation on Amazon ECS has enabled RayIntelligence to easily ingest data streams of any size, from predictably small chunks coming throughout the day, to bulk quantities that arrive once a week. And the onboarding of data from new clinics, which used to take around 2 days, can now be completed in less than 10 hours. “Regardless of when there’s a surge of data processing activity, the Amazon ECS clusters easily and automatically scale to ingest the workloads,” Asaf said.

Because the AWS compliance program aligns with strict guidance set forth in HIPAA, HITRUST, and General Data Protection Regulation, RaySearch can focus on developing software solutions that advance cancer treatment. The company’s product roadmap currently includes a list of features that stretch out well into the future. “We operate in a highly regulated industry where data protection is absolutely essential,” Asaf said. “Out of the box, Windows Containers on Amazon ECS offered a level of data portability and traceability that was fully compliant.”