We use cookies to help provide you with the best possible online experience. Learn more

Machine learning is one of the fastest-growing areas of technology today. It has had a key role in advances in many fields, and its significance for the future of healthcare is potentially enormous. RaySearch already has a strong focus on automation and machine learning brings this to a new level. Through machine learning, smarter and faster software is created and automatic treatment plan generation* and deep-learning organ segmentation* are the first applications in RayStation 8B*.

*Subject to regulatory clearance in some markets.

Key Features

  • Generate contours of organs in less than 45 seconds with deep neural network models
  • Generate personalized treatment plans in minutes 
  • Benefit from trained models from leading cancer clinics
  • Train your own models
  • Share models with other clinics

Watch the movie to learn more.

“Machine learning is rapidly transforming many areas of technology. It will be a cornerstone of healthcare software in the future, and a core element of both RayStation and RayCare.”

Fredrik Löfman
Head of Machine Learning,
RaySearch

The machine learning model deployment process is independent from the RayStation version. This means that machine learning models provided by RaySearch will be added continuously and you won’t need to wait for a new release to access them. Clinics will also be able to train their own models for both segmentation and planning and share models with other clinics. The nature of machine learning makes it possible to share models without the inclusion of personal data and thus creates fantastic opportunities for knowledge sharing between cancer centers. 

Auto-segmentation of organs in RayStation is set to reach new heights with the introduction of deep learning segmentation. The algorithm uses models that have been trained and evaluated on clinical data for different body sites. The GPU-powered algorithm is fast and produces consistent segmentation results.

How does it work? Select a pre-trained deep learning model and the organs are segmented automatically in less than 45 seconds. The output is standard geometries that can be manually adjusted if needed.

WHITE PAPER: DEEP LEARNING SEGMENTATION
Fill in your information and the document will be sent to you.

Your data is handled with discretion, read our privacy statement 

RaySearch has partnered with Princess Margaret Cancer Center to develop world’s first machine learning treatment plan generation module. Clinics can now achieve personalized treatment plans, benefiting from the experience of one of the world’s leading cancer centers, generated in minutes by selecting a pre-trained machine learning model. One or multiple deliverable treatment plans can be automatically generated with varying target/OAR tradeoffs.

RaySearch licenses UHN Automation Technologies

In this video, UHN outlines the impact of the automatic treatment planning innovations it recently licensed to RaySearch, enabling machine learning algorithms from Princess Margaret Cancer Center to be integrated into RayStation.

“Machine learning is a natural fit for automating the complex treatment-planning process. It will enable us to generate highly personalized radiation treatment plans more efficiently, thereby allowing clinical resources or specialist technical staff to dedicate more time to patient care. We know that the RayStation algorithm generates high quality treatment plans that are deemed clinically acceptable by world experts with the majority of cases we have formally studied, showing automated plans are preferred or deemed equivalent to clinical plans.”

 

Tom Purdie
Medical Physicist, Princess Margaret Cancer Center, Canada

WHITE PAPER: MACHINE LEARNING AUTOMATED PLANNING
Fill in your information and the document will be sent to you.

Your data is handled with discretion, read our privacy statement 

MACHINE LEARNING IN RADIATION ONCOLOGY

During the first part of the webinar, Fredrik Löfman, head of machine learning at RaySearch, discuss how RaySearch is pioneering machine learning for smarter and faster oncology software. In the second part, Tom Purdie, medical physicist at Princess Margaret Cancer Center, talks about automatic treatment planning and machine learning from a clinical perspective.

RaySearch will implement both classical machine learning techniques as well as deep learning methods, as they are appropriate for different problem settings. For instance, certain deep-learning approaches are very well suited for extracting information from medical images, while classical machine learning approaches have proven to be suitable for predicting dose for automatic plan generation. RaySearch’s machine learning applications will be developed in close collaboration with clinics to incorporate clinical knowledge into the algorithms.

Data analytics and machine learning will be cornerstones of both RayCare and RayStation, empowering the user by presenting relevant information at the right time and enabling the clinics to make use of their data and to build learning models. This has the potential to enable efficient workflows and highly consistent treatments, where every new piece of data contributes to constantly ongoing improvement.

Large amounts of data are generated for every patient who is treated, and machine learning can help support clinicians to make care smoother and more consistent.

This technology means clinicians can spend less time on repetitive tasks and free up time for patient consultations and complex cases. Furthermore, machine learning models could be shared or co-trained between clinics to help disseminate clinical knowledge between hospitals.

Interview with Fredrik Löfman, Head of Machine Learning at RaySearch