AI-enabled Radiomics Revolutionizes Cancer Research and Treatment on AWS
Guest post by Derek Cooper, Senior Vice President Customer Success, and Nik King, Lead Dev Ops Engineer, HealthMyne
At HealthMyne, we are pioneers in applied radiomics, the cutting-edge field of extracting novel data and predictive biomarkers from medical images. Through our AI-enabled radiomic solutions, we help organizations access and easily translate ground-breaking radiomic insights into use in cancer research, treatment planning, and clinical management. By leveraging radiomics, our clients and partners can evaluate disease progression, monitor therapy response, and determine clinical outcomes.
We believe that every cancer patient’s story starts with an image, and images are more than pictures– they are data. For example, a 69-year-old male lung cancer screening patient was routinely scanned using low-dose CT. A ground glass nodule in the right upper lobe was deemed stable using traditional measurement techniques. When measured and analyzed in HealthMyne using radiomic data, the nodule was determined to be growing. The patient was biopsied, presented with low grade adenocarcinoma, and went on to a successful surgery with a positive prognosis.
As an advanced medical imaging analytics solution, HealthMyne is critically dependent on highly available, scalable, and reliable infrastructure with very strong storage and graphics capabilities. In particular, HealthMyne puts client workstations under intense computing pressure for its image processing needs, including rendering images, enabling image review and analytics, extracting thousands of radiomics metrics per image, and applying machine learning models and deep learning–all in real time. This problem is often exacerbated in the clinical setting by complex radiology and imaging workstations already executing a number of additional software packages, each with unique and intense demands.
To address this need, we sought a highly scalable solution requiring minimal maintenance from HealthMyne and minimal investment from our customers.
After reviewing several solutions from most of the major players in the virtual workstation and imaging industry, we selected Amazon AppStream 2.0 as the primary distribution mechanism for the HealthMyne imaging workstation.
AppStream 2.0 g4dn instances offer the hardware acceleration needed to render medical images at high frame rate and image quality. Running the AppStream 2.0 fleet in “desktop stream view” allows users to run HealthMyne on multiple high-resolution monitors. Utilizing SAML 2.0 single-sign-on (SSO) for authentication, HealthMyne’s customers can log into AppStream 2.0 using their preferred identity provider.
AppStream 2.0 enables HealthMyne to support capabilities we simply did not have previously: On demand access to workstations, the ability to ramp-up and ramp-down as needed, access to GPUs to handle the graphics processing, and ultimately to support our global reach, across the US, EU, and Asia. As a result, HealthMyne has been able to accelerate our delivery of our applied radiomics solutions to our life sciences and clinical customers and help bridge the gap between medical imaging and personalized care. To learn more visit us at www.healthmyne.com.