-
Gordon Haff authoredGordon Haff authored
Edge Medical Diagnosis
Updated November 2023.
Some details will differ based on the requirements of a specific implementation but all Red Hat architectures generalize one or more successful deployments of a use case.
Use cases
-
Accelerating medical diagnosis using condition detection in medical imagery with (artificial intelligence/machine learning) AI/ML in edge computing at medical facilities
-
Specific use case of skin cancer detection
Background
The current detection process for many types of medical conditions, including skin cancers, involves imaging. For example, photographing suspect or at risk patches of skin is becoming common to be able to check their development over time.
AI lends itself to being effectively a medical assistant in medical imaging in a variety of ways, including:
-
Automating tasks: AI can be used to automate tasks such as image segmentation, which is the process of identifying and isolating different parts of an image. This can free up radiologists to focus on more complex tasks, such as interpreting images.
-
Detecting abnormalities: AI can be used to detect abnormalities in medical images that may be invisible to the human eye. This is especially useful for detecting small or subtle changes that can be early signs of disease.
-
Personalizing treatment: AI can be used to personalize treatment for patients based on their individual medical imaging data. This can help to ensure that patients receive the most effective treatment possible.
-
Providing clinical decision support: AI can be used to provide clinical decision support to radiologists and other healthcare professionals. This can help them to make more informed decisions about patient care.
There are challenges to using AI in this manner such as explainability of results, the use of training data while maintaining confidentiality, and algorithm validation. Nonetheless, AI has the potential to make a significant impact on medical imaging. As the technology continues to develop, we can expect to see even more innovative and beneficial applications of AI in this field.
Solution overview
This architecture, for which Figure 1 provides an overview, covers the use case around Edge Medical Diagnosis in the healthcare industry and drills down on skin cancer detection in particular. The solution provides a way to increase efficiency of a medical facility by reducing time spent on routine diagnosis and medical diagnosis of disease. This gives medical experts more time to focus on the difficult cases.
Why Edge Medical Diagnosis?
-
Need for faster medical diagnosis of diseases
-
Provide expert-assist to medical experts to reduce time spent on routine diagnosis
-
Focus medical expert time on most difficult to interpret cases
Figure 1. Overview of Edge Medical Diagnosis architecture.
Logical diagram
Figure 2 shows a logical diagram of the solution including the frontend application, the intelligent services, storage, and deployment and management tools.
Figure 2. Logical diagram of Edge Medical Diagnosis solution.
The technology
The following technology was chosen for this solution:
Red Hat OpenShift is a unified platform to quickly build, modernize, and deploy both traditional and cloud-native applications at scale. It is packaged with a complete set of services for bringing apps to market on your choice of infrastructure. It’s based on an enterprise-ready Kubernetes container platform built for an open hybrid cloud strategy. It provides a consistent application platform to manage hybrid cloud, public cloud, and edge deployments. Try It >
Red Hat OpenShift Serverless provides event-driven functions and scales up based on an event trigger. The application scale down to zero for resource optimization, while starting up with minimal bootstrap time when it is required. Try It >
Red Hat OpenShift GitOps automates the deployment of the edge medical diagnosis elements, pick up changes from code repository into the CI/CD pipelines, and triggers image builds and deploys into clouds. It’s delivered via an operator through the Operator Hub.
Red Hat OpenShift Data Foundation is software-defined storage for containers. Red Hat OpenShift Data Foundation helps teams develop and deploy applications quickly and efficiently across clouds. For the purposes of this solution, it provides storage services for medical imagery, continuous deployment models, analytics, and AI/ML datasets and models, allowing them to be provisioned across multiple cloud environments. Diagnosis models can be continuously trained and updated, resulting in a streamlined workflow for a more rapid, agile application lifecycle. Try It >
Red Hat AMQ is a lightweight messaging platform for real-time integration. Based on open source communities like Apache ActiveMQ and Apache Kafka, it reliably and scalably delivers information between distributed endpoints.
Red Hat Ansible Automation Platform provides an enterprise framework for building and operating IT automation at scale across hybrid clouds including edge deployments. It enables users across an organization to create, share, and manage automation—-from development and operations to security and network teams. Try It >
Red Hat Enterprise Linux is the world’s leading enterprise Linux platform. It’s an open source operating system (OS). It’s the foundation from which you can scale existing apps—and roll out emerging technologies—across bare-metal, virtual, container, and all types of cloud environments. Try It >
IBM Storage Ceph is a software-defined storage solution for block storage, file storage, and object storage used for images, continuous deployment models, analytics, AI/ML datasets and models.
IBM Watson Studio develops, trains, and tests for AI/ML modeling and visualization in sandbox environment. Diagnosis models are being continuously trained and updated, this streamline workflow allows a more rapid, agile application lifecycle.
Architectures
Figure 3 provides a network-centric view of the general Edge Medical Diagnosis solution. Figure 4 shows into GitOps deliver and deployment. Figure 5 looks specifically at the skin cancer detection use case which includes IBM Watson Studio technology.
Edge Medical Diagnosis overview
Figure 3. Network-centric schematic view of Edge Medical Diagnosis architecture.
The solution shown in Figure 3 has two distinct locations: the diagnostic facility where the medical staff and the edge x-ray devices are located and the medical data center where development and monitoring of the solution takes place.
Initial images are sent into the diagnostic facility image receiver and an event is registered to start the processing for automated diagnosis. These images are stored locally, anonymized, and automatically evaluated for possible disease detection. A notification is generated for the medical staff, whether automated detection, non-detection, or an edge case requiring qualified medical staff review. (In many cases, medical staff will want to give at least a quick review of detection and non-detection cases as well.)
In the process of image capture and processing, the images are sent back to the medical data center to be added to the collection used for model training and development. The applications, machine learning models, data science development, and dashboards for monitoring the processes are all in constant evolution. Developers and operations teams maintain code and infrastructure manifests for full GitOps deployment of the architectural elements.
Edge Medical Diagnosis with GitOps
Figure 4. Schematic view of Edge Medical Diagnosis architecture with a focus on GitOps delivery and deployment.
GitOps delivery and development are essential to a fully automated cloud hosted solution. Figure 4 features the elements focusing only on development and deployment of the Edge Medical Diagnosis elements needed for this solution. It does not show the patient-facing medical staff and the edge image capturing; rather it features developer and IT operations staff on the back end.
In the medical data center, developers deliver code projects into the CI/CD pipelines and trigger eventual container image builds that are placed into the registry. The same thing happens on the IT operations side, where system configuration and manifest code is maintained in their repository.
The developer image registry is replicated out to the image registry in the remote diagnostic facility and the source code repository for IT operations is also replicated out to the remote location. These both are setup to trigger the GitOps pipelines to sync updates to the image registry and the operation’s source code repository to the OpenShift platform. This means it’s deploying, configuring, and applying manifests to the applications and services used to process the medical diagnosis imaging solution.
Skin Cancer Detection use case
Figure 5 shows a Skin Cancer Detection use case that includes IBM data science and storage components.
Figure 5. Schematic view of Skin Cancer Detection architecture that includes IBM data science and storage components.
The schematic in Figure 5 focuses on the datacenter side of the system. This specific focus is needed since the images can be acquired in different ways from the various devices that can be used to acquire such images. In case those devices are connecting to smartphones or tables, an application can be installed on those devices. If those devices communicate with a computer, it can host such an application.
The chosen device connected to the image acquirer tool transfers the images to the Image Upload Application. The Image Upload Application saves the image metadata in a database; the image itself is saved in an object storage provided by IBM Storage Ceph. The database is backed by IBM Storage Ceph block storage. Additionally, the Image Upload Application will also place a message into AMQ to ensure the image will be processed.
IBM Watson watches the AMQ (Kafka) incoming-images queue and processes incoming images; it returns the result to the doctor via the notification service.
The doctors' diagnoses based on the images and the results of biopsies, where available, can be used to retrain the AI/ML model periodically to improve the accuracy and the precision of the model. The applications, machine learning models, data science development, and dashboards for monitoring the processes are all in constant evolution. Developers and operation teams maintain code and infrastructure manifests for full GitOps deployment of the architectural elements. The installation and management of all components in the environment is done with automation using Ansible, which helps create a predictable and auditable enviornment.
Provide feedback on this architecture
You can offer to help correct or enhance this architecture by filing an issue or submitting a merge request against this architecture product in our GitLab repositories.