To the surprise of no one who has attended medical imaging conferences in recent years, artificial intelligence (AI) again figures prominently in the exhibition area at this year’s congress. Key industry trends include an increasing focus on bringing AI into the clinical workflow for radiologists and using AI to improve image acquisition and processing.
TeraRecon is highlighting its AI Sync and AI-Key-Workflows technologies. A software enhancement to its EnvoyAI platform, AI Sync prepares results and presents them for verification before, during, or after radiologist interpretation, enabling delivery of interactive findings into the report and other information systems.
In addition, the firm is showcasing three versions of its workflow offerings: Stroke and trauma AI Key-Workflow; Key-Workflow for Advanced EVAR measurement protocols; and Cardiac MR AI Key-Workflow. All are designed to provide near-zero-click iNtuition workflows and clinical experiences that are accelerated by AI-enhanced functionality and automation, the company said.
Elsewhere, Philips Healthcare is unveiling the IntelliSpace AI Workflow Suite, which is designed to help providers integrate radiology AI technology into the imaging workflow. The platform features a set of applications for integration and centralised workflow management of AI algorithms, according to the vendor. Initial algorithm partners for this launch include Aidoc, MaxQ AI, Quibim, Riverain Technologies, and Zebra Medical Vision.
Philips is also detailing how AI technology such as its VitalEye, VitalScreen, and SmartExam applications can contribute to workflow improvements in MRI. In addition, the company is focusing on how intelligence and analytics capabilities embedded in its next-generation enterprise imaging system can support clinical and operational workflows.
At its booth, Canon Medical Systems is introducing AI-driven workflow and clinical enhancements based on its Automation platform. For example, the vendor is demonstrating the ability to automatically analyse images and tags and categorise exams based on anatomical information. In addition, the firm will show ‘smart’ algorithms for detecting, segmenting, and characterising abnormal conditions, as well as the use of AI to deliver prioritised results and alerts, according to the vendor.
The company is also emphasising its Collaborative Imaging strategy for bringing multiple imaging modalities and clinical applications together with integrated advanced visualisation, AI, and analytics. This approach features ‘intelligent’ image acquisition, as well as image processing with integrated advanced visualisation algorithms to enable detection, classification, and characterising of conditions at the point of image acquisition, according to the vendor.
The initiative, which aims to deliver holistic and optimised patient information to healthcare providers at the point of care, also includes image and information sharing across networks, as well as analytics for detailed analysis of imaging data throughout the organisation, the company said.
Siemens Healthineers is presenting two additions to its AI-Rad Companion line of software assistants. AI-Rad Companion Brain MR for morphometric analysis automatically segments MR brain images, measures brain volume, and marks volume deviations in tables for diagnosis and treatment, according to the vendor.
Meanwhile, AI-Rad Companion Prostate MR for Biopsy Support segments MR images of the prostate so radiologists can mark lesions and the organ’s outer contour for more precise procedures and enhanced patient outcome, the company said.
GE Healthcare is presenting Edison Open AI Orchestrator, an algorithm management system designed to integrate clinical applications into the radiology reading workflow. Part of the Centricity Open PACS AI package, it also includes Centricity PACS, Universal Viewer V7, and optional AI-based clinical applications from GE as well as third parties.
The vendor is also spotlighting several of its applications, including its work-in-progress AIR Recon DL deep-learning technology designed to improve signal-to-noise and image sharpness in MRI while shortening scan times. In CT, it is pointing to the Auto Positioning technology, which employs a deep-learning algorithm to better position patients. The deep-learning image reconstruction engine – available on the Revolution Evo Gen 3 CT scanner – employs a deep neural network to generate what the company calls TrueFidelity CT images.
The work-in-progress Critical Care Suite is a collection of triage algorithms embedded on the Optima XR240amx mobile x-ray device. One algorithm flags cases with suspected pneumothorax for radiologists to triage, while others – such as Intelligent Auto Rotate AI, Intelligent Field of View, and Intelligent Protocol Check – provide automated quality checks and workflow improvement features for technologists.
Ziosoft is showcasing a multimodality suite of advanced visualisation and postprocessing applications on a vendor-neutral AI platform that includes PhyZiodynamics in Vivo, a 4D processing system that enables the automated registration of DICOM-based images into true fidelity 3D and 4D organ datasets. These datasets can then be interactively interrogated throughout the entire structure while still or in motion, according to the vendor.
Sectra is displaying several offerings, including integrated tools in its unified diagnostic workspace and the latest developments in its marketplace for AI applications. Other featured technology includes new lesion detection and quantification capabilities, as well as work-in-progress AI-powered display protocols.
At the Fujifilm booth, the focus is on the REiLI medical imaging and informatics AI initiative, which it introduced at last year’s congress. The company is showing a range of applications, including Region Recognition for recognising and extracting organ regions, computer-aided detection for lowering image interpretation time and supporting clinical decision-making, and Workflow Support for optimal study prioritisation, alerts for AI findings, and automated report population.
Meanwhile, a key theme for Kheiron Medical Technologies at ECR 2020 is: ‘What is generalisability and why is it essential for Breast AI: Learnings from a multi-centre, multi-vendor U.K. trial’. The firm is organising a workshop and presentations to investigate this topic.
This is just a small sample of the AI technology being displayed this week. A slew of dedicated AI firms are also on hand to share the latest advances in algorithms for a wide range of clinical applications.