Vascular and Molecular Imaging (VMI)

IMAGIO: Imaging and Advanced Guidance for Workflow Optimization in Interventional Oncology

The goal of this IHI EU project is to develop tools for an end-to-end solution for patient selection, treatment planning, needle guidance and confirmation of technical success. Furthermore, the workflow should be simplified for thermal liver ablations and the effectiveness and clinical outcome of thermal liver tumour ablations will be evaluated.

ASSIST: Automation, Surgery Support and Intuitive 3D visualization to optimize workflow in image guided therapy systems to simplify clinical workflow during image guided therapy procedures

The goal of the ASSIST project is to develop and integrate AI-based solutions for precision diagnosis, personalized treatment, intuitive 3D visualizations and robotic assisted interventions. A major improvement for the diagnosis and treatment planning is the combination of different imaging modalities (CT and MRI) through image fusion and feeding this data to a multi-modal deep learning network. The innovations will be in the advanced fusion of multi-modal imaging and the development of multi-modal deep learning networks for segmentation and characterization of tumor regions. The proposed solution will improve the accuracy of tumor detection, especially in identifying smaller tumors at an early stage.

After segmentation of the tumors and detecting the surrounding major blood vessels, this information is combined into 3D models for treatment planning. The innovative part is that the surgeon will see the model in 3D, which will help getting a better understanding of the distance between the tumor regions and the surrounding blood vessels and organs at risk. It also will provide a proposal for resection planes. This information will be available to the surgeon during robotic interventions.  The proposed solution will speed up both the treatment planning and the procedure itself since it gives better insight to the surgical team and thus allows for easier communication.

For this project, we also look into the possibilities for federated learning and synthetic data generation.

IWISH: Intelligent Workflow optimization and Intuitive System interaction in Healthcare

The goal IWISH is to work on a solution to realize automated data gathering combined with AI-enabled data analytics applications to deliver valuable insights about the clinical workflow. Consequently, these insights will be used to realize new or improved software applications for optimal workflow scheduling, and services to optimize the user experience of existing imaging equipment and surgical devices. The solutions will be based on innovation challenges in the areas of (1) automatic procedure tracking, (2) procedure scheduling optimization and (3) data driven usability research. The development of applications in these three innovation areas will rely on Artificial Intelligence and data analytics as a powerful method that offers transformative potential in healthcare applications, where (4) synthetic data generation tools will be developed to address data accessibility and quality issues.

SYMPHONY: An Orchestrated Open IT Eco-System enabling End-to-End Solutions for Disease Specific Clinical Decision Support

Goal of the ITEA project is to create an ecosystem that will empower medical professionals to have all relevant data for a particular patient at their disposal based on disease-centric and intuitive applications.  It addresses the yet unresolved challenges of the previous Harmony proposal including data  interoperability, data aggregation and clinical workflow and adds new element to cater for AI Based Inference workflows and Result Representation.

The role of the LUMC is developing AI supported tools which support the optimization of the Aortic Aneurysm workflow by developing tools for segmentation, data exchange and visualization

PAN-HI-AI: Improving early detection of PANcreatic cancer in HIgh-risk individuals through Artificial Intelligence methods

The goal of this project is to develop machine learning and deep learning techniques for the early detection of pancreatic tumours in MRI images. There is a need to improve the modalities used to screen these patients, as only a minority of tumours are discovered as high grade precursor lesions or early-stage cancers. Artificial intelligence techniques applied to magnetic resonance imaging, specifically radionics and deep learning models, could aid by complex pattern recognition of subtle changes associated with progression to malignancy in these patients, enabling intervention at an earlier stage.

QCU-CMS: Quantitative intravascular measurements.

The goal of the QCU-CMS project is to provide research tools to researchers with quantitative and qualitative information from Intravascular Ultrasound pullbacks (IVUS) and Intravascular Optical coherence Tomography pullbacks. In is a long running software project for which involves contour detection, stent detection, 3D reconstruction, and image registration and tissue classification. For this, the echogenicity algorithm for IVUS has been extended and the attenuation and backscatter algorithms have been implemented for OCT. At the moment, parts of the existing algorithms are being updated to deep learning approaches, which allow for more automatic analysis of the images. Several clinical research groups in the field of interventional cardiology collaborate on the development of QCU-CMS.