A typical cardiac ultrasound examination takes 30-40 minutes; additional image analysis and reporting will often double this time. Thus, cardiologists are calling for simplification and automation to optimize clinical workflows towards more time with patients and less time on analysis/reporting tasks.
To bridge the gap between clinical needs and currently available technology forthe assessment of cardiac diseases, there is a need to develop competence in EU. New tools, capable of assisting the user to increased efficiency and accelerated decision making, are needed. Such tools can be crafted by introducing intelligent algorithms that can exploit knowledge from expert users and previously acquired data and learn from this. However, several factors hinder progress:
• Emerging tools for automated image classification cannot go beyond structure recognition in order to also provide cardiac performance- and diagnostic metrics
• Lack of large training databases with curated echocardiographic data linked to pathophysiological findings
• Lack of researchers capable of operating across the integrated fields of ultrasound image processing, machinelearning, cardiology and physiology.
We will build on recent research breakthroughs in machinelearning (deep Learning), which have boosted the performance substantially for speech recognition and image classification. The ambition is to develop tools that will lead to increased diagnostic confidence as well as significantly enhanced operational efficiency.
Through training and knowledge development in a focused research project - to be carried under the MSCA-ITN-EID framework, MARCIUS will provide researchers trained with the cross-disciplinary understanding and skills necessary to develop enabling technologies in an industrial and clinical setting. The longer-term outcome ofthe project is new products, which will benefit patients across Europe and worldwide directly by advancing the state of care within cardiology.