Apply
This is the page description.
— Apply —
Clinical knowledge collection and validation
Background
Both Leuven University Hospital and Oslo University Hospital (OUH) have been pioneers in the field of strain imaging by speckle tracking echocardiography. Among the many clinical research areas where strain-based diagnostic indices have provided new mechanistic insights are dyssynchronous heart failure and its therapy with cardiac resynchronization, myocardial ischemia and infarction, and ventricular arrhythmias. Both hospitals possess significant amounts of clinical (strain) data which – combined with the database of the partner Jessa Hospitals in Hasselt – will results in a database of over 30.000 cases that can be used for MARCIUS.
Approach
In MARCIUS, we plan to set up and curate a single large clinical database by combining the data available at the different partner institutions (similar to what was done before between Leuven University Hospital and Oslo University Hospital for cardiac resynchronization patients). The quality of the (clinical) strain annotations and the patient outcome data will be adjudicated under supervision of senior clinical experts (and redone if necessary) and will be extended by other clinically available imaging studies that could confirm/strengthen the outcome data (CMRI, X-ray angiography) (ESR6). The latter will facilitate validation of the computer-aided diagnostic DNN’s as misclassified patients can then more easily be verified in terms of the imposed clinical outcome labels. In MARCIUS, patients with ventricular dysfunction as a result of myocardial infarction and/or left bundle branch block will be targeted. Importantly, this clinical database will not only be used for training the analysis and diagnosis DNN’s but also for validation of the DNNs and for generating new insights into the pathophysiology of these complex cardiac diseases. At the end of MARCIUS, we will be in the best possible position to design a prospective study aiming at independent validation of the technology developed, as well as a companion in silico trial to validate the potential of virtual patients to introduce a paradigm shift in the current environment of clinical testing activities. However, the execution of the prospective study falls outside the scope of MARCIUS and will require separate funding
Outcomes & Synergies
Computer-aided echocardiographic data analysis in a large cohort of well-documented patients with dyssynchronous heart failure will improve the diagnostic accuracy of this complex disease. Moreover, the DNN’s will enable linking the image data (or the parameters on regional cardiac tissue properties automatically derived therefrom) to the underlying myocardial substrate (e.g. contractility, fibrosis, conduction delay) thereby generating new pathophysiologic insights, paving the way to optimization of existing treatment options, development of entirely new therapies, and more advanced and accurate monitoring of treatment effects during follow-up.
