Artificial Intelligence in New Cardiac MR Markers for Congenital Heart Disease

Background:

An increasing number of children with complex congenital heart disease (CHD) are today treated and survive into adulthood. They are in need of a lifelong surveillance for cardiovascular complications and an individualised treatment.

In paediatric aortic and pulmonary valve disease, the long-term complications are related to valvular dysfunction with myocardial fibrosis and ventricular dysfunction as a result of longstanding volume or pressure overload. Many patients will need surgery or interventional treatment already in childhood, where the optimal timing of the intervention is crucial, since early treatment may prevent irreversible injury to the heart.

Moreover, to enable detection of recurrent or new problems early, a long-term diagnostic follow-up is needed where cardiovascular magnetic resonance (CMR) has become important for a serial reproducible assessment of morphology, function and recently advanced methods for flow and tissue characterisation. Still the the clinical decision of the perfect time-point and type of treatment for a child, may be difficult in many cases, and better measures are needed.

Aim:

This project will investigate children with aortic and pulmonary valve disease treated or untreated longitudinally using established CMR measures with additional newly developed, promising, highly refined and clinically applicable quantitative imaging biomarkers, as compared to the conventional CMR estimates. Investigators aim to evaluate risk stratification and investigate the haemodynamic and morphological changes as well as physical capacity before and after interventional or surgical treatment at short and long term.

CMR will be analysed for established imaging biomarkers as well as the potential value of new data-driven computer tools searching for new information in the CMR images i.e. shape/texture analysis of tissue change, regional haemodynamics, blood flow in relation to metallic implants and evaluation of myocardial deformation over time.

CMR measures will be compared to data from echocardiography, catheterization and serological markers.

Clinical importance:

If the new refined CMR biomarkers can reveal improved disease characterisation, they may reveal feasible in aiding the clinical decision making to optimise timing of intervention and predicting long term complications. This could contribute to individualise the treatment and improve health in children with CHD. The study may lead to improved patient care and use of socio-economical resources.

Inclusion criteria

• aortic valve disease
• or pulmonary valve disease
• clinical indication for valvular surgery
• or clinical indication for valvular intervention

Exclusion criteria

• contraindications for MRI
• need of sedation or general anesthesia, unless there is a clinical indication for a CMR