David Marlevi

I am a research lead in quantitative cardiovascular imaging, developing
data-driven image analysis tools to tackle urgent clinical challenges across
the heart, aorta, and brain.
I am a biomedical engineer with a focus on translational cardiovascular
imaging. Specifically, I am intrigued by the translation of image-based
engineering into clinical practice, using data-driven utilities to enhance
diagnosis, improve prognosis, and provide fundamental mecanistic
understanding of cardiovascular disease.
After graduating from the joint doctoral program in Medical Technology from
the Royal Institute of Technology (KTH) and KI with a thesis entitled
"Non-invasive imaging for improved cardiovascular care", I spent two 
years as a postdoctoral fellow at the Massachusetts Institute of Technology
(MIT), funded by a Knut and Alice Wallenberg foundation scholarship and
working under the tutelage of Prof. Elazer R. Edelman (edelmanlab). At 
MIT, I worked on AI-driven image analysis to monitor intravascular
interventions, as well as lead lab efforts on novel vascular drug delivery
systems. In 2021, I returned to Sweden and KI as a research lead in
quantitative cardiovascular imaging, working closely with clinical and
technical fellows at both KI and the Karolinska University Hospital to
translate advanced image technologies into clinical practice. Specific
focuses has been on hemodynamic mapping by full-field phase-contrast magnetic
resonance imaging (4D Flow MRI), with coupled physics-informed analysis
allowing for regional hemodynamic quantifications across the cardiovascular
system.

 
* Early Career Award – Translational Science, Society for Cardiovascular
Magnetic Resonance (SCMR), 2021
* Runner-up, Best presentation in Basic science, 1st Annual Marvin M. Kirsh
Resident Research Symposium, University of Michigan, 2021
* Potchen-Pasariello Award – Best presentation in Clinical Science,
Society for Magnetic Resonance Angiography (SMRA), 2020
* Trainee grant, IEEE Nuclear Science Symposium and Medical Imaging
Conference, 2015
* Travel award, IEEE International Ultrasonics Symposium, 2015
* KTH Best graduate student of the year, KTH Royal Institute of Technology,
2014
* Endeavour Research Award, Australian Government Research Award fellowship
* Henrik Göransson Sandviken scholarship, 2011
* Hjalmar Berwalds minne för framstående matematiska studier, 2010

Non-invasive estimation of cardiovascular pressure gradients:
Regional quantification of cardiovascular pressure gradients is critical for
diagnosis, treatment planning, and risk prediction of many cardiovascular
disease. Still, for a large number of conditions, non-invasive assessment is
obstructed by inherent method limitations, and a wide range of clinical
instances exist where regional pressure behaviour remains unexplored. To
tackle this, we have recently deployed a combination
of physics-informed image analysis (invoking fundamental fluid mechanical
description of blood flow) and full-field flow imaging (4D Flow MRI) to allow
for arbitrary probing of pressure gradients across previously inaccessible
compartments. Here, we seek to extend these utilities to further understand
early hemodynamic changes indicative of later physiological impairement,
including validation, implementation, and clincial utility across spatial
(large / small vessels), temporal (fast / slow flows) and flow (laminar /
turbulent) scales.

 
Super-resolution 4D Flow MRI:
The advent of full-field flow imaging by 4D Flow MRI has fundamentally
changed our ability to interrogate complex hemodynamic behaviour in a direct
clinical setting. However, spatiotemporal limitations exist based on the
clinical time frames in which the systems can be used, obstructing assessment
of regional or highly transient flow events. To tackle this, we have recently
employed deep residual networks to enhance spatial image resolution,
effectively pushing quantitative 4D imaging into challenging intracranial
vessels. Now, we seek to extend the same utilities into temporally
challenging flows such as in the heart, or through complex aortic disease.
Further inclusion of so called physics-informed networks are also expected to
expand clinical impact and versatility of 4D Flow MRI across a wide
cardiovascular application range.