Alexandros Kontogiannis has been awarded a National Fellowship in Fluid Dynamics to extend the capabilities of magnetic resonance imaging (MRI) scanners, by generating high-resolution imaging of blood flow using Digital Twin methods developed during his PhD.
Digital Twins can enable patient-specific cardiovascular modelling, reduce patient scan times, replace invasive techniques, and permit the imaging of smaller vessels such as those found in neonatal and fetal cardiology.Alexandros Kontogiannis
These quantitatively-accurate computer simulations can enable patient-specific cardiovascular modelling and treatment planning, for example, and could ultimately lead to increased adoption of four-dimensional (4D) flow-MRI by clinicians in order to detect early-stage cardiovascular disease.
Alexandros is one of 11 new postdoctoral research fellows funded by the Engineering and Physical Sciences Research Council (EPSRC) with an investment of £3.7 million. The fellows will work together to solve common research problems in fluid dynamics and drive industrial and academic collaboration across the UK.
Flow-MRI scans depict blood flow velocity but they are ‘noisy’ and have limited resolution in space and time. Digital Twins can enable the high-resolution imaging of flows and provide quantitative estimates of haemodynamic biomarkers (used to diagnose cardiovascular disease) that are difficult or impossible to measure otherwise.
Alexandros said: “During my PhD, I developed a mathematically rigorous framework that automatically generates the most probable Digital Twin by assimilating ‘noisy’ and sparse flow-MRI data into the physical laws of fluid motion.
“Unlike AI algorithms, this method is interpretable, has high explanatory power (can reconstruct velocity fields from ‘noisy’ and sparse data), and can extrapolate to new flow conditions.”
Alexandros’ research focus lies at the intersection of fluid dynamics with applied mathematics. He is particularly interested in inverse problems in fluid dynamics (flow reconstruction), and adaptive (multiresolution) methods such as wavelets (the building blocks of convolutional neural nets). Wavelets are a sophisticated mathematical tool that can be used to efficiently reconstruct details of the flow, learn patterns in turbulent flow, and crack the computational complexity of fluid flow simulations, thus accelerating the process of Digital Twin generation.
“Over the course of the next three years, I will be extending the Digital Twin methods that I have developed from simple steady flows in rigid geometries to time-varying flows in flexible geometries,” he said. “I will also study and develop multiresolution methods to reduce the computational complexity of fluid flow simulations, and scope out challenges posed by in vivo cardiovascular haemodynamics (the dynamics of blood flow).
“My ultimate goal is to introduce Digital Twins for 4D flow-MRI. Digital Twins can enable patient-specific cardiovascular modelling, reduce patient scan times, replace invasive techniques such as cardiac catheterisation (the insertion of a catheter into a chamber or vessel of the heart), and permit the imaging of smaller vessels such as those found in neonatal and fetal cardiology.
“By formulating adaptive Digital Twin algorithms, these can be used to accelerate academic research in fluid dynamics and flow-MRI. In medical imaging (e.g. cardiovascular and cerebrovascular flows), these methods will enable patient-specific modelling and treatment planning and, if successful, will lead to increased adoption of 4D flow-MRI by clinicians in order to detect early-stage cardiovascular disease.”
As part of the Fellowship, Alexandros will collaborate with the other new fellows in the upcoming Data in Fluids Summer Programme, due to take place in Cambridge, from 10 July to 18 August 2023.
Professor Matthew Juniper, Principal Investigator, Hub for the National Fellowships in Fluid Dynamics, and Alexandros’ PhD supervisor at Cambridge, said: “The UK is a world leader in fluid dynamics research, which is central to numerous societal and industrial challenges.
“These fellowships have grown out of sustained investment in this crucial discipline. They will help the UK to remain a global leader, while also looking to a future that combines recent data-driven methods with centuries of accumulated physics-based knowledge.
“Fluid dynamics research has the potential to grow the UK economy and help us address some of the key challenges of our age, from health and medicine to industrial development and sustainability.”
Dr Rachel Bishop, Deputy Director for Research Base at EPSRC, said: “We are delighted to be able to give some of the brightest young engineers and scientists the chance to develop their research careers in fluid dynamics. The National Fellowships in Fluid Dynamics are a great way to bring together academia and industry to give new momentum to a vibrant sector.”
We asked Alexandros: What is it that you love about fluid dynamics?
He said: “Fluid flows, be they biological flows (e.g. blood flow); aerodynamic flows (e.g. flow over aeroplane wings); or astrophysical flows (e.g. gas flow in star formation) are quite satisfying to observe in nature and really challenging to model and simulate in silico.
“In physics, turbulence, the seemingly random and chaotic fluid motion, remains an old and unsolved problem of great importance for science, engineering and many industrial applications.
“In maths, proving that the dynamic laws of fluid motion produce a unique and well-posed solution is one of the unsolved Millennium Prize Problems (seven well-known complex mathematical problems selected by the Clay Mathematics Institute).
“These unanswered questions in fluid dynamics drive, and will continue to drive, the development of new physics and maths for the years to come.”
Adapted from a UKRI news release.