Talks

 

Day 1: US Imaging - State of the art and novel acquisition protocols: From simulation to experiments

 

The lectures of the first day were focused on state-of-the-art acquisition methods, how these can be simulated to understand ultrasound physics and novel advances on ultrafast imaging.

 

Olivier Bernard, CREATIS, Lyon

Modeling of ultrasound waves and image reconstruction. The role of simulation in design and validation. (9:00-10:30)

Olivier explained the physical background of echography. He started with the wave equation in 2D and the different phenomena that occur when the sound waves traverse materials with different acoustic conductivity: refraction, dispersion and diffraction. He also explained the fundamentals of Doppler imaging.

 

Mathieu Pernot, ESPCI, Paris

Ultrafast imaging: can ultrasound get the whole heart picture? (11:00-12:30)

 

Mathieu Pernot spoke to the audience about new advances in ultrafast imaging, a very appealing concept given the limited temporal resolution of classical acquisitions, which hinders the understanding of potentially important cardiac events that occur at smaller time-scales.

 

Day 2: US Image Processing - How to extract deformation and flow measurements from US images?

 

The second day was centered on  image processing, focused on blood flow and tissue deformation.

 

Damien Garcia, CREATIS, Lyon

From Doppler to understanding hemodynamics (9:00-10:30)

 

Damien Garcia introduced the audience to Vector Flow Mapping (i.e. vector fields of the displacement of blood cells inside the chambers) and how it can be used for the visual (e.g. vortex formations) and quantitative analysis (e.g. hemodynamic parameters) of ultrasound sequences for the assessment of left-ventricular diastolic function. 

 

Jan D'hooge, KU Leuven

Strain imaging. Should we opt for Doppler or speckle tracking or both? (11:00-12:30)

 

Dr. Jan D’hooge explained how the interference between different sound waves produces speckles that can be tracked during short time steps to compute small displacements. He compared that with Tissue Doppler, which uses the Doppler effect to compute the same property.

 

Day 3: Imaging the individual - How is ultrasound used to diagnose and guide interventions?

 

The third day focused on clinical challenges in diagnostics and interventional ultrasound imaging.

 

Eric Saloux, Centre Hospitalier Universitaire de Caen

Clinical relevance of ultrasound to understand the patient's hemodynamics  (9:00-10:30)

 

Eric Saloux introduced the audience to the main hemodynamic measurements in clinical practice, how to use ultrasound imaging to obtain non-invasive quantifications and commented on their reliability against the ground-truth invasive methods.

 

Marta Sitges, Hospital Clínic de Barcelona

Clinical challenges in interventional imaging (11:00-12:30)

 

Dr. Marta Sitges discussed the challenges of interventional imaging using echocardiography. Concretely, she walked the audience through the various cardiac interventions, such as transapical valve implantation for the aorta, mitral valve repair, mitral/tricuspid clip technology, and left-atrial appendage closure.

 

Day 4: Imaging a population - How to design clinical trials? When population analysis meets machine learning.

 

The fourth day of the summer school was dedicated to how to leverage the knowledge obtained from trials, clinical practice and available data to understand markers of pathology, and create consensus among the experts.

Alan Fraser, University Hospital of Wales, Cardiff

Ultrasound in decision imaging and clinical trials. (9:00-10:30)

 

The thought-provoking lecture by Dr. Alan Fraser showed an audience of mostly engineers the achievements and challenges of evidence-based medicine, with a focus on ultrasound in cardiology. Despite efforts for a practice based on scientific results, many of current parameters in international guidelines are not directly supported by clinical trials. Instead, the appropriateness of a parameter is often measured in terms of the expert consensus.

 

 

Sergio Sanchez, Universitat Pompeu Fabra, Barcelona

Machine learning for patterns analysis in US-based clinical studies (11:00-12:30)

In his talk, Sergio Sanchez introduced the audience to machine learning and showed a comprehensive set of examples of applications in cardiac imaging data.

 

Hands-on sessions

 

Regarding the hands-on sessions, three practical workshops were proposed to offer a unique training opportunity.

Olivier Bernard, CREATIS, Lyon

US Image formation and reconstruction

 

In this hands-on session Olivier prepared daily exercises to learn about the formation of US images by simulating different probes.

 

The first day we learned the importance of the different physical parameters of the probes such as the number of piezoelectric elements, the distance between them (pitch), the frequency of the images or how the delay is computed. A matlab code was provided where the wave equation was solved for the problem of reconstructing an image from energy sent from a probe. By tuning the parameters inside the wave equation, participants were able to understand their influence on the directivity of the transmission and the formation of grating lobes.

 

The second day, participants performed the reconstruction of ultrafast ultrasound images based on the emission of steered plane waves generated by linear transducers. It helped us understand the different steps involved to reconstruct a B-mode image. The matlab code was again provided so that we could understand which were the most important parameters to have a clean plane wave that would allow us to reconstruct ultrafast ultrasound images.

 

The last day we worked with raw ultrasound signals measured in the CREATIS laboratory (www.biomecardio.com). It was useful to understand the nature of ultrasonic RF (radio frequency) signals used in medical imaging. Participants also observed how to create an ultrasound image and measure Doppler velocities from ultrafast ultrasonic imaging.

 

Todo: keep editing with the information from the laptop. tomorrow.

 

Jan D’hooge,  KU Leuven

Strain Imaging

In this session, Jan D’hooge addressed the basics of speckle tracking, a popular approach for motion/deformation estimation in cardiac ultrasound. Two types of methods were discussed: those based on block matching and optical flow. Awareness was drawn to critical steps when putting the algorithms in practice, for example, the definitions of kernel size, search-window size, similarity metrics, filtering and regularization in the block matching algorithm. Then, the participants had the opportunity to implement their own matlab/python version of the block matching algorithm, and test it on synthetic and real echocardiographic sequences for the estimation of the myocardial deformation along the cardiac cycle.

 

Ada Doltra,  Hospital Clínic de Barcelona

Clinical cases

Dr. Ada Doltra explained how cardiac echocardiography measurements are obtained in daily clinical practise. She started by describing the standard echo views, and which are the common pathologies that can assessed through echo. The participants learned how to obtain the distinct views in practice in healthy volunteers, how to identify the anatomical landmarks and healthy behaviour, and how to evaluate the common markers of dysfunction. In the next session, Dr. Doltra discussed several interesting cases, and the participants could observe the differences in function and geometry between patients and healthy subjects. In particular, the participants investigated the images of normal athlete hearts, and compared them with pathological cases, stressing the importance of the integration of clinical context and other sources of information to reach a diagnosis. Through the course of the next sessions, the participants examined images of several pathologies including mitral or tricuspid regurgitation, heart failure due to increased myocardial stiffness or thickness, right-ventricular dysfunction due to increased pulmonary afterload, and heart failure with preserved ejection fraction. Dr. Eric Saloux also showed us how to use color Doppler and 3D echo to further evaluate the degree of tricuspid or mitral regurgitation, and how leverage the potential of the echo parameters to improve the image quality for a given objective. Finally, the participants observed the adaptation of the heart to increased output through an exercise test. Echo images were acquired while the volunteers maintained a steady handgrip movement, followed by the recovery the baseline. Finally, a hand-grip stress test was performed on a volunteer, so as to illustrate the acute adaptation of the heart to exercise.

 

A total of 30 researchers (excluding the invited speakers) took part in the winter school. In addition to researchers working at UPF and Philips, the winter school attracted students from several other local and international organisations (Centre Hospitalier Universitaire de Caen, Ecole Supérieure de Physique et de Chimie Industrielles, Institut National de Recherche en Informatique et en Automatique, Univ. Pierre et Marie Curie, Institut National des Sciences Appliquées de Lyon and L’Institut de Rythmologie et Modélisation in France; Katholieke Universiteit Leuven, University College London, Skolkovo Institute of Science and Technology, Technical University of Munich, Charité University Hospital Berlin and German Cancer Research Center from the rest of Europe).

 

 

Survey on the CardioFunXion winterschool 2017

A survey was conducted to obtain the feedback of the participants and invited speakers regarding the winter school. The following questions were asked:

Q1: How did you value the morning sessions?

Q2: How did you value the afternoon sessions?

Q3: How valuable are such events for your scientific training?

Q4: Will you encourage colleagues in enrolling for similar events?

Q5: Are you keen to encourage your colleagues to watch the recorded sessions?

The possible answers ranged from 1 to 5 where 5 where 5 = best. The feedback of the participants and invited speakers was very positive, most of them having considered this type of event of great value for their scientific training.

 

 

Moreover, participants were asked the components of the winter school that pleased them the most and the least, as well as what other topics they would like to see covered, so that we can take that into account in the organization of a future edition.