Program

Program NFBI Symposium October 13, 2011:

12:00 – 13:00    Lunch (In front of Lecture Room 5)

13:00 – 13:05    Opening by Max Viergever (Lecture Room 5)

13:05 – 13:25    Holger Grüll (TU Eindhoven)

13:25 – 13:45    Coenraad Hendriksen (NVI/Utrecht University)

13:45 – 14:05    Andrew Webb (LUMC)

14:05 – 14:25    Marcel van Herk (NKI/AvL)

14:25 – 14:45    Kenneth Gilhuijs (ISI – UMC Utrecht)

 14:45 – 15:15    Coffee Break

15:15 – 15:35    Clarisa Sánchez (Radboud University Medical Centre Nijmegen)

15:35 – 15:55    Charl Botha (TU Delft)

15:55 – 16:15    Fedde van der Lijn (Erasmus MC)

16:15 – 16:45    Keynote: Hans Reiber (LUMC)

16:45 – 16:50    Closing by Max Viergever

16:50                  Drinks (Boerhaaveplein)

Abstracts

Abstracts:

13:05 – 13:25

Temperature induced Drug Delivery under MRI guidance

Holger Grüll

Eindhoven University

Diagnostic imaging is indispensible in clinical routine for diagnosis and staging of diseases. Over the last years, diagnostic imaging moved also into the therapeutic setting to guide therapeutic interventions. Examples range from x-ray guided placement of stents, radiofrequency or high intensity focused ultrasound (HIFU) based thermal ablation of tumors to catheter-based embolization procedures. Besides these pure device based interventions, image guided interventions are immerging that comprise a drug formulation. In general the aim is to achieve as high as possible concentrations in the lesion, while sparing the healthy tissue. One example are temperature activatable drug formulations that were proposed already thirty years ago but are only reaching clinical approval. Here, a small molecule drug, doxorubicin, is encapsulated in a temperature sensitive liposome, which prevents the drug from leaving vascular system reducing for example cardiac toxicity. For tumor treatment, the lesion is heated with radiofrequency or high intensity focused ultrasound to temperatures around 42° C. At this temperature, the liposomes become porous and quickly release the drug payload. This rapid release inside the tumor leads to high local concentrations and improves the therapeutic window. Magnetic resonance imaging plays a pivotal role in guidance of thermal ablation procedures, hyperthermia treatment of high focused ultrasound induced drug delivery as it provides morphological pictures with superb resolution and allows real time temperature mapping. The latter provides a feedback to ensure that an adequate thermal dose is applied. For image guided drug delivery, MRI is also able to provide a non-invasive way to quantify the drug concentration in vivo. This talk will give an overview of recent developments in image guided interventions with detailed examples of applications of MRI- guided high intensity focused ultrasound in thermal ablation and drug delivery.

13:25 – 13:45

Animal experimentation and the use of 3R alternatives

Coenraad F.M. Hendriksen

Netherlands Vaccine Institute (NVI) &  Department Animals in Science & Society, Faculty of Veterinary Medicine, Utrecht University, the Netherlands.

Animal models have been and still are instrumental to progress in biomedical research and testing. Effective vaccination programmes such as against poliomyelitis, innovative imaging technologies or R&D of new anti-cancer drugs wouldn’t have been possible without laboratory animal use. This link between animal research and medical progress is reflected in the annual figures on laboratory animal use. The statistics of the Netherlands on laboratory animal use ‘Zo doende’ show a total figure of 692,665 for 2009. Although no detailed figures are available, worldwide animal use is roughly estimated to be about 100 million.

Despite all its benefits for human and veterinary healthcare, animal experimentation is also an area which, for various reasons, is heavily discussed. The consequence of subjecting living sentient animals to painful and/or stressful procedures has resulted in a debate about the moral acceptability of animal experiments and these discussions tend to intensify. Animals are generally used as surrogate models for what we want to learn about human diseases. Data from these animal studies have to be extrapolated to the human situation, a process which is poorly understood. Apart from scientific and ethical limitations, performing animal experiments can also be expensive, labour intensive and difficult to standardise.

A powerful tool for a change was introduced by Russell and Burch in 1959. They were the first to describe the concept of the Three Rs: Replacement, Reduction and Refinement. Replacement means replacing animals by non-sentient beings or material, Reduction means reducing the number of animals in a particular experiment and Refinement is limiting pain and distress to the animals involved and/or optimising welfare of these animals. Although it took several decades, the Three R concept is now widely accepted and is generally known as ‘alternative to laboratory animal use’. It has been the dominant theme in the revised European regulations on animal experimentation (Directive 2010/63/EU) that was adopted recently.

This presentation will discuss the effect of the Three Rs in animal research and testing. General trends with respect to Reduction and Refinement can be summarised in two one-liners: ‘More data, less animals’ and ‘Happy animals make good science’. The contribution of bio-imaging to these Rs will be illustrated. Next, examples will be given of successful strategies to replace animal use altogether.

13:45 – 14:05

Compressed sensing in high field MRI

Andrew Webb

Department of Radiology, Leiden University Medical Center

High field magnetic resonance imaging represents a major investment in Holland, with four ultra-high scanners currently in place or planned. The clinical utility of 7 Tesla MRI has developed very rapidly, with funded applications in neurodegenerative diseases, cardiovascular conditions, and musculoskeletal pathologies. In selected applications, image and spectroscopic data quality are much improved at 7 Tesla compared to studies at clinical field strengths of 3 and 1.5 Tesla, allowing new types of clinically relevant information to be gathered. The higher field strength has been used to acquire data with higher spatial resolution and/or higher tissue contrast. However, the temporal resolution, ie how fast the images are acquired, has in general not been increased, and this is highly significant for clinical studies in which patient scanning time is limited. One of the most promising approaches towards the aim of reducing total scanning time is compressed sensing, in which the Nyquist sampling criterion can be overcome for data which is sparse in a chosen domain. This talk will concentrate on present and future applications of compressed sensing in high field MRI.

14:05 – 14:25

Image guided radiotherapy – achievements and limitations

Marcel van Herk

Radiation oncology department, the Netherlands Cancer Institute / Antoni van Leeuwenhoek Hospital Amsterdam

In the past 25 years, the technology of external beam radiotherapy has improved tremendously. First, due to the development of advanced 3D and 4D imaging such as multislice CT, MRI and PET, tumor localization has greatly improved. Next, the recognition of the presence of large internal organ motion has lead to development of image guidance systems based on 2D, 3D or 4D imaging systems integrated with the medical accelerators. The achievable localization accuracy is so high, that the necessity for invasive fixation has disappeared. For treatment of brain metastases, for instance, localization of the skull with 3D image guidance is accurate to well within one mm. We have developed an image guidance system based on cone beam CT that includes 4D image acquisition and 4D image analysis allowing effective image-guidance for lung tumors.

The main factors that affect the accuracy of treatment after image guidance has been implemented are: uncertainties in target volume definition, the quality of the surrogate (if any) used to localize the tumor, intrafraction movement, and movement that is too complex to be corrected by the image guidance solution (e.g., deformations). These uncertainties may lead to uncertainties in the mm range (for intrafraction motion) to cm’s (for target volume definition uncertainties).

We conclude that, in spite of modern IGRT, there are still uncertainties that need to be covered by safety margins. Margins for intrafraction motion can often be small, the most important uncertainties relate to imaging and biology that are not corrected by IGRT.

14:25 – 14:45

PET/MR decision support systems for non-invasive assessment of breast cancer prognosis and therapy response

Kenneth Gilhuijs

Image Sciences Institute, UMC Utrecht

Approximately 13.000 women in the Netherlands are diagnosed with breast cancer each year. Some cancers will metastasize, leading to patient death, while others will not. Because this distinction currently cannot be made beforehand, 75% of women with early-stage breast cancer may receive chemotherapy while they do not need it, leading to adverse side effects and unnecessary cost.  Moreover, when chemotherapy is given, uncertainty exists as to which kind is most effective in individual patients.

Prognosis: Fully automated systems are being built to analyze MR images of the breast, correlating them with the underlying genotype of breast cancers of metastatic potential. These analyses include automated segmentation, extraction of spatio-temporal features of enhancing lesions and classification into low-risk or high-risk of developing metastases from the primary breast tumor.

Response monitoring: Automated decision support systems are being built to infer whether the chosen type of chemotherapy will be sufficiently effective to eradicate the whole tumor.  For this purpose 18F-FDG PET and MR images of the breast are automatically registered using deformable registration.  Changes in PET/MR features before and during chemotherapy are mapped to a likelihood of complete tumor removal, providing decision support which patients would benefit from switching to an alternative chemotherapy regimen.

15:15 – 15:35

Contextual computer-aided detection: Improving bright lesion detection in retinal images and coronary calcification identification in CT scans

Clarisa Sánchez

Radboud University Medical Centre Nijmegen

Contextual information plays an important role in medical image understanding. Medical experts make use of context to detect and differentiate pathologies in medical images, especially when interpreting difficult cases. The majority of computer-aided diagnosis (CAD) systems, however, employ only local information to classify candidates, without taking into account global image information or the relation of a candidate with neighboring structures. In this talk, I would like to present a generic system for including contextual information in a CAD system. Context is described by means of high-level features based on the spatial relation between lesion candidates and surrounding anatomical landmarks and lesions of different classes (static contextual features) and lesions of the same type (dynamic contextual features). The added value of contextual CAD is demonstrated for two real-world CAD tasks: the identification of exudates and drusen in 2D retinal images and coronary calcifications in 3D computed tomography scans. Results show that in both applications contextual CAD is superior to a local CAD approach with a significant increase of the figure of merit of the Free Receiver Operating Characteristic curve from 0.84 to 0.92 and from 0.88 to 0.98 for exudates and drusen, respectively, and from 0.87 to 0.93 for coronary calcifications.

15:35 – 15:55

Recent advances in interactive medical volume visualisation

Charl Botha

Graphics Group, Delft University of Technology

In 1987, Lorensen and Cline published their landmark paper “Marching

Cubes: A high resolution 3D surface construction algorithm”. In this paper they described an algorithm that was able to extract, very efficiently, triangle mesh isosurfaces from volumetric datasets. In 1988, Levoy published his paper “Display of Surfaces from Volume Data”, and Drebin, Carpenter and Hanrahan then at Pixar their paper “Volume Rendering”, both presenting the idea of direct volume rendering, that is simulating the traversal of light rays through volumetric data to generate 2D images.

Volume visualisation is a textbook example of graphics and visualisation techniques that have gone very much mainstream. Surface extraction and rendering, and also direct volume rendering, often play an important role in interactive medical imaging applications, in research, diagnosis and treatment.

In this talk, I briefly recap direct volume rendering techniques, focusing on the challenges that have been overcome to turn these inherently processing-intensive techniques into interactive tools. I then discuss in more depth two recent advances from the Medical Visualisation group in Delft and its collaborators: First, a new breed of direct volume rendering techniques that integrate heavily GPU-optimised raytraced lighting and are therefore able to generate photo-realistic volume renderings at interactive speeds. Second, a direct volume rendering approach that is able to render in real-time huge correlation matrices in a 3D spatial context, applied to the visualisation of full-brain voxel-wise resting-state fMRI functional connectivity data.

15:55 – 16:15

Brain structure segmentation, Ford Model T style

Fedde van der Lijn

Biomedical Imaging Group Rotterdam, Erasmus MC – University Medical Center Rotterdam

The image processing community has worked for over fifteen years on automated brain structure segmentation in MR images. However, until recently none of these methods were actually being used outside the labs they were developed. Fortunately, this is changing and segmentation methods are increasingly being applied.

Contrary to early expectations, automated brain structure segmentation is rarely used in the clinic, but rather for fundamental research on the anatomy of neurological conditions like Alzheimer’s Disease or schizophrenia. This development is driven in particular by the advent of large-scale neuroimaging studies like the Rotterdam Scan Study and the Alzheimer’s Disease Neuroimaging Initiative, which include several thousands of subjects. Measuring brain structure volume in MR images and relating it to variables like cognitive performance can provide in-vivo information on the origin and development of neurological diseases. Moreover, brain structure atrophy can also act as a biomarker of disease progression and is therefore starting to be used in clinical medication trials. Automated segmentation techniques have become a crucial tool to extract biomarkers from the massive amounts of image data generated by these large neuroimaging studies.

At the Biomedical Image Group Rotterdam of Erasmus MC we have been working for several years on the analysis of MR images from the Rotterdam Scan Study. This presentation will give an overview of our experiences with large-scale brain structure segmentation. What have been the most valuable methods. How have these measurements been used? And how can the results from population studies be introduced in the clinic after all?

16:15 – 16:45

Keynote : Innovation and valorization in cardiovascular applications

Hans Reiber

Division of Image Processing, Leiden University Medical Center/Medis Medical Imaging Systems

The Division of Image Processing (LKEB), Department of Radiology, LUMC has carried out research in image processing over the past 34 years in various domains, congruent with the current LKEB-sections being vascular and molecular imaging, knowledge-guided image processing, magnetic resonance imaging, neuro-image processing and orthopedics and pulmonology. However, the emphasis of the work has always been in the cardiovascular domain. Many of its algorithms and applications have been distributed over these years through various partnerships. Over the past 20 years, the dominant partner has been Medis medical imaging systems bv in Leiden. This partnership has been a model for the current paradigm of the government, in that more of the research carried out at the universities should find its way into products, supporting the health care system in a broad sense, and thereby also stimulating setting up new enterprises. Stimulating this partnership between the universities, the health care system, and industry, and defining core businesses in these areas at universities has been the ultimate goal of the Innovative Medical Devices Initiative (IMDI) that has been shaped over the past 4 years by ZonMW/NWO and in which medical imaging plays a dominant role. In this presentation, I would like to describe in particular the valorization part of my work and describe the collaboration between LKEB and Medis, the position of Medis worldwide and the challenges and opportunities. Medis concentrates its activities in three areas being X-ray angiography (QAngio XA), magnetic resonance imaging (QMassMR and QFlow) and intravascular imaging (QIVus). In addition to the quality of the segmentation algorithms, workflow, robustness, validation, innovation and of course support are the important ingredients for users to use medical imaging software in their daily practice and clinical research. In summary, an overview will be given of all aspects of cardiovascular imaging from research to the roll-out in the real world.

Venue map

Please click on the red address pin (shown on the map) to get directions.

NS Route planner