Molecular Image Analysis

Analysis and visualization of heterogeneous molecular imaging data

ENCITE consortium, Medical Delta, Delft University of Technology

Martin Baiker, PhD student (m.baikerATlumc.nl)
Artem Khmelinskii, PhD student (a.khmelinskiiATlumc.nl)
Peter Kok, PhD student (p.kok.lkebATlumc.nl)

Introduction 

With the rapid progress in molecular imaging technology, biomedical imaging is now covering a broad scale range, from the molecular level through the cellular level to the scale of the whole organism. Molecular processes can be visualized with optical, nuclear and MR imaging using targeted contrast agents, anatomical details with structural modalities (CT, MR, ultrasound), and functional information with e.g. specialized MR acquisitions. Due to their non-invasive nature, disease and development processes can be monitored over time and correlated with macroscopic anatomical changes at a later stage. This adds a new dimension to pre-clinical research, since the traditional cross-sectional studies can be extended to follow-up studies.

The increasing wealth of imaging possibilities from molecule to organism is generating a new problem: the heterogeneity and sheer bulk of imaging data make it virtually impossible for a human observer to interpret and quantify the complex relationships between molecular processes and the structural and functional changes they cause. Molecular imaging studies typically involve data from multiple imaging modalities, covering a wide range of spatial and temporal scales. Differences in imaging geometry, subject posture, and information content occur between modalities, but also between time points in follow-up studies using the same modality. As a result, there is a great demand for methods to integrate from different modalities like e.g. Micro-CT, Micro-MRI or diffuse light photography. At the KGB section, we develop methods to automatically bring datasets from different modalities (CT, MRI, PET, SPECT, BLI, etc.) into the same reference frame to exploit complementary information and to facilitate data comparison.

Objectives

The goal of this project is to develop methods for fusion of heterogeneous pre-clinical molecular imaging data. We especially focus on the following technical challenges:

  • Registration of whole-body follow-up data (Martin Baiker)
  • Registration of heterogeneous multi-modal whole-body imaging data (Artem Khmelinskii)
  • Visualization methods for exploring complex molecular imaging follow-up studies (Peter Kok)

Methods

The interrelation of whole-body scans obtained with different imaging modalities and at different time points requires versatile image registration techniques, which take into account the variations caused by differences in limb and joint positions. Existing nonrigid registration techniques applications are not capable of recovering the large differences in body pose and joint positions typically encountered in pre-clinical imaging. Novel registration algorithms will be developed to separate deformations caused by random variations in articulations from disease-specific structural changes. This involves the development and incorporation of anatomical atlases and hierarchical part-tree models. Novel algorithms will be developed to allow matching of anatomical to multi-angle and 3D optical imaging data. We also develop new techniques to allow integration, visualization, and quantification of the complementary imaging data (see also this link visualisation.tudelft.nl/Projects/BioluminescenceImaging).
The goal here is to exploit the complementarity of the imaging modalities and to achieve a more accurate quantification of the concentrations, distributions, migration and differentiation of cells and cell clusters, in relation to structural and functional changes. This will facilitate the monitoring of therapeutic effects and the development of tissue- and pathology-specific pharmacokinetic models.

Publications

Journal/Conference Papers

  • M. Wildeman, M. Baiker, M.J.T. Reinders, C.W.G.M. Löwik, J.H.C. Reiber, B.P.F. Lelieveldt, “2D/3D Registration of Micro-CT data to Multi-view Photographs based on a 3D Distance Map”, paper #1564, proc. ISBI 2009, 2009.     
  • M. Baiker, B. Vastenhouw, W. Branderhorst, J.H.C. Reiber, F. Beekman, B.P.F. Lelieveldt, “Atlas-Driven Scan Planning for High-Resolution µSPECT Scanning from Multi-View Photographs: a Pilot Study”, proc. SPIE Medical Imaging 2009, paper # 72611L, 8 pages.
  • M.Baiker, J. Dijkstra, I. Que, C.W.G.M. Lowik, J.H.C. Reiber, B.P.F. Lelieveldt, “Organ approximation in micro-CT data with low soft tissue contrast using an articulated whole-body atlas”, paper nr 1191, proc. ISBI 2008, pp 1267-1270, 2008
  • P. Kok, J. Dijkstra, C. P. Botha, F. H. Post, E. Kaijzel, I. Que, C. Lowik,  J. Reiber, B. P. F. Lelieveldt “Integrated visualization of multiangle bioluminescence imaging and micro CT“ Proc SPIE Medical Imaging 2007, paper nr 6509‑65, 2007.
  • M. Baiker, J. Milles, A.M. Vossepoel, I. Que, E.L. Kaijzel, C.W.G.M. Löwik, J.H.C. Reiber, J.Dijkstra, B.P.F. Lelieveldt, “Fully automated whole-body registration in mice using an articulated skeleton atlas”, paper nr 1430, proc. ISBI 2007.

 Abstracts/ Book Chapters

  • M. Baiker, J. Dijkstra, J.R. Milles, C.W.G.M. Löwik, B.P.F. Lelieveldt, “Atlas-based whole-body registration in mice”, in “Biomedical Image Analysis: Methodologies and Applications” (Eds J. Duncan, N. Paragios, N. Ayache), in press, 2009

Contact

B.P.F. Lelieveldt PhD
Division of Image Processing
Department of Radiology, 1-C2S
Leiden University Medical Center
P.O. Box 9600
2300 RC Leiden
The Netherlands
Tel. +31 (0)71 526 1130
Fax. +31 (0)71 526 6801