Visualization & Exploration of Perfusion Data

Perfusion data are dynamic medical image data which characterize the regional blood flow in tissue. We focus on the visualization and exploration of perfusion data for diagnostic purposes in three major application areas: ischemic stroke diagnosis, breast tumor diagnosis and the diagnosis of coronary heart disease.

For a reliable diagnosis, the multi-dimensional space of perfusion parameters needs to be explored. Therefore, we aim at different exploration and visualization techniques for an improved diagnosis. Furthermore, visual analytics approaches have been developed and integrated in the diagnosis frameworks.

During the evaluation of perfusion data, clustering is often carried out to reduce the influence of noise or image artifacts when analyzing the temporal variation of the tissue’s signal intensity values. Thus, the exploration of perfusion data comprises the adaption of clustering methods as well as the evaluation of these clustering methods based on an automatic classification into benign and malignant tumors.

This project was a subproject of the DFG funded Scalable Visual Analytics priority programme.

Publications

2014
Visualization of 3D Cluster Results for Medical Tomographic Image Data

Sylvia Glaßer, Kai Lawonn, Bernhard Preim

Visualization of 3D Cluster Results for Medical Tomographic Image Data Inproceedings

In Proc. of Conference on Computer Graphics Theory and Applications (VISIGRAPP/GRAPP), pp. 169–176, 2014.

BibTeX | Links:

Adapted Spectral Clustering for Evaluation and Classification of DCE-MRI Breast Tumors

Sylvia Glaßer, Sophie Roscher, Bernhard Preim

Adapted Spectral Clustering for Evaluation and Classification of DCE-MRI Breast Tumors Inproceedings

Bildverarbeitung für die Medizin (BVM), pp. 198–203, 2014.

BibTeX | Links:

2013
Visual analysis of longitudinal brain tumor perfusion

Sylvia Glaßer, Steffen Oeltze, Uta Preim, Atle Bjornerud, Helwig Hauser, Bernhard Preim

Visual analysis of longitudinal brain tumor perfusion Inproceedings

Proc. of the SPIE Medical Imaging, pp. 86700Z, 2013.

BibTeX | Links:

Can we Distinguish Between Benign and Malignant Breast Tumors in DCE-MRI by Studying a Tumors Most Suspect Region Only?

Sylvia Glaßer, Uli Niemann, Bernhard Preim, Myra Spiliopoulou

Can we Distinguish Between Benign and Malignant Breast Tumors in DCE-MRI by Studying a Tumors Most Suspect Region Only? Inproceedings

Proc. of Symposium on Computer-Based Medical Systems (CBMS), pp. 59-64, 2013.

BibTeX | Links:

Classification of Benign and Malignant DCE-MRI Breast Tumors by Analyzing the Most Suspect Region

Sylvia Glaßer, Uli Niemann, Uta Preim, Myra Spiliopoulou, Bernhard Preim

Classification of Benign and Malignant DCE-MRI Breast Tumors by Analyzing the Most Suspect Region Inproceedings

Bildverarbeitung für die Medizin (BVM), pp. 45–50, 2013.

BibTeX | Links:

Method for the Evaluation of US Perfusion for Brain Tumor Surgery

Claire Chalopin, Steffen Oeltze, Bernhard Preim, Andreas Müns, Jürgen Meixensberger, Dirk Lindner

Method for the Evaluation of US Perfusion for Brain Tumor Surgery Inproceedings

12. Jahrestagung der Deutschen Gesellschaft für Computer- und Roboter Assistierte Chirurgie, pp. 198–202, Innsbruck, 2013.

BibTeX | Links:

2011
Local similarity measures for lesion registration in DCE-MRI of the breast

Sebastian Schäfer, Uta Preim, Sylvia Glaßer, Bernhard Preim, Klaus-Dietz Tönnies

Local similarity measures for lesion registration in DCE-MRI of the breast Journal Article

Annals of the BMVA, 3 , pp. 1–13, 2011.

BibTeX | Links:

The File-Card-Browser View for Breast DCE-MRI Data

Sylvia Glaßer, Kathrin Scheil, Uta Preim, Bernhard Preim

The File-Card-Browser View for Breast DCE-MRI Data Inproceedings

Bildverabeitung für die Medizin (BVM), pp. 314–318, Lübeck, 2011.

BibTeX | Links:

2010
Visuelle Analyse medizinischer Daten

Bernhard Preim, Steffen Oeltze, Klaus-Dietz Tönnies

Visuelle Analyse medizinischer Daten Journal Article

Informatik Spektrum, 33(6) , pp. 569–579, 2010.

BibTeX | Links:

Sylvia Glaßer, Uta Preim, Klaus-Dietz Tönnies, Bernhard Preim

A visual analytics approach to diagnosis of breast DCE-MRI data Journal Article

Computer and Graphics, 34(5) , pp. 602–611, 2010.

BibTeX

2009
A Visual Analytics Approach to Diagnosis of Breast DCE-MRI Data

Sylvia Glaßer, Sebastian Schäfer, Steffen Oeltze, Uta Preim, Klaus-Dietz Tönnies, Bernhard Preim

A Visual Analytics Approach to Diagnosis of Breast DCE-MRI Data Inproceedings

Proc. of Vision, Modeling, and Visualization (VMV), pp. 351–362, Braunschweig, 2009.

BibTeX | Links:

Visual Analysis of Cerebral Perfusion Data -- Four Interactive Approaches and a Comparison

Steffen Oeltze, Helwig Hauser, Jarle Rorvik, Arvid Lundervold, Bernhard Preim

Visual Analysis of Cerebral Perfusion Data -- Four Interactive Approaches and a Comparison Inproceedings

Proc. of the 6th International Symposium on Image and Signal Processing and Analysis (ISPA), pp. 588–595, 2009.

BibTeX | Links:

Survey of the Visual Exploration and Analysis of Perfusion Data

Bernhard Preim, Steffen Oeltze, Matej Mlejnek, Eduard Gröller, Anja Hennemuth, Sarah Behrens

Survey of the Visual Exploration and Analysis of Perfusion Data Journal Article

IEEE Transactions on Visualization and Computer Graphics, 15(2) , pp. 205–220, 2009.

BibTeX | Links:

2008
A Comprehensive Approach to the Analysis of Contrast Enhanced Cardiac MR Images

Anja Hennemuth, Achim Seeger, Ola Friman, Stephan Miller, Bernhard Klumpp, Steffen Oeltze, Heinz-Otto Peitgen

A Comprehensive Approach to the Analysis of Contrast Enhanced Cardiac MR Images Journal Article

IEEE Transactions on Medical Imaging, 27(11) (11), pp. 1592–1610, 2008.

BibTeX | Links:

A Four-level Focus+Context Approach to Interactive Visual -- Analysis of Temporal Features in Large Scientific Data

Philipp Muigg, Johannes Kehrer, Steffen Oeltze, Harald Piringer, Helmut Doleisch, Bernhard Preim, Helwig Hauser

A Four-level Focus+Context Approach to Interactive Visual -- Analysis of Temporal Features in Large Scientific Data Journal Article

Computer Graphics Forum (EuroVis), 27 (3) , pp. 775–782, 2008.

BibTeX | Links:

Glyph-Based Visualization of Myocardial Perfusion Data and Enhancement with Contractility and Viability Information

Steffen Oeltze, Anja Hennemuth, Sylvia Glaßer, Caroline Kühnel, Bernhard Preim

Glyph-Based Visualization of Myocardial Perfusion Data and Enhancement with Contractility and Viability Information Inproceedings

VCBM, pp. 11–20, 2008.

BibTeX | Links:

Intuitive Mapping of Perfusion Parameters to Glyph Shape

Steffen Oeltze, Arvid Malyszczyk, Bernhard Preim

Intuitive Mapping of Perfusion Parameters to Glyph Shape Inproceedings

Tolxdorff, T; Braun, J; Deserno, T M; Handels, H; Horsch, A; Meinzer, H P (Ed.): Bildverarbeitung für die Medizin (BVM), pp. 262–266, 2008.

BibTeX | Links:

Survey of the Visual Exploration and Analysis of Perfusion Data

Bernhard Preim, Steffen Oeltze, Matej Mlejnek, Eduard Gröller, Anja Hennemuth, Sarah Behrens

Survey of the Visual Exploration and Analysis of Perfusion Data Inproceedings

IEEE Transactions on Visualization and Graphics, pp. to appear, 2008.

BibTeX | Links:

CT Late Enhancement Segmentation for the Combined Analysis of Coronary Arteries and Myocardial Viability

Anja Hennemuth, Andreas Mahnken, Caroline Kühnel, Steffen Oeltze, Heinz-Otto Peitgen

CT Late Enhancement Segmentation for the Combined Analysis of Coronary Arteries and Myocardial Viability Inproceedings

VCBM, pp. 1–10, 2008.

BibTeX | Links:

Enhanced Cardio Vascular Image Analysis by Combined Representation of Results from Dynamic MRI and Anatomic CTA

Caroline Kühnel, Anja Hennemuth, Steffen Oeltze, Tobias Boskamp, Heinz-Otto Peitgen

Enhanced Cardio Vascular Image Analysis by Combined Representation of Results from Dynamic MRI and Anatomic CTA Inproceedings

Proc. of SPIE Conference on Medical Image Computing, 2008.

BibTeX | Links:

2007
Interactive Visual Analysis of Perfusion Data

Steffen Oeltze, Helmut Doleisch, Helwig Hauser, Philipp Muigg, Bernhard Preim

Interactive Visual Analysis of Perfusion Data Journal Article

IEEE Transactions on Visualization and Computer Graphics (TVCG), 13 (6), pp. 1392–1399, 2007.

BibTeX | Links:

Integrierte Visualisierung kardialer MR-Daten zur Beurteilung von Funktion, Perfusion und Vitalität des Myokards

Lydia Paasche, Steffen Oeltze, Anja Hennemuth, Caroline Kühnel, Frank Grothues, Bernhard Preim

Integrierte Visualisierung kardialer MR-Daten zur Beurteilung von Funktion, Perfusion und Vitalität des Myokards Inproceedings

Bildverarbeitung für die Medizin (BVM), pp. 212–216, Springer, 2007.

BibTeX | Links:

Novel Methods for Parameter Based Analysis of Myocardial Tissue in MR-Images

Anja Hennemuth, Sarah Behrens, Caroline Kühnel, Steffen Oeltze, Olaf Konrad, Heinz-Otto Peitgen

Novel Methods for Parameter Based Analysis of Myocardial Tissue in MR-Images Inproceedings

SPIE Conference on Medical Image Computing, 2007.

BibTeX | Links:

2006
Integrated Visualization of Morphologic and Perfusion Data for the Analysis of Coronary Artery Disease

Steffen Oeltze, Frank Grothues, Anja Hennemuth, Anja Kuß, Bernhard Preim

Integrated Visualization of Morphologic and Perfusion Data for the Analysis of Coronary Artery Disease Inproceedings

IEEE/Eurographics Symposium on Visualization, pp. 131–138, Springer, 2006.

BibTeX | Links:

New Software Assistants for Cardiovascular Diagnosis

Caroline Kühnel, Anja Hennemuth, Susanne Bock, Steffen Oeltze, Tobias Boskamp, Stefan Krass, Bernhard Preim, Heinz-Otto Peitgen

New Software Assistants for Cardiovascular Diagnosis Inproceedings

GI-Workshop - Softwareassistenten - Computerunterstützung für die medizinische Diagnose und Therapieplanung, pp. 491–498, Springer, 2006.

BibTeX | Links:

2005
Multiparametervisualisierung zur Exploration dynamischer Bilddaten

Steffen Oeltze, Christian Bendicks, Sarah Behrens, Bernhard Preim

Multiparametervisualisierung zur Exploration dynamischer Bilddaten Inproceedings

Bildverarbeitung für die Medizin (BVM), pp. 317–321, Springer, 2005.

BibTeX | Links:

2003
Mehrdimensionale Visualisierung dynamischer Bilddaten am Beispiel der Durchblutungsquantifizierung

Bernhard Preim, Sven Kohle, Olaf Konrad-Verse, Richard Rascher-Friesenhausen, Jonathan Wiener, Ronald Leppek, Heinz-Otto Peitgen

Mehrdimensionale Visualisierung dynamischer Bilddaten am Beispiel der Durchblutungsquantifizierung Inproceedings

Simulation und Visualisierung, pp. 77–88, 2003.

BibTeX | Links:

2002
Exploration of time-varying data for Medical Diagnosis

Sven Kohle, Bernhard Preim, Jonathan Wiener, Heinz-Otto Peitgen

Exploration of time-varying data for Medical Diagnosis Inproceedings

Vision, Modelling und Visualization (VMV), pp. 31–38, Erlangen, 2002.

BibTeX | Links:

Images

  • Identification of areas with rapid contrast agent washout (e,f). Smooth brushing (h) reveals subtle spikes along the border of Slarge (g).
  • Exploiting the PCA results also reveals the affected tissue. Result is very similar as compared to perfusion parameter based analysis (1/2).
  • Intensity difference based (b), local (c) analysis reveals a suspicious structure (a). Its characteristics (d,e) indicate a high-grade glioma.
  • Curve view showing time-intensity curves of the entire dataset (d). Brushing may separate cortex (a), medulla (b), and pelvis (c).
  • PCA results for 2 datasets, Heart 1/2 (a). Brushing the 1st principal component (pc) (b) reveals ischemic tissue in Heart 1 (green area (c)).
  • Brushing TTP and Up-Slope (f) reveals ischemic tissue in Heart 2 (green area in (e)). Brushing the 2nd pc (h) indicates preprocessing failures (g).
  • Brushing high intensity differences over time (b) highlights strong contrast agent wash-in (a). Analysis is focused on local region (c,d).
  • Curve views show the time-intensity curves of each voxel. Brushes facilitate a differentiation of infarction core (l) and penumbra (r).
  • The user-defined brush resembles a curve pattern which is typical for suspicious tissue (left). Two suspicious structures are revealed (right).
  • Feature specification in a scatterplot (b) reveals the infarction core (a). Adding a near-focus region (d) indicates penumbral tissue (c).
  • Smooth gradient brushing (f) adds context information (e). Smooth brushing of PCA results (h) yields similar results (g) as compared to (c).
  • A glyph labeling, a slice of the original perfusion data and an orientation cube improve the spatial orientation.
  • A 3D TIC resembles the averaged signal intensity time course. The gray reference glyphs represent the entire myocardium.
  • The view along the left ventricular long-axis resembles a Bull’s Eye Plot view. It gives an overview of the perfusion in all segments.
  • The right ventricle is provided as context. The left ventricular wall is emphasized at intersections with the perfusion imaging planes.
  • 3D TICs encode the perfusion and isolines encode the transmurality of the scar. The wall thickening is color-coded on the endocardium.
  • Focus and near-focus regions (bluish rectangles) have been defined (inset) resulting in different degrees-of-interest (DOI).
  • Blue, horizontally oriented glyphs indicate a diminished and decelerated perfusion caused by an infarction.
  • Each ring of the plot has been bisected. The resulting inner and outer rings facilitate a comparison of stress and rest perfusion (parameters).
  • Four slices have been acquired for evaluating the perfusion. Small bluish glyphs indicate a disturbance in the scar region (brown surface).
  • Division according to the AHA 17-segment model. Glyph height equals the slice thickness of the original perfusion data.
  • Tissue affected by an ischemic stroke is characterized by a delayed (high Time-To-Peak) and diminished (low Integral) blood supply.
  • The contrast agents wash-out velocity is color-coded and projected through the lens in the context of the original perfusion data.
  • Synchronized lenses support the comparison of both hemispheres. The initially defined lens is mirrored on a relocatable line of symmetry.
  • The rightmost image represents an integrated visualization of the four parameter maps by means of color icons (template shown as inset).
  • One glyph represents 4×4 voxels in order to gain screen space for glyph drawing. The user may interactively adjust this ratio.
  • The triangular glyphs resemble the original concentration-time curves generated by plotting contrast agent concentration over time.
  • A parameter sub-domain has been defined in a feature specification component. The glyph display is restricted to the corresponding tissue.