Medical Imaging Cardiac CT Imaging Medical Image ProcessingComputer Assisted SurgeryRadiation Therapy
Osteoporosis Research  Dose Assessment/Dose Reduction


Contact: Prof. Dr. Willi A. Kalender, Ph.D
      Background of Cardiac Imaging
The rotation of the acquisition system, the gantry, comes along with a time-dependence of the projection data used for image reconstruction. When performing a standard image reconstruction e.g. a filtered backprojection, all these projection data are taken into account without considering the time. This limited temporal resolution in the case of cardiac CT leads to motion artifacts when using standard image reconstruction algorithms. To overcome the restricted image quality the reconstruction algorithm can be synchronized with the object's movement under the assumption of a quasi-periodic motion of the object. Only projection data of the same motion phase are used for the image reconstruction, while the projection data acquired during other phases do not contribute to the image. In this way, the temporal resolution can be increased artificially and the object can be imaged in a quasi-static way.
  Example of cardiac images using phase-correlated reconstruction: coronal view (left) (C 150/W 700) and volume rendered image (right)
      History of Cardiac Imaging at the IMP
The Institute of Medical Physics shares in the academic mission of providing sophisticated solutions for the ever evolving field of non-invasive cardiac imaging. The institute’s research activities focus on improving all aspects of quality of care and the system’s safety, while also making it more effective, patient-centered, timely, efficient and equitable. To reach this ambitious goal the fields of research are dedicated to the development and implementation of new concepts and algorithms and the design and testing of phantoms in order to assure and maintain a high standard in image quality.
  October 1995 Start of cardiac CT imaging at the IMP
January 1996 Installation: SOMATOM PLUS4 (prototype) by Siemens Medical Solutions, CT-Division at the IMP
1997/1998 Presentation of the first clinical results obtained at conferences in Vienna, Chicago and Rotterdam [1, 2, 3, 4]
1998-2001 Grant by the Bavarian Research Foundation (Bayerische Forschungsstiftung):
    “Imaging of the heart with sub-second spiral CT” (6/1998 – 11/1999)
    “Non-invasive coronary CT angiography and functional studies of the heart on a subsecond-spiral CT” (01/2000 – 07/2001)
1998 Installation: SOMATOM Volume Zoom (prototype) by Siemens Medical Solutions, CT-Division at the IMP.

Greenfield Award of the American Association of Physicists in Medicine for dedicated cardiac reconstruction algorithms [5].

1997-2003 New concepts for quality assurance and according phantom developed [22] which are now being considered as a de facto standard in industry.
1999 Establishing ECG independent cardiac image reconstruction (kymogram-based) [10].
1999/2000 The concepts for quality assurance have been awarded twice with the highest given recognition: “magna cum laude” at the ECR [7, 14]
2000 Publication of reconstruction algorithms for multi-slice spiral CT [12, 13].
2001 Development of algorithms for 16-slice spiral CT scanners (Siemens SOMATOM Sensation 16) [17].
2001 Installation: SOMATOM Sensation 16 (prototype) by Siemens Medical Solutions, CT-Division at the IMP.
2002 Further development of kymogram-correlated cardiac image reconstruction [23].
July 2002 Arrangement of the “International Conference on Cardiac Spiral CT” (chairman: W. A. Kalender).
2002 Development of a phase-correlated cone-beam algorithm for scanners with 128 slices or more [24].
2003 Three page Editorial in European Radiology about published concepts for quality assurance [25].
2004 Installation: SOMATOM Definition by Siemens Medical Solutions, CT-Division at the IMP.
2005/2006 Further improvements of kymogram-based cardiac image reconstruction, including the detection of the optimal reconstruction phase [28, 30, 31].
2009 Installation: SOMATOM Definition Flash by Siemens Healthcare, CT-Division at the IMP.
      Projects concerning Cardiac Imaging at the IMP
Kymogram-correlated Image Reconstruction
For cardiac imaging usually the ECG is used as a synchronization function representing the cardiac cycle, i.e. the periodic electrophysical activity. As an alternative synchronization signal the kymogram function [23] can be used. It is generated by analyzing the temporal variation of the mass distribution in the patient. Caused by the contraction of the heart the mass of the projected slices varies with time and correlates with the heart rate, offering the possibility of a synchronization of the reconstruction algorithm [23, 30]. Thus, the kymogram can be computed without additional devices using directly the CT rawdata. In contrast to the ECG, it captures the true mechanical heart movement and does not represent an indirect electrical signal.
Optimal Reconstruction Phase
Since the heart's contractile motion is a non-uniform quasi-periodic motion, not all phase intervals in the cardiac cycle are equally well suited with respect to the phase-correlated reconstruction algorithm. The diastolic phase represents the optimal cardiac reconstruction phase for most cases. However, a general prediction of the optimal reconstruction phase showing a minimum of motion is not possible. Hence, it is common practice to approach the optimal image quality iteratively by multiple image reconstructions at different phase points. To overcome these problems, a fully automatic algorithm has been proposed providing the optimal reconstruction phase [28, 32]. Hereby, cardiac images with an optimal image quality can be achieved with a single image reconstruction process.
  Phase-correlated cardiac images in transaxial view for the empirical reconstruction phase 60% R-R (left) and computed optimal reconstruction phase 76% R-R (right) (C 150/W 700)
Assessment of the Temporal Resolution
For the introduced CT systems the acquisition of the Radon transform of the object comes along with an acquisition time. Therefore, the rotation time had been increased with every new generation of CT devices. To overcome the mechanical restrictions of rotation time and the limited temporal resolution multi-segment reconstruction algorithms and finally multi-tube-detector system had been introduced. Unfortunately, the improvement which can be achieved show a complex dependency on multiple variables [26, 27]. By using a dedicated motion robot the temporal resolution of the complete imaging system can be assessed and the corresponding improvement of new approaches can be validated [31].
  Theoretical temporal resolution for different rotation times and number of tube-detector systems for a multi-segment reconstruction algorithm
Calcium Scoring
In addition to general diagnostic imaging of the heart, there is also a particular interest in the quantification of coronary calcium, i.e. calcium scoring as an early marker for the presence of coronary artery disease. Since 1990 the only quantification method for coronary calcium, present in the form of hydroxyapatite (HA) was the Agatston score [22, 25]. With its many limitations and drawbacks a new quantification method that corresponds to a physical measure and also allows the comparison between different scanners, protocols and segmentation approaches has also been introduced. In addition a phantom with calibration inserts for calcium scoring for quality assurance has been designed for setting up a new quality assurance concept in coronary calcium scoring.
Pulmonary Imaging in the Pericardiac Region
Image quality in pulmonary CT imaging is commonly degraded by motion artifacts in pericardial areas of the lung. Phase-correlated image reconstruction techniques can reduce these artifacts by restricting the data used for reconstruction to certain cardiac phases. However, phase-correlated image reconstructions suffer from a higher image noise as a consequence of the reduced amount of data used for reconstruction and therefore reduce image quality in non-moving regions. However, an ECG signal is commonly not recorded in thoracic imaging the kymogram function can be used to synchronize the image reconstruction. Based on phase-correlated reconstructed images, the motion-free merging technique is used to combine motion free areas of standard reconstructions yielding low image noise with moving but artifact free image parts of phase-correlated reconstructed images. Thereby, an optimal trade-off is achieved for all image regions [29, 33].
  Example of pulmonary images using standard reconstruction (left) and motion free merging (right) (C -400/W 1300)
Cardiac Motion Robot
Anthropomorphic motion curves including the possibility of 3D arrhythmic trajectories are essential for constancy testing and quality assurance of medical devices. A 3D motion robot (QRM, Möhrendorf, Germany) was designed offering the possibility to simulate motion profiles for universal medical applications as rhythmic and arrhythmic motion of coronary arteries in high repeat accuracy. The system provides a motion range of 100 mm and repetition rates up to 200 bpm. Different sample probe can be connected to a bracket which can be positioned in a water filled PMMA-cylinder centrally placed in an anthropomorphic thorax phantom. Hence, the motion robot offers the possibility to analyze CT imaging systems with respect to moving objects either in an anthropomorphic or freely chosen way.
  Sketch of the thorax phantom used (left) and foto of the 3D motion robot (right, courtesy of QRM)
Respiratory Gating in Micro CT Imaging
Small animal imaging emerged as an important tool for preclinical research. Dedicated micro-computed tomography (micro-CT) systems have been developed for three dimensional visualization. Common micro-CT systems provide higher spatial resolution than clinical CT systems but suffer from reduced temporal resolution. Thereby, images of in-vivo studies are affected by respiratory motion causing motion artefacts mainly in the diaphragm region. Retrospective phase-correlated image reconstruction procedure for respiratory gating in micro CT imaging can improve these limitations [34]. An automatic detection of the optimal data window known from cardiac imaging [28] can assure least amount of motion blurring. Thereby, optimal image quality can be provided without the need for performing more than the final image reconstruction by an object-specific identification of the optimal reconstruction phase.
  Example of different images in coronal view of the same object: standard reconstruction (left) and for a phase identified by the proposed algorithm (right).
1. Kalender W.A. and Polacin A. (1997): "Imaging of the heart by ECG-oriented reconstruction from subsecond spiral CT scans". European Radiology 7: 237
2. Kachelrieß M. and Kalender W. A. (1997): "ECG-based phase-oriented reconstruction from subsecond spiral CT scans of the heart". Radiology 205(P): 215
3. Kachelrieß M., Kalender W. A., Karakaya S., Achenbach S., Nossen J., Moshage W. and Bautz W. A. (1998): "Imaging of the heart by ECG-oriented reconstruction from subsecond spiral CT scans". Advances in CT IV, G. Glazer and G. Krestin, Eds. Heidelberg, Berlin, New York, Springer Verlag: 137-143
4. Kachelrieß M. and Kalender W. A. (1998): "Imaging of the heart by ECG-oriented reconstruction from subsecond spiral multi-row detector CT scans". Radiology 209(P): 323
5. Kachelrieß M. and Kalender W. A. (1998): "Electrocardiogram-correlated image reconstruction from subsecond spiral CT scans of the heart". Med Phys 25(12): 2417-2431
6. Kachelrieß M. and Kalender W. A. (1999): "ECG-based phase oriented image reconstruction from subsecond spiral multi-row CT scans of the heart". European Radiology 9 Suppl. 1: S 277
7. Ulzheimer S., Kachelrieß M. and Kalender W. A. (1999): "Improvements of cardiac CT using ECG-oriented image reconstruction in subsecond spiral multirow scanning". European Radiology 9 Suppl. 1: S 419
8. Kachelrieß M. and Kalender W. A. (1999): "ECG-correlated image reconstruction from subsecond multirow spiral CT scans of the heart: Theoretical considerations, phantom measurements and patient studies". Radiology 213(P): 401
9. Ulzheimer S., Kachelrieß M. and Kalender W. A. (1999): "New phantoms for quality assurance in cardiac CT". Radiology 213(P): 402
10. Kalender W. A. and Kachelrieß M. (1999): "Computertomograph mit objektbezogener Bewegungsartefaktreduktion und Extraktion der Objektbewegungsinformation (Kymogramm)". European Patent Office (Patent pending)
11. Kachelrieß M., Kalender W.A. (2000): "Kymogram-correlated image reconstruction from subsecond multi-slice spiral CT scans of the heart". Radiology 217(P): 439
12. Kachelrieß M., Ulzheimer S. and Kalender W. A. (2000): "ECG-correlated image reconstruction from subsecond multi-slice spiral CT scans of the heart". Med. Phys. 27(8): 1881-1902
13. Kachelrieß M., Ulzheimer S. and Kalender W. A. (2000): "ECG-correlated imaging of the heart with subsecond multi-slice spiral CT". IEEE Transactions on Medical Imaging 19(9): 888-901
14. Achenbach S., Ulzheimer S., Baum U., Kachelrieß M., Ropers D., Giesler T., Bautz W., Daniel W. and Kalender W. A. (2000): "Noninvasive coronary angiography by retrospectively ECG-gated multislice spiral CT". Circulation 102: 2823-2828
15. Ulzheimer S., Kachelriess M., Fuchs T., Decker R., Achenbach S., Moshage W. and Kalender W. A. (2000): "4D cardiac imaging with spiral CT: Algorithms, scan protocols, quality assurance and clinical studies". European Radiology 10(2 Suppl. 1): 339
16. Ulzheimer S., Decker R., Kachelriess M. and Kalender W. A. (2000): "Quality assurance in cardiac CT: Phantoms for scanner calibration and comparison of different imaging parameters". European Radiology 10(2) Suppl. 1: 303
17. Kachelrieß M., Fuchs T., Lapp R., Sennst D.-A., Schaller S. and Kalender W. A. (2001): "Image to volume weighting generalized ASSR for arbitrary pitch 3D and phase-correlated 4D spiral cone-beam CT reconstruction". Proceedings of the 2001 Int. Meeting on Fully 3D Image Reconstruction: 179-182
18. Achenbach S., Giesler T., Ropers D., Ulzheimer S., Derlien H., Schulte C., Wenkel E., Moshage W., Bautz W., Daniel W., Kalender W. A. and Baum U. (2001): "Detection of coronary artery stenoses by contrast-enhanced, retrospectively electrocardiographically-gated, multislice spiral computed tomography". Circulation 103: 2535-2538
19. Kalender W. A., Ulzheimer S. and Kachelrieß M. (2001): "Cardiac imaging with multislice spiral computed tomography. Scan reconstruction principles and quality assurance. Multislice CT: A Practical Guide". (Hrsg.: B. Marincek et al.) Springer Heidelberg, Berlin, New York: 79-89
20. Kachelrieß M., Sennst D.-A., Maxlmoser W. and Kalender W. A. (2001): "Kymogram - vs ECG-correlated imaging from subsecond multi-slice and cone-beam spiral CT scans of the heart". Radiology 221(P): 458
21. Ropers D., Ulzheimer S., Wenkel E., Baum U., Giesler T., Derlien H., Moshage W., Bautz W., Daniel W., Kalender W. A. and Achenbach S. (2001): "Investigation of aortocoronary artery bypass grafts by multislice spiral computed tomography with electrocardiographic-gated image reconstruction". Amer J of Cardiology 88: 792-795
22. Ulzheimer S. (2001): "Cardiac imaging with x-ray computed tomography: New approaches to image acquisition and quality assurance". In: Berichte aus dem Institut für Medizinische Physik. Bd 6. (Hrsg.: W. A. Kalender). Aachen, Shaker Verlag
23. Kachelrieß M., Sennst D.-A., Maxlmoser W. and Kalender W. A. (2002): "Kymogram detection and kymogram-correlated image reconstruction from sub-second spiral computed tomography scans of the heart". Med Phys 29(7): 1489-1503
24. Kachelrieß M. and Kalender W. A. (2002): "Extended parallel backprojection for cardiac cone-beam CT for up to 128 slices". Radiology 225(P): 310
25. Ulzheimer S. and Kalender W. A. (2003): "Assessment of calcium scoring performance in cardiac computed tomography". European Radiology 13: 484-497
26. Kachelrieß M., Knaup M., Kalender W. A. (2005): "Phase-correlated Imaging from Multi-threaded Spiral Cone-beam CT Scans of the Heart". Proc. 8th International Meeting on Fully 3D Image Reconstruction: 159-162
27. Kachelrieß M., Knaup M., Kalender W.A. (2006): "Multithreaded cardiac CT". Med. Phys. 33(7): 2435-2447
28. Ertel D., Kachelrieß M., Pflederer T., Achenbach S., Lapp R. M., Nagel M., Kalender W. A. (2006): "Rawdata-based detection of the optimal reconstruction phase in ECG-gated cardiac image reconstruction". Lecture Notes in Computer Science 4191: 348-355
29. Ertel D., Kachelrieß M., Kalender W. A. (2007): "Histogram-Driven Multi-Dimensional Adaptive Filtering (HD-MAF)". IEEE Medical Imaging Conference, (Honolulu, USA)
30. Ertel D., Pflederer T., Achenbach S., Kachelrieß M., Steffen P., Kalender W. A. (2008): "Validation of a Rawdata-based Synchronization Signal (Kymogram) for a Phase-Correlated Cardiac Image Reconstruction". European Radiology 18(2): 253-262.
31. Ertel D., Kröber E., Kyriakou Y., Langner O., Kalender W. A. (2008): "MTF-based Assessment of Temporal Resolution: Validation for Single Source CT and Dual Source CT". Radiology 248(3): 1013-1017
32. Ertel D., Pflederer T., Achenbach S., Kachelrieß M., Kyriakou Y., Kalender W. A. (2009): "Technical Note: Rawdata-based Approach to Identify the Optimal Reconstruction Phase in Coronary CT Angiography". Journal of Computer Assisted Tomography 33(1) 26-31
33. Lapp R. M., Kachelrieß M., Ertel D., Kyriakou Y., Kalender W. A.: "Cardiac Phase-correlated Image Reconstruction and Advanced Image Processing in Pulmonary CT Imaging". European Radiology, in DOI: 10.1007/s00330-008-1237-x
34. Ertel D., Kyriakou Y., Lapp R. M., Kalender W. A. (2009): "Respiratory Phase-Correlated Micro CT Imaging of Free-Breathing Rodents". Physics in Medicine and Biology, submitted