DE102009039986A1 - Method for movement correction of three dimensional digital subtraction angiography, during visualization of e.g. blood vessel tree in heart of human body, involves producing subtracted volume that is reduced at subtraction artifacts - Google Patents

Abstract

The method involves implementing movement correction by two dimensional or three dimensional registration (27) and subtraction of two-dimensional image data sets, which comprise mask image data set and filling image data set. The mask image data set and the filling image data set are subtracted from each other. Registration process is applied for registration of image data sets. The process is based on additional two-dimensional image data sets so that subtracted volume is produced, where the subtracted volume is reduced at subtraction artifacts. The volume is represented using a display unit. An independent claim is also included for a device for movement correction of the three dimensional digital subtraction angiography.

Description

The invention relates to a method and a device motion correction for 3-D digital subtraction angiography using the projection data.

Background of the invention

In digital subtraction angiography (DSA), a widely used method of visualizing the blood vessels in the human body, multiple X-ray images are acquired with different image information. In general, an X-ray image is taken before the administration of contrast agent, a so-called mask image, and subtracted from an X-ray image after the administration of contrast agent, a so-called filling image with contrast agent. In the resulting image, only the contrasted structure, such as the blood vessels, can be seen.

By recording two image sequences with a C-arm system before and after administration of a contrast agent, mask and fill projections are generated from several view directions by means of two acquisition runs. From this, three-dimensional subtracted images of the contrasted structure, for example of the vascular tree, can be reconstructed without disturbing bone structures.

Each projection data set taken individually allows, ideally, an artifact-free reconstruction of a 3-D image. For the doctor, the difference between the two 3-D images is interesting, so that, for example, only the contrasted blood vessels are visible and the bones are suppressed.

The capture of mask and fill projections often results in image defects in the reconstructed volume, since movements during the acquisition in the medical environment are generally unavoidable. Mechanical inaccuracies of the recording system or patient movements between the recording runs lead to artifacts, since not all disturbing structures in the subtracted image can be completely removed. Furthermore, movements within the contrast run interfere with the assumption of static objects needed for standard reconstruction algorithms, causing motion reconstruction artifacts. In the medical field in particular, such movement-related reconstruction artifacts occur frequently, so that the non-contrasted structures, for example the bones, can not be completely removed.

In the field of 2-D imaging, the DSA method has been possible for a long time. There are a variety of methods for correcting motions that register the 2-D mask image directly with the 2-D fill image (2-D / 2-D registration for frames). There are approaches that assume pure shifts (so-called "pixel shift") and those that take distortions into account, such as "flexible pixel shift".

Analogously, methods are possible that register a reconstructed 3-D mask volume with a reconstructed 3-D fill volume (3-D / 3-D registration) and use this information to improve the subtracted volume.

The use of reconstructed 3D volumes for registration has the disadvantage that they can show reconstruction artifacts that make registration difficult. Such reconstruction artifacts are particularly noticeable when the motion is not limited to the time between shots of the mask and fill projections, but has also occurred within the mask or fill shot run.

Such a 3-D / 3-D registration is for example from the US 7,315,605 B2 in which, in order to reconstruct several volumes, the input data for the individual reconstructions are determined via an external signal, for example the ECG. Subsequently, the reconstructed 3-D volumes are registered with each other. The individual 3-D volumes essentially correspond to different states of motion of the examined object.

In the DE 41 37 652 A1 a correction method is described with which a 3-D image of a moving object can be reconstructed. In doing so, information about the movement between individual "views" of a set of projection data is used to reconstruct an artifact-free 3-D image. The information about the movement is provided by an additional sensor.

From the US 6,125,164 For example, a method is known for registering a 3-D diagnostic image with 2-D projections of a patient in a radiotherapy device.

There are other known methods which are described, for example, in the publications of Venema et al. [Ven01] Rohlfing et al. [Roh01] and van Straten et al. [vS04] are described. There it is envisaged that the reconstructed 3-D volumes will be registered with each other.

Presentation of the invention

The invention is based on the task of a method of the type mentioned above to train that the described motion artifacts can be corrected image-based.

The object is achieved with the method and the device according to the independent claims. Advantageous embodiments of the method and the device are the subject of the dependent claims or can be taken from the following description and the embodiments.

The invention relates to a method for motion correction for a three-dimensional digital subtraction angiography of at least two at least two-dimensional image data sets, from which optionally three-dimensional volumes can be reconstructed, wherein a motion correction by registration and then subsequent subtraction is performed, characterized in that applied to the registration of a registration technique which is based on at least one two-dimensional and at least one further one of the image data records, so that, as a result, a subtraction volume reduced by subtraction artifacts is generated.

In order to correct the described movement artifacts image-based, it is provided according to the invention to use different registration techniques for the compensation of subtraction artifacts, which preferably operate on 2-D / 2-D or 2-D / 3-D image data.

One aspect of the invention is that the registration is not performed between two 3-D volumes, which may be corrupted by inconsistent input data. Instead, either 2-D projection data of one run is registered with 2-D projection data of another run, or 2-D projection data of one run with the reconstructed volume is registered from a second run.

The approach is used both for correcting rigid movements of the entire cerebral vascular tree and for local motion estimation of diseased blood vessels.

A further aspect of the invention is a movement correction device for a three-dimensional digital subtraction angiography of at least two at least two-dimensional image data sets from which three-dimensional volumes can be reconstructed, characterized by means for motion correction by registration and subsequent subtraction, wherein a registration technique applicable to at least a two-dimensional and at least one further of the image data sets, resulting in a subtraction volume reduced by subtraction artifacts.

By performing a calibration, the projection geometry obtained by calibration can be incorporated into the reconstruction.

Advantageous effects of the invention

A motion-corrected 3-D volume is not generated solely based on the 3-D mask and 3-D fill volumes.

A 2-D / 2-D registration is based solely on the projection data, the registration in this case is completely independent of the reconstruction.

Although a 2-D / 3-D registration requires a consistent mask capture run but not a fully consistent fill pick, or a consistent fill pick, it does not require a fully consistent mask pick-up run.

Description of one or more embodiments

Further advantages, details and developments of the invention will become apparent from the following description of embodiments in conjunction with the drawings. In the drawing show:

1 a known X-ray C-arm system with an industrial robot as a supporting device,

2 a view of the trajectory of a detector and a radiation source according to 1 around an object to be examined in the axial direction and

3 an enhanced image processing chain for the correction of DSA artifacts with a 2-D / 2-D registration,

4. An advanced image processing chain for the correction of DSA artifacts with a 2-D / 3-D registration.

Fig. 1

In the 1 an X-ray diagnostic device is shown, the one on a stand in the form of a six-axis industrial or articulated robot 1 rotatably mounted C-arm 2 has at its ends an X-ray source, for example an X-ray source 3 , and an X-ray image detector 4. are mounted as an image recording unit.

By means of the example of the DE 10 2005 012 700 A1 known articulated robot 1 , which preferably has six axes of rotation and thus six degrees of freedom, the C-arm 2 be spatially adjusted, for example, by turning around a center of rotation between the X-ray source 3 and the X-ray detector 4. is turned. The X-ray system according to the invention 1 to 4. is in particular about centers of rotation and axes of rotation in the plane of the X-ray image detector 4. rotatable, preferably around the center of the X-ray image detector 4. and around the center of the X-ray image detector 4. cutting axes of rotation.

The well-known articulated robot 1 has a base frame, which is for example fixedly mounted on a floor. It is rotatably mounted about a first axis of rotation a carousel. On the carousel is pivotally mounted about a second axis of rotation a rocker arm, on which is rotatably mounted about a third axis of rotation, a robot arm. At the end of the robot arm, a robot hand is rotatably mounted about a fourth axis of rotation. The robot hand has a fastener for the C-arm 2 which is pivotable about a fifth axis of rotation and about a perpendicular thereto extending sixth axis of rotation rotatable.

The X-ray image detector 4. may be a rectangular or square semiconductor flat detector, preferably made of amorphous silicon (a-Si).

In the beam path of the X-ray source 3 is located on a patient table 5 for receiving, for example, a heart, a patient to be examined 6 , At the X-ray diagnostic facility is a system control unit 7 with an image system 8th connected to the image signals of the X-ray image detector 4. receives and processes. The x-rays can then be viewed on a monitor 9 to be viewed as.

At the system control unit 7 is a calibration device 10 connected, which is designed such that it performs a calibration, evaluates and calculates the correct projection geometry, so that a highly accurate reconstruction by the image system 8th can be carried out.

Fig. 2

The X-ray source 3 emits one of a beam focus 11 its beam emanating from its X-ray source 12 pointing to the x-ray image detector 4. meets, as shown schematically in the 2 is illustrated. If 3-D data sets are to be created according to the so-called DynaCT method, the rotatably mounted C-arm is used 2 with beam focus 11 of the X-ray source 3 and X-ray image detector 4. turned so that, like the 2 schematically in view of the center of rotation 13 shows, the X-ray emitter pictured here by its beam focus 3 as well as the X-ray image detector 4. in the beam path of the X-ray source 3 located object to be examined 14 in an orbit 15 move. The orbit 15 can be completely or partially traversed to create a 3-D data set.

The C-arm 2 with X-ray source 3 and X-ray image detector 4. In this case, according to the DynaCT method, it preferably moves by at least one angle range of 180 °, for example 180 ° plus fan angle, and takes up projection images from different projections in rapid succession. The reconstruction can only take place from a subarea of this recorded data.

In the object to be examined 14 it may be, for example, an animal or human body but also a phantom body.

The X-ray source 3 emits a beam emanating from a beam focus of its X-ray source 12 pointing to the x-ray image detector 4. meets.

The X-ray source 3 and the X-ray image detector 4. each run around the object 5 around, that's the X-ray 3 and the X-ray image detector 4. on opposite sides of the object 14 are opposite.

In normal radiography or fluoroscopy by means of such an X-ray diagnostic device, the medical 2-D data of the X-ray image detector 4. in the picture system 8th if necessary cached and then on the monitor 9 played.

Fig. 3

In the 3 For example, the two-dimensional registration of mask and fill images is schematic in the steps shown 20 to 26 shown.

In this case, 2-D projection data, preferably the mask images, of an image acquisition run with 2-D projection data, preferably the fill images, of another image acquisition run are registered. The beginning of both image acquisition runs will be in 3 with the two input arrows before step 20 indicated.

In step 22 a motion estimation is performed by means of flexible "pixel shift" FPS of the movement range on the basis of in step 20 logarithmically transformed images (logarithmic Transformation LOG). Preferably, the application 24 the motion correction z. B. by deformation DEFORM means "pixel shift" using the estimated range of motion or deformation area 23 after preprocessing PREPRO in step 21 performed in one of the two image acquisition runs. It is also conceivable (not shown in the figure) to perform the deformation before the processing precursor or elsewhere in the 2D processing precursor. In step 25 For example, a reconstruction algorithm EFA is applied to the projection data, for example the Feldkamp algorithm. In the last step 26 a subtraction of the projection data is performed from each other.

These steps can be repeated or restarted as needed.

Fig. 4

In 4. presented an advanced image processing chain for correcting DSA artifacts with 2-D / 3-D registration.

Here, 2-D projection data of one run with the reconstructed volume from a second run is registered. The procedure is similar to the procedure described above 3 , After the volume reconstruction in step 25 of the first run, a projection data set, taken from different directions / angles, with the reconstructed volume in step 27 registered. After the registration REG, a transformation matrix for movement correction is supplied to the reconstruction algorithm EFA, with the aid of which a volume is reconstructed from the 2D projection data. In step 26 the volume data is subtracted from each other.

These steps can be repeated or restarted as needed.

Either 3 as well 4. show integration of a 2-D / 2-D or 2-D / 3-D registration into an image processing chain to determine and estimate a motion between the images of the 2-D projections from the same direction.

A completely subtracted, d. H. Movement-corrected 3-D volume is calculated by not using only the reconstructed 3-D volumes (mask and fill volume) for registration. For registration, the 2-D projection data is included, from which the 3-D images can be reconstructed.

• a) Wie beispielsweise vorstehend zu 3 erläutert: 2-D/2-D-Registrierung korrespondierender Masken- und Füllungsprojektionen, Subtraktion der Projektionsbilder, Rekonstruktion des subtrahierten Volumens aus den subtrahierten registrierten Projektionen.
• b) Wie beispielsweise vorstehend zu 4 erläutert: Rekonstruktion des Maskenvolumens, 2-D/3-D-Registrierung einer oder mehrerer Füllungsprojektionen mit dem Maskenvolumen (Bestimmung einer 3-D-Transformation), Rekonstruktion des Füllungsvolumens, Subtraktion von Masken- und Füllungsvolumen unter Berücksichtigung der bestimmten 3-D-Transformation.
• c) Wie beispielsweise vorstehend zu 4 erläutert: Rekonstruktion des Füllungsvolumens, 2-D/3-D-Registrierung einer oder mehrerer Maskenprojektionen mit dem Füllungsvolumen (Bestimmung einer 3-D-Transformation), Rekonstruktion des Maskenvolumens, Subtraktion von Masken- und Füllungsvolumen unter Berücksichtigung der bestimmten 3-D-Transformation.

The following scenarios are conceivable example:

• a) As for example above 3 explains: 2-D / 2-D registration of corresponding mask and fill projections, subtraction of the projection images, reconstruction of the subtracted volume from the subtracted registered projections.
• b) As for example above 4. Explanation: reconstruction of the mask volume, 2-D / 3-D registration of one or more filling projections with the mask volume (determination of a 3-D transformation), reconstruction of the filling volume, subtraction of mask and filling volume taking into account the determined 3-D transformation Transformation.
• c) For example, as above 4. Explanation: Reconstruction of the filling volume, 2-D / 3-D registration of one or more mask projections with the filling volume (determination of a 3-D transformation), reconstruction of the mask volume, subtraction of mask and filling volume taking into account the determined 3-D transformation Transformation.

By the method according to the invention, a movement-corrected 3-D volume is not generated exclusively on the basis of the 3-D mask and 3-D filling volumes.

The above scenario a) is based on the projection data; the registration in this case is completely independent of the reconstruction.

In scenario b), although a consistent masks take-up run is required, not a completely consistent fill take-up run is required. In scenario c) it is exactly the opposite; a consistent padding run is needed, but not a fully consistent mask picking run.

The present invention preferably relates to the correction of movements that occurred between the two picking runs (or, more precisely, between the pickups of the 2-D projections from the same direction). This makes the procedure robust against inconsistencies in the recording runs. A temporal resolution of the movement is not required, as occurs for example in a 4-D image of the heart.

Furthermore, the inventive method requires only the image data as input. No further signals (such as ECG, motion sensors, external camera systems) are required to perform the motion correction.

By virtue of the procedure according to the invention, movement artifacts can be reduced and thus the image quality can be improved accordingly.

literature

• [Ven01] HW Venema, FJH Hulsmans, GJ the Heeten: CT Angiography of the Circle of Willis and Intracranial Internal Carotid Arteries: Maximum Intensity Projection with Matched Mask Bone Elimination - Feasibility Study, Radiology, Vol. 218 (3), 2001, pp. 893-898 ,
• [Roh01] T. Rohlfing, CR Maurer Jr: Intensity-Based Non-rigid Registration Using Adaptive Multilevel Free-Form Deformation with an Incompressibility Constraint, in MICCAI '01: Proceedings of the 4th International Conference on Medical Image Computing and Computer-Assisted Intervention, Springer -Verlag, London, UK, 2001, pages 111-119 ,
• [VS04] M. van Straten, HW Venema, GJ Streekstra, CBLM Majoie, GJ den Heeten, CA Grimbergen: Removal of bone in CT angiography of the cervical arteries by piecewise matched mask bone elimination, Medical Physics, Vol. 31 (10), 2004, Pages 2924-2933 ,

LIST OF REFERENCE NUMBERS

1

industrial robots

2

C-arm

3

X-ray

4

X-ray image detector

5

Patient table

6

patient

7

Control unit

8th

image system

9

monitor

10

calibration

11

Turning center / axis of rotation

12

ray beam

13

object

14

orbit

20

logarithmic transformation (LOG)

21

physical preprocessing (PREPRO)

22

Motion estimation using flexible pixel shift (FPS)

23

deformation area

24

Deformation by pixel shift (DEFORM)

25

extended Feldkamp algorithm (EFA)

26

subtraction

27

2-D / 3-D registration (REG)

28

transformation matrix

QUOTES INCLUDE IN THE DESCRIPTION

This list of the documents listed by the applicant has been generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Cited patent literature

• US 7315605 B2 [0009]
• DE 4137652 A1 [0010]
• US 6125164 [0011]
• DE 102005012700 A1 [0030]

Cited non-patent literature

• Venema et al. [0012]
• Rohlfing et al. [0012]
• Straten et al. [0012]
• HW Venema, FJH Hulsmans, GJ the Heeten: CT Angiography of the Circle of Willis and Intracranial Internal Carotid Arteries: Maximum Intensity Projection with Matched Mask Bone Elimination - Feasibility Study, Radiology, Vol. 218 (3), 2001, pp. 893-898 [0057]
• T. Rohlfing, CR Maurer Jr: Intensity-Based Non-rigid Registration Using Adaptive Multilevel Free-Form Deformation with an Incompressibility Constraint, in MICCAI '01: Proceedings of the 4th International Conference on Medical Image Computing and Computer-Assisted Intervention, Springer -Verlag, London, UK, 2001, pp. 111-119 [0057]
• M. van Straten, HW Venema, GJ Streekstra, CBLM Majoie, GJ den Heeten, CA Grimbergen: Removal of bone in CT angiography of the cervical arteries by piecewise matched mask bone elimination, Medical Physics, Vol. 31 (10), 2004, Pages 2924-2933 [0057]

Claims (7)

Method for motion correction for a three-dimensional digital subtraction angiography of at least two at least two-dimensional image data sets, from which optionally three-dimensional volumes can be reconstructed, wherein a movement correction ( 8th ) by registration (REG) and then following subtraction ( 26 ), characterized in that for the registration (REG) a registration technique is applied, which is based on at least one two-dimensional and at least one further one of the image data sets, so that as a result a subtraction volume reduced by subtraction artifacts is generated. Method according to the preceding claim, characterized in that a registration technique is used which works on 2-D / 2-D or 2-D / 3-D image data sets. Method according to one of the preceding claims, characterized in that registration registers 2-D projection data of a two-dimensional image data set with 2-D projection data of another two-dimensional image data set or 2-D projection data of a two-dimensional image data set with a reconstructed volume from another two-dimensional image data set become. Method according to one of the preceding claims, characterized in that the at least two at least two-dimensional image data sets comprise at least one mask image data set and at least one filling image data record. Method according to the preceding claim, that the at least one mask image data set and the at least one filling data set are subtracted from each other. Contraption ( 7 ) for motion correction for a three-dimensional digital subtraction angiography of at least two at least two-dimensional image data sets from which optionally three-dimensional volumes can be reconstructed, characterized by means ( 8th ) for movement correction by registration (REG) and then following subtraction ( 26 ) using a registration technique based on at least one two-dimensional and at least one further one of the image data sets, resulting in a subtraction volume reduced by subtraction artifacts. Device according to the preceding claim, characterized by display means ( 9 ), through which the volume can be displayed.

Images