Method for determining a tissue parameter, X-ray equipment and computer program

Abstract

Method for determining a tissue parameter of the tissue that can be determined from the passage of a contrast medium through the tissue on the basis of a series of temporally successive two-dimensional X-ray images (9) of digital subtraction angiography showing the temporal spread of the contrast medium in the tissue and a vascular system (12) present in the area of ​​the tissue, characterized in that - at least some of the vessels of the vascular system (12) are located by segmentation in the X-ray images (9), - segmented vessels are assigned an interpolation intensity determined by interpolation from the intensities of at least some of the pixels surrounding the segmented vessel, so that vessel-corrected X-ray images (9) are produced, and - the tissue parameters for at least some of the pixels of the series of vessel-cleaned X-ray images (9) are determined, - wherein, in the segmentation of the vessels, at least one time-dependent intensity threshold is determined and used to distinguish between tissue intensities and vessel intensities.

Description

[0001] The invention relates to a method for determining a tissue parameter, as determined from the passage of a contrast medium through the tissue, based on a series of temporally successive two-dimensional X-ray images of digital subtraction angiography showing the temporal spread of the contrast medium in the tissue and a vascular system present in the tissue. The invention also relates to an X-ray device and a computer program.

[0002] In angiography, it is common practice to administer contrast agents that are clearly visible in images, such as X-rays, thus allowing for the assessment of blood flow and perfusion within a patient's vascular system and tissues. A classic technique for tracking the spread of the contrast agent is digital subtraction angiography. In this procedure, a mask image without contrast agent is acquired. Subsequently, often as a time series, raw images are acquired within a specific timeframe as the contrast agent traverses the target area or vascular system of interest. These raw images also show the contrast agent. To eliminate interfering anatomical details during evaluation, digital subtraction angiography X-ray images are generated by subtracting the mask image from the raw image, leaving essentially only image information related to the contrast agent.To obtain better orientation, two-dimensional digital subtraction angiography is often performed from several projection directions simultaneously, for example, by using a biplane system. This usually results in raw images and mask images of the target area, particularly from mutually perpendicular projection directions, so that X-ray images for these mutually perpendicular projection directions can be obtained and viewed and evaluated for diagnostic purposes.

[0003] From two-dimensional subtraction angiography X-ray images, a multitude of time parameters that show or are derived from the behavior of the contrast agent can be determined by examining the time-intensity curves (TICs) in the X-ray images. For at least some image points / areas of interest, the image datum (the intensity) of the X-ray image is plotted against the acquisition times for all time points in the series for which an X-ray image is available, resulting in a time-intensity curve, often also referred to as the contrast agent curve. This curve is amenable to classical evaluation methods. For example, the time until the maximum contrast agent concentration at the image point can be determined, which is usually referred to as the "time to peak" (TTP).Another frequently considered time parameter is the mean transit time (MTT), which can be defined in various ways, especially relative to the maximum value of the time-intensity curve.

[0004] Contrast-enhanced examinations are particularly important for the human brain, especially regarding parenchymal blood flow. To conduct these investigations, regions of interest (ROIs) are typically defined in subtraction angiography images, ideally without obscuration by larger vessels. The underlying principle is that integrating the time-intensity curve provides information about the amount of contrast agent that has flowed through a given point in the image. If the time-intensity curves are considered for an entire region of interest, this applies to all structures traversed by the X-rays within that region. From this, cerebral blood volume (CBV) and cerebral blood flow (CBF) can be derived, particularly relative to a reference range.Specifically, the procedure can be described as follows: first, the area A under the time-intensity curve I(t) is determined as. A=∫0TI(t)dt.

[0005] The contrast agent concentration C is proportional to A. The cerebral blood volume CBV is defined as being proportional to... CBV∝CROIC artery.

[0006] If a reference area (reference ROI) is added, the relative blood volume rCBV is calculated as follows: rCBV∝AROIAref.

[0007] The reference ROI does not necessarily have to be taken from the X-ray image itself, although left / right comparisons are frequently performed, but can also be derived from an angiography dataset acquired before an intervention, or similar sources. With appropriate conversion, the corresponding rCBF can also be determined.

[0008] Methods for determining parameters relating to cerebral perfusion are known in the prior art, e.g., from publications US 2008 / 0247503A1 and US 2013 / 0131507A1. Tools are also known in the prior art to assist users in evaluating two-dimensional subtraction angiography (SAT) X-ray images, particularly those acquired with a biplane X-ray system. These tools propose determining time parameters for each individual X-ray image from the time-intensity curves and displaying them, for example, using color coding or a grayscale. Color coding systems are known, for instance, in which early TTP is assigned a red color, mid-TTP a yellow to green color, and high TTP a blue color.Such visualization options also exist for tissue parameters, as it can be relevant in certain medical contexts to visualize only the perfusion or blood flow in the parenchyma, particularly in the brain parenchyma. Examples include examinations following a stroke or in cases of vasospasm. However, with current techniques, the display of tissue parameters is not reliable or even possible for the entire X-ray image, since the large vessels of the vascular tree overlay the parenchyma and high contrast agent concentrations occur there. Other methods for obtaining such tissue parameters include three-dimensional perfusion imaging, but this is not always desirable due to its complexity.

[0009] The invention is therefore based on the objective of providing a way to improve the determination of tissue parameters from two-dimensional DSA x-ray images.

[0010] To solve this problem, a method of the type mentioned at the outset provides according to the invention that - at least some of the vessels of the vascular system are located by segmentation in the X-ray images, - segmented vessels are assigned an interpolation intensity determined by interpolation from the intensities of at least some of the pixels surrounding the segmented vessel, so that vessel-corrected X-ray images are produced, and - the tissue parameters for at least some of the pixels in the series of vessel-cleaned X-ray images are determined.

[0011] The tissue parameters, which are then available as a map, can subsequently be displayed, for example, as a color-coded representation. Relative color-coded blood volume maps and relative color-coded blood flow maps are conceivable, for instance. Thus, a relative blood volume and / or a relative blood flow are preferably determined as tissue parameters. The determination of these parameters has already been described in detail; however, within the scope of the present invention, they are based on an image series that has been at least largely purified of the proportions of vessels, and thus on a time-intensity curve.

[0012] Therefore, a key aspect of the present invention is to remove the interfering portions of large vessels as much as possible, so that tissue parameters can be determined more reliably across the entire series of X-ray images. In this way, the influence of large vessels is reduced and more emphasis is placed on the parenchymal components. In principle, a similar approach to that used for metal artifact correction is proposed, whereby the vessels are segmented and interpolation is then performed for the corresponding pixels. However, the temporal component is a crucial difference here: vessels are only clearly (and disturbingly) visible in some X-ray images due to the contrast agent, which is taken into account in the present invention by segmenting over time.Therefore, what matters is not so much whether a vessel is present at a particular location, but rather whether it is filled with contrast medium, thus interfering with the determination of the tissue parameter. It can therefore be said that contrast-filled vessels of the vascular system are segmented. Corresponding to this segmentation, the result is not just a 2D interpolation, but a (2D+t) interpolation. This reduces the influence of vessel filling on the measured blood volume, which is used to generate maps of the tissue parameter that are as free of superimposition as possible.

[0013] The method according to the invention is advantageously fully automated and applicable to any DSA-2D series. It produces a representation of vascular parameters that correlates with three-dimensional perfusion images.

[0014] When determining tissue parameters referenced to a specific area in the X-ray images of a series, it is advisable to use a reference area containing few or no segmented vessels as the reference area for determining the relative tissue parameters. This minimizes the impact of interpolation on the reference area and provides inherently highly reliable values.

[0015] During vessel segmentation, at least one intensity threshold is used to differentiate between tissue and vessel intensities. This means that intensity thresholds are determined by which vessel filling can be distinguished from tissue filling. It is preferred that the intensity threshold be determined at least partially, and in particular fully, automatically, and / or relative to a maximum intensity in a feeding artery. A feeding artery, for example, the carotid artery in the case of imaging of the human brain, can be easily located automatically and exhibits the highest contrast agent concentration that will occur. Therefore, the intensity threshold can be determined particularly easily relative to the filling in the largest artery.Advantageously, the intensities occurring over time and the spread of the contrast agent are analyzed to automatically determine the intensity threshold. This is because, for example, it is known that initially quite high intensities occur in the arteries due to the contrast agent, while lower intensities occur during outflow in the veins, resulting in a lower threshold value. This allows for the definition of intensity ranges that can be used to assign the contrast agent to the parenchyma or to a filled vessel. Overall, a preferred embodiment of the invention provides that the intensity threshold is automatically determined from a temporal analysis of the occurring intensities, particularly based on the identification of an arterial and a venous phase.

[0016] According to the invention, a time-dependent intensity threshold is determined, in particular different intensity thresholds for an arterial phase, a venous phase, and a tissue phase. As already mentioned, the highest contrast agent concentration occurs when the contrast agent bolus enters the target area via the supplying artery, whereupon the contrast agent distributes itself into the decreasing arteries, flows through the tissue, and is transported away again via the veins at a significantly lower intensity. This means, however, that different intensity thresholds over time may be advantageous in order to distinguish the tissue from the vessels filled with contrast agent. For example, a temporal division into three categories is conceivable: an arterial phase, a tissue phase, and a venous phase.However, finer divisions are also conceivable, in particular the determination of an intensity threshold for each time point at which an X-ray image of the series is available.

[0017] Specifically, the segmentation and interpolation can ultimately be performed such that for each pixel or, if another division of the X-ray images is chosen, image segment, a maximum time of maximum intensity is determined, whereby if the maximum intensity exceeds the intensity threshold, or, if a time-dependent intensity threshold is used, the intensity threshold assigned to the maximum time, then, for each time at which an X-ray image is available, within a time interval extending around the maximum time, at least two neighboring pixels or image segments as closely as possible to the pixel or image segment under consideration, whose image data lie below the intensity threshold, or, if a time-dependent intensity threshold is used, the intensity threshold assigned to the time, are identified.The pixel or image segment under consideration is located, and an intensity interpolated from its neighboring pixels or image segments is assigned to the pixel or image segment being considered. Thus, for each pixel or image segment (or, if larger image segments containing multiple pixels are to be considered, the maximum time point at which the highest intensity is present across the entire series of X-ray images is determined). This pixel or image segment is then considered a vascular pixel or vascular image segment if the maximum intensity exceeds the intensity threshold, possibly at the maximum time point. If this is the case, local environmental segmentation is performed for this pixel or image segment, for example, using the "region growing" method, again taking into account the intensity threshold assigned to the maximum time point, if applicable. As soon as at least two, and possibly more, pixels or image segments have been identified,If the values ​​are no longer attributable to a contrast-filled vessel but to tissue, the values ​​of these neighboring pixels or image segments can be used to perform a 2D interpolation and replace the value for the pixel or image segment under consideration. Since the intensity threshold corresponding to a neighboring time point may still be exceeded at adjacent time points, the procedure is also extended to a time interval around the maximum time point to ensure correct replacement there as well. This results in a 2D+t interpolation.

[0018] An efficient embodiment of the present invention provides that interpolation is also performed for vessel pixels or vessel image segments located between the considered pixel or image segment and its neighboring pixels or neighboring image segments, with the interpolated values ​​being used when evaluating the next pixel or image segment to be considered. In this way, interpolation always takes place at least in the local vicinity of a detected vessel point, thus avoiding the need to individually consider the remaining pixels later on, since they have already been processed together with a previously considered pixel or image segment and should no longer exceed the intensity threshold. Region-growing segmentation, as already mentioned, proves to be extremely useful for locating pixels or image segments for which interpolation is to be performed.Nevertheless, it is always advisable to record which pixels or image segments were detected as belonging to (contrast-filled) vessels during the segmentation and interpolation process.

[0019] A particularly advantageous choice of time interval is achieved when the time interval is defined as the transit time of the contrast agent determined from the time-intensity curve of the pixel or image segment under consideration, or as all time points at which the intensity of the pixel or image segment under consideration exceeds the intensity threshold, or, if a time-dependent intensity threshold is used, the intensity threshold assigned to the time point. In this way, all time points at which the pixel or image segment can be identified as representing a vessel filled with contrast agent are captured. Thus, the segmentation is ultimately extended in the time dimension, which then also applies to the interpolation.

[0020] It should be noted at this point that segmentation according to the TTP (which corresponds to the maximum time point) is also conceivable in principle, but is less preferred since there are overlapping areas in which the arteries have already partially delivered the contrast agent to the tissue, while more distant arteries still carry contrast agent themselves, and so on.

[0021] Furthermore, it is pointed out that other segmentation methods, such as segmentation methods oriented to the vascular tree, segmentation methods taking into account a three-dimensional image data set registered with the X-ray images, at least partially manual segmentation methods and the like, are also conceivable in principle to locate vascular pixels or vascular image segments, after which interpolation can then take place.

[0022] Interpolation can be linear or performed using a spline function, particularly a thin-plate spline function. However, other interpolation methods are also conceivable, which can at least partially attempt to accurately represent even more complex relationships and structures.

[0023] A particularly advantageous embodiment of the invention provides that a vessel model is created from pixels segmented to represent a vessel. This model is displayed together with tissue parameters, particularly when these parameters are shown, for orientation purposes. In this way, the identification of vessel pixels (or vessel image segments) is used not only for interpolation and thus "vessel correction," but also, since vessels are identified, for creating a vessel model. For this purpose, during the interpolation process, all pixels or image segments that are identified as belonging to a vessel at at least one point in time can be recorded and form the vessel model. In a further processing step, it is conceivable, for example, to determine vessel boundaries and / or their midlines to enable a suitable representation of the vessel model.An additional advantage is that, due to the temporal segmentation, it is also conceivable to differentiate between different vessels in the vascular model, for example by examining the TTP in a local neighborhood for jumps, and so on.

[0024] It may be possible to divide the vessel-corrected X-ray images into image segments using a grid to determine tissue parameters. To average out local noise fluctuations and / or minimize other measurement errors for individual points, it is already known in the art to determine tissue parameters for larger regions of interest (ROIs). For example, prior art methods are known in which ROIs are manually selected that, in the user's opinion, are least affected by superimposed blood vessels.Since the method described here allows for a significant reduction in the influence of blood vessels across the entire image, it can be advantageous to define the division into areas using a grid overlaid on the vessel-corrected X-ray images as image segments. These segments could, for example, be square and contain a specific number of pixels. However, it should be noted that it is also conceivable to use image segments for interpolation, as indicated there, that are likewise defined by such a grid, possibly even the same grid.

[0025] A particularly advantageous embodiment of the present invention arises when tissue parameters are determined for different grid positions, wherein the final tissue parameter assigned to a pixel is determined by combining, in particular averaging, the tissue parameters determined at different grid positions. It is therefore conceivable to determine different grid positions, and thus differently defined image segments, which accordingly overlap at least partially. If mean values ​​of the image segments contributing to the overlapping areas are now used, a smoothed final result for the tissue parameters is obtained, one that is less affected by statistical fluctuations. The grid movement between different grid positions can occur in a specific direction; however, different directions of movement with corresponding grid positions are also conceivable.In this design, each image segment ultimately forms a kind of sliding window, so that overall one can speak of a technique using a sliding grid.

[0026] In addition to the method, the invention also relates to an X-ray device comprising a control unit designed for carrying out the method according to the invention. All aspects relating to the method according to the invention can be applied analogously to the X-ray device according to the invention. The X-ray device is preferably an angiography X-ray device, for example, an X-ray device with a C-arm that can be positioned at a specific angulation relative to the patient in order to acquire the mask and raw images underlying the X-ray images of digital subtraction angiography. The use of so-called biplane X-ray devices is also particularly advantageous, in which two imaging arrangements, each with an X-ray detector and an X-ray source, are provided, which can be arranged at different angulations, for example, each assigned to its own C-arm.This allows X-ray images from different projection directions to be generated, whereby both series of X-ray images produced in this way can be evaluated using the method according to the invention.

[0027] The control unit, in its specific configuration, can therefore comprise a segmentation unit for segmenting vessels of the vascular system in the X-ray images, an interpolation unit for interpolating vessel pixels or vessel segments, and a tissue parameter determination unit for determining the tissue parameters. Further conceivable functional units include acquisition units for the actual acquisition of the X-ray images and a display unit that can show a representation of the tissue parameters, in particular superimposed on a vascular model.

[0028] Finally, the invention also relates to a computer program that performs the steps of the inventive method when executed by a computing device. The computing device can, for example, be the control unit of an X-ray device according to the invention. All aspects of the inventive method can also be applied analogously to the inventive computer program. The computer program can be stored on a non-transient data carrier, for example, a CD-ROM.

[0029] Further advantages and details of the present invention will become apparent from the exemplary embodiments described below and from the drawings. These show: Fig. 1 a drawing for the biplane survey, Fig. 2. An X-ray image with an associated time-intensity curve in a projection direction, Fig. 3 a flow chart of the inventive method, Fig. 4. A schematic diagram for the segmentation of vessel pixels, Fig. 5 a sketch for interpolation, Fig. 6. A division of an X-ray image into a grid, Fig. 7. A sketch showing different grid positions. Fig. 8 a possible representation of the tissue parameters, and Fig. 9 an X-ray device according to the invention.

[0030] The embodiment of the method according to the invention, which will be discussed below, concerns the determination of tissue parameters in the human brain, specifically relative cerebral blood volume and relative cerebral blood flow (rCBV and rCVF). Thus, the present case involves an examination of a patient's head, with the target being the parenchyma. The method according to the invention requires a series of digital subtraction angiography X-ray images. In particular, before the administration of the contrast agent, a mask image is first acquired for each projection direction considered (for example, two projection directions in biplane examinations) with the patient in position. After administration of the contrast agent, a time series of raw images is acquired, showing the spread of the contrast agent in the vascular system, here the cerebral vascular system, and in the tissue.In digital subtraction angiography, the X-ray images are obtained by subtracting the mask image taken in the same projection direction from each raw image of the time series, so that the anatomical component is largely removed and only the components originating from the contrast medium remain, thus making the contrast-filled vessels or tissue clearly visible.

[0031] One possible geometry for capturing such X-ray images shows Fig. 1. The aim is to image an object 1, specifically the vascular system and tissue of interest as the target area, with a contrast-filled vessel 2 being highlighted as an example. The biplane X-ray system used here as an example comprises two imaging arrangements, each with an X-ray source 3, 4 and an X-ray detector 5, 6, whose projection directions are preferably perpendicular to each other for the purpose described here. As indicated by arrows 7, the vessel 2 is projected onto different locations of the respective X-ray detectors 5, 6, where it therefore also appears in the X-ray image.

[0032] Fig. Figure 2 shows a rough schematic diagram of the basic structure of an X-ray image 9, where the possible outer outline 8 of a patient's skull is shown only for orientation, as the corresponding signal is no longer present in the X-ray image 9 due to digital subtraction angiography. The X-ray image 9 shows both the parenchyma 11, in which the contrast agent is distributed and therefore present in a lower concentration, and the vascular system indicated by Figure 12. If one considers a specific point or pixel 13 and the intensity measured there over the time series, a time-intensity curve 14 is obtained, which can be evaluated for time parameters and can also serve as a basis for determining tissue parameters.Examples of time parameters are the time to maximum contrast agent concentration (TTP) and the mean transit time (TTP), which can be defined in various ways, in this case as half the intensity at TTP (TTP), which will also be referred to as the maximum time point for pixel 13. The methods for determining tissue parameters such as relative cerebral blood volume and relative cerebral blood flow for areas in the parenchyma have already been described.

[0033] The inventive method, which will be referred to below with reference to Fig. 3, which will be explained in more detail, provides a way to determine the tissue parameters as much as possible across the entire X-ray images, thereby reducing / minimizing the influence of contrast-filled vessels that overlay the tissue.

[0034] In step S1, a time-dependent intensity threshold is first determined for the time series of X-ray images. This intensity threshold is intended to differentiate between intensities found in vessels and intensities found in tissue at the time points at which X-ray images are available. Such an intensity threshold can be determined for each time point at which X-ray images are available, or for specific time intervals, for example, for an arterial phase, a venous phase, and a tissue phase. The highest contrast agent concentration occurs at the time the contrast agent bolus enters the target area via the feeding artery. Subsequently, the contrast agent spreads into the decreasingly smaller arteries, flows through the tissue, and is then transported away again by the veins at a lower intensity.This progression of contrast agent spread can be automatically analyzed, starting from the arrival of the contrast agent bolus in the supplying artery, in this case, for example, the carotid artery. Over time, this reveals different intensity thresholds, which make it possible to distinguish the tissue from the vessels filled with contrast agent.

[0035] In step S2, the first pixel (or, if applicable, an image segment comprising several pixels, if such a division is chosen) of the X-ray images in the series is selected, and the corresponding time-intensity curve 14 is examined to locate the maximum time point, i.e., time 15 of maximum contrast agent concentration. In step S3, the intensity at the maximum time point is compared with the intensity threshold assigned to that maximum time point, as determined in step S1. If this comparison indicates that the pixel is likely a tissue pixel, the procedure continues with the next pixel in step S2; however, if a vascular pixel is identified, the procedure continues in step S4.

[0036] The following steps S4 and S5 are performed for each X-ray image within a time interval around the maximum time point. This time interval is defined as encompassing all time points at which the intensity of the pixel under consideration exceeds the intensity threshold assigned to that time point. Therefore, the time interval includes all time points at which the pixel would be classified as a vascular pixel.

[0037] In step S4, a region-growing segmentation is performed starting from the pixel under consideration, which has been identified as a vessel pixel, until at least two neighboring pixels are found that are located as close as possible to the pixel under consideration and that are assigned to the tissue, thus not exceeding the intensity threshold assigned to the time point. This is done with regard to Fig. Figure 4 explains in more detail, showing a section 17 of an X-ray image 9 at a specific time. The currently viewed pixel 18 is marked with an "X" and lies within a vessel course indicated by dashed lines 19. Starting from the viewed pixel 18, further vessel pixels in the vicinity are now found in the region-growing algorithm that also exceed the corresponding intensity threshold and which are displayed in Fig. 4 are shown hatched. Pixels not shown hatched are potential neighboring tissue pixels in which the intensity value consequently falls below the intensity threshold for the time of section 17.

[0038] For example, the neighboring pixels for the interpolation following in step S5 at the considered pixel 18 can be those in Fig. Four pixels marked with a circle are used. The intensity value of the considered pixel 18 is then interpolated from the intensity values ​​of the neighboring pixels, in this case by interpolation using thin-plate splines. Of course, other interpolation methods are also conceivable.

[0039] It may also be intended that the other vessel pixels found in step S4 (hatched in) Fig. 4) Interpolation is performed now to obtain a more efficient approach, whereby, of course, further neighboring pixels can also be used. Each pixel that has been identified as a vessel pixel at any time is also marked and inserted into a vessel model that describes the course of the vessels.

[0040] To indicate that segmentation and interpolation also occur in the temporal direction, the following are shown in Fig. 4 also temporally adjacent image sections 17' and 17" are shown.

[0041] Fig. Figure 5 shows, as a further schematic diagram, the result of the interpolation in step S5, from which it can be seen that none of the pixels are shown hatched anymore, meaning that the intensity threshold has been undercut everywhere and the overlay by the container has been largely eliminated.

[0042] After this interpolation has been performed for all X-ray images within the time interval, step S6 checks whether all pixels of the X-ray images have already been processed; if this is not the case, the process continues for the next pixel in step S3.

[0043] After all vessel pixels have been located and interpolated in this way, the tissue parameters are determined in step S7. This is not always done pixel by pixel, but rather, at least as an intermediate step, the determination is carried out using larger image segments (often also called sectors). In this case, the image is divided into segments by using a grid 20, as shown in Fig. Figure 5 shows the image overlaid with a roughly sketched X-ray image 9. The grid defines image segments 21, each containing a specific number of pixels. One of these image segments 21, in this case image segment 21a, is used as a reference image segment for determining the tissue parameters, specifically rCBV and rCBF. Image segment 21a is selected to contain as few vessels as possible, thus minimizing the impact of interpolation.

[0044] In Fig. Figure 6 already shows the vessel-cleaned X-ray image 9, which means that only the parenchyma 11, i.e. the tissue of interest, can ultimately be seen as filled with contrast medium.

[0045] For each image segment 21, the rCBV and the rCBF can now be determined, as is generally known in the prior art, via the area under the time-intensity curve 14 for the image segment under consideration and the reference image segment 21a.

[0046] However, in step S7, a more spatially precise determination of the tissue parameters is also carried out by using a sliding grid 20, as in Fig. 7 is shown, which is next to grid 20 in the already in Fig. In the position shown (here depicted as a dashed line), grid 20' is also shown in a position shifted relative to the position shown, thus forming overlapping image segments 21. This can be carried out for a multitude of possible positions and thus a multitude of possible image segments that overlap in various ways, so that for each pixel, for example, an improved value for the tissue parameter can be determined by averaging the image segments containing this pixel from the different grid positions.

[0047] In step S8, the tissue parameter can then be displayed as a two-dimensional representation, for example, color-coded. Figure 22 shows a section of such a representation. Fig. 8. As indicated by the different hatching patterns, pixels representing different tissue parameters are displayed in different colors. For orientation, similar to a "map," the representation also includes the boundaries of the segmented vessels, which were saved as a vessel model in step S4. Since temporal information is also available, connected vessels can be identified, and consequently, their boundaries and / or midlines can be determined. In this way, the representation of the tissue parameter can be usefully expanded.

[0048] Fig.Figure 9 shows an X-ray device 24 according to the invention. This is a biplane X-ray device comprising two C-arms 25, 26, on which, as explained above, an X-ray source 3, 4 and an X-ray detector 5, 6 are arranged opposite each other. The C-arms 25, 26 are rotatable in a plane 28 about the axis of rotation 29 via a pivot bearing 27. The C-arms 25, 26 are held by a support 30, which is shown here only in outline. A patient table 31 is provided for positioning a patient.

[0049] The operation of the X-ray device 24 is controlled by a control unit 32, which is designed to carry out the method according to the invention, and on which, for example, a computer program according to the invention is available.

[0050] Although the invention has been illustrated and described in detail by the preferred embodiment, the invention is not limited by the disclosed examples and other variations can be derived by the person skilled in the art without leaving the scope of protection of the invention.

Claims

[1] Method for determining a tissue parameter of the tissue that can be determined from the passage of a contrast medium through the tissue on the basis of a series of temporally successive two-dimensional X-ray images (9) of digital subtraction angiography showing the temporal spread of the contrast medium in the tissue and a vascular system (12) present in the area of ​​the tissue, characterized by , that - at least some of the vessels of the vascular system (12) are located by segmentation in the X-ray images (9), - segmented vessels are assigned an interpolation intensity determined by interpolation from the intensities of at least some of the pixels surrounding the segmented vessel, so that vessel-corrected X-ray images (9) are produced, and - the tissue parameters for at least some of the pixels of the series of vessel-cleaned X-ray images (9) are determined, - wherein, in the segmentation of the vessels, at least one time-dependent intensity threshold is determined and used to distinguish between tissue intensities and vessel intensities. [2] Method according to claim 1, characterized by that a relative blood volume and / or a relative blood flow are determined as tissue parameters. [3] Method according to claim 2, characterized by , that a reference area containing no or as few segmented vessels as possible is used for determining the relative tissue parameters. [4] Method according to any of the preceding claims, characterized by that the intensity threshold is determined at least partially, in particular completely automatically, and / or relative to a maximum intensity in a feeding artery. [5] Method according to claim 4, characterized by, that the intensity threshold is automatically determined from a temporal analysis of the occurring intensities, in particular based on the identification of an arterial and a venous phase. [6] Method according to any of the preceding claims, characterized by , that different intensity thresholds are determined for an arterial phase, a venous phase and a tissue phase. [7] Method according to any of the preceding claims, characterized by, that for each pixel (18) or, if a different division of the X-ray images (9) is chosen, image segment, a maximum time of maximum intensity is determined, wherein, if the maximum intensity exceeds the intensity threshold, or, if a time-dependent intensity threshold is used, the intensity threshold assigned to the maximum time, at least two neighboring pixels or image segments as closely as possible to the pixel (18) or image segment under consideration, whose image data are below the intensity threshold, or, if a time-dependent intensity threshold is used, the intensity threshold assigned to the time, are located, and an intensity interpolated from the neighboring pixels or image segments is assigned to the pixel (18) or image segment under consideration. [8] Method according to claim 7, characterized by , that interpolation is also performed for vessel pixels or vessel image segments located between the considered pixel (18) or image segment and the neighboring pixels or neighboring image segments, whereby the interpolated values ​​are used when evaluating the next pixel (18) or image segment to be considered. [9] Method according to claim 7 or 8, characterized by , that the time interval is defined as a transit time of the contrast agent determined from the time-intensity curve of the pixel (18) or image segment under consideration, or as all time points at which the intensity of the pixel (18) or image segment under consideration exceeds the intensity threshold, or, if a time-dependent intensity threshold is used, the intensity threshold associated with the time point. [10] Method according to any of the preceding claims, characterized bythat the interpolation is linear or performed using a spline function, in particular a thin-plate spline function. [11] Method according to any of the preceding claims, characterized by , that a vessel model is created from pixels segmented to represent a vessel, which is displayed together with tissue parameters for orientation, especially when tissue parameters are displayed. [12] Method according to any of the preceding claims, characterized by , that the vessel-corrected X-ray images are divided into image segments (21, 21a) using a grid (20, 20') for determining the tissue parameters, for which the tissue parameter is determined. [13] Method according to claim 12, characterized by, that the determination of tissue parameters for different positions of the grid (20, 20') is carried out, wherein the final tissue parameter assigned to a pixel is determined by combination, in particular by averaging, from the tissue parameters determined in different grid positions. [14] X-ray apparatus (24) comprising a control device (31) designed to carry out a method according to one of the preceding claims. [15] Computer program which performs the steps of a method according to any one of claims 1 to 13 when executed on a computing device.

Images