Evaluation method for a temporal sequence of x-ray images

This application claims priority of German application No. 10 2007 028 226.7 filed Jun. 20, 2007, which is incorporated by reference herein in its entirety.

The present invention relates to an evaluation method for a temporal sequence of x-ray images of an examination region of an examination object.

Such an evaluation method is known for instance from the technical paper “Quantitative Analyse von koronarangiographischen Bildfolgen zur Bestimmung der Myokardperfusion” [Quantitative analysis of coronary-angiographic image sequences for determining myocardial perfusion] by Urban Malsch et al., published in “Bildverarbeitung für die Medizin 2003” [Image processing for medical science 2003], Springer Verlag. With this evaluation method, a computer receives the temporal sequence of x-ray images and determines a perfusion of the tissue supplied with blood in a myocardial region.

DE 10 2005 039 189 A1 likewise discloses an evaluation method of the afore-described type. With this method, a computer also receives the temporal sequence of x-ray images. In each instance the computer determines a characteristic value for each pixel of an evaluation image for each of the x-ray images on the basis of the data values of the pixels of the respective x-ray image, which lie in an evaluation core which is standard for all x-ray images and is defined by the respective pixel of the evaluation image. On the basis of the temporal sequence of the characteristic value, it assigns the respective pixel to one of the categories comprising vessel, perfusion region and background. The computer performs further evaluations for the pixels which the computer has assigned to the perfusion region.

DE 10 2004 057 026 A1 discloses an evaluation method for several images of an examination region of an examination object which have been detected at different times. The images may be x-ray images. A computer receives the images, determines the deviations of the images from one another and performs further evaluations as a function of the “change images” thus determined.

The evaluation methods of the prior art already function pretty effectively. They can however still be further improved.

The object of the present invention consists in creating possibilities, by means of which the temporal information relating to the blood flow into the vessels can be seen more easily.

The object is achieved by an evaluation method for a temporal sequence of x-ray images of an examination region of an examination object,

with the examination region including a blood vessel system and tissue supplied with blood by way of said blood vessel system, with a detection time being assigned in each instance to the x-ray images, the x-ray images corresponding locally with one another in terms of pixels and the x-ray images each displaying a distribution of a contrast agent in the examination region provided at the respective detection time, with a computer receiving the sequence of x-ray images, determining the temporal course of the average value of the data values of the pixels located in the evaluation region for at least one evaluation region which is standard for all x-ray images and assigning a type to the evaluation region as a function hereof.

In accordance with the invention, the computer receives the sequence of x-ray images. The computer determines the temporal course of the temporal derivation of the data values of the pixels located in the evaluation region for at least one evaluation region which is standard for all x-ray images.

Alternatively or in addition, the computer determines the temporal course of the temporal derivation of the average value of the data values of the pixels located in the evaluation region. The computer assigns a type to the evaluation region as a function of the determined temporal course and/or the determined temporal courses.

The computer program is correspondingly embodied such that the computer proceeds as above.

The core of the present invention thus consists in no longer operating with the temporal courses of the data values themselves but instead with the temporal course of the temporal derivation of the data values.

Within the scope of the evaluation, it is possible for instance for the computer to determine whether the temporal course of the temporal derivation of the average value of the data values of the pixels located in the evaluation region comprises a local maximum and a local minimum and assigns a type of a first type group and on the other hand a type of a second type group to the evaluation region in the event of the absence of the local maximum and of the local minimum.

If the local maximum and/or the local minimum are not present, the computer can

assign a type zero to the evaluation region for instance if the absolute value of the temporal course remains below a fluctuation limit, and can assign a type one to the evaluation region for instance if the temporal derivation of the average value of the data values of the pixels located in the evaluation region is globally greater than zero and the maximum value of the temporal derivation of the average value of the data values of the pixels located in the evaluation region exceeds the fluctuation limit.

In the event that the local maximum and the local minimum are present, the computer can determine the position and/or width of the local maximum and of the local minimum for instance and can assign a type two or a type three to the evaluation region as a function hereof.

The types zero and one thus form the first type group. The types two and three form the second type group.

The fluctuation limit can be fixedly predetermined. Alternatively, it can be predetermined to the computer by the user (if necessary also interactively). Alternatively, it is possible for the computer to automatically determine the fluctuation limit on the basis of the sequence of x-ray images.

To decide whether the computer assigns the type two or the type three to the evaluation region, it is possible for the computer to determine the temporal sequence of the temporal derivation of the average value of the data values of the pixels located in the reference region for a reference region which is standard for all x-ray images, to determine the position and/or width of a local maximum and a local minimum of this temporal course as reference values and to decide whether to assign the type two or the type three to the evaluation region as a function of the position and/or width of the local maximum and of the local minimum of the temporal derivation of the average value of the data values of the pixels located in the evaluation region relative to the corresponding reference values.

In a particularly preferred embodiment, provision is made for the computer to

assign the type two to the evaluation region if the local maximum and/or the local minimum of the temporal derivation of the average value of the data values of the pixels located in the evaluation region are temporally delayed and/or propagated relative to the corresponding reference values, and on the other hand to assign the type three.

It is possible for the evaluation region to only include one single pixel. It is alternatively possible for the evaluation region to include several pixels, for instance a block of 4×4 pixels. Other variables are however also possible.

If the evaluation region includes several pixels, it is possible for the computer, in the event that it assigns type two to the evaluation region, to determine the temporal course of the temporal derivation of the data values of these pixels for the pixels located in the evaluation region, and to evaluate these temporal courses so as to decide on the basis of this evaluation whether it retains the type two for the evaluation region or whether it assigns a type two-a or two-b to the evaluation region.

The last-mentioned procedure can be configured such that the computer determines whether the temporal courses of the temporal derivation of the data values of the pixel located in the evaluation region each comprise a local maximum and a local minimum, in the event that the local maximum and the local minimum are present in each instance, and determines the position and/or width of the respective local maximum and of the respective local minimum and

assigns the type two-a to the evaluation region if the computer assigns the type two as a maximum to a first portion of the pixels located in the evaluation region and assigns a type of the first type group to the remaining pixels on the basis of a procedure in line with assigning the types of the first type group and of the type two and assigns the type two-b to the evaluation region if the computer assigns the type two to at least a second portion of the pixels located in the evaluation region on the basis of a procedure in line with assigning the type two.

The computer generally assigns a display property to a display region of an evaluation image which corresponds to the at least one evaluation region as a function of the type assigned to the evaluation region and outputs the evaluation image to a user. For instance, the evaluation image may be a color-coded display of the types assigned to the different evaluation regions.

In a preferred embodiment, provision is made for the computer to receive a closed surface as the preliminary evaluation region and to determine the evaluation region as a result such that it separates the pixels of the x-ray images which lie within the preliminary evaluation region and which correspond to the blood vessel system. This procedure is on the one hand user-friendly and enables a simple specification of the evaluation region, which on the other hand still provides for a high-quality evaluation.

Further advantages and details result from the description below of exemplary embodiments in conjunction with the drawings, in which;

FIG. 1 shows a schematic diagram of a block diagram of an evaluation arrangement,

FIG. 2 shows a schematic diagram of a flow chart,

FIG. 3 shows a schematic diagram of an x-ray image,

FIGS. 4 to 11 show schematic diagrams of time diagrams and

FIGS. 12 to 18 show schematic diagrams of flow charts.

In accordance with FIG. 1 , a computer has conventional components 1 to 6 . In particular, the computer has a microprocessor 1 , a working memory (RAM) 2 , a mass storage device 3 (for instance a hard drive), a user interface 4 , a data interface 5 and a programming interface 6 . The components 1 to 6 are embodied in a conventional manner and also interact with one another in a conventional manner. The user interface 4 can include conventional input and output facilities for instance, like for example a keyboard, a mouse, a display device etc. The data interface 5 can be an internet or a LAN interface or a USB interface for instance. An embodiment as a drive for a removable storage device (for instance a CD-ROM or a DVD) is also possible. Similar embodiments are possible for the programming interface 6 . If necessary, the data interface 5 and the programming interface 6 can be combined to form a common interface.

A computer program 7 is supplied to the computer by way of the programming interface 6 . A data carrier 8 , upon which the computer program 7 is stored in machine-readable form, is connected to the computer for instance. The computer program 7 is consequently read out from the data carrier 8 and is copied into the mass storage device 3 of the computer and is thus likewise stored there.

The computer program 7 contains machine codes 9 , in other words, program instructions, which can be run directly and immediately by the computer. The computer program 7 can be called by a user 10 by means of conventional input commands (for instance a double mouse click). If the computer program 7 is called, it is loaded into the working memory 2 of the computer and run by the computer. The execution of the computer program 7 by the computer causes the computer to run a determination method, which is described in detail below in conjunction with the additional FIGS.

According to FIG. 2 , in step S 1 , the computer receives a temporal sequence of x-ray images B. Based on the fact that the sequence is a temporal sequence, a detection time t is assigned to each x-ray image B, at which time the respective x-ray image B was detected. A temporal distance between two directly consecutive detection times t is in this case generally less than a second. It is in most cases even less than a quarter of a second, for instance less than a tenth of a second.

Each x-ray image B generally displays a two-dimensional distribution of a contrast agent in an examination region of an examination object provided at the respective detection time t. The contrast agent is herewith preferably injected into an artery 11 supplying the examination region (see FIG. 3 ). The injection is preferably carried out rapidly, in other words with a short duration and at the beginning of a high leading edge and at the end of a short trailing edge.

The examination object is generally a human. The examination region can be the brain region of the human for instance. By way of example, FIG. 3 shows an x-ray image B of an examination object.

In accordance with FIG. 3 , the examination region includes a blood vessel system with arteries 11 and veins 12 as well as tissue 13 , with the tissue 13 being supplied with blood by way of the blood vessel system. Capillary regions K, in which the tissue anabolism between the blood and the tissue 13 takes place, thus exist between the arteries 11 of the blood vessel system and the veins 12 of the blood vessel system, see FIG. 3 .

The x-ray images B are generally detected using constant imaging geometry. The x-ray images B thus correspond locally with one another in terms of pixels.

In step S 2 , the computer receives a selection (at least) of an evaluation region 14 (see also FIG. 3 ) from the user 10 . The evaluation region 14 is standard for all x-ray images B. The specification of the evaluation region 14 results in the computer only executing subsequent steps within the evaluation region 14 .

Step S 2 is only optional and could thus be omitted. For this reason, it is only shown in FIG. 2 with a dashed line. When it is omitted, the evaluation region 14 is either fixedly predetermined or is automatically determined by the computer.

As appropriate, the computer also receives further input parameters within the scope of step S 2 . Reference is made again to examples of parameters of this type.

In step S 3 , the computer determines a temporal course of the data values of the respective pixel 15 and/or of the average value of the pixels 15 of the evaluation region 14 for each pixel 15 of the evaluation region 14 and/or for the average value of the pixels 15 of the evaluation region 14 . FIGS. 4 to 7 show exemplary typically possible temporal courses of the data values.

In step S 4 , the computer executes a smoothing operation, for instance a low pass filtering or a fitting of a parameterizable curve. Step S 4 is optional here. For this reason, it is only shown in FIG. 2 with a dashed line. It can be omitted.

In step S 5 , the computer determines the temporal derivation of the temporal courses determined in step S 3 . In the event of the absence of step S 4 , the computer assumes the temporal curves smoothed in step S 4 . By way of example, FIGS. 8 to 11 illustrate the temporal derivations of the temporal courses shown in FIGS. 4 and 7 .

In step S 6 , the computer performs further evaluations on the basis of the temporal derivations determined in step S 5 . In step S 7 , the computer assigns a display property to a display region of an evaluation image A, which corresponds to the evaluation region 14 , as a function of the type which the computer has assigned to the evaluation region 14 . For instance, the computer can assign a type-specific color to the display region. In step S 8 , the computer outputs the evaluation image A to the user 10 .

In step S 9 , the computer verifies whether the additional implementation of the evaluation method is to be terminated. If this is not the case, the computer returns to step S 2 . On the other hand, the execution of the evaluation method is terminated.

One possible embodiment of step S 2 in FIG. 2 is shown in more detail below in conjunction with FIG. 12 .

According to FIG. 12 , in step S 11 , the computer receives the specifications of a preliminary evaluation region 14 ′. The preliminary evaluation region 14 ′ corresponds here, see FIG. 3 , to a closed surface. Step S 11 corresponds to one part of step S 2 in FIG. 2 .

In step S 12 , the computer receives the additional input parameters.

In step S 13 , the computer determines those pixels 15 which lie within the preliminary evaluation region 14 ′, and which correspond to the blood vessel system 11 , 12 . Step S 14 allows separating the pixels determined in step S 13 from the preliminary evaluation region 14 ′. In this embodiment, the evaluation region 14 thus only includes the pixels 15 of the preliminary evaluation region 14 ′, which do not form an integral part of the blood vessel system 11 , 12 .

Step S 13 in FIG. 12 is only optional and is only shown with a dashed line in FIG. 12 for this reason. If it is omitted, within the scope of the embodiment in FIG. 12 , the corresponding pixels 15 must be made known to the computer in another manner.

Different procedures are possible in order to realize the evaluation of the temporal derivations. Preferred evaluation methods which are based on one another are described in more detail below in conjunction with FIGS. 13 to 17 .

According to FIG. 13 , in step S 21 , the computer firstly determines a logical variable OK. The logical variable OK then assumes the value “TRUE” and only then if the temporal course of the temporal derivation of the average value of the data values of the pixels 15 located in the evaluation region 14 has both a local maximum 16 and also a local minimum 17 . The term “local maximum” is to mean here that the temporal derivation has lower values prior to and after the maximum value of the temporal derivation. The term “local minimum” is to be understood in a similar fashion.

In step S 22 , the computer verifies the value of the logical variable OK. Depending on the result of the verification of step S 22 , the computer subsequently performs a step S 23 or a step S 24 .

In step S 23 , the computer assigns a type of a first type group to the evaluation region 14 . In step S 24 , the computer assigns a type of a second type group to the evaluation region 14 . For instance, the computer can assign a type of the first type group to temporal derivations according to FIGS. 8 and 9 , and a type of the second type group to temporal derivations according to FIGS. 10 and 11 .

According FIG. 14 , in other words in the event that the local maximum 16 and/or the local minimum 17 are not present on the one hand and that the local maximum 16 and the local minimum 17 (both) are present on the other hand, the steps S 23 and S 24 in FIG. 14 are embodied as follows:

In step S 31 , the computer redetermines the value of the logical variable OK. The logical variable OK then in this case assumes the value “TRUE” and only then if the absolute value of the temporal course remains below a fluctuation limit S (see FIG. 8 ).

In step S 32 , the computer verifies the value of the logical variable OK. If the logical variable OK has the value “TRUE”, the computer executes a step S 33 , in which it assigns a type zero to the evaluation region 14 . If the logical variable OK has the value “FALSE”, the absolute value of the temporal course is forced to exceed the fluctuation limit S (see FIG. 9 ) at least once (in other words locally). For systematic reasons, the temporal derivation is in this case also globally (in other words for all values) larger than zero. In this case, the computer assigns a type one to the evaluation region 14 in step S 34 .

In step S 35 , the computer can also determine the position T 1 of the local maximum 16 and the position T 2 of the local minimum 17 (see FIGS. 10 and 11 ). Alternatively or in addition, in step S 36 , the computer can determine the width T 3 of the local maximum 16 and the width T 4 of the local minimum 17 . In step S 37 , the computer can then assign a type two or a type three to the evaluation region 14 as a function of the position T 1 , T 2 and/or the width T 3 , T 4 of the local maximum 16 and of the local minimum 17 .

Step S 37 in FIG. 14 can be embodied in accordance with FIG. 15 as follows for instance.

According to FIG. 15 , in step S 41 , the computer determines the temporal course of the temporal derivation of the average value of the data values of pixels which are located in the reference region 18 for a reference region 18 . The step S 41 corresponds to steps S 3 and S 5 in FIG. 2 in respect of its basic procedure. Here the reference region 18 can be predetermined to the computer by the user 10 .

In step S 42 , the computer determines the positions T 5 , T 6 of a local maximum 19 and a local minimum 20 . The local maximum 19 and the local minimum 20 are related here to the temporal derivation of the course, which was determined on the basis of the reference region 18 . Similarly, the computer determines widths T 7 , T 8 of this local maximum 19 and this local minimum 20 within the scope of step S 42 . The values T 5 and T 8 thus determined form reference values T 5 to T 8 . Step S 42 corresponds to steps S 35 and S 36 in FIG. 14 in respect of its procedure.

In step S 43 , the computer alternatively assigns the type two or the type three to the evaluation region 14 . The computer decides here whether to assign the type two or the type three to the evaluation region 14 , as a function of the position T 1 , T 2 and/or the width T 3 , T 4 of the local maximum 16 and the local minimum 17 of the temporal derivation of the data values of the pixels 15 located in the evaluation region 14 relative to the corresponding reference values T 5 and T 8 .

Step S 43 can be embodied in accordance with FIG. 16 as follows for instance:

In steps S 51 and S 52 , the computer determines logical variables OK 1 and OK 2 . The logical variable OK 1 then assumes the value “TRUE” here and only then if the positions T 1 , T 2 lie downstream of the positions T 5 , T 6 to a noticeable extent, for instance more than 5% to 10%. The logical variable OK 2 then assumes the value “TRUE” and only then if the widths T 3 , T 4 are greater than the widths T 7 , T 8 to a noticeable extent—by more than 5% to 10% for instance.

One of steps S 51 and S 53 can be omitted. If both steps are present, the computer determines the logical variable OK by means of a logical link (alternatively AND gating or OR gating) of the logical variables OK 1 and OK 2 in step S 53 .

In step S 54 , the computer verifies the value of the logical variables OK. Depending on the result of the verification, the computer executes a step S 55 or a step S 56 as the next step. In step S 55 , the computer assigns the type two to the evaluation region 14 and the type three in step S 56 .

The previously described procedures can be realized irrespective of the size of the evaluation region 14 . In particular, they can be realized irrespective of whether the evaluation region 14 has only one single pixel 15 or whether the evaluation region 14 has several pixels 15 . In the event that the evaluation region 14 has several pixels 15 , it is possible (however not imperative) to subsequently assign step S 55 with steps S 61 and S 63 , which are described in more detail below in conjunction with FIG. 17 .

According to FIG. 17 , in step S 61 , the computer can determine the temporal course of the temporal derivation of the data values of these pixels 15 for the pixels 15 located in the evaluation region 14 . Step S 61 is implemented here separately for each individual pixel 15 .

In step S 62 , the computer evaluates the temporal courses determined in step S 61 . In step S 63 , the computer decides on the basis of this evaluation whether or not to retain the type two for the evaluation region 14 . If the computer does not retain the type two, it assigns either type two-a or type two-b to the evaluation region 14 . Within the scope of this embodiment, a detailed account is shown of how this is explained below in conjunction with FIG. 18 .

According to FIG. 18 , in step S 71 the computer determines whether the temporal courses of the temporal derivation of the data values of the pixels 15 located in the evaluation region 14 has a local maximum and a local minimum in each instance. Step S 71 corresponds in its approach to step S 21 in FIG. 13 . The difference is that it is implemented individually for each pixel 15 .

In step S 72 , the computer determines a first portion a 1 . The portion a 1 corresponds to the number of pixels 15 of the evaluation region 14 , to which the computer would assign the type zero or the type one, if it were to execute the same procedure for the respective pixel 15 which was performed within the scope of the evaluation of the temporal course of the temporal derivation of the average value.

For those pixels 15 , in which the temporal course of the temporal derivation has a local maximum and a local minimum, in step S 73 , the computer determines the positions and/or the widths of the respective local maximum and/or local minimum. Step S 73 corresponds in its approach to steps S 35 and S 36 in FIG. 14 .

In step S 74 , the computer determines a portion a 2 . The portion a 2 corresponds to the number of pixels 15 of the evaluation region 14 , to which the computer would assign the type two, if it were to adopt the same procedure which it adopted in respect of the temporal course of the temporal derivation of the average value of the data values for the temporal courses of the temporal derivations of the data values of the individual pixels 15 .

In step S 75 , the computer determines the value of the logical variable OK. In this case, the logical variable OK then receives the value “TRUE” and only then if the number a 2 , divided by the sum of the numbers a 1 and a 2 , lies below a first portion A 1 and the sum of the numbers a 1 and a 2 corresponds to the total number of the pixels 15 located in the evaluation region 14 . In step S 76 , the computer verifies the value of the logical variable OK. If the logical variable OK in this case has the value “TRUE”, the computer assigns the type two-a to the evaluation region 14 in step S 77 . The portion A 1 can lie between 60 and 70% for instance, in particular at approximately ⅔.

In step S 78 , the computer likewise determines the value of the logical variable OK. In this case, logical variable OK then assumes the value “TRUE” and only then if the number a 2 , in respect of the total number of pixels 15 located in the evaluation region 14 , exceeds a second portion A 2 . In step S 79 , the computer verifies the value of the logical variable OK. In this case, the computer assigns the type two-b to the evaluation region 14 in step S 80 . The second portion A 2 is considerably larger here than the first portion A 1 . It generally lies above 90%, in particular above 95%. It can even reach 100%.

If the computer assigns either the type two-a or the type two-b to the evaluation region 14 , it retains type two for the evaluation region 14 .

The teaching of the present invention was described above in conjunction with a single evaluation region 14 . It is however naturally possible to specify several evaluation regions 14 to the computer simultaneously and to implement the afore-described procedures independently of one another for each of the evaluation regions 14 . By way of example, the computer can subdivide the x-ray images B into a number of equally sized lots, with each lot corresponding to its own, if necessary preliminary, evaluation region 14 , 14 ′.

The present invention has many advantages. In particular, a considerably more precise evaluation than in the prior art is possible in a simple fashion.

The above description is used exclusively to explain the present invention. By contrast, the scope of protection of the present invention is to be determined exclusively by means of the appended claims.