Method for determining collateral information describing blood flow in collaterals, medical imaging device, computer program and electronically readable data carrier

Abstract

Method for determining collateral information describing the blood flow in collaterals (4) of a blood vessel system in a target area of ​​a patient from a four-dimensional vascular data set, which describes in the image values ​​the temporal flow of a contrast agent and / or labeled blood components through the target area and which was acquired using at least one medical imaging device (9), comprising the following steps: - Segmentation of the blood vessel system in the vessel dataset, - Identification of collaterals (4) among the segmented blood vessels using a collateral classifier, - for all identified collaterals (4) Determination of a diameter of the collaterals (4) taking into account the segmentation as well as a filling parameter describing the filling of the collaterals (4) and a time parameter describing the time behavior relative to a reference point in the blood vessel system, each from a time course (5) of the image values ​​in a considered proportion of the collaterals (4), - Determination of collateral information as or depending on the filling parameter and the time parameter.

Description

[0001] The invention relates to a method for determining collateral information describing the blood flow in collaterals of a blood vessel system in a target area of ​​a patient, a medical image recording device, a computer program and an electronically readable data carrier.

[0002] A system and method for diagnosing strokes is known from WO 2014 / 036638 A1. In particular, the system describes the timely acquisition of patient status information, which is intended to enable a physician to make diagnostic and treatment decisions regarding ischemic and hemorrhagic strokes. It enables an efficient and quantitative assessment of arterial collaterals in the brain to support decisions in ischemic strokes. In hemorrhagic strokes, it is possible to determine whether a leak is present and the leak rate.

[0003] Many medical indications are caused by a critical disruption of the blood supply. One example is a stroke, in which the blood supply to the brain is critically disrupted. In strokes, medical image data is often acquired as a vascular dataset showing the vascular system in order to assess intervention options and make convalescence prognoses. It should be noted that each patient's vascular system is unique, especially above a certain diameter, and can certainly contain cross-connections between individual blood vessels, so-called anastomoses. Blood vessels of this size are generally referred to as collaterals, meaning that collaterals generally ensure the blood supply to an area of ​​tissue in the event of blockage or injury to individual blood vessels.If the patient's target area is the brain, the condition of the leptomeningeal anastomoses and collaterals in particular has been shown to be one of the decisive factors in predicting the success of mechanical recanalization.

[0004] For such assessments, vascular datasets from four-dimensional angiography are typically used. In addition to the three spatial dimensions, such four-dimensional vascular datasets describe the temporal flow of a contrast agent and / or labeled blood components through the target area in the image values. In the case of X-ray imaging, contrast agents are typically used; in magnetic resonance imaging, in addition to the administration of contrast agents, spin labeling and similar techniques are also known. Vascular datasets from 4D angiography can thus provide both anatomical and physiological information relating to blood flow. As indicated, the status of the collaterals is extremely important for the clinical use of such vascular datasets, particularly in the context of stroke treatment.To assess the status of collaterals, it is known, for example, to visually assess two-dimensional images from digital subtraction angiography (DSA) or computed tomography angiography (CTA) based on the fill level of the respective blood vessels. A suitably trained physician performs the analysis manually. The disadvantage of this approach is that a subjective assessment and classification of the severity of the occlusion or flow disturbance is required for assessing therapeutic measures for stroke treatment.

[0005] The invention is therefore based on the task of determining an objective statement about the collateral status.

[0006] To achieve this object, the invention provides a method for determining collateral information describing the blood flow in collaterals of a blood vessel system into a target area of ​​a patient from a four-dimensional vessel data set describing the temporal flow of a contrast agent and / or marked blood components through the target area in the image values, which is recorded by means of at least one medical image recording device, which comprises the following steps: - Segmentation of the blood vessel system in the vessel dataset, - Determination of collaterals among the segmented blood vessels using a collateral classifier, - for all identified collaterals, determination of a diameter of the collateral taking into account the segmentation as well as a filling parameter describing the filling of the collaterals and a time parameter describing the time behavior relative to a reference point in the blood vessel system, each from a temporal progression of the image values ​​in a considered portion of the collaterals, - Determination of the collateral information as or depending on the filling parameter and the time parameter.

[0007] The present invention is based on several fundamental findings. Ultimately, the blood flow through the collaterals is to be assessed, with a fundamental measure for evaluating collaterals being determined by two main factors: first, the intensity of the collateral filling—described according to the invention by the filling parameter—and second, the temporal filling profile for the collateral filling—described according to the invention by the time parameter. These criteria can be considered sufficient to describe the status of the collateral supply and thus, in a subsequent step, to better assess the success of the therapy or to plan the therapy.However, especially with regard to the time parameter, it is advisable to use an optimally suitable four-dimensional vascular data set instead of the two-dimensional or three-dimensional vascular data sets used so far, since this allows information on the flow through the collaterals (as long as flow is present) to be derived automatically and reliably.

[0008] The obtained collateral information can therefore be evaluated in a subsequent step based on clinical thresholds already established in studies, possibly even automatically within the framework of a computer-aided diagnosis (CAD) after establishing a suitable data basis. However, this is no longer part of the procedure described here.

[0009] The invention thus allows for the automatic determination of collateral information describing the status of collaterals based on a multiparametric evaluation of four-dimensional angiography images. Effects that could previously only be recorded manually are now calculated and evaluated automatically. In particular, the collateral information can be used to evaluate, for example, whether a patient has sufficient and hemodynamically intact collaterals that would positively influence stroke treatment. This allows for faster diagnosis and objectification regarding treatment preparation and treatment success for stroke patients.

[0010] It should be noted at this point that the target area within the scope of the present invention does not necessarily have to be the patient's brain; applications in other areas are also conceivable. For example, in the context of tumor embolization, such as of the liver, the collateral status can be checked to determine whether the embolization was successful, since the goal is to prevent blood flow to the tumor. Other areas of application are also conceivable in which the collateral information determined automatically according to the invention, for example, by a control device of the medical image acquisition device, can be profitably used for subsequent evaluations / diagnostic steps.

[0011] In this case, it can be provided that an image dataset from four-dimensional subtraction angiography (DSA) and / or a magnetic dataset is used as the vascular dataset. Although aspects of the method according to the invention are often presented below using four-dimensional X-ray angiography (in particular 4D DSA) without limitations on general validity, the method according to the invention is fundamentally applicable to any 4D-capable imaging modality. Furthermore, disadvantages of an imaging modality, such as the limited resolution of magnetic resonance imaging, can be eliminated by appropriate fusion of vascular data from different imaging modalities.For example, in the case of magnetic resonance imaging to obtain vascular data, a combination with digital X-ray 2D / 2D vascular data can be used to compensate for resolution deficits and obtain a suitable four-dimensional vascular data set.

[0012] In the case of four-dimensional digital subtraction angiography (4D-DSA), a contrast agent is typically measured, which is administered in a step not belonging to the inventive method before the image data underlying the vascular data are acquired, for example, either intravenously or selectively intra-arterially in several stages / injections. The basic principle of digital subtraction angiography is to subtract an image dataset without contrast agent (mask image dataset) from image datasets showing contrast agent (fill image datasets) in order to retain only the contrast agent signal. A time series of such subtraction image datasets then yields the four-dimensional vascular dataset.

[0013] It is expedient to determine the collateral information, in particular the time parameter, for both arterial and venous collaterals. Arteries and veins can be differentiated or classified as such in a known manner based on their temporal progression, although other differentiation techniques are also conceivable. Specifically, it is conceivable, for example, to adapt a 4D DSA image so that both the arterial and venous phases can be differentiated in order to be able to determine, for example, an artery-artery delay (AAD) or an artery-venous delay (AVD) as a time parameter, which will be discussed in more detail below. Adjustments can also be made during reconstruction, for example, to ensure the consistency of two-dimensional projections for determining the three-dimensional image, and the like.

[0014] In a first step of the analysis of the physical-technical measurement data from the vascular dataset described here, the blood vessel system in the vascular dataset is segmented. Corresponding segmentation options and suitable algorithms are already largely known in the state of the art, so they will not be discussed in detail here. Segmentation is particularly easy to perform, especially when using digital subtraction angiography.

[0015] Vessels of the segmented blood vessel system, as contained in the vessel dataset, must then be classified according to whether a collateral or another vessel is present. To implement the collateral classifier, it can be specifically provided that a segmented vessel is classified as a collateral if its diameter is smaller than a threshold or lies within a collateral interval. In many cases, it is sufficient to consider blood vessels whose diameter is smaller than a threshold, because small vessels in the parenchyma, such as capillaries, cannot usually be directly resolved or visualized using conventional imaging. In this way, a simple threshold analysis can focus on the anastomoses or visible collaterals, not on the parenchymal filling.Additionally or alternatively, it can be provided that vessels branching off from the main vessels, particularly those identified using an anatomical atlas, to the parenchyma are classified as collaterals. An analysis can therefore also be performed to locate certain vessels that should be present in all people, for example, by registration with an anatomical atlas. Once these vessels have been identified, the collaterals are often defined as blood vessels branching off from these vessels to the parenchyma, which can also be done accordingly within the scope of the present invention through automatic evaluation.

[0016] During segmentation, it may be provided that centerlines of the segmented blood vessels of the vascular system are determined and used to define the portions for which the filling parameter and the time parameter are to be determined. Once the center of the blood vessels is known in this way, these centerlines can also be expediently used to define the portions to which the diameter, the filling parameter, or the time parameter refer. In particular, it may be provided that the collateral information is determined as a portion for voxels lying on the centerlines and / or for voxel groups comprising the extent of the collaterals perpendicular to the centerline and / or for vessel sections defined along the centerline.While in many cases it may be sufficient to consider a central voxel located on the midline of a blood vessel classified as a collateral, in particular to use its temporal course of the image values, wherein the resulting collateral information is then preferably also transferred to neighboring voxels of the blood vessel perpendicular to the course of the midline, in some embodiments it may be advantageous to use several voxels of the blood vessel, be it along the midline and / or perpendicular to it, wherein then in particular the temporal courses of the image values ​​are averaged or generally viewed statistically, for example in order to be able to filter out measurement errors more effectively and / or to optimize the signal-to-noise ratio.

[0017] The filling parameter can expediently be determined as a function of a maximum over time and / or a flow rate determined, in particular, by integration over time and / or a gradient of the flow rate during the inflow phase and / or a combination of these values. It is also conceivable, in particular, to determine the filling parameter as these values ​​or the combination of values. However, it is preferable to place the filling parameter in a suitable reference to facilitate comparison based on different diameters / sub-areas of the target area, which will be discussed in more detail below.The maximum in the temporal progression thus corresponds to the maximum image value over time. If the temporal progression of the image values, often referred to as the contrast agent curve in relation to a contrast agent, is integrated, a measure of the flow rate through the corresponding collateral is obtained. A maximum gradient of the temporal progression during the inflow phase, i.e., when the contrast agent concentration increases, is also a suitable measure of filling. Of course, other values ​​are also conceivable / definable that describe the filling during the flow of the contrast agent or the labeled blood components.

[0018] As already indicated, a particularly advantageous embodiment of the present invention provides that the filling parameter is determined based on the diameter and / or on at least one first reference value of the same patient and / or on at least one second reference value related to a patient-external reference. The second reference value thus originates from a reference outside the patient. Reference to the diameter of the vessels ultimately allows for better comparability, particularly if the collateral information, as will be explained below, is preferably displayed as a collateral map.With regard to filling, the amount of contrast agent / labeled blood components that can flow through the collateral often depends on the diameter of the vessel. While the diameter can generally be determined at least roughly from a representation of the vessel dataset, establishing the correlations would be more complex. Therefore, the diameter, which is already determined as collateral information, is at least implicitly incorporated into the filling parameters in this case.

[0019] The interpretation of the collateral information or a comparison is further simplified by reference to at least one first reference value and / or at least one second reference value, the latter relating to a patient-external reference. A patient-internal reference can be created, for example, by comparing a healthy sub-area of ​​the target area with a sub-area of ​​the target area exhibiting a circulatory disorder, for example, by comparing different, standardized brain areas. The at least one first reference value can then enable a comparison of filling parameters, preferably for collaterals of the same diameter, between diseased and healthy sub-areas to be integrated into the collateral information.A similar approach is of course also possible with an external reference, where, for example, second reference values ​​are used for collaterals of different diameters, which indicate which filling would be expected in the case of a healthy collateral / perfusion, especially in a corresponding anatomical region.

[0020] It is preferred if the first and / or second reference value are determined by statistical analysis, in particular broken down by the diameter of the collaterals, for the target area and / or a sub-area of ​​the target area. For example, if the at least one first and / or at least one second reference value are statistically analyzed in a healthy sub-area of ​​the patient themselves and / or for healthy target areas / sub-areas of other patients, an excellent expected value for the filling of collaterals in the healthy case results, which, as explained, is preferably broken down by diameter.

[0021] Several approaches are conceivable for reference. For example, it is possible to determine the filling parameter, for example, as a mathematical function of a value determined from the temporal progression of the image values, for example, the maximum filling value, and the respective reference (diameter / reference value), for example, to "normalize" it by dividing it by the respective reference. However, within the scope of the present invention, it is preferred to perform an evaluation as a function of the reference, in particular the diameter and / or the at least one reference value. For this purpose, a characteristic curve / characteristic map and / or a look-up table can be used, for example. The at least one first or second reference value can then define intervals, for example, in which the filling can be evaluated as "normal," "too low," "too high," and the like, in particular also in finer gradations.Ultimately, such a characteristic curve or map and / or a look-up table maps a value derived from the temporal progression of the image values ​​in the considered portion to the filling parameter as collateral information, which then explicitly describes an evaluation of the filling. In the case of first reference values, a look-up table (or a map / curve) can also be dynamically generated for patients, for example, as already described, by statistically examining a healthy sub-area of ​​the target area. Such mapping options targeting second reference values ​​external to the patient can be permanently stored in a database of the computing device implementing the method.

[0022] The time parameter can be conveniently determined as a function of a bolus arrival time and / or a time until the maximum image value. Here, too, various concrete definitions of values ​​derivable from the temporal progression of the image values ​​in the considered portion of the collateral, which can form a basis for determining the time parameter, are conceivable. For example, a bolus arrival time can be defined as the time at which the image value exceeds a threshold value and / or as the time of maximum gradient in the inflow phase; the time at which the maximum image value is reached, i.e., maximum filling, can also be conveniently used. Other definitions are also conceivable in principle.

[0023] The comparable time reference is established using the aforementioned reference point. Specifically, it can be provided that a predefined reference point is located in the vascular dataset, taking the segmentation into account. A time used as the time parameter and / or used to determine it is determined relative to a reference time determined at the reference point. Preferably, an analogous determination of the time can be performed at the reference point as a value derived from the temporal progression, for example, as a bolus arrival time and / or a time to the maximum image value.Thus, comparable times initially exist from all collaterals, whereby a reference point in a supply vessel of the blood vessel system of the target area is preferably used as the reference point, in the case of the brain as the target area, for example, as a reference point in the carotid artery, which can also be determined reproducibly for different patients in order to allow a comprehensible comparison between different patients.

[0024] When comparing times derived from the temporal progression of the image values, however, it should be noted that paths through the vascular system can be quite different in length, even if the observed portions of collaterals appear to be equally distant from the reference point. This aspect can also be particularly advantageously taken into account within the scope of the present invention if it is provided that, based on the segmentation, in particular on centerlines through the segmented blood vessels, a spatial distance through the vascular system between the reference point and the currently observed portion is determined, wherein the time parameter is determined relative to the distance.Similar to the diameter with regard to the filling parameter, this distance, which would be extremely difficult to determine by purely visual observation, enables easier comparability of collateral information, for example if it is recognized that a bolus arrives in a certain collateral significantly later than in another collateral, although the same distance, i.e. the same path length, through the blood vessel system starting from the reference point is given.

[0025] To further improve the comparison, it may also be expedient to determine the time parameter relative to a third reference value for the same patient and / or an external fourth reference value. Ultimately, the statements regarding the filling parameter apply accordingly, so it is also particularly expedient if the third and / or fourth reference value are determined by statistical analysis, particularly broken down by distance from the reference point. Here, too, it is preferable to use look-up tables orTo use characteristic maps / characteristic curves in order to derive the final time parameter as an evaluation variable from the times derived from the temporal course of the image values ​​as preliminary values, whereby comparisons are again made with healthy sub-areas of the target area in the same patient using third reference values ​​and / or with healthy sub-areas / target areas of other patients using patient-external fourth reference values, whereby the breakdown is not based on the diameter but on distances through the blood vessel system from the reference point, which can be reproducibly determined in the case of a comparison between different patients.

[0026] A preferred development of the method according to the invention generally provides that a collateral quantity of detectable collaterals for the target area and / or sub-areas of the target area is determined as additional collateral information. This ultimately results in a type of overview value that indicates how well the target area or, preferably, certain sub-areas of the target area are supplied with blood. For example, the standardized, generally known division into brain regions (M1, M2, ...) can be used to achieve a useful subdivision. The corresponding collateral information can represent rough information, in particular again related to a patient-internal and / or patient-external reference, as described above, but can also be specified with any desired level of precision, insofar as the vascular data set allows.For example, to "count" collaterals, the number of voxels assigned to collaterals can be set in relation to the number of total voxels or voxels not assigned to collaterals, etc. Patient-internal / external fifth or sixth reference values ​​can then, as already explained, be determined in a similar manner, preferably statistically at least for patient-external sixth reference values, and used to evaluate the final collateral quantity.

[0027] In a particularly advantageous development of the present invention, it is provided that a two-dimensional or three-dimensional, in particular sub-area-specific collateral map is determined and displayed to represent the spatially resolved collateral information. Such a collateral map is an excellent means of outputting the collateral information, since it is determined in a spatially resolved manner, for example, by portions defined along the center line of the collaterals. An overview display then not only reveals which sub-areas of the target area or sub-area have perfused collaterals, but also allows the localized reading of the quality of the perfusion, with the filling parameter and the time parameter preferably being read off as separate information in the collateral map.Less preferred, but also conceivable, is for the image elements of the collateral map to contain, as image values, a display variable derived from the fill parameter and the time parameter. For example, the fill parameter can be weighted with the time parameter, or vice versa. Preferably, each image element contains the fill parameter and the time parameter as display information assigned to different display channels, in particular in different color channels and / or a brightness channel. In this way, a two-dimensional representation of collateral image elements in the collateral map is achieved, which visually conveys both the fill parameter and the time parameter. In principle, known visualization options can be used, for example the use of different color channels or the use of brightness variations for one of the parameters.This creates a visual, accurate representation of the collateral information, which is easy to interpret and can be evaluated for further diagnosis / assessment. If the determination of collateral information is standardized to such an extent that it can be reproducibly applied to different patients, this is a particularly advantageous design, as it allows comparisons between different patients and allows for a particularly rapid accumulation of experience in assessing such collateral maps.

[0028] Such a collateral map can be further improved if, when determining a subregion-resolved collateral volume, this is integrated into the collateral map as a marker of the background not occupied by collaterals, particularly by coloring the corresponding subregion. For example, areas with a low collateral volume can be displayed in red, areas of critical collateral volume in yellow, and areas with a normal collateral volume, indicating sufficient blood flow, in green. Thus, the collateral map intuitively visualizes another piece of collateral information, namely the collateral volume.

[0029] Preferably in addition to, but also alternatively to, determining a collateral map, it is of course also conceivable to determine a collateral score for sub-areas of the target area and / or the entire target area from or as collateral information, taking into account at least the time parameters and the filling parameters. Thus, calculating an overall assessment as the final evaluation result is also conceivable, which is preferably determined resolved for sub-areas. For example, if a sub-area affected by a stroke can be defined when examining the brain, a general statement can be made using a collateral score, which is preferably supplemented by a spatially accurate, spatially resolved representation in a collateral map.

[0030] In addition to the method, the invention also relates to a medical image recording device comprising a control device designed to carry out the method according to the invention. This can preferably be an angiographic image recording device, for example an X-ray device suitable for 4D DSA. The use of both X-ray devices with a C-arm and CT devices is conceivable here. However, the medical image recording device can also be a magnetic resonance device, whereby, of course, individual image recording devices can also be designed to combine image data from different modalities to form the vascular data set, as explained above. All statements regarding the method according to the invention continue to apply to the medical image recording device, so that the aforementioned advantages can also be achieved with this.

[0031] The control device can expediently comprise a segmentation unit for segmenting the blood vessel system in the vessel data set, a classification unit that implements the collateral classifier, and a collateral information determination unit. Additionally, a visualization unit can also be provided, in which, for example, a collateral map can be determined.

[0032] The invention further relates to a computer program that performs the steps of a method according to the invention when executed on a computing device. The computing device can be the aforementioned control device of the medical image acquisition device. The computer program according to the invention can, for example, be loaded directly into a memory of a control device of a medical image acquisition device and has program means for executing the steps of a method described herein when the computer program is executed in the control device of the medical image acquisition device. The computer program can be stored on an electronically readable data carrier according to the invention.An electronically readable data carrier according to the invention comprises electronically readable control information stored thereon, which comprises at least one mentioned computer program and is designed such that, when the data carrier is used in a control device of an image recording device or another computing device, it carries out a method described herein. The data carrier is preferably a non-transient data carrier, for example a CD-ROM or a DVD. Naturally, the statements regarding the method according to the invention also apply accordingly to the computer program and the electronically readable data carrier.

[0033] Further advantages and details of the present invention will become apparent from the exemplary embodiments described below and from the drawings. In the drawings: Fig. 1 a flow chart of an embodiment of the method according to the invention, Fig. 2 a schematic section of a blood vessel system in a sub-area of ​​the brain, Fig. 3 a temporal progression of image values, and Fig. 4 a medical imaging device.

[0034] Fig. 1 shows a flowchart of an exemplary embodiment of the method according to the invention. This is based on a vascular dataset showing the blood vessels of a blood vessel system in a target region of a patient. The present exemplary embodiment relates, without limitation of generality, to a four-dimensional vascular dataset of digital subtraction angiography (4D-DSA) of a patient's brain, acquired using X-ray as the imaging modality. To generate the 4D-DSA vascular dataset 1, a contrast agent was administered to the patient, and three-dimensional filling image datasets of the target region were acquired at various times during the flow of the contrast agent, comprising both the arterial and venous phases.A mask data set without contrast agent was subtracted from each of these filling data sets, so that in the resulting four-dimensional vessel data set 1, which is created by concatenating the individual 3D-DSA subtraction data sets, the temporal course of the concentration of the contrast agent in the corresponding voxel is encoded in each image value.

[0035] Accordingly, in a step S1 of the exemplary embodiment described here, it is possible to easily segment the blood vessels through which the contrast agent has flowed in the target area due to the presence of the contrast agent. Known segmentation algorithms can generally be used. As part of the segmentation, the centerlines of the blood vessels and their diameters can also be determined. In particular, however, a reference point is also located in step S1, which, in this case, is located within the carotid artery as the supplying arterial blood vessel and can be reproducibly found in different patients.

[0036] Fig. Figure 2 shows a purely schematic, principle sketch of a portion of the course of the middle cerebral artery 2, which clearly exhibits various branches. Only one subregion 3 of the brain, in this case a cortical region (M3), has been depicted in more detail as the target region. It is evident that smaller branching blood vessels and tributaries exist towards the parenchyma, which are not depicted in detail here and are not resolved in the vascular dataset. These branches run differently in each patient and are referred to as collaterals 4. The collaterals 4 can be defined as blood vessels with a diameter below a threshold value and / or as blood vessels branching off from a segment of the artery 2 that is identifiable in all individuals.

[0037] In a step S3, see again Fig. 1, a collateral classifier is used to identify and select collaterals 4 in the target area. The collateral classifier can check the diameter of blood vessels for a threshold value. A distinction to lower diameters is not necessarily required, since the capillaries in the parenchyma cannot be resolved using X-ray imaging. Additionally or alternatively, an anatomical atlas can also be used within the collateral classifier to identify certain segments of the large arteries present in all humans and to identify blood vessels branching off from them as collaterals 4. Collaterals also include anastomoses connecting blood vessels of one type (arteries or veins).

[0038] In a step S3, spatially resolved collateral information is then determined for the identified and selected collaterals 4, for which purpose portions of the respective blood vessels are first defined. For this purpose, the aforementioned centerlines determined during segmentation are used by defining the portions relative to voxels that contain the centerline. It is conceivable to limit the evaluation to such voxels (or to voxels following one another along the centerline), but it is also conceivable to fully or partially consider the diameter of the blood vessel perpendicular to the centerline. Whenever groups of voxels are considered, the temporal progression of the image values ​​is statistically analyzed, for example, averaged.

[0039] For each of these portions containing one or more collateral voxels, a filling parameter and a time parameter are determined for the known diameter. To do this, parameters of the temporal course of the image value—in this case, the contrast agent concentration—are correlated with the values ​​that classify the currently observed portion and also with values ​​that serve as a reference.

[0040] Fig. For a more detailed explanation, Figure 3 shows an exemplary temporal progression 5 of image values ​​for a portion, here as a progression of the contrast agent concentration c(t) versus time t. The time values ​​discussed below are considered relative to corresponding time values ​​at the aforementioned reference point in the carotid artery, thus forming delays. Both arterial and venous blood vessels, and thus delays, are considered.

[0041] The contrast agent concentration clearly increases during an inrush phase, beginning with the bolus arrival time (BAT), which was defined here as the time at which a threshold value for the image value / contrast agent concentration is exceeded, but for simplicity, was shown as zero at the beginning of the peak, since previously only noise was supposed to be contained in the image values. The position 6 of greatest gradient, symbolized by arrow 7, is also highlighted and can define a time value derived from the time course 5. Another important time value characterizing the time course 5 is the so-called "time to peak" (TTP), i.e. the time at which the maximum contrast agent concentration (and thus the maximum image value) is reached. This is shown here as c max designated.

[0042] In addition to the time-domain parameters BAT, TTP, and time of maximum gradient, another important criterion for assessing blood flow in collaterals 4 is the filling behavior, i.e., the question of how much contrast agent actually reaches or flows through the observed portion. One measure of this is, for example, the maximum image value, i.e., the maximum contrast agent concentration c max, whereby additionally or alternatively, the area 8 under the contrast agent curve, i.e., the time course 5, can also be determined as such a filling value. This area is obtained by integration and ultimately describes the total amount of contrast agent flowing through. However, the time values ​​and filling values ​​mentioned as examples, which are obtained from the evaluation of the time course 5 in the portion of a collateral under consideration, are not sufficiently meaningful / intuitively assessable on their own, so that the actual collateral information, i.e., the filling parameter and the time parameter, can only be derived from them by referring to references.

[0043] With regard to the filling value, the diameter of collateral 4 in the section under consideration must first be considered, as this ultimately describes the "capacity." Therefore, only collaterals 4 of the same diameter should be compared. To further improve the interpretation of the collateral information regarding the filling parameter, reference is also made to a reference value for the filling value that describes the behavior of a healthy collateral 4. For this purpose, this exemplary embodiment provides look-up tables sorted by diameter. These tables assign a filling value for a specific diameter to the filling parameter that indicates the extent to which the filling value corresponds to that of a healthy collateral 4.The reference values ​​describing this in the look-up table can be determined statistically from a healthy sub-area 3 of the target area, in this case the brain, of the same patient or by statistically examining healthy areas of other patients. In this case, a comparison with second, patient-external reference values ​​that refer to the same sub-area with healthy blood flow is preferred.

[0044] Regarding the aforementioned time values, which can be derived from the temporal progression 5 of the image values, in this case, a distance of the observed portion from the reference point through the vascular system, i.e., a path length through the vascular system, is determined based on the determined center lines, since these distances obviously do not necessarily correlate with the purely geometric distances for distances outside the vascular system. Similar to the diameter, a distinction is made here between different distances or distance intervals, although look-up tables are also used to determine the reference values ​​describing the time parameter by comparing them with time values ​​of healthy collaterals 4 at the same distance from the reference point (which, as described, can be reproducibly found regardless of the patient).

[0045] Further collateral information is provided in step S3 ( Fig. 1) In relation to sub-areas 3 into which the brain is divided as a target area, a collateral quantity (hence a collateral quantity parameter) is also determined, which describes how many perfused collaterals 4 are present in the corresponding sub-area 3. This can be defined, for example, as a collateral density (voxels identified as belonging to a collateral 4 to the total voxels of sub-area 3), but also in other ways.

[0046] The collateral information thus resolved for the collaterals 4 and their various components can be further processed in various ways. In step S4, a collateral map is determined, which can be determined three-dimensionally or two-dimensionally (using appropriate projection or rendering methods). Each image element, i.e., voxel or pixel, of the collateral map is assigned an image value resulting from the fill parameter and the time parameter. Independent display channels are used to reproduce these values, for example, different color channels or a color channel for one of the parameters and a brightness channel for the other, so that both parameters are recognizably included in the visual representation using the collateral map.Furthermore, the amount of collaterals is also processed by coloring image elements not occupied by collaterals, i.e. the background, for example in red for a low amount of collaterals, in yellow for a critical amount of collaterals, and in green for a normal amount of collaterals. It should be noted that a reference to patient-external or patient-internal references is also useful here.

[0047] The collateral map is displayed on a display device to a user who can then interpret it and, in particular, make diagnoses, assess therapeutic success and / or plan therapeutic measures.

[0048] In step S5, a collateral score is determined at sub-area resolution, taking into account at least the filling parameters and the time parameters, preferably also the collateral quantity. This score thus contains an easily understandable general statement for a sub-area. This score can also be output accordingly, preferably together with the collateral map.

[0049] Fig.Finally, Figure 4 shows a medical image acquisition device 9 according to the invention, which in this case is designed as an angiographic X-ray device with a C-arm 10. An X-ray emitter 11 and an X-ray detector 12 are arranged opposite one another on the C-arm 10. The C-arm 10 can be pivoted around a patient bed 13, on which a patient can be placed, for acquiring projection images that form the basis of a three-dimensional image data set. However, the medical image acquisition device 9 can also be designed as a CT device or a magnetic resonance device.

[0050] The operation of the medical image acquisition device 9 is controlled by a control device 14, which is also designed to carry out the method according to the invention. For this purpose, the control device 14 has, in particular, a segmentation unit for carrying out step S1, a classification unit for carrying out step S2, and a collateral information determination unit for carrying out step S3. Furthermore, a visualization unit for determining a collateral map (step S4) can be provided.

[0051] A method described herein may also be in the form of a computer program that implements the method on the control device 14 when executed on the control device 14. Likewise, an electronically readable data carrier (not shown) with electronically readable control information stored thereon may be present, which includes at least one described computer program and is configured to carry out a described method when the data carrier is used in the control device 14.

[0052] Although the invention has been illustrated and described in detail by the preferred embodiment, the invention is not limited to the disclosed examples and other variations may be derived therefrom by those skilled in the art without departing from the scope of the invention.

Claims

[1] Method for determining collateral information describing the blood flow in collaterals (4) of a blood vessel system in a target area of ​​a patient from a four-dimensional vessel data set describing the temporal flow of a contrast agent and / or marked blood components through the target area in the image values, recorded by means of at least one medical image recording device (9), comprising the following steps: - Segmentation of the blood vessel system in the vessel dataset, - Determination of collaterals (4) among the segmented blood vessels using a collateral classifier, - for all determined collaterals (4) determining a diameter of the collateral (4) taking into account the segmentation as well as a filling parameter describing the filling of the collaterals (4) and a time parameter describing the time behavior relative to a reference point in the blood vessel system, each from a time course (5) of the image values ​​in a considered portion of the collaterals (4), - Determination of the collateral information as or depending on the filling parameter and the time parameter. [2] Method according to claim 1, characterized by that a four-dimensional subtraction angiography image dataset and / or a magnetic resonance dataset is used as the vascular dataset and / or that the target area is the head. [3] Method according to claim 1 or 2, characterized bythat a segmented vessel is classified as a collateral (4) if its diameter is smaller than a threshold value or lies in a collateral interval and / or vessels branching off towards the parenchyma from main vessels identified in particular by means of an anatomical atlas are classified as collaterals (4). [4] Method according to one of the preceding claims, characterized by that center lines of the vessels of the vascular system are determined and used to define the proportions. [5] Method according to claim 4, characterized by that the collateral information for voxels lying on the center lines and / or the extent of the collateral (4) comprising voxel groups perpendicular to the center line and / or for vessel sections defined along the center line are determined as a proportion. [6] Method according to one of the preceding claims, characterized bythat the filling parameter is determined as a function of a maximum in the time course (5) and / or a flow rate determined in particular by integration over the time course (5) and / or a gradient of the time course in the flooding phase and / or a combination of these values. [7] Method according to one of the preceding claims, characterized by that the filling parameter is determined in relation to the diameter and / or to at least one first reference value of the same patient and / or to at least one external second reference value. [8] Method according to claim 7, characterized by that the first and / or the second reference value are determined by statistical analysis, in particular broken down according to the diameters of the collaterals (4), for the target area and / or a sub-area (3) of the target area. [9] Method according to one of the preceding claims, characterized bythat the time parameter is determined depending on a bolus arrival time and / or a time to the maximum image value. [10] Method according to one of the preceding claims, characterized by that a predetermined reference point is located in the vessel data set taking into account the segmentation, wherein a time used as the time parameter and / or used to determine it is determined relative to a reference time determined at the reference point. [11] Method according to claim 10, characterized by that a reference point is used in a supply vessel of the vascular system of the target area. [12] Method according to claim 10 or 11, characterized bythat based on the segmentation, in particular on center lines through the segmented blood vessels, a spatial distance through the blood vessel system between the reference point and the currently viewed portion is determined, whereby the time parameter is determined relative to the distance. [13] Method according to one of the preceding claims, characterized by that the time parameter is determined relative to a third reference value of the same patient and / or an external fourth reference value. [14] Method according to claim 13, characterized by that the third and / or fourth reference value are determined by statistical analysis, in particular broken down by distance from the reference point. [15] Method according to one of the preceding claims, characterized by that as further collateral information a collateral quantity of detectable collaterals (4) is determined for the target area and / or sub-areas (3) of the target area. [16] Method according to one of the preceding claims, characterized by that a two-dimensional or three-dimensional, in particular sub-area-related collateral map is determined and displayed to represent the spatially resolved collateral information. [17] Method according to claim 16, characterized by that the image elements of the collateral map contain, as image value, a display variable derived from the fill parameter and the time parameter, or that in each image element the fill parameter and the time parameter are contained as display information assigned to different display channels, in particular in different color channels and / or a brightness channel. [18] Method according to claim 16 or 17, characterized bythat when determining a sub-area-resolved collateral set, this is integrated into the collateral map as a marking of the background not occupied by collaterals (4), in particular coloring, of the corresponding sub-area. [19] Method according to one of the preceding claims, characterized by that a collateral score is determined for sub-areas (3) of the target area and / or the entire target area, taking into account at least the time parameters and the filling parameters. [20] Medical image recording device (9), comprising a control device designed to carry out a method according to one of the preceding claims. [21] A computer program which, when executed on a computing device, performs the steps of a method according to any one of claims 1 to 19. [22] An electronically readable data carrier on which a computer program according to claim 21 is stored.

Images