Method, arrangement and computer program for quantitative bifurcation analysis in angiographic images

Abstract

A method for performing a quantitative analysis of medical image data of a bifurcated tubular organ having at least one proximal part, a first distal part, and a second distal part, the method comprising: - processing said medical image data to identify outlines of said bifurcated tubular organ along its longitudinal extent, wherein the outlines have first and second outer outlines and a middle outline between them; - using said outlines to determine a confluence polygon in the middle of said bifurcated tubular organ, wherein the confluence polygon is bounded by a plurality of lines intersecting at vertices, and wherein the plurality of lines identify both a proximal start of the proximal part of the bifurcated tubular organ and distal ends of the first and second distal parts of the bifurcated tubular organ.wherein a first line of the majority of the lines extends along the first outer outline and a second line extends along the second outer outline,- using the said confluence polygon to determine at least one parameter value that characterizes the geometry of the said bifurcated tubular organ, and- delivering the said at least one parameter value to a user.

Description

BACKGROUND OF THE INVENTION AREA OF INVENTION The present invention relates to a method for quantitative bifurcation analysis of medical images, in particular angiographic images. STATE OF THE ART Angiography, particularly coronary angiography, refers to the process of examining coronary arteries to determine the severity of any narrowing, such as by identifying stenotic arteries. Quantitative coronary analysis (QCA) of individual vessels has become standard practice for accompanying interventions and conducting trace studies during coronary revascularization. However, angiographic analysis of injuries at or near a bifurcation presents a significant challenge, as QCA for individual vessels is unable to handle the rather complex geometries. The definition of a bifurcation in this context is the division of a main tubular vessel into two or more smaller tubular vessels. For example, the left coronary artery bifurcates into the left anterior descending artery and the left circumflex artery. QCA of a bifurcation involves the automatic segmentation of the bifurcation. This can be followed by a reconstruction of the vessel's health status, encompassing the bifurcation area itself. The diameter of the reference vessel, which represents the diameter of the healthy vessel as calculated by QCA, is typically based on the mean values ​​of the "normal" portions of the vessel before and after the bifurcation. The greatest challenge for bifurcation injury analysis is extracting the true reference vessel size of the nearby vessel and its tributaries. Conventional QCA detects vessel outlines under the assumption of minimal vessel narrowing and cannot handle large diameter variations caused by the bifurcation itself. Most conventional QCA methods allow the input of a user-defined reference value, which could eliminate incorrect reference definitions. However, this reference diameter would still only be valid on the side of the bifurcation where the user defined the reference value. Furthermore, this option has limited reproducibility due to inaccuracies and operator subjectivity. Currently, no detailed publication has resolved the aforementioned limitations. O. Goktekin et al., “A new quantitative analysis system for the evaluation of coronary bifurcation lesions: Comparison with current conventional methods,” Catheterization and Cardiovascular Interventions 69:172-180 (2007), evaluates a bifurcation package where the bifurcation is divided into three parts, with conventional QCA applied to all three. Goktekin describes a procedure for solving the reference value problem by eliminating the central bifurcation region from the reference calculations. However, the central bifurcation is still disregarded for calculating a diameter and defining a reference value. From US patent 2004 / 0120561A1, a device is known that is intended to support image diagnosis, distinguishing between, on the one hand, shadows produced by a carcinoma that are in the viewer's actual focus (focus shadows), and, on the other hand, shadows produced by common bodily structures, such as blood vessels or bronchi. Digital data processing assigns a probability to each shadow as to whether it is a focus shadow, with the probability being indicated by a marker, e.g., different colors. For the differentiation, it can be assumed that a carcinoma in a plane section closely approximates a circular shape, while this is usually not the case for a blood vessel in the area of ​​a bifurcation.For this purpose, a ratio is calculated between the sum of the areas adjoining the concave edges of the shadow and the area of ​​the shadow itself. Under the aforementioned assumption, this ratio is significantly larger in images of blood vessel bifurcations than in images of carcinomas, thus providing a measure for distinguishing between the two sources of shadows. Modified methods for determining the area ratio are presented. For example, a circle is placed inside the shadow, or a circle enclosing the shadow. In each case, the area of ​​the circle is compared to the total area of ​​the shadow or to the area of ​​the parts of the shadow lying outside the circle. US 2004 / 0120561A1 also proposes placing a triangle within the shadow, its vertices touching the shadow's boundaries, and determining the ratio of the total area of ​​the shadow's outer portion to the triangle's area. Another example uses a pentagon, its vertices also touching the shadow's outline. Here, too, the ratio of the total area of ​​the shadow's outer portion to the pentagon's area is determined. The aforementioned publication does not offer a solution for identifying pathological constrictions at bifurcations in bifurcated tubular organs. SUMMARY OF THE INVENTION Therefore, one of the objects of the present invention is to provide a more accurate and reproducible method and system for performing a quantitative analysis of a bifurcation to overcome the limitations of the current state of the art. This object is achieved according to the present invention by a methodology for the quantitative analysis of medical image data of a bifurcated tubular organ according to claim 1. The methodology comprises processing the medical image data to identify the outline of the bifurcated tubular organ. The outline is used to determine a confluence polygon located within the bifurcated tubular organ. The confluence polygon is used to determine at least one parameter value that characterizes the geometry of the bifurcated tubular organ. The at least one parameter value is made available to a user for angiographic purposes.The at least one parameter value can contain at least one diameter value of the forked tubular organ, at least one angle value between parts of the forked tubular organ, and at least one reference diameter value for the forked tubular organ, at least one reference diameter value as compensation for damage to the forked tubular organ. The methodology according to the present invention has been shown to deliver stable and reliable results for many complex situations, while requiring relatively unambiguous procedures. A significant advantage of determining an angle or diameter according to the methodology of the present invention is that it provides a surgeon with an improved tool for finding a suitable surgical component or device, such as a stent or a yolk element, in practice. The present invention also relates to a data processing capability and to a program storage arrangement for carrying out the methodology according to the present invention. Several advantageous aspects of the present invention are defined in the dependent claims. Furthermore, although the above generally describes coronary arteries, the principle of the present invention is also applicable to other arteries, veins, and generally to several tubular organs that can benefit from angiographic imaging, with the aid of the improvements of the present invention. Consequently, the term artery should be considered in a broader sense in the context of the present invention. Additional tasks and advantages of the present invention should be obvious to the person skilled in the art from the detailed description in conjunction with the accompanying figures. BRIEF DESCRIPTION OF THE DRAWING Figure 1 shows a flowchart of the processing of the bifurcation data, Figure 2 shows an example of a bifurcation angiographic image, Figures 3a and 3b each show a line extending from a corrected user point, Figure 4 shows a representation of several maxima within areas bounded as shown, Figures 5a, 5b and 5c each show multiscale echoes of an intensity profile, Figure 6 shows points used to identify the left and right vessels, Figure 7 shows an example of a confluence pentagon according to the present invention, Figure 8 shows a representation of the calculation of the angles between the respective branches of a bifurcation according to the methodology of the present invention, Figures 9a and 9b each show a representation of the calculation of the diameter of a bifurcation according to the methodology of the present invention.10 a representation of the modeling of part of a vessel using a wreath, Fig. 11a and Fig. 11b each a representation of the modeling of a bifurcation using a set of wreaths. DETAILED DESCRIPTION OF THE PREFERRED VERSION Fig. 1 shows a flowchart for processing the bifurcation data. In this flowchart, block 20 represents the detection of the boundary, block 22 the combination of the boundary, block 24 the definition of a confluence pentagon, block 26 the definition of the angles between the bifurcations, block 28 the measurement of the diameter, and block 30 the definition of the reference value. The measurements from blocks 26, 28, and 30 are communicated to the user, for example, as part of a display screen or as a printed output. Each of blocks 20–28, as defined in the flowchart, is described in detail below.The methodology of the present invention is preferably implemented by a data processing system, such as a computer workstation, comprising a data processing platform (for example, a CPU, a storage system, a read-only memory, a playback adapter) as an interface between user input devices (such as a keypad and a display device) and one or more playback devices and / or a printer for user data output. The methodology of the present invention is preferably implemented as a software application stored on one or more optical disks (or on another form of non-volatile memory) or, optionally, downloaded from a remote computer system and loaded into the data processing platform for execution. Figure 2 shows a typical example of a bifurcation, or more generally, of a tubular organ, in an angiographic image. For quantitative bifurcation analysis, the vessels associated with the bifurcation must be detected. In this case, the three vessels are the proximal vessel, the first distal vessel (i.e., the distal part of the main branch), and the second distal vessel (i.e., the distal part of the collateral branch). The two distal vessels may be of equal size, and the steps to be taken to achieve optimal visual representation of the bifurcation are considered known. Instead of attempting to depict the bifurcation as two or three vessels sharing common parts, the bifurcated vessel is viewed as a single object, outlined by a left, a middle, and a right highlighted outline without further assumptions. The detection of the bifurcation can begin in three different ways (see Block 20). The first procedure for initiating bifurcation detection is for the user to roughly define the arterial bifurcation midline using a number of points, such that all lines connecting the points lie approximately within the main branch (proximal and distal 1) and the collateral branch (distal 2 from the midpoint of the bifurcation). An example of this approximation is given in Gronenschild and Tijdens: “A second Generation System for Off-line and On-line Quantitative Coronary Angiography”, Catheterization and Cardiovascular Diagnosis 33: 61-75 (1994). The second method for initiating bifurcation detection involves the user specifying a starting point at the proximal branch and endpoints at distal 1 and distal 2. From these points, three line segments are automatically calculated using a wavefront algorithm. This algorithm simulates wavefront propagation, similar to a water wave traveling through a river. See: "Introduction to Algorithms," Thomas H. Cormen, Charles E. Leiserons, Ronald L. Rivest, ©1990 The MIT Press, Cambridge, Massachusetts, London, England. Chapter 25.2, Dijkstra's algorithm, pages 527–532. The third method for initiating bifurcation detection is for the user to specify the midpoint of the bifurcation, which triggers automatic route line extraction at the three branches. This method is also based on the wavefront algorithm. Considering that the main drawback of current bifurcation violation analysis systems is the lack of a reliable reference value definition, our preferred method automatically terminates route line propagation either when a branch is detected or when a predefined length is reached, as at points 31 and 39 in Fig. 3a. This stabilizes the definition of a reference value at each branch, mitigates user variability, and produces a highly reproducible definition of the vessel segment(s) in question. Fig. 3a, Fig.Figure 3b illustrates a path extending from a corrected user point and the image intensity used to locate the path lines for the three branches (proximal, distal 1, and distal 2). To limit propagation (as long as the predefined length is not reached), wavefront propagation is extended to include branch detection. First, at point i (36), the density profile is considered along the circle with center i and radius n times r, where r is the vessel radius at point i, and n is a suitable factor greater than one, such as between 1.5 and 5. If i is indeed a branch point, the density profile will have three indentations (like points 32, 34, and 38 in Figure 3b), corresponding to the three vessels at the branch. Each indentation has a size proportional to the local vessel size.Two of the indentations (32, 34) can already be coupled to a vessel because the line segment through point P passes through two vessels that are connected at the junction. It is now assumed that the index values ​​on the circle of intersection points with the line segment are i1 and i2 (the black and gray points 32, 34 in Fig. 3a and Fig. 3b). The size of these two indentations is also known because the radius of the two vessels is known from the adapted wavefront algorithm for each point along the line segment. Fig. 4 shows the amplitudes in areas bounded as depicted. First, attenuation 1 and attenuation 2 are indicated as horizontal arrows projected as dashed lines onto the intensity profile. Outside the attenuations at 32 and 34, two index regions lie along the circle. In each region, the index with the lowest intensity is indicated as i3 and i4, respectively. Now, the index sets {i1, i2, i3} and {i1, i2, i4} are considered individually. We want to know whether i3 or i4 (or both) is the center of a vessel whose inlet size is comparable to i1 and i2. First, we consider {i1, i2, i3}. The algorithm for the second set is the same. We search for the intensity maxima m1, m2, m3 in the regions bounded by {i1, i2, i3}. These maxima define the left and right background intensity of the vessels. From this, we define the minimum background intensities bmin1, bmin2, bmin3 and the maximum background intensities bmax1, bmax2, bmax3 of the three vessels. It should be noted that in Figures 3b and 4, the profile on the 360° circle does not have a true end; therefore, m1 is located at the end of the curve. From the figure, it follows that bmin1 = MIN(m1, m2), bmin2 = MIN(m2, m3), bmin3 = MIN(m3, m1), and bmax1 = MAX(m1, m2), bmax2 = MAX(m2, m3), bmax3 = MAX(m3, m1). The procedure may change somewhat, for example, if i2 were located before i1. The minimum contrast for the vessels can then be defined as follows: • cmin1 = bmin1 - intensity at i1 • cmin2 = bmin2 - intensity at i2 • cmin3 = bmin3 - intensity at i3 The maximum contrast for the vessels can then be defined as follows: • cmax1 = bmax1 - intensity at i1 • cmax2 = bmax2 - intensity at i2 • cmax3 = bmax3 - intensity at i3 It is now assumed that i3 is a point within the third vessel if the following conditions are met: • cmin1, cmin2, cmin3 are all above a threshold (n), the image noise level. • cmax3 has a contrast at least equal to n1 * MIN(cmax1, cmax2), where n1 is a threshold. This allows soft background structures and irrelevant details, such as a much smaller vessel branching off from the main vessel, to be ignored. • If the contrast to the left and right of the vessels differs, the contrast of the vessel with the background is higher than if the contrast at the left and right vessels is the same but lower than in the first case.To take this into account, we define the improved contrast of the three vessels by multiplying the minimum contrast value by 1.5, while ensuring that the result is at most the maximum contrast value: • enh1 = MAX(cmax1, n2*cmin1) • enh2 = MAX(cmax2, n2*cmin2) • enh3 = MAX(cmax3, n2*cmin3). Now, let's assume that enh3 is greater than 0.4*MAX(enh2, enh3). Consequently, the improved contrast of the third vessel must be at least approximately half the maximum value of the improved contrast of the other two vessels. If i3 (or i4) is the third vessel, then there are 3 vessels at point i, and it is assumed that i(36) is a branch point. The methodology described above for determining the distance from a position can be applied to a single vessel. In that case, the distance line will either end at a proximal / distal bifurcation or at a predetermined distance from the specified user point. The methodology described above can be extended to multiple bifurcations, such as a vascular tree. The procedure should then be adapted to proceed at the proximal or distal bifurcation, while defining a different limiting number. The final distance can be retained. The methodology described above can be extended to allow the user to pre-edit the position of the point specified by the user in order to correct any incorrect placement, resulting in even higher reproducibility. This could even allow for a starting point outside the vessel. The preprocessing generates four bidirectional scan lines of predetermined length that intersect the user point horizontally, vertically, and diagonally. Applying a multiscale approximation allows the vessel means to be found along these scan lines. A combination of the first derivative and the negated second discrete derivative of the density along each scan line is calculated at a set of different scales. This is done by convolving the scan lines with the negated first and second derivatives of the one-dimensional Gaussian function.A point c on the scanning line k is the center of a vessel with diameter z if and only if: - There is a zero crossing in the negation of the first derivative of the scanning line k at scale s = ceil(z / 2) - The negation of the second derivative at c at scale s is a positive value and is maximal over all indices j with a zero crossing at every scale in S with j ∈ [c - z, c + z] on the scanning line k. Figures 5a to 5c, for example, show multiscale responses of the intensity profile along a sampling line. Figure 5a shows an unprocessed profile, where the horizontal scale in pixels is along the sampling line. Figure 5b shows the negated Gaussian first derivative scale space of the intensity profile at scale parameter values ​​1, 2, 3, and 4 (80, 82, 84, 86). At the midpoint of the vessel (approximately at the horizontal index 46), we see a zero crossing in the first derivative at all scales. Similarly, Figure 5c shows the negated Gaussian second derivative scale space of the intensity profile at scale parameter values ​​1, 2, 3, and 4 (90, 92, 94, 96). There is a maximum value on scale 3 (94) that corresponds to a vessel diameter of 6 (in units of the horizontal axis). No higher negated second-order derivatives have a zero crossing within this diameter on either side of the maximum value.This area is indicated by arrows 88 and 98. See also 108. This shows the detection of a vessel at index 46 with a diameter of 6 (scale values ​​are essentially arbitrary). This concludes our example. Now, if the Euclidean distance of the c defined above from the user point is at most z, where z is the vessel diameter at c, all other candidate centers are eliminated. The maximum negated second derivative D is then determined among the remaining candidates. Finally, we shift the user point to the candidate vessel center closest to the user point among all candidates with a negated second derivative of at least 0.5*D. Checking the second derivative prevents the user point from being shifted to a background structure: due to the presence of contrast fluid, the second derivative in a vessel center will be much higher than that of a background structure. There are now three route lines and the detection of the bifurcation can be reduced to three conventional edge detections, for example by applying the minimum cost algorithm described by the reference material of Gronenschild and Tijdens. The midlines should now meet effectively at the bifurcation point. Each midline originates from the algorithm used for edge detection. In short, it is the midpoint between the detected outlines at each point along the vessel. The starting point for the bifurcation is the last point on the midline of the proximal vessel. Initially, the next points lie on the bifurcation boundaries: bleft, i; bright, j; bmid, k. A circle is drawn through these points, and the center of this circle is used as a better approximation for the bifurcation. This is repeated until the bifurcation no longer changes significantly. At the end of these repetitions, the bifurcation point is equal to the center of the largest circle that can be made within the bifurcation. The border combination (block 22 in Fig. 1) is described below. The outlines around the three vessels, generated using a method as described in the reference material by Gronenschild and Tijdens, must now be combined. In this respect, Fig. 6 shows the three points c0, c1, c2, which are used to determine the left and right vessels. Fig. 7 shows an example of a confluence pentagon. In Fig. 6, points c0, c1, and c2 represent the bifurcation point and any midline points near the confluence pentagon of the first distal vessel and the second distal vessel, respectively. Orientation is determined by the sine of the angle c1c0c2. If positive, c2 lies to the left of c1; otherwise, c2 lies to the right of c1. After determining the orientation of the bifurcation, it is known which outlines must be combined: - The left outline of the proximal vessel should be combined with the left outline of the left distal vessel. - The right outline of the proximal vessel should be combined with the right outline of the right distal vessel. - The left outline of the right distal vessel is swapped and combined with the right outline of the left distal vessel. The swap is necessary because the outlines must run in the same direction. The combination itself is performed by checking whether the two outlines intersect or are close to each other. For each point along one outline (denoted by b1,n), the vector to three subsequent points on the other outline (denoted by b2,m+{-1,0,1}) is calculated, and the inner product is used to determine whether the vectors point in the same direction. If this is not the case, then the outline has intersected the other outline. Furthermore, if the two outlines are closer to each other at a certain point than a certain number of pixels, they are also considered combinable: The results of the bifurcation segmentation are three sides, shown in Fig. 2. Block 24 in Fig. 1 represents the definition of the "confluence pentagon." To determine the end of the bifurcation and the start of the bifurcation region, we first construct the largest circle that fits inside the bifurcation. The intersections of this circle with the means are the positions that represent the bifurcation region. The point in the proximal part represents the start of the bifurcation, and the points in the distal parts represent the ends of the bifurcation; this region is called the "confluence pentagon." Three sides of the pentagon, 110, 112, and 114, each lie across one of the channels. Two sides of the pentagon, 116 and 118, lie along the outer sides of the channels. Where the two sides 112 and 114 meet, there is only a small space, which is approximated by a vertex of the pentagon. In rare cases, a hexagon might fit better into the measured points. Block 26 in Fig. 1 shows how the angles between the bifurcations are calculated. Fig. 8 shows the calculation of the angles between the respective bifurcations; the dots and crosses are the intersection points for determining the sides of the pentagon. The angles between the arterial bifurcations in conjunction with the confluence pentagon are calculated according to the following highly reproducible procedure. The influence of the position of the midpoint of the bifurcation (bifurcation point) is eliminated in this way. Within each bifurcation, a portion of the vessel near the confluence pentagon is defined, representing the direction of that particular bifurcation. This direction is indicated by a line segment based on the midline of that bifurcation. For the starting point of these line segments, we use the circle from the definition of the "confluence pentagon." The points of intersection (50, 52, 54 in Fig. 7) of the circle within the center lines are the starting points of the line segments. We then use each of these intersection points as the center of a new circle (56, 58, and 60, respectively). These new circles have the same radius as the circle that lies within the "confluence pentagon," or a radius that depends on the main vessel diameter of the branch in question. The points of intersection of these new circles with the center lines are the endpoints of the line segments (62, 64, and 66, respectively). These latter points of intersection are the starting and ending points of the line segments for use in determining the angle. Now we construct three line segments between the points of intersection. For each of these lines, we can calculate the tangent and consequently the angle. The diameter along the bifurcation is measured according to block 28 in Fig. 1. A bifurcation consists of left, middle, and right boundaries. The diameter measurement within the confluence pentagon is based on the approximation of "minimal freedom." This approximation is performed at each point, moving from the boundaries of the confluence pentagon toward the center point within the confluence pentagon. Each sampling point is centered within the boundaries. For each sampling point ci, the approximation of "minimal freedom" is performed as follows: 1. Find the boundary points on the left (bleft,i), (bright,j), and middle (bmid,k) boundary closest to ci. 2. Calculate the distance between the three points. 3. If the vessel is proximal ⇒ di = dlr; if the vessel is left distal ⇒ di = min(dlr, dlm); if the vessel is right distal ⇒ di = min(dlr, drm) leftAt the point of the left boundary right At the point of the right boundary middle At the point of the middle boundary lrFrom the left boundary to the right boundary point closest to the sampling point From the left boundary to the middle boundary point closest to the sampling point From the right boundary to the middle boundary point closest to the sampling point bThe most nearby point at the boundary dDistance Fig. 9 illustrates the calculation of the bifurcation diameter by introducing the new metric system described above. Within the confluence pentagon, indicated by the shaded portion in Fig. 9a, a single diameter at any midline point 70 within the confluence pentagon is calculated as the distances between boundary points 72 and 74, which are found by identifying the boundary point with the shortest distance from midline point 70. It should be noted that the diameter D, as shown in Fig. 9b, is identical from proximal to bifurcation point 76 of the main branch and the lateral branch, since the midline splits into distal1 and distal2 at bifurcation point 76.Diameters outside the confluence pentagon are determined as the distance between the left and right boundaries of the branch in question, as described in the reference material by Gronenschild and Tijdens. Block 30 in Fig. 1 represents the reference definition along the bifurcation. Acquired injuries can be expressed by quantities such as the percentage of narrowing, and therefore a healthy vessel is reconstructed by defining its diameter. The state of the art is limited to defining the true "reference value" that compensates for acquired injuries in the proximal vessel and its tributaries, as well as within the confluence pentagon. Reconstructing the confluence pentagon, in particular, is a challenging task. Now, a reference set is calculated for each branch, as described in the prior art by Gronenschild and Tijdens. Based on the reconstructed sides derived from these sets, the reference set within the pentagon is interpolated using a new curvature-based interpolation technique described below. Several prerequisites allow for the reconstruction of the sides of the "confluence pentagon." The reconstruction procedure logically follows from this. Prerequisites (bifurcation): B1. Blood flows smoothly from the proximal vessel to the two distal vessels. B2. The curvature of the "confluence pentagon" is constant. The following are the equivalent conditions for a single vessel: V1. Blood flows smoothly through a healthy vessel. V2. The curvature of a healthy vessel is locally constant. Assumption V1 implies that the diameter of a vessel does not change much. Since we are only considering a local model, this holds true for a healthy vessel. Assumption V2 implies that we can use a model of constant curvature, where the associated curve is a circle. Extending this idea to an object with a constant diameter produces a bulge in 3D and a ring in 2D. In 2D, this allows it to be modeled locally as part of a ring, as shown in Fig. 10. The ring is described by the inner radius (r, inverse of the curvature), the width (d, equal to the width of the vessel), and by a segment for which the local model applies between specified angles φprox(100) and φdist(102). This leads to the following model: The orientation of the vessel can be modeled by changing the sign of d. If d is negative, the right outline will be shorter, resulting in a vessel that curves to the right: the absolute value of d, however, is still equal to the diameter of the vessel. To fulfill requirements B1 and B2, the model is extended to include a bifurcation. For a bifurcation, we have three widths: dprox, dleft, and dright. Due to requirement B2, the curvature should not change at the transition from the proximal to one of the distal vessels. Consequently, the three widths should be combined with two inner radii: rleft and rright, resulting in three rings. The first two rings ensure condition B1 for the left branch. The last two rings ensure condition B1 for the right branch. The first two share the left boundary, while the last two share the right boundary. The boundaries are found by fitting the set of four rings to certain reference points in the bifurcation. It is now assumed that we have two proximal points, namely 130 and 132, two left distal points, 134 and 136, and two right distal points, 138 and 140. Figures 11a and 11b show the modeling of a bifurcation using a set of rings. For the right boundary (Fig. 11a), the centers lie somewhere on the line perpendicular to the two rightmost points 130 and 138 (φprox, right and φright, right). The position on the line is determined by making the "proximal left" point 132 fit the ring 〈mright, rright, dprox〉, and by making the "correct left" point 140 fit the ring 〈mright, rright, dright〉. The same is done for the left boundary, as shown in Fig. 11b (right image). The centers of the rings are denoted by 148 and 150, respectively. For each of the right and left boundaries, four of the six points are used to make the rings fit. In general, four points and two distances are used to find the parameters of two rings, which is done as follows: 1. Define a line of possible centers: 2. Calculate the distances from the other two points: 3. Find the center m (λ) and the inner radius r(λ) by minimizing d3(λ)2+d4(λ)2 to λ. The reference diameter within the "confluence pentagon" is determined by applying the diameter measurement method as described above. The tubular organs can include an artery, a vein, a coronary artery, a carotid artery, a pulmonary artery, a renal artery, a hepatic artery, a femoral artery, a Messenian artery, or another tubular organ acquired from angiographic image processing. This allows for a wide range of applications for the present invention. Often the polygon is a confluence pentagon with vertices on each side of the aforementioned tubular organs, and with a common vertex below any two distal tubular organs. This is essentially a very general situation, which lends itself to a clear analysis. A clear and rapid procedure has been identified by starting at a starting point in a proximal tubular organ (50) and proceeding to the endpoints (31, 39) in the respective distal tubular organs (52, 54). In particular, the diameter values ​​along the bifurcation are determined. Advantageously, a new unit of measurement is defined by determining a bifurcation diameter within the confluence pentagon as it extends between two arterial vertices (72-74) that are closest to a midline point of a single bifurcation (70 in Fig. 9). The angle between proximal and distal arteries is advantageously determined from lines extending between points on the midlines outside the confluence pentagon (50, 52, 54, 60, 64, 66). This is a quick procedure. Advantageously, the input for identifying a bifurcation is a single point (36) that approaches a midpoint of the bifurcation. A procedure is advantageously applied, enabling the tracking of a single tubular organ segment between a proximal and a distal bifurcation from a single starting point until a bifurcation is encountered or a predetermined distance is traversed. The method can also be used to detect a combination of multiple vessel tree bifurcations. Advantageously, a reference bifurcation is modeled by a set of rings with an inner circular edge matching an inner arterial margin curve and with an outer circular edge matching the opposite arterial margin curve (Fig. 11a , Fig. 11b). The present invention has now been described with the aid of preferred embodiments. Naturally, those skilled in the art will be able to think of several additions and modifications. Consequently, the description should be considered illustrative rather than limiting, and no limitations should be interpreted other than those explicitly stated in the accompanying claims.

Claims

A method for performing a quantitative analysis of medical image data of a bifurcated tubular organ having at least one proximal part, a first distal part, and a second distal part, the method comprising: - processing said medical image data to identify outlines of said bifurcated tubular organ along its longitudinal extent, wherein the outlines have first and second outer outlines and a middle outline between them; - using said outlines to determine a confluence polygon in the middle of said bifurcated tubular organ, wherein the confluence polygon is bounded by a plurality of lines intersecting at vertices, and wherein the plurality of lines identify both a proximal start of the proximal part of the bifurcated tubular organ and distal ends of the first and second distal parts of the bifurcated tubular organ.wherein a first line of the majority of the lines extends along the first outer outline and a second line extends along the second outer outline,- using the said confluence polygon to determine at least one parameter value that characterizes the geometry of the said bifurcated tubular organ, and- delivering the said at least one parameter value to a user. Method according to claim 1, wherein:- the at least one parameter value comprises at least one diameter value of the said forked tubular organ. Method according to claim 2, wherein:- the at least one diameter value is selected from the group comprising a diameter value of the proximal part of the bifurcated tubular organ, a diameter value of the first distal part of the bifurcated tubular organ and a diameter value of the second distal part of the bifurcated tubular organ. Method according to claim 2, wherein:- the at least one diameter value is derived from a distance between two boundary points on the confluence polygon. Method according to claim 4, wherein:- the said two boundary points are identified by scanning points on the closed surface, in order to identify the two points that are closest to a centerline point within the confluence polygon. Method according to claim 1, wherein:- the at least one parameter value has at least one angle value between parts of the said forked tubular organ. Method according to claim 6, wherein: - the at least one angle is selected from the group comprising an angle between the proximal and the first distal part of the bifurcated tubular organ, an angle between the proximal and the second distal part of the bifurcated tubular organ, and an angle between the first and the second part of the bifurcated tubular organ. Method according to claim 7, wherein:- the at least one angle is derived from a line extending between a centerline point on the confluence polygon and a centerline point outside the confluence polygon. Method according to claim 1, wherein:- the at least one parameter has at least one reference diameter value for the said forked tubular organ, wherein the at least one reference diameter value compensates for damage to the forked tubular organ. Method according to claim 9, wherein: - the at least one reference diameter value is selected from the group comprising a reference diameter value for the proximal part of the bifurcated tubular organ, a reference diameter value of the first distal part of the bifurcated tubular organ, and a reference diameter value of the second distal part of the bifurcated tubular organ. Method according to claim 9, wherein: - the at least one reference diameter value is derived from a set of rings with an inner circular rim matching an inner boundary curve and with an outer circular rim matching an opposite boundary curve. Method according to claim 1, wherein:- the said confluence polygon is derived from a bifurcation point located in the middle of a circle that fits into the said bifurcated tubular organ. Method according to claim 12, wherein:- the said confluence polygon is derived from line segments that extend through an intersection point between the said circle and center lines of the said proximal part, the said first distal part or the said second distal part. Method according to claim 1, wherein:- the said confluence polygon has five vertices, with one vertex adjacent to the intersection of the first and second distal part of the bifurcated tubular organ. Method according to claim 1, wherein:- the outlines of the bifurcated tubular organ are identified as beginning at a starting point in the proximal part to the endpoints in the respective first and second distal parts of the bifurcated tubular organ. Method according to claim 1, wherein:- the outlines of the bifurcated tubular organ are derived from the user identification of a single point that approximates a bifurcation center of the bifurcated tubular organ. Method according to claim 1, wherein: - the outlines of the bifurcated tubular organ are derived from a single-point trigger for tracking a single segment of a tubular organ, starting at a detected proximal bifurcation or at a predefined distance from the starting point, which may be closer to the starting point, and ending at a detected distal bifurcation or at a predefined distance from the starting point, whichever may be. Method according to claim 1, wherein:- the method is provided to influence a multiple bifurcation analysis to detect a combination of multiple vessel tree bifurcation. Method according to claim 1, wherein:- the bifurcated tubular organ is selected from the group comprising an artery, a vein, a coronary artery, a carotid artery, a pulmonary artery, a renal artery, a hepatic artery, a femoral artery and a Messenian artery. Data processing system for the quantitative analysis of medical image data of a bifurcated tubular organ with at least one proximal part, a first distal part and a second distal part, the data processing system comprising: - means for processing the said medical image data to identify the outlines of the said bifurcated tubular organ along its longitudinal extent, wherein the outlines have first and second outer outlines and a middle outline in between, - means for determining a confluence polygon in the middle of the bifurcated tubular organ based on the said outlines;wherein the confluence polygon is bounded by a plurality of lines intersecting at vertices, and wherein the plurality of lines identify both a proximal start of the proximal part of the bifurcated tubular organ and distal ends of the first and second distal parts of the bifurcated tubular organ, wherein a first line of the plurality of lines extends along the first outer outline and a second line extends along the second outer outline, and means for determining at least one parameter value characterizing the geometry of said bifurcated tubular organ based on said confluence polygon; and means for delivering said at least one parameter value to a user. Data processing system according to claim 20, wherein: - the at least one parameter value is selected from the group comprising at least one diameter value of the said forked tubular organ, at least one angle value between parts of the said forked tubular organ, and at least one reference diameter value for the said forked tubular organ, wherein the at least one reference diameter value compensates for damage to the forked tubular organ. A program memory arrangement that can be read by a machine, comprising a program with instructions that can be executed by the machine, for carrying out the process steps for the quantitative analysis of medical image data of a bifurcated tubular organ having at least one proximal part, a first distal part, and a second distal part, wherein said process steps comprise: - processing said medical image data to identify outlines of said bifurcated tubular organ along its longitudinal extent, wherein the outlines have first and second outer outlines and a middle outline between them; - using said outlines to determine a confluence polygon in the middle of said bifurcated tubular organ, wherein the confluence polygon is bounded by a plurality of lines intersecting at vertices.and wherein the majority of lines identify both a proximal start of the proximal part of the bifurcated tubular organ and distal ends of the first and second distal parts of the bifurcated tubular organ, wherein a first line of the majority of lines extends along the first outer outline and a second line extends along the second outer outline, - using the said confluence polygon to determine at least one parameter value that characterizes the geometry of the said bifurcated tubular organ, and - delivering the said at least one parameter value to a user. Program memory arrangement according to claim 22, wherein: - the at least one parameter value is selected from the group comprising at least one diameter value of the said forked tubular organ, at least one angle between parts of the said forked tubular organ, and at least one reference diameter value for the said forked tubular organ, wherein the at least one reference diameter compensates for damage to the forked tubular organ.

Images