WO2017109662A1 - Heart model guided coronary artery segmentation - Google Patents

Heart model guided coronary artery segmentation Download PDF

Info

Publication number
WO2017109662A1
WO2017109662A1 PCT/IB2016/057758 IB2016057758W WO2017109662A1 WO 2017109662 A1 WO2017109662 A1 WO 2017109662A1 IB 2016057758 W IB2016057758 W IB 2016057758W WO 2017109662 A1 WO2017109662 A1 WO 2017109662A1
Authority
WO
WIPO (PCT)
Prior art keywords
coronary artery
heart
segmentation
boundary
distance
Prior art date
Application number
PCT/IB2016/057758
Other languages
French (fr)
Inventor
Cristian Lorenz
Tobias Klinder
Holger Schmitt
Hannes NICKISCH
Original Assignee
Koninklijke Philips N.V.
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority to US201562270981P priority Critical
Priority to US62/270,981 priority
Application filed by Koninklijke Philips N.V. filed Critical Koninklijke Philips N.V.
Publication of WO2017109662A1 publication Critical patent/WO2017109662A1/en

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00Image analysis
    • G06T7/10Segmentation; Edge detection
    • G06T7/149Segmentation; Edge detection involving deformable models, e.g. active contour models
    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00Image analysis
    • G06T7/10Segmentation; Edge detection
    • G06T7/11Region-based segmentation
    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00Image analysis
    • G06T7/10Segmentation; Edge detection
    • G06T7/12Edge-based segmentation
    • GPHYSICS
    • G16INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16HHEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/50ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for simulation or modelling of medical disorders
    • AHUMAN NECESSITIES
    • A61MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61BDIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00Detecting, measuring or recording for diagnostic purposes; Identification of persons
    • A61B5/02Detecting, measuring or recording pulse, heart rate, blood pressure or blood flow; Combined pulse/heart-rate/blood pressure determination; Evaluating a cardiovascular condition not otherwise provided for, e.g. using combinations of techniques provided for in this group with electrocardiography or electroauscultation; Heart catheters for measuring blood pressure
    • A61B5/026Measuring blood flow
    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00Indexing scheme for image analysis or image enhancement
    • G06T2207/10Image acquisition modality
    • G06T2207/10072Tomographic images
    • G06T2207/100764D tomography; Time-sequential 3D tomography
    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00Indexing scheme for image analysis or image enhancement
    • G06T2207/10Image acquisition modality
    • G06T2207/10072Tomographic images
    • G06T2207/10081Computed x-ray tomography [CT]
    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00Indexing scheme for image analysis or image enhancement
    • G06T2207/10Image acquisition modality
    • G06T2207/10072Tomographic images
    • G06T2207/10088Magnetic resonance imaging [MRI]
    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00Indexing scheme for image analysis or image enhancement
    • G06T2207/30Subject of image; Context of image processing
    • G06T2207/30004Biomedical image processing
    • G06T2207/30048Heart; Cardiac
    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00Indexing scheme for image analysis or image enhancement
    • G06T2207/30Subject of image; Context of image processing
    • G06T2207/30004Biomedical image processing
    • G06T2207/30101Blood vessel; Artery; Vein; Vascular

Abstract

A system (100) for segmenting a coronary artery vessel tree (182) of a patient heart in a three dimensional (3D) cardiac image (120) includes a coronary volume definition unit (150) and a coronary artery segmentation unit (180). The coronary volume definition unit (150) sets a spatial boundary (210, 220) from internal and external surfaces of heart tissues in the 3D cardiac image based on a fitted heart model (200). The coronary artery segmentation unit (180) segments the coronary artery vessel tree (182) in the 3D cardiac image using a segmentation algorithm with a search space limited by the spatial boundary set from the internal and external surfaces of the heart tissues.

Description

HEART MODEL GUIDED CORONARY ARTERY SEGEMENTATION

FIELD OF THE INVENTION

The following generally relates to medical imaging with specific application to coronary artery segmentation.

BACKGROUND OF THE INVENTION

Accurate computation of virtual fractional flow reserve (FFR) using a digital representation of a segmented coronary artery vessel depends on the accuracy of the segmented vessel. This virtual method is non-invasive in contrast with a more conventional method using an invasive catheter based actual FFR measurement. FFR measures pressure differences across a coronary artery stenosis to determine a likelihood that a stenosis impedes oxygen delivery to the heart muscle.

Current approaches to segmentation of the coronary artery use a three dimensional (3D) volume image of the cardiac region, such as generated from a computed tomography (CT) image, or a magnetic resonance (MR) image. The modalities provide a resolution sufficient to identify at least the larger diameter portions of the coronary artery tree which includes the coronary artery and the vessels that branch from the coronary artery and supply oxygen to the tissues of the heart. Due to limits on resolution and artifacts, such as motion artifacts, the boundaries of vessels are not always clearly delineated, e.g. blurred, indistinct, etc. These imaging aspects in combination with variability in anatomical structure of the coronary artery tree from subjects make accurate representations of the coronary vessel tree difficult.

The current approaches use various algorithms to define a coronary vessel tree, such as seed growing algorithms with parameters that are difficult to control. For example, the algorithms generally use a comparison of adjacent or neighboring voxels with the current growing region to determine whether to add the compared voxel to the current growing region. The algorithms rely on image properties, such as a change in intensity to make the determination of whether or not to add the compared voxel, e.g. if difference is less than a threshold amount determined by a parameter, then add otherwise do not add.

One of the problems encountered in conventional approaches to segmentation of the coronary artery tree is leakage of the segmentation into nearby structures. For example, in region growing algorithms, this typically occurs where the boundaries are blurred or less distinct, and the growing algorithm adds voxels of nearby structures, such as the heart chambers, myocardium, or pulmonary artery, into the current growing region. Some algorithms try to control for this leakage based on adjustments of parameters. The parameters control the intensity thresholds or measured intensities in combinations of voxels, which becomes computationally expensive, e.g. long run-times, and can miss smaller branches of the vessel tree, such as those close to image resolution, e.g. only a few millimeters in diameter.

Heart models are known in the art, which are typically used for heart segmentation or heart motion correction. For example, a digital representation of a heart can include tissues of the heart represented as triangular meshes. The model is fit to the imaged heart, and motion correction is determined from matching the model at different phases to the image heart. However, the heart models typically include (if at all) only the large diameter vessels, and do not include entire vessel trees.

SUMMARY OF THE INVENTION

Aspects described herein address the above-referenced problems and others.

The following describes a method and system for a heart model guided coronary artery segmentation. The heart model is fit to internal and external surfaces of a heart in image data and the fitted heart model is used to provide spatial limitations for the coronary artery segmentation. A fitted coronary artery probability map can be used to guide the segmentation.

In one aspect, a system for segmenting a coronary artery vessel tree of a patient heart in a three dimensional (3D) cardiac image includes a coronary volume definition unit and a coronary artery segmentation unit. The coronary volume definition unit sets a spatial boundary from internal and external surfaces of heart tissues in the 3D cardiac image based on a fitted heart model. The coronary artery segmentation unit segments the coronary artery vessel tree in the 3D cardiac image using a segmentation algorithm with a search space limited by the spatial boundary set from the internal and external surfaces of the heart tissues.

In another aspect, a method of segmenting a coronary artery vessel tree of a patient heart in a three dimensional (3D) cardiac image includes setting a spatial boundary from internal and external surfaces of heart tissues in the 3D cardiac image based on a fitted heart model. The coronary artery vessel tree in the 3D cardiac image is segmented using a segmentation algorithm with a search space limited by the spatial boundary set from the internal and external surfaces of the heart tissues.

In another aspect, a system for segmenting a coronary artery vessel tree of a patient heart in a medical image includes one or more processors configured to segment the coronary artery vessel tree in the image using a segmentation algorithm with a search space limited by a spatial boundary set from internal and external surfaces of heart tissues in the image using a fitted heart model.

BRIEF DESCRIPTION OF THE DRAWINGS

The invention may take form in various components and arrangements of components, and in various steps and arrangements of steps. The drawings are only for purposes of illustrating the preferred embodiments and are not to be construed as limiting the invention.

FIGURE 1 schematically illustrates an embodiment of a heart model guided coronary artery segmentation system.

FIGURE 2 illustrate an exemplary cross section of a heart model fitted to internal and external heart surfaces in a patient image and a spatial boundary determination.

FIGURE 3 illustrates an exemplary coronary artery probability map.

FIGURE 4 flowcharts an embodiment of a method of segmenting a coronary artery vessel tree based on the fitted heart model.

DETAILED DESCRIPTION OF EMBODIMENTS

Initially referring to FIGURE 1, a heart model guided coronary artery segmentation system 100 is schematically illustrated. A medical imaging device 110, such as a CT scanner, a MR scanner, combinations and the like, generates a three dimensional (3D) volumetric cardiac image 120. The 3D cardiac image 120 is of sufficient spatial resolution to show the coronary artery, such as a millimeter or better. The generated image can include the use of an administered contrast agent, which contrasts the coronary artery lumen. The generated 3D cardiac image 120 can be prospectively or retrospectively gated to reduce motion artifacts. The 3D cardiac image 120 can include the heart and surrounding tissues or portions of the heart and surrounding tissues.

A heart model fit unit 130 fits an anatomical heart model 140 known in art to the heart in the 3D cardiac image. The heart model 140 is a digital spatial representation of the tissues of the heart. The heart model can include labeled anatomical points or regions. The heart model fit unit 130 elastically fits the heart model to the imaged heart, e.g.

individualizes the heart model to a specific patient. For example, the surfaces of the tissues can be represented as triangular meshes. The fit can be based on matching of anatomical points or regions in heart model 140 to anatomical points in the heart in the 3D cardiac image 120, such as positions of valves, apexes, septum, chordae tendineae, etc. Points or regions of the elastic mesh can be anchored at the matched anatomical points or regions and the elastic mesh can be adjusted to match the surface of the corresponding tissue.

With the surface of the patient heart fitted with the heart model, a coronary volume definition unit 150 sets a spatial boundary from the surface of the heart tissues. The search space includes myocardial tissues between the surfaces of the heart tissues along with the arterial feeds disposed in that spatial region. In some instances this ensures that arterial flow that goes deeply into the heart tissues is included in the search space. A first bounding distance di outside heart is set for external heart surfaces outside, which excludes tissues, such as the pulmonary vasculature. A second bounding distance d2 is set for surfaces inside heart, which excludes volumes such as heart chambers. For example, the first distance can extend externally 1.5 centimeters (cm) perpendicular to the external heart surface and the second distance extends internally 1.0 cm perpendicular to the internal heart surfaces.

Another example uses 3.0 cm and 2.0 cm as the external and internal distances. In another example, the distances are the same externally and internally of 1.5 cm or 3.0 cm. Other distances are contemplated. The bounding distance inside and outside of the organ can be different or the same. In one embodiment, the bounding distances can be graduated according to a distance along the heart surface decreasing in distance toward the ventricles. In one embodiment, the graduate distances include a function of a spatial distance from a start of the coronary artery along the heart surface. For example, as the artery decreases in diameter, such as by branching, the search space can be moved closer to the heart surface.

A coronary ostia finder unit 160 finds initial seed points from left and right coronary artery ostium based on the fitted heart model. For example, the coronary ostia finder unit 160 uses marked triangles from the mesh model to begin search through ascending aorta model wall identifying a transition in contrast from medium to high intensity corresponding to the left and right coronary ostia.

A coronary artery auxiliary unit 170 generates and maintains the coronary artery probability map 172, which includes a probability that each voxel is part of the coronary artery vessel tree. Based in part on the larger vessels of the fitted heart model, and the probability map 172 can generated for the coronary artery and branches based on a distribution of the vessel tree across a sample population and matched to or fused with the heart model. For example, a sample of segmented coronary arteries from a healthy patient population are deformably registered to the heart model, and a probability computed as a function of a frequency of occurrences of the segmented coronary arteries in each voxel. The probabilities according to each voxel or spatial location according to the heart model form the probability map 172. In one embodiment, the probability map can be stored as part of the heart model, in which each surface element stores the probabilities for a set of voxels surrounding the surface element. As the heart model is elastically fitted, the probability map is elastically fitted as well. In some embodiments, the fitted probability map 172 can be used to exclude loops in the segmentation, such as at the vessel segment level. A loop occurs when a vessel appears to feed itself. In some embodiments, the fitted probability map 172 can be used with thresholds values, such as in bifurcation of branches. For example, arteries branch within angles of 0-70 degrees, which can be used as a threshold to limit a local search space of neighboring voxels.

The coronary artery segmentation unit 180 segments the coronary artery vessel tree 182 with a segmentation algorithm bounded by the spatial boundaries determined by the coronary volume definition unit 150. Suitable algorithms can include tree

segmentation frame works or distance map approaches, such as, T. Buelow et al "A general framework for tree segmentation and reconstruction from medical volume data" using a fast marching as region expansion with measures including vessel voxels, vessel segments, and a whole vessel tree. Another example, is a method described by T. Klinder et al. "Robust Peripheral Airway Segmentation" that can be adapted to arterial vessels, and uses a filter based vessel candidate detection with segment connection functionality. Examples of distance map approaches can include S. Svensson et al. "Digital Distance Transforms in 3D images using information from neighborhoods up to 5x5x5" and R. Calvin et al. "A linear time algorithm for computing exact Euclidean distance transforms of binary images in arbitrary dimensions." Other algorithms are contemplated. In some instances, the spatial boundary reduces segmentation run-time by decreasing search space. In some instances, segmentation robustness is increased by eliminating structures typically causing leakage, such as heart chambers, and pulmonary vasculature. Acceptance criteria for inclusion in the segmented vessel tree 182 can include voxel based acceptance, such as a distance measure, a probability and/or a directional flow, a vessel-segment acceptance, such as a probability of a next branch, and/or a whole vessel tree level acceptance. The heart model fit unit 130, the coronary volume definition unit 150, the coronary ostia finder unit 160, the coronary artery auxiliary unit 170, and the coronary artery segmentation unit 180 comprise one or more configured processors 190, e.g., a

microprocessor, a central processing unit, a digital processor, and the like. The one or more configured processors 190 are configured to execute at least one computer readable instruction stored in a computer readable storage medium, which excludes transitory medium and includes physical memory and/or other non-transitory medium to perform the techniques described herein. The one or more processors 190 may also execute one or more computer readable instructions carried by a carrier wave, a signal or other transitory medium. The one or more processors 190 can include local memory and/or distributed memory. The one or more processors 190 can include hardware/software for wired and/or wireless

communications over a network 192. For example, the lines indicate communications paths between the various components which can be wired or wireless. The one or more processors 190 can comprise the computing device 194, such as a desktop, a laptop, a body worn device, a smartphone, a tablet, and/or cooperative/distributed computing devices including one or more configured servers (not shown). The computing device 194 can include a display device 196, which can display the segmented vessel tree 182 or computations made from the segmented vessel tree 182, such as FFR. The computing device 194 can include one or more input devices 198 which receive commands, such as identifying the 3D cardiac image 120, confirming seed points by the coronary ostia finder unit 160, displaying the coronary probability map 172, and/or segmenting the coronary artery vessel tree 182.

The 3D cardiac image 120, the heart model 140, the segmented coronary artery tree 182 are represented as digital data sets stored on an electronic storage medium or computer memory. The 3D cardiac image 120 can include a Digital Imaging and

Communications in Medicine (DICOM) format or other suitable image format.

With reference to FIGURE 2, an exemplary cross section of a fitted heart model 200 fitted to internal and external surfaces of a heart from the 3D cardiac image 120 and spatial boundary determination is illustrated. Tissues, such as myocardial tissues 205, represented in the fitted heart model 200 are illustrated superimposed over the tissues represented in the 3D cardiac image 120. From the exterior surface of the fitted heart model 200, the first boundary 210, di, is determined. The first boundary is defined as a surface of distance di determined by orthogonal projections from the mesh surface of the fitted heart model 200. Tissues external to the first boundary 210 are excluded from the search space of the segmentation algorithm. From the interior surface of the fitted heart model 200, the second boundary 220, d2 is determined. Tissues enclosed by a surface defined by the second boundary 220 are also excluded from the search space of the segmentation algorithm, e.g. chamber volumes. The two boundary determinations are made with distances such that search space includes the coronary artery vessel tree including myocardial tissues between the external surface and internal surface with their associated arterial supply. In some embodiments, the boundary determinations excludes tissues identified from the fitted heart model 200 that have a low probability of including the coronary artery vessel tree based on the fitted probability map.

From the fitted model 200, a lumen 230 of the coronary artery can be located and initial seed points for the segmentation algorithm determined, such as the ostia of the left and right coronary artery. The ostia can be automatically identified and confirmed visually by a healthcare practitioner.

With reference to FIGURE 3, an exemplary coronary artery probability map 172 is illustrated. The probability map 172 is illustrated visually as a two dimensional (2D) representation of the 3D map with 2D pixels darkened according to corresponding probability, e.g. darker pixels with higher probability of belonging to coronary artery vessel tree. The probabilities of the coronary artery probability map 172 are represented

corresponding to spatially located 3D voxels.

The coronary artery probability map 172 can include probabilities for conflicting structures, such as the coronary venous system. The coronary artery probability map 172 can include distance values, such as typical distances values between a nearest artery and an endocardial surface, such as represented with a triangle of the fitted mesh model. The coronary artery probability map 172 can include a directional weighted probabilities based on arterial flow, e.g. higher probability of adjacent voxels along centerline of lumen with decreasing probability as angle increases away from the centerline.

With reference to FIGURE 4, an embodiment of a method of segmenting a coronary artery vessel tree based on the fitted heart model is flowcharted. At 400, the 3D cardiac image 120 is received. The 3D cardiac image 120 can be received from a storage subsystem, such as a Picture Archiving and Communication System (PACS), departmental Radiology Information System (RIS), Electronic Medical Record (EMR), and the like. The 3D cardiac image 120 can be generated by the medical imaging device 110 and received directly from the medical imaging device 110.

At 410, the heart model 140 is fitted to the heart in the 3D cardiac image 120. The fitted heart model 200 includes defined tissue surfaces that are fitted to the tissue surfaces of the heart in the 3D cardiac image 120. The first boundary 210 and the second boundary 220 are determined from the fitted heart model 200. Surfaces of the boundaries limit the search space. The coronary artery probability map 172 is fit based on the fitted heart model 200 and included in the search space. The coronary artery probability map 172 is constructed from a sampling of segmented coronary artery vessel trees deformably registered to the heart model 140. Probabilities can be computed from the sampling distribution. The coronary artery probability map 172 can include distance measures to the nearest artery, a directional indicator of the arterial flow and/or a branching threshold.

The coronary artery vessel tree 182 is segmented at 420. The segmentation can include identifying seed points with a search for the seeds points based on the fitted heart model 200. The segmentation can include revising or updating the probability map 172 based on acceptance level at the voxel, segment, or whole vessel tree. The segmentation includes spatially limiting the search space to the volume between the surfaces defined by the first boundary 210 and the second boundary 220. The segmented coronary artery vessel tree 182 includes a digital representation that for each spatially located voxel defines whether a voxel is included in the vessel lumen. In one embodiment, the digital representation can include stenosis, such as calcium, plaque, etc.

At 430, the coronary artery vessel tree 182 can be stored in a computer memory and/or displayed on the display device 196. The coronary artery vessel tree 182 or portions thereof can be displayed as 2D projections. The projections can include interior navigation, different perspectives, color contrast, and the like. The projections can include measurements, such as diameters, which are contrasted in the display. The projections can include visual contrasts of stenosis and/or materials, such as plaque, calcium, etc.

At 440, a fractional flow reserve value (FFR) can be computed based on the segmented coronary artery vessel tree 182. The FFR can be used to identify and rate stenosis in the vessel tree 182.

The invention has been described with reference to the preferred embodiments. Modifications and alterations may occur to others upon reading and understanding the preceding detailed description. It is intended that the invention be constructed as including all such modifications and alterations insofar as they come within the scope of the appended claims or the equivalents thereof.

Claims

CLAIMS:
1. A system (100) for segmenting a coronary artery vessel tree (182) of a patient heart in a three dimensional (3D) cardiac image (120), comprising:
a coronary volume definition unit (150), comprising one or more processors (190) configured to set a spatial boundary (210, 220) from internal and external surfaces of heart tissues in the 3D cardiac image based on a fitted heart model (200); and
a coronary artery segmentation unit (180), comprising the one or more processors configured to segment the coronary artery vessel tree (182) in the 3D cardiac image using a segmentation algorithm with a search space limited by the spatial boundary set from the internal and external surfaces of the heart tissues.
2. The system according to claim 1, wherein the spatial boundary includes a first boundary of a first distance (210) from external heart surfaces and excludes from
segmentation tissues external to the first distance and a second boundary of a second distance (220) from internal heart surfaces and excludes from segmentation tissues enclosed by the second distance.
3. The system according to either one of claims 1 and 2, further including:
a coronary artery auxiliary unit (170), comprising the one or more processors configured to fit a coronary artery probability map (172) based on the fitted heart model; and wherein the segmentation algorithm uses the probability map to guide segmentation based on the probability map.
4. The system according to claim 3, wherein the coronary artery probability map is constructed from a sampling of segmented coronary arteries deformably fit to a heart model (140) of the fitted heart model (200).
5. The system according to any one of claims 1-4, further including:
a coronary ostia finder unit (160) comprising the one or more processors configured to identify initial seed points for the segmentation algorithm located based on predetermined points identified based on the fitted heart model.
6. The system according to any one of claims 2-5, wherein the first boundary distance and the second boundary distance are different.
7. The system according to any one of claims 1-6, wherein the external surfaces and the internal surfaces are different surfaces.
8. The system according to any one of claims 2-7, wherein the first boundary distance is graduated and decreasing toward an apex of the patient heart.
9. The system according to any one of claims 3-8, wherein the coronary artery probability map includes a probability that a corresponding spatially located voxel is included in the coronary artery vessel tree.
10. The system according to any one of claims 3-9, wherein the coronary artery probability map includes a probability based on a directional arterial flow that a
corresponding spatially located voxel is included in the coronary artery vessel tree.
11. A method of segmenting a coronary artery vessel tree (182) of a patient heart in a three dimensional (3D) cardiac image (120), comprising:
setting (410) a spatial boundary (210, 220) from internal and external surfaces of heart tissues in the 3D cardiac image based on a fitted heart model (200); and
segmenting (420) the coronary artery vessel tree (182) in the 3D cardiac image using a segmentation algorithm with a search space limited by the spatial boundary set from the internal and external surfaces of the heart tissues.
12. The method according to claim 11, wherein the spatial boundary includes a first boundary of a first distance (210) from external heart surfaces and excludes from segmentation tissues external to the first distance and a second boundary of a second distance (220) from the internal heart surfaces and excludes from segmentation tissues enclosed by the second distance.
13. The method according to either one of claims 1 land 12, wherein segmenting includes fitting a coronary artery probability map (172) based on the fitted heart model; and wherein the segmentation algorithm uses the probability map to guide segmentation based on the probability map.
14. The method according to any one of claims 11-13, wherein setting includes: constructing the coronary artery probability map from a sampling of segmented coronary arteries deformably fit to a heart model (140) of the fitted heart model (200)
15. The method according to any one of claims 11-14, wherein segmenting includes identifying initial seed points for the segmentation algorithm located based on predetermined points identified based on the fitted heart model.
16. The method according to any one of claims 11-15, wherein the first boundary distance and the second boundary distance are different.
17. The method according to any one of claims 11-16, wherein the internal and external surfaces are different surfaces.
18. The method according to any one of claims 14-17, wherein the coronary artery probability map includes a probability that a corresponding spatially located voxel is included in the coronary artery vessel tree.
19. The method according to any one of claims 14-18, wherein the coronary artery probability map includes a probability based on a directional arterial flow that a
corresponding spatially located voxel is included in the coronary artery vessel tree.
20. A system (100) for segmenting a coronary artery vessel tree (182) of a patient heart in a medical image (120), comprising:
one or more processors configured to segment the coronary artery vessel tree in the image using a segmentation algorithm with a search space limited by a spatial boundary set from internal and external surfaces of heart tissues in the image using a fitted heart model (200).
PCT/IB2016/057758 2015-12-22 2016-12-19 Heart model guided coronary artery segmentation WO2017109662A1 (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
US201562270981P true 2015-12-22 2015-12-22
US62/270,981 2015-12-22

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
US16/060,453 US20180365838A1 (en) 2015-12-22 2016-12-19 Heart model guided coronary artery segmentation
JP2018532617A JP2019500960A (en) 2015-12-22 2016-12-19 Heart model guided coronary artery aneurysm segmentation
EP16823055.5A EP3394829A1 (en) 2015-12-22 2016-12-19 Heart model guided coronary artery segmentation
CN201680075741.3A CN108475428A (en) 2015-12-22 2016-12-19 Heart model guided coronary artery segmentation

Publications (1)

Publication Number Publication Date
WO2017109662A1 true WO2017109662A1 (en) 2017-06-29

Family

ID=57755416

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/IB2016/057758 WO2017109662A1 (en) 2015-12-22 2016-12-19 Heart model guided coronary artery segmentation

Country Status (5)

Country Link
US (1) US20180365838A1 (en)
EP (1) EP3394829A1 (en)
JP (1) JP2019500960A (en)
CN (1) CN108475428A (en)
WO (1) WO2017109662A1 (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2018037005A1 (en) * 2016-08-22 2018-03-01 Koninklijke Philips N.V. Model regularized motion compensated medical image reconstruction

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20060056694A1 (en) * 2004-09-09 2006-03-16 Daniel Rinck Method for segmenting anatomical structures from 3D image data by using topological information
WO2007072363A2 (en) * 2005-12-19 2007-06-28 Koninklijke Philips Electronics, N.V. Method for facilitating post-processing of images using deformable meshes

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20060056694A1 (en) * 2004-09-09 2006-03-16 Daniel Rinck Method for segmenting anatomical structures from 3D image data by using topological information
WO2007072363A2 (en) * 2005-12-19 2007-06-28 Koninklijke Philips Electronics, N.V. Method for facilitating post-processing of images using deformable meshes

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
R. CALVIN ET AL., A LINEAR TIME ALGORITHM FOR COMPUTING EXACT EUCLIDEAN DISTANCE TRANSFORMS OF BINARY IMAGES IN ARBITRARY DIMENSIONS
S. SVENSSON ET AL., DIGITAL DISTANCE TRANSFORMS IN 3D IMAGES USING INFORMATION FROM NEIGHBORHOODS UP TO 5X5X5
T. KLINDER ET AL., ROBUST PERIPHERAL AIRWAY SEGMENTATION

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2018037005A1 (en) * 2016-08-22 2018-03-01 Koninklijke Philips N.V. Model regularized motion compensated medical image reconstruction

Also Published As

Publication number Publication date
CN108475428A (en) 2018-08-31
US20180365838A1 (en) 2018-12-20
EP3394829A1 (en) 2018-10-31
JP2019500960A (en) 2019-01-17

Similar Documents

Publication Publication Date Title
Manniesing et al. Level set based cerebral vasculature segmentation and diameter quantification in CT angiography
Li et al. Optimal surface segmentation in volumetric images—A graph-theoretic approach
US20050111719A1 (en) Method for registration of an image applied to digital subtracted angiography
Hernandez et al. Non-parametric geodesic active regions: method and evaluation for cerebral aneurysms segmentation in 3DRA and CTA
Van Uitert et al. Subvoxel precise skeletons of volumetric data based on fast marching methods
JP6302922B2 (en) Fractional flow reserve (ffr) index
US20050110791A1 (en) Systems and methods for segmenting and displaying tubular vessels in volumetric imaging data
US9323887B2 (en) Device and computed tomography scanner for determining and visualizing the perfusion of the myocardial muscle
US20070165917A1 (en) Fully automatic vessel tree segmentation
US9907527B2 (en) Vascular data processing and image registration systems, methods, and apparatuses
Tu et al. In vivo comparison of arterial lumen dimensions assessed by co-registered three-dimensional (3D) quantitative coronary angiography, intravascular ultrasound and optical coherence tomography
CN1620990A (en) Method and apparatus for segmenting structure in CT angiography
JP4785371B2 (en) Extraction method and system of the multidimensional structure using dynamic constraints
US7603154B2 (en) Non-invasive left ventricular volume determination
Frangi Three-dimensional model-based analysis of vascular and cardiac images
Vauclin et al. Development of a generic thresholding algorithm for the delineation of 18FDG-PET-positive tissue: application to the comparison of three thresholding models
Bogunović et al. Automated segmentation of cerebral vasculature with aneurysms in 3DRA and TOF‐MRA using geodesic active regions: An evaluation study
Frangi et al. Quantitative analysis of vascular morphology from 3D MR angiograms: in vitro and in vivo results
CA2974349A1 (en) Systems and methods for estimating blood flow characteristics from vessel geometry and physiology
Jandt et al. Automatic generation of 3D coronary artery centerlines using rotational X-ray angiography
Ukwatta et al. Three‐dimensional ultrasound of carotid atherosclerosis: Semiautomated segmentation using a level set‐based method
WO2012174495A2 (en) Physics based image processing and evaluation process of perfusion images from radiology imaging
US8620040B2 (en) Method for determining a 2D contour of a vessel structure imaged in 3D image data
Egger et al. Pituitary adenoma volumetry with 3D Slicer
CA2723670A1 (en) A method for tracking 3d anatomical and pathological changes in tubular-shaped anatomical structures

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 16823055

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase in:

Ref document number: 2018532617

Country of ref document: JP

Kind code of ref document: A

NENP Non-entry into the national phase in:

Ref country code: DE

WWE Wipo information: entry into national phase

Ref document number: 2016823055

Country of ref document: EP