WO2008078265A2

WO2008078265A2 - Medical imaging system

Info

Publication number: WO2008078265A2
Application number: PCT/IB2007/055161
Authority: WO
Inventors: Pascal Allain; Olivier Gerard
Original assignee: Koninklijke Philips Electronics N.V.
Priority date: 2006-12-26
Filing date: 2007-12-17
Publication date: 2008-07-03
Also published as: WO2008078265A3; JP2010514486A; RU2009128709A; KR20090098839A; CN101568941A

Abstract

The invention relates to a medical imaging system. First, determination of the borders of at least one region of a feature of interest is performed. Second, a confidence level which is representative of the determination of the borders of at least one region of the feature of interest is computed. Third, an information representative of at least one region of the feature of interest where the associated borders have a confidence level which is lower than a predefined threshold is displayed.

Description

Medical imaging system

FIELD OF THE INVENTION

The present invention relates to a medical imaging system, and to a corresponding method. The invention finds, in particular, its application in the domain of ultrasound imaging.

BACKGROUND OF THE INVENTION

A known medical imaging system makes it possible to acquire a sequence of 3D images of a feature of interest of a body, such as the left ventricle of the heart, and to display it on a screen, to determine the borders of such feature of interest and display them on the screen, to detect visually on the screen errors of border determination, and to correct manually such errors. In order to detect visually such errors, the user chooses a region of interest on a 3D sequence and looks at all the 2D images composing a 3D image. Then, when the user sees that there is an error of border determination, he may correct it manually. One drawback of said imaging system is that the user of said system is loses if he wants to see and correct if necessary all the borders of the regions of the feature of interest of the 3D sequence because he has to extract all the 2D corresponding images, one 3D image being composed of about a hundred 2D images. He has to look at about 3000 images in a cardiac cycle, which is very tedious.

SUMMARY OF THE INVENTION

It is an object of embodiments of the invention to propose a system which permits a user to save time and help him for the correction of the borders of regions of a feature of interest.

To this end, the system comprises, in an embodiment, controlling means for controlling the following operations: - automatic determination of the borders of at least one region of a feature of interest in a sequence of images of a part of a body,

- computation of a confidence level which is representative of the determination of the borders of at least one region of the feature of interest, - display of an information representative of at least one region of the feature of interest where the associated borders have a confidence level which is lower than a predefined threshold.

Although the invention is well adapted for a sequence of images comprising a plurality of images, such as a sequence of images representative of a whole cardiac cycle, the invention may be used with a sequence of images comprising a single image. Therefore, the expression "sequence of images" should also be understood as meaning "at least one image".

The display of an information representative of a border with its confidence level permits the user to save time, as he will see automatically the borders where he has to focus on and which may need to be corrected.

According to a not limited embodiment, the displayed information is a map of the confidence levels associated respectively to the borders of a plurality of regions of the feature of interest. It permits the user to have a global view of the regions of a feature of interest and their associated confidence level.

According to a not limited embodiment, the controlling means permit the control of the display of a second information representative of at least one region of the feature of interest whose borders have been corrected. It permits the user to follow his own modifications.

According to a not limited embodiment the controlling means permit the control of the automatic display of a 2D slice view of one region of low confidence based on the information. It permits to guide the user in his corrections. The present invention also relates to a method for medical imaging which comprises the steps of :

- determining automatically the borders of at least one region of a feature of interest in a sequence of images of a part of a body, - computing a confidence level which is representative of the determination of the borders of at least one region of the feature of interest,

- displaying an information representative of at least one region of the feature of interest where the associated borders have a confidence level which is lower than a predefined threshold.

The present invention finally relates to a computer program product comprising program instructions for implementing said method when said program is executed by a processor.

These and other aspects of the invention will be apparent from and will be elucidated with reference to the embodiments described hereinafter.

BRIEF DESCRIPTION OF THE DRAWINGS

The present invention will now be described in more detail, by way of not limited examples, with reference to the accompanying drawings, wherein:

- Fig.l is a schematic diagram of a system according to an embodiment of the invention which cooperates with a probe ;

- Fig.2 is a schematic drawing of a feature of interest such as the left ventricle of a heart, from which a sequence of images is acquired via a system according to an embodiment of the invention ;

- Fig.3 is a first view of a segmentation of a feature of interest such as the left ventricle of a heart, which may be used by the system according to an embodiment of the invention;

- Fig.4 is a second view of a segmentation of a feature of interest such as the left ventricle of a heart, which may be used by the system according to an embodiment of the invention ;

- Fig.5 is a display of a feature of interest such as the left ventricle of a heart with regions having borders with low confidence levels, performed by the system according to an embodiment of the invention ;

- Fig.6 is another display of the borders of some regions of a feature of interest such as the left ventricle of a heart, performed by the system according to an embodiment of the invention ;

- Fig.7 is a first display, of a map of confidence levels associated to different regions of a feature of interest such as the left ventricle of a heart, performed by the system according to an embodiment of the invention ;

- Fig.8 is a second display, of a map of confidence levels associated to different regions of a feature of interest such as the left ventricle of a heart, performed by the system according to an embodiment of the invention ; and

- Fig.9 represents a diagram of a method for medical imaging according to an embodiment of the invention.

DETAILED DESCRIPTION OF THE INVENTION

A system SYS in accordance with an embodiment of the invention is described in Fig.l.

It cooperates with a transducer's array TAR and its associated electronics, the whole forming a probe PRB. The system SYS comprises a controller CTRL for controlling the following operations :

- acquisition of a sequence of images SQ of a part of a body;

- automatic determination of the borders B of at least one region RI of a feature of interest FI in a sequence of images SQ of a part of a body, - computation of a confidence level CL associated to the borders B of at least one region RI of the feature of interest FI which is representative of the determination of said borders,

- display of an information IN representative of at least one region RI of the feature of interest FI where the associated borders have a confidence level CL which is lower than a predefined threshold TH.

The system SYS further optionally comprises a screen SCR for displaying the sequences SQ of images acquired, such as a LCD screen, and a user interface M USER.

The system SYS may comprise a memory MEM in order to save the images I acquired.

In an embodiment, the controller CTRL is further arranged to control the display of the sequence of images SQ, and the automatic display of a 2D slice view of one region RI of low confidence based on the information IN;

It is to be noted that the controller CTRL comprises a microprocessor that can be pre-programmed by means of instructions or that can be programmed by a user of the system SYS, for instance via the interface M USER.

It is to be noted that an image I is a 3D grey level image that may be split up in 2D slices which is usually called a MPR "Multiplanar Reconstruction" view.

Such a system SYS may be used in ultrasound, in particular, where organ measurements need to be performed, such as the left ventricle LV of a heart.

One reminds that a heart is composed of a left and a right ventricles LV and RV, an aorta AO, and a left and right atrium LA and RA as shown in Fig.2, and that the arterial blood goes from the left ventricle LV to the aorta AO while the right ventricle RV exits the venous blood received from the right atrium RA to the pulmonary artery. As the way the left ventricle LV is working is indicative of the health of the heart, one focus more particularly on said left ventricle LV when using the ultrasound imaging system SYS.

Referring now to Fig.3, the inner wall of left ventricle LV of the heart may be segmented in seventeen segments SG as defined in the standard "Standardized Myocardial Segmentation and Nomenclature for Tomographic Imaging of the Heart" by the Cardiac Imaging Committee of the Council on Clinical Cardiology of the American Heart Association. Thus, Fig.3 is a display on a circumferential polar plot of such a segmentation called "bulls eye" and Fig.4 is a 3D view of such segmentation. The seventeen segments are named by the standard. For example, the segment number 17 is the apex, and the segments number 1 and 7 which identify the locations of the anterior wall at the base and mid-cavity are named basal anterior and mid-anterior. Such a segmentation may be used by the ultrasound imaging system as described below. In order to acquire images of the left ventricle LV, the ultrasonic probe PRB is applied on the body of a patient, at the apex near the heart in a not limited embodiment, and the imaging system SYS performs the operations described hereinafter.

1) Acquisition of a sequence of three-dimensional images SQ.

The user of the system SYS moves the probe PRB on the part of the body which is of interest, here the heart, and more particularly the left ventricle LV. A sequence of grey level three-dimensional images is acquired. The sequence of images SQ is displayed on the screen SCR. It is to be noted that a sequence SQ of three- dimensional images is performed at about 20Hz and a sequence SQ is composed of about 20 three-dimensional images. It is to be noted that in order to view the entire volume of the left ventricle LV, the images acquisition is performed during four cardiac cycles, wherein one fourth of the left ventricle LV is acquired at each cardiac cycle. This 3D acquisition permits to obtain some volumes.

It should be noted that acquisition of the sequence of images SQ is not necessary to the invention. In the embodiment of Fig.1, the controller CTRL also controls this acquisition, however this acquisition may be controlled by a separate system. For instance, the acquisition may be performed by an acquisition system and the sequence of images sent, for instance by means of a wireless connection, to a system comprising means for controlling automatic determination of borders of at least one region RI of a feature of interest FI in the sequence of images SQ, for ccomputing a confidence level associated to the borders of at least one region of the feature of interest which is representative of the determination of said borders, and for displaying an information representative of at least one region of the feature of interest where the associated borders have a confidence level which is lower than a predefined threshold.

2) Automatic determination of the borders B of at least one region RI of a feature of interest FI in a sequence of images SQ of a part of a body. This operation permits to determine the position of the borders B of different regions RI of the left ventricle LV in each image I of the sequence SQ, usually by a method called segmentation of the left ventricle LV. Such a segmentation is well described, for instance, in the document "Efficient Model-Based Quantification of Left Ventricular Function in 3- D Echochardiography - O. Gerard, A. Collet-Billon, J-M. Rouet, M. Jacob, M. Fradkin and C. Allouche- IEEE transactions on medical imaging, Vol.21, N°9, September 2002".

It is to be noted that a region RI is composed of at least one voxel and may be composed of a plurality of voxels. The automatic determination may for instance be based on a measurement of the movement of the left ventricle LV. In this case, it may be based for instance on features characterization in the image such as edge based on gradient, density level, texture tracking.

It is to be noted that the type of features characterization used is chosen according to an anatomic model function of the regions of the left ventricle LV seen on the sequence SQ of images.

Of course any method for automatic determination of the borders B of a region RI of a feature of interest FI in an image I may be used.

When the borders B have been estimated, one may determine the evolution of the surface of the left ventricle LV by acquiring a parametric image IP based for instance on the velocity information of different regions RI. A color is associated to the velocity information of the regions RI of the left ventricle LV in order to be displayed as a parametric image IP on the screen SCR. For example, a red color can be used when the regions contract whereas a blue color can be used when the regions relax. When the left ventricle LV is working correctly, the whole left ventricle LV should be displayed in red when it contracts, and in blue when it relaxes. If it is not the case, the left ventricle LV is displayed for some parts in red and for other parts in blue. The colors are not uniform. It means that some regions RI contract or relax later than other regions RI because their speed peaks are different. Hence, the evolution of the surface shows if there is any asynchronism in the left ventricle LV. Of course, other information than the velocity may be used for the parametric image IP such as, for instance, deformation, displacement or acceleration information. The acquisition of a parametric image is well described, for instance, in the document " Strain and strain rate parametric imaging. A new method for post processing to 3-/4- dimensional images from three standard apical planes. Preliminary data on feasibility, artefact and regional dyssynergy visualisation" from A. Stoylen, CB. Inggul, H. Torp - Cardiovascular Ultrasound 2003, 1 :11 doi:10.1186/1476-7120-1-11 - Department of Circulation and Medical Imaging, Faculty of Medicine, Norwegian University of Science and Technology, Trondheim. Norway.

3) Computation of a confidence level CL which is representative of the determination of the borders B of at least one region RI of the feature of interest FI.

Because the image quality of an ultrasound acquisition depends of several factors such as for instance the patient echogenicity, the line density, the limited field of view, ...etc., a confidence level CL is associated to the determination of the borders B of the different regions RI of the left ventricle LV. This confidence level depends on local estimation of the feature characterization used to determine the borders B, as described before, such as threshold on grey level for the density level, gradient level for the edge based on gradient, and global/local statistic for the texture tracking.

The computation of a confidence level is performed for each image I of the acquired sequence SQ of images.

Such a confidence level associated to a measurement is well described, for instance, in the document "A confidence measure based moving object extraction system built for compressed domain - R. Wang, H.J. Zhang, Y. Q. Zhang - IEEE International Symposium on Circuits and Systems, May 28-31, 2000, Geneva, Switzerland - ISCAS2000".

4) Display of an information IN representative of at least one region RI of the feature of interest FI where the associated borders B have a confidence level CL which is lower than a predefined threshold TH. The threshold TH may be for instance of 60%. Of course, any other values of threshold may be defined.

The display is performed for each image I of the sequence SQ of images acquired. For instance, the information IN is the 3D image of the left ventricle LV with colors on regions RI as illustrated in Fig.5, each color being associated to a value of confidence level CL. The regions RIl and RI2 have a low confidence level CL. Of course, another color may be used for the other regions which have an associated confidence level greater than the threshold TH. Hence, a map of confidence levels CL associated respectively to the borders B of a plurality of regions RI of the feature of interest FI which have a low or high confidence level may be displayed.

Hence, with the information IN displayed, the user may either validate or correct manually the borders B of the regions RI which have a low confidence level CL.

In order for the user to manually modify the borders B, the user interface M USER comprises manual editing tools.

5) Automatic display of a 2D slice view of one region of low confidence based on the information IN.

In order to help the user to correct manually the borders B of the regions of low confidence level, the system SYS, based on the confidence map MP, may slice automatically the 3D images at positions of low confidence. For instance a MPR "Multiplanar Reconstruction" 2D slice view through low confidence part of the image will be automatically displayed to the user, also with color indicating regions (for instance the regions RIl, RI2) to review, as illustrated in Fig.6. More than one MPR view may be displayed. Three perpendicular planes slicing through a region of interest RI may be displayed for instance.

If the region RI shown by the system SYS is of interest for the user, he may correct its borders. The system SYS will then automatically move the region of interest RI toward the region with the second lowest confidence, etc. If not, the user indicates to the system that he wants to move to the next region with low confidence and so on. In another embodiment, the user may himself choose the MPR views with the use of an orthoviewer P as illustrated in Fig.5. The user may use the orthoviewer P and move it until he sees the 2D slices of regions of interest, such as the 2D slices for the regions RIl and RI2 for instance.

Automatic positioning on regions RI of low confidence permits to guide the user in his corrections. The user thus saves time.

6) Update the information IN and display.

Whenever the user corrects a border B, the information may be updated in order to show the corrections of the user and to apply a confidence level CL which is reliable on these corrections. Another color may be used in order to indicate to the user which parts are effectively covered by his corrections.

Hence, the originated confidence information IN called intrinsic may always be displayed, and an updated confidence information INu called extrinsic may be displayed in parallel. Hence, this extrinsic confidence information INu will store all the regions and extent of the user interaction, along its history. Hence, the extrinsic confidence information INu will allow displaying to the user where he has made modifications. With this information, a change in the density of the intervention (e.g. lots of interventions in one part of the image and none in other parts) will be a warning sign to explore the non-modified regions.

It is to be noted that after a correction is performed by the user, one may perform the step 3) of computation of local estimation of a confidence level again. It permits to redo an estimation taking into account the corrections of the user.

7) Merge of a parametric image IP with the intrinsic IN and extrinsic INu confidence information. As a final result, the user may have a final parametric image IF which is a merge between the parametric image IP, which shows for instance the velocity information as described before, and the intrinsic IN and extrinsic INu confidence information, as illustrated in Fig.7 and Fig.8. Fig.7 is a representation in 2D of the final parametric image IF. Regions at low confidence level have been erased from the parametric image IP. Therefore, the user may see which segments SG of the parametric representation is covered by a region or a plurality of regions RI with a low confidence level for instance.

Fig.8 is a representation in 3D of the final parametric image IF where regions at low confidence level have been erased from the parametric image IP.

Hence, the resulting parametric image IF permits to identify and to show the region where final measurements (wall dyssynchrony) are really reliable.

It is to be noted that the intrinsic and extrinsic confidence information are here a combination of all the intrinsic and extrinsic information of each image I.

In order for the user to obtain the final parametric image IF, the user interface M USER comprises adequate means such as a button for example.

Fig.9 illustrates the method for medical imaging according to an embodiment of the invention where the different operations controlled by the system SYS are shown.

It should be noted that the above-mentioned embodiments illustrate rather than limit the invention, and that those skilled in the art will be capable of designing many alternative embodiments without departing from the scope of the invention as defined by the appended claims. The examples used have been described for an echocardiographic sequence of images, but of course they may be extended to any images coming for other imaging modalities.

In the claims, any reference signs placed in parentheses shall not be construed as limiting the claims. The word "comprising" and "comprises", and the like, does not exclude the presence of elements or steps other than those listed in any claim or the specification as a whole. The singular reference of an element does not exclude the plural reference of such elements and vice-versa.

The invention may be implemented by means of hardware comprising several distinct elements, and by means of a suitably programmed computer. In a device claim enumerating several means, several of these means may be embodied by one and the same item of hardware. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage.

Claims

1. A medical imaging system, comprising controlling means (CTRL) for controlling the following operations :

- automatic determination of the borders (B) of at least one region (RI) of a feature of interest (FI) in a sequence of images (SQ) of a part of a body, - computation of a confidence level (CL) which is representative of the determination of the borders (B) of at least one region (RI) of the feature of interest

(FI),

- display of an information (IN) representative of at least one region (RI) of the feature of interest (FI) where the associated borders (B) have a confidence level (CL) which is lower than a predefined threshold (TH).

2. A system as claimed in claim 1, wherein the displayed information (IN) is a map of the confidence levels (CL) associated respectively to the borders (B) of a plurality of regions of the feature of interest (FI).

3. A system as claimed in claim 1, wherein the controlling means (CTRL) permits the control of the display of a second information (INu) representative of at least one region (RI) of the feature of interest (FI) whose borders (B) have been corrected.

4. A system as claimed in claim 1, wherein the controlling means (CTRL) permits the control of the automatic display of a 2D slice view of one region (RI) of low confidence based on the information (IN).

5. A method for medical imaging, comprising the steps of : - Determining automatically the borders (B) of at least one region (RI) of a feature of interest (FI) in a sequence of images (SQ) of a part of a body,

- Computing a confidence level (CL) which is representative of the determination of the borders (B) of at least one region (RI) of the feature of interest (FI), - Displaying an information (I) representative of at least one region (RI) of the feature of interest (FI) where the associated borders (B) have a confidence level (CL) which is lower than a predefined threshold (TH).

6. A computer program product comprising program instructions for implementing, when said program is executed by a processor, the method as claimed in the preceding claim.