WO2005070298A1

WO2005070298A1 - Ultrasound imaging system for calculating a compression ratio

Info

Publication number: WO2005070298A1
Application number: PCT/IB2005/000164
Authority: WO
Inventors: Nicolas Villain
Original assignee: Koninklijke Philips Electronics N.V.
Priority date: 2004-01-23
Filing date: 2005-01-19
Publication date: 2005-08-04
Also published as: CN1909838A; EP1711108A1; JP2007518508A

Abstract

The present invention relates to an ultrasound imaging system for calculating a compression ratio of an object of interest located in an area of the patient body scanned by an ultrasonic probe. The system in accordance with the invention comprises compression means (1) for compressing said area of the patient body, acquisition means (3) for acquiring a first echographic image and a second echographic image of the object of interest before and during compression respectively, measuring means (11) for measuring a first height of said object of interest in said first echographic image and a second height h2 of said object of interest in said second echographic image and computation means for deriving said compression ratio from said first and second heights. Said compression ratio is for instance used for identifying fat lobules in the breast.

Description

ULTRASOUND IMAGING SYSTEM FOR CALCULATING A COMPRESSION RATIO

FIELD OF THE INVENTION The present invention relates to an ultrasound imaging system for calculating a compression ratio of an object of interest. The invention further relates to a method for use in such a system. The invention finally relates to a computer program implementing such a method.

BACKGROUND OF THE INVENTION Several methods are commonly used for the early detection of breast cancer. Historically, the first one is palpation by the patient herself on a regular basis in order to detect any change in the structure of her breasts (commonly known as self breast examination or SBE) or by a doctor who looks for abnormal nodules in the breasts. More recently, several countries, including the U.S.A., have introduced systematic screening programs using X-ray mammography. X-ray mammography is a very sensitive technique, except in dense breasts, which can detect masses as small as 5 mm, i. e. cancers at a very early stage. However, X-ray mammography detects a lot of false alarms which can be normal tissues with ambiguous morphology such as fat lobules or benign lesions such as cysts or fibroadenoma. Ultrasound is commonly used as an adjunct to mammography to reject those false alarms without resorting to biopsy. Figs. 1A, IB and 1C respectively show examples of echographic images comprising a fibroadenoma FA, a fat lobule FL and a cancer C. In a course from the American Institute of Ultrasound in Medicine (AIUM), available for sale on CD-ROMs on the Internet site http://www.aium.orR/products/store/ productDetail.asp?id=02CTBUCD&cat=B&words=&p g=l , the state of the art in breast sonography is presented by expert sonographers and radiologists. In one of those presentations, a sonographer explains that normal structures like fat lobules can be mistaken for lesions. Since fat is very common in the breast, it is of paramount importance to have a fast and reliable method to tell fat lobules apart. The sonographer describes an approach based on a principle applied in palpation, i. e. the difference in elastic properties of the tissues. As a matter of fact, fat lobules are made of normal and deformable tissue and, therefore, present a higher compression ratio than abnormal tissues like fibroadenoma and even more so, cancers. The computation of such a compression ratio is performed manually by the sonographer. In practice, the method of the sonographer comprises the steps of: acquiring a first echographic image of the breast in which the abnormal nodule is visible, - in this first echographic image selecting a superior edge point belonging to a superior edge of the abnormal nodule and an inferior edge point belonging to an inferior edge of the abnormal nodule, calculating a first height of the abnormal nodule as a distance between the superior and inferior edge points, - applying a compression to the breast in an area including the abnormal nodule, acquiring a second echographic image of the breast during compression, measuring a second height of the abnormal nodule in the second echographic image as described above, - calculating a compression ratio as a ratio between the second and first heights. An abnormal nodule with a compression ratio greater than predetermined threshold, typically 30 %, is considered as a fat lobule and rejected. A drawback of such a manual calculation of the compression ratio is that it is tedious and slow for the sonographer.

SUMMARY OF THE INVENTION It is an object of the invention to provide a solution for automatically calculating the compression ratio of an object of interest using an ultrasound imaging technique, which is more efficient. This is achieved by an ultrasound imaging system comprising:

- compression means for compressing an area of a patient body,

- acquisition means for acquiring a first echographic image before compression and a second echographic image during compression of an object of interest located in said area of the patient body, using an ultrasonic probe,

- measuring means for measuring a first height of said object of interest in said first echographic image and a second height of said object of interest in said second echographic image,

- computation means for deriving a compression ratio from said first and second heights. With the invention, the user has no more to manually select a first couple of superior and inferior edge points in the first echographic image, to measure a distance between these edge points to get the first height, to repeat the operation in the second echographic image in order to get the second height and to calculate the compression ratio from the first and second heights. The user only has to apply a compression to the area of interest of the patient body and to indicate, for instance by pushing a button, when the acquisitions of the first and the second echographic images have to be performed. From the first and second echographic images, a compression ratio is immediately calculated and displayed by the ultrasound imaging system in accordance with the invention.

Therefore, the contribution of the sonographer is strongly reduced and the operation is quickened. Moreover, the procedure of measuring the first and second heights, based on image processing techniques, is objective and performed in the same way for any couple of first and second echographic images. On the contrary in the prior art, the manual selection of the first and second heights is subject to the interpretation of the sonographer. Therefore, with the invention, the measurement of heights is made more reproducible. hi a first embodiment of the invention, a local approach is proposed. The measuring means in accordance with the invention comprise detection sub-means for detecting a superior edge point and an inferior edge point on a profile in accordance with a direction of the ultrasound beam. The height is measured as an Euclidean distance between the superior and the inferior edge point. Such a procedure is repeated in the second echographic image. Such a local approach imitates in an automatic way the manual procedure described in the prior art and involves image processing techniques. An advantage of this first embodiment of the invention is to be simple. In a first alternative of the first embodiment of the invention, a plurality of profiles are considered in the first and second echographic images and therefore a plurality of heights are measured in both echographic images. The system in accordance with the invention further comprises means for averaging said plurality of measured heights. An advantage is to get a more robust measurement of the height and therefore a more accurate calculation of the compression ratio of the object of interest. In a second alternative of the first embodiment of the invention, the system in accordance with the invention comprises fitting means for fitting a surface curve to said plurality of edge points. The surface curve model advantageously depends on a priori knowledge about the object of interest. For fat lobules, for instance, a model of ellipse is used. A first advantage of this second alternative is that aberrant detected edge points can be rejected. A second advantage is that additional geometrical measurements can be performed on the object of interest like a surface measurement, in order to make the description of the deformation more precise. In a second embodiment of the invention, a region-based approach is proposed. The measuring means in accordance with the invention comprise segmentation means for segmenting an object of interest having a predetermined shape in an echographic image. For fat lobules, the predetermined shape is for instance an ellipse. An advantage of the second embodiment of the invention is that it facilitates a matching of the first shape and the second shape corresponding to the object of interest in the first and respectively in the second echographic images, when a plurality of objects of interest is present.

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 example, with reference to the accompanying drawings, wherein: - Figs. 1A to 1C show examples of echographic images comprising a fibroadenoma, a fat lobule and a cancer,

- Fig. 2 is a schematical drawing of an ultrasound imaging system in accordance with the invention,

- Figs. 3A to 3D a schematic representation of how the first and the second heights of the object of interest are measured by the system in accordance with a first embodiment of the invention,

- Figs 4A and 4B a schematic representation of several profiles for measuring and averaging the first and second heights of the object of interest in accordance with a first embodiment of the invention, - Figs. 5A and 5B a schematic representation of an approximation of the contour of the object of interest by an elliptic curve in accordance with the first embodiment of the invention, - Figs. 6A and 6B a schematic representation of segmentations of objects of interest with a predetermined shape model in the first and second echographic images in accordance with a second embodiment of the invention,

- Fig. 7 is a schematic representation of successive compressions of the object of interest in accordance with a third embodiment of the invention,

- Fig. 8 is a schematic representation of a method in accordance with the invention.

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an ultrasound imaging system for measuring a compression ratio of an object of interest. Such a system is particularly adapted to medical image processing, in particular to diagnosis of breast cancer in an echographic image of the breast. Fig. 2 gives an overview of an ultrasound imaging system in accordance with the invention. An ultrasonic probe 1 includes a multielement transducer array 2, which transmits waves of ultrasonic energy into the body of a patient and receives ultrasonic echoes returning from structures in the body. The structures in the body are, for instance, objects of interest of the breast. The ultrasonic probe 1 is connected to acquisition means 3 comprising a transmitter/receiver 4 which alternately pulses individual elements of the transducer array 2 to shape and steer an ultrasonic beam and receives, amplifies, digitizes echo signals received by the transducer elements following each pulse transmission. The transmitter/receiver 4 is coupled to a beamformer 5 which controls the times of activation of the elements of the transducer array 2 by the transmitter/receiver 3. The beamformer 5 also receives the digitized echo signals produced by the transmitter/receiver 4 during echo reception and appropriately delays and sums them to form coherent echo signals. The beamformer 5 is connected to an image processor 6, which processes the amplitude information of the echo signals on a spatial basis for the formation of an echographic image Ii of the tissue in the area of the patient being scanned. The intensity value of the echographic image are applied to a scan converter and display processor 7 which spatially arranges the intensity values in the desired image format. The echographic image Ii is stored in an image sequence memory 7 and displayed on a display screen 8. The system in accordance with the invention further comprises means for applying a compression on a the patient body in the area which has just been scanned. In the present application, the ultrasonic probe 1 is used for applying a pression on the breast and therefore the pression is applied in the scanning direction of the ultrasonic beam. It should be noted however that a mechanical device could have been used as well. In this lattest case, the direction of compression can be different from the one of the ultrasonic beam. The system in accordance with the invention further comprises user interaction means 10, which allow a user interacting with the system. In particular, the user can click on a button in order to actuate the acquisition means 3 and get a new echographic image. In the present application, the user has to order the acquisition of a second echographic image I₂ during compression of the same area of the patient body. The system in accordance with the invention further comprises measuring means 11 for measuring a height of an object of interest in an echographic image. Once a first echographic image Ii before compression and a second echographic image I₂ after compression of a same area of the breast have been acquired, the measuring means 11 are able to measure a first height hi and a second height h₂ of the object of interest in the first and respectively the second echographic images (Ii, I₂). Said first and second heights hi and h₂ are used by computation means 12 for calculating a compression ratio of the object of interest. Fig. 3 A is a schematical drawing of the ultrasonic probe 1 which has been placed in contact with an area of the breast skin 20. At this position, the ultrasonic beam produced by the ultrasonic probe 1 scans a scanning area 21 comprising at least an object of interest 22, which is, for instance a nodule. Fig. 3B is a schematical drawing of the ultrasonic probe 1 during compression of the same area of the breast. As already mentionned above, the compression is advantageously applied via the ultrasonic probe 1 to the breast area of interest. Therefore, in this case, the direction of compression corresponds to the direction of the ultrasonic beam. In a first embodiment of the invention shown in Figs. 3 A to 3D, the measuring means 11 comprise detection sub-means for detecting a first couple of edge points (E_{l ls} E₁ ) in the first echographic image Ii and a second couple of edge points (E₂₁, E₂₂) in the second echographic image I₂. It should be noted that the first edge point (En, E₂₁) corresponds to the edge point reached at a first time by the transmitted ultrasonic beam and the second edge point (E₁₂, E ₂) the edge point reached at a second time by the ultrasonic beam. Under the effect of compression, the object of interest 22 is more or less deformed. The degree of deformation mainly depends on the strength of compression and of the constitution of the object of interest 22. As shown in Fig. 3B, a second height h₂ is measured during compression in the second echographic image I₂ along a profile corresponding, for instance, to the location of the profile T?_\ with respect to a referential (O, x, y) of the ultrasound probe 1. Figs. 3C and 3D show the intensity curves ICi and Id corresponding to the profiles ?ι and P'j respectively. Advantageously the measuring means 11 further comprise smoothing sub-means for smoothing the intensity curves IC_ls Id before detection of the edge points (En, E₁₂, E ₁, E ₂). The smoothing sub-means for instance involve a low-pass filter using a Gaussian kernel. Smoothed intensity curves SICi, Sid are obtained. An advantage is these smoothed intensity curves are less noisy. The detection sub-means for instance involve a technique of the maximum of the Gradient, which is well known to those skilled in the art. In an alternative, the local edge point detection provided by the detection sub-means can be advantageously replaced by fitting a predetermined shape to the intensity curves IQ, Id, for instance a negative Gaussian kernel . Such a fitting is achieved by adjusting the characteristics of the Gaussien kernel, which are the mean and the standard deviation. Once the amplitude of the profile is normalized, the position of best fitted Gaussian kernel indicates the position of the object of interest and the edge points are easily derived from it. An advantage of such an alternative is to be less local. In particular, this solution takes into account the fact that the intensity curve of the object of interest first comprises a down step followed by an up step. The detection may be made more robust for echographic images comprising a plurality of objects of interests, which are likely to appear in a same intensity curve. The height (h_ls h₂) is calculated by subtracting a location (y₁₂, y₂₂) of the inferior edge point E₁₂, E₂ from a location (yn, y₂₁) of the superior edge point En, E₂₁ on the profile P_l5 P'i, as follows:

h₂ = (y₂₂-y2i) where yn, yi₂, yn, yn are the y-coordinates of the edge points En, E₁₂, E₂₁ and E₂₂ in a referential (O, X, Y) of the ultrasonic probe.

The computation means 12 are intended to calculate a compression ratio CR from the measured first and second heights h_ls h₂ in the following way:

CR =^- h The compression ratio CR is for instance expressed in purcentage and displayed with the first and/or second echographic images L;, 1₂. As already mentioned above, such a compression ratio CR can advantageously be compared with normative figures related to a priori knowledge and be used for identifying an object of interest. In the particular case of breast echographic images, the goal is to distinguish between breast cancers and other types of object of interests, namely fat lobules and fibroademas, which may have a very similar appearance in an echographic image of the breast. The knowledge of the compression ratio allows rejecting fat lobules with a high level of confidence. As a matter of fact, fat lobules are made of fat tissue, that is normal and deformable tissue. On the contrary, cancers are lesions made of hard tissue, also called solid masses, which cannot be deformed easily. Consequently, it has been empirically established that nodules having a compression ratio greater than a thresholding value of 30 % were fat lobules and should be rejected. In an alternative to the first embodiment of the invention, the measuring means 11 are intended to measure the first and second heights on a plurality of profiles instead of one profile. Referring to Figs. 4A to 4D, three profiles P_lr P₂, P and three profiles P'_l5 P'₂, P'₃ are considered in the first and second echographic images Ii_. and I₂ respectively. The measurements of heights hn, h₂₁, h₃₁, h₁₂, h₂₂ and h₃₂ are repeated by the measuring means 11 for each profile P_1? P₂, P₃ and P'ι, P'₂, P'₃. The measuring means 11 further comprise averaging sub-means for averaging the first heights hn, h₂₁, lι₃₁ measured in the first echographic image I_\ and the second heights h₁₂, I ₂₂, h₃₂ measured in the second echographic image I₂ for the object of interest. Average heights h_la and h_2a are obtained, which further contribute to the calculation of the compression ratio CR. A first advantage of this alternative is that it is more robust to local variations of deformation. As shown in Fig. 4B, an object of interest may undergo locally different deformations when a compression is applied to the patient body. A second advantage of this alternative is to cope with a lateral displacement that the object of interest may have under the effect of compression. Therefore, the profile

before compression may not intersect the object of interest at the same location as the profile P] during compression, but the plurality of profiles may cover the whole object of interest. If the object of interest completely leaves the scanning area of the ultrasonic beam, it is supposed that the user will repeat the acquisition procedure in order to get a couple of echographic images (I_1} I₂) in which the object of interest is included.

In another alternative of the first embodiment of the invention, the measuring means 11 further comprise fitting means for fitting a predetermined shape contour curve to the detected edge points in the first and second echographic images Ii, I₂. The predetermined shape contour curve is chosen in accordance with a priori knowledge on the searched objects of interests. Referring to Figs. 5 A and 5B, for breast cancers, an ellipse curve is chosen. A first ellipse curve 41 is fitted to the contour of the object of interest in the first echographic image ϊ_\ and a second ellipse curve 42 is fitted to the contour of the corresponding object of interest in the second echographic image I₂. A first advantage is that the compression ratio is more accurately estimated even if the object of interest has moved under compression. For instance, the compression ratio can be calculated as the ratio of the maximum thicknesses of the fitted ellipses, which can be at different locations under the probe in the echographic images Ii and I₂. A second advantage is that an estimation of a first surface S_\ and a second surface S₂ of the object of interest in the first and second echographic images ϊ_\ and I₂ respectively can be deduced and a deformation model can be applied. Therefore, such surface estimations not only allow to compute the compression ratio but also to derive other deformation measurements such as the elasticity of the object of interest. a second embodiment of the invention, the measuring means 11 comprise segmentation means for segmenting regions with a predetermined shape model in an echographic image. In the particular application of breast echographic imaging, objects of interests are advantageously searched as dark elliptic shapes on a brighter background as shown in Figs 6 A and 6B. For instance, a segmentation technique based on the generalized Hough transform would detect all possible ellipses in the images. Another possibility would be to apply a region growing technique in every local minimum of intensity of the image that would represent dark regions and select those with elliptic shapes. A third possibility would be to design an active contour submitted to internal forces favoring elliptical shapes and external forces attracting it toward high gradient boundaries and make it evolve to match the boundaries of all dark regions. With any of those methods or a similar one, a plurality of ellipses may be detected in the echographic images I\ and I₂. Advantageously, the measuring means 11 further comprise association means for associating a first segmented shape Shi, Sh₂, Sh₃ or Sr^ of the first echographic image Ii with a second segmented shape Sh , Sh'₂, Sh'₃ or Sh'₄ of the second echographic image. Said association means are mainly based on a set of conditions expressed as localisation and geometrical parameters derived from a priori clinical knowledge. This set of conditions may include and is not limited to the following assumptions: - a fat lobule is wider than tall, so the main dimension of the first and second segmented shapes (Shi, Sh'i₎, (Sh₂, Sh'₂₎, (Sh₃, Sh'₃)have to be horizontal,

- the second segmented shapes Sh'i, Sh'₂, Sh'₃ represent the object of interest after compression, so they must be flatter than the first segmented shapes Shi, Sh₂, Sh₃, - the object of interest should not move too much during compression so the positions of the first and second segmented shapes (Sh_1; Sh'i), (Sh_2; Sh'₂₎, (Sh₃₎ Sh'₃₎ should be close to each others. A first advantage of approximating the shape of the object of interest with a predetermined shape model is that the geometrical parameters of the predetermined shape model are well-known. Therefore, they can be used for choosing the most appropriate first and second heights (hi, h'i) in the first and second echographic images (I_l5 1₂). For instance, the first height hi is the maximum thickness of the first segmented ellipse Shi and the second height h 'i is the maximum thickness of the second segmented ellipse Sh'i. Therefore the estimation of the compression ratio is more robust to any displacement of the object of interest due to compression. A second advantage of the second embodiment of the invention is to that it is more robust with respect to a number of objects of interest in the echographic images. As a matter of fact, if several objects of interest are close to each other in the first and second echographic images, the detection means in accordance with the first embodiment of the invention are likely to malce some errors, because a profile will contain edges belonging to several objects of interest, whereas the detection means in accordance with the second embodiment of the invention are capable of distinguishing between the different objects of interest. A third advantage of the second embodiment of the invention is that it allows to perform other measurements of the compression of the object, such as elasticity.

In a third embodiment of the invention, the compression means are intended to apply successive pressions to a region of interest of the breast. Referring to Fig. 7, a new second echographic image I₃, is acquired for each new pression and a measurement of second heights h , t _t is performed in the new second echographic images. The system in accordance with the invention further comprises compression averaging means for averaging the second heights h₂, h and h An average second height h2_34a is obtained, which is used for calculating the compression ratio. An advantage is that successive pressions, which are manually applied by the user to the region of interest, are averaged. This compensates for the variation of strength applied to a manual compression from one time to another. Therefore, comparisons of values of compression ratios are made more reliable. It should be noted that the time needed for computing a compression ratio from a succession of compression is increased, but this is made acceptable by the time gain brought by the system in accordance with the invention.

Referring to Fig. 8, the present invention also concerns an ultrasound imaging method of measuring a compression ratio of an object of interest, comprising the steps of:

- acquiring 70 a first echographic image II of an object of interest 71 located in an area of a patient body scanned by an ultrasound probe 1, - compressing 72 said object of interest,

- acquiring 73 a second echographic image 12 during compression of said object of interest,

- measuring a first height hi of said object of interest in said first echographic image Ii and a second height h₂ of said object of interest in said second echographic image I₂,

- calculating 74 a compression ratio CR of said object of interest from said first and second heights (hi, h₂).

The drawings and their description hereinbefore illustrate rather than limit the invention. It will be evident that there are numerous alternatives, which fall within the scope of the appended claims. In this respect the following closing remarks are made: there are numerous ways of implementing functions by means of items of hardware or software, or both. In this respect, the drawings are very diagrammatic, each representing only one possible embodiment of the invention. Thus, although a drawing shows different functions as different blocks, this by no means excludes that a single item of hardware or software carries out several functions, nor does it exclude that a single function is carried out by an assembly of items of hardware or software, or both.

Any reference sign in a claim should not be construed as limiting the claim. Use of the verb "to comprise" and its conjugations does not exclude the presence of elements or steps other than those stated in a claim. Use of the article "a" or "an" preceding an element or step does not exclude the presence of a plurality of such elements or steps.

Claims

CLAIMS 1. An ultrasound imaging system comprising: compression means for compressing an area of a patient body, acquisition means for acquiring a first echographic image before compression and a second echographic image during compression of an object of interest located in said area of the patient body, using an ultrasonic probe, - measuring means for measuring a first height of said object of interest in said first echographic image and a second height of said object of interest in said second echographic image, - computation means for deriving a compression ratio from said first and second heights.

2. An ultrasound imaging system as claimed in claim 1, wherein said measuring means comprise detection sub-means for detecting a superior edge point and an inferior edge point in said first and second echographic images, on at least one profile line oriented in accordance with a scanning direction of said ultrasonic probe and said first and second heights are calculated as a distance between said inferior and said superior edge points in said first and second echographic images, respectively.

3. An ultrasound imaging system as claimed in claim 2, wherein said detection means are intended to detect said superior and inferior edge points on a plurality of profile lines in order to provide a plurality of first and second heights and said measuring means comprise averaging sub-means for calculating an average first height and an average second height from said plurality of measured first and second heights.

4. An ultrasound imaging system as claimed in claim 3, wherein said measuring means comprise fitting means for fitting a predetermined shape curve contour to said plurality of detected first and second edge points.

5. An ultrasound imaging system as claimed in claim 1, wherein said measuring means comprise segmentation sub-means for segmenting a first predetermined shape surface of said object of interest in said first echographic image and a second predetermined shape surface of said object of interest in said second echographic image.

6. An ultrasound imaging system as claimed in claim 5, wherein said measuring means comprise association sub-means for associating said first predetermined shape surface with said second first predetermined shape surface.

7. An ultrasound imaging system as claimed in one of claims 4 and 5, wherein said computation means are intended to calculate a surface-based compression ratio from said first and second predetermined shape surfaces.

8. An ultrasound imaging system as claimed in claim 1, wherein said compression is applied to said area of the patient body using said ultrasonic probe as a piston.

9. An ultrasound imaging method of measuring a compression ratio of an obj ect of interest, comprising the steps of: acquiring a first echographic image of an object of interest located in a patient body using an ultrasound probe placed on said patient body,

- compressing said object of interest,

- acquiring a second echographic image during compression of said object of interest,

- measuring a first height of said object of interest in said first echographic image and a second height of said object of interest in said second echographic image, - deriving a compression ratio of said object of interest from said first and second heights.

10. A computer program comprising a set of instructions which, when loaded into a processor, causes said processor to carry out the method as claimed in claim 9.