Measuring the hip

Abstract

In a method for hip measurement in an ultrasound image, an ultrasound image of the hip is obtained. Also obtained is a spatial coherence map associated with one or more lags of ultrasound waves associated with the ultrasound image; a quality metric is then determined based on the hip ultrasound image and the spatial coherence map, where the quality metric is indicative of the suitability of the ultrasound image for use in measuring a hip. If the quality metric indicates that the ultrasound image is above a threshold quality, then the method includes then indicating that the ultrasound image is suitable for hip measurement.

Description

TECHNICAL FIELD

[0001] The disclosure herein relates to ultrasound imaging. In particular, but not exclusively, embodiments herein relate to systems and methods for making measurements in ultrasound images of the hip. BACKGROUND

[0002] Ultrasound imaging (US) is used for a range of medical applications, such as fetal monitoring. Medical ultrasound imaging involves moving a probe containing ultrasound transducers which produce high frequency sound waves on the skin. The high frequency sound waves are transmitted into soft tissue and pass / propagate through the tissue and reflect from internal surfaces, such as tissue boundaries. The reflected waves, or echoes, from the interrogated medium are detected and used to build up an image of the internal structure of interest.

[0003] Ultrasound imaging can be used to create two-dimensional or three-dimensional images. In a typical workflow, a user, such as an ultrasound physician, radiologist, clinician or other medical professional, can use two-dimensional imaging to locate an anatomical feature of interest. Once the feature is located in two dimensions, the user can activate a three-dimensional mode to take a three-dimensional image.

[0004] It is an object of embodiments herein to provide systems and methods that can improve ultrasound imaging efficiency. SUMMARY

[0005] Developmental dysplasia of the hip (DDH) is an umbrella term describing all forms of abnormal development of the hip in infants and young children. The overall incidence is 1-3 per 1000 live births, with ethnic and geographical variations. DDH has a wide spectrum of presentation, with mild findings often resolving spontaneously, while the most severe examples can lead to disability if not diagnosed early in life. Successful management of infants with DDH depends on early diagnosis and appropriate treatment.

[0006] There are several methods of diagnosing DDH in clinical practice, including 1) physical examination of newborns (especially those with a higher risk of DDH due to, for example, ethnicity, family history, and / or pregnancy history) and 2) medical imaging (e.g.: ultrasound imaging and magnetic resonance imaging (MRI), X-ray radiography or computed tomography (CT).

[0007] Physical examination methods are often less reliable and highly dependent on the experience of the operator. Medical imaging is increasingly used for screening / diagnosis and follow-up (e.g. to confirm hip reduction after surgery) throughout the diagnosis and management cycle of DDH infants. MRI can often provide detailed information about the hip structures. However, there are several problems with using MRI in the screening and diagnosis of DDH, including: a) high cost; b) it can be difficult to obtain adequate MRI images unless the infant is sedated (infants older than 4-6 weeks are often too mobile to scan without sedation); c) long scan times compared to ultrasound imaging (MRI takes tens of minutes per scan, compared to a few minutes for 2D ultrasound or a few seconds for 3D ultrasound); d) the pre-MRI settling process for small infants can be time-consuming for staff and parents, potentially making MRI scans impractical; e) if sedation is required to obtain diagnostic quality images beyond the neonatal period, this carries the risk of brain damage or even death.

[0008] Therefore, it is desirable to use ultrasound imaging to detect and monitor DDH in infants where possible. While ultrasound imaging is widely used for selective screening of infants at high risk of DDH, it is a challenge to identify in real-time (e.g. during an ultrasound examination) ultrasound image frames of sufficient quality to enable measurement of key anatomical structures of the infant’s hip (e.g. the iliac bone and the femoral head). This can result in inadequate images being obtained, leading to poor quality of measurement and diagnosis, and / or wasting time and resources when ultrasound examinations must be repeated. It is an object of embodiments herein to improve this situation by helping sonographers and other users of ultrasound imaging devices to obtain real-time feedback and / or guidance to help them obtain high quality ultrasound images suitable for DDH measurements.

[0009] According to a first aspect, therefore, there is provided a system for hip measurements in ultrasound images. The system comprises a memory and a processor, the memory comprising instruction data representing a set of instructions, the processor being configured to communicate with the memory and run the set of instructions. The set of instructions, when executed by the processor, cause the processor to: obtain an ultrasound image of the hip; obtain a spatial coherence map associated with one or more lags of ultrasound waves associated with the ultrasound image; determine a quality metric based on the hip ultrasound image and the spatial coherence map, wherein the quality metric is indicative of the suitability of the ultrasound image for making measurements of the hip; and indicate that the ultrasound image is suitable for making measurements of the hip if the quality metric indicates that the ultrasound image is above a threshold quality.

[0010] In this way, the quality of the image frame / appropriateness of the hip measurement can be evaluated in real time while an ultrasound physician or other user performs an ultrasound examination on the hip. This provides an objective metric for determining whether an image has appropriate quality. Without this method, the quality of the obtained ultrasound image depends on the operator's skill or intuitive "feel" for whether the image is suitable for real-time examination. Furthermore, as will be described in more detail below, various structures of the hip, such as the femoral head and / or the boundary between the iliac bone and nearby tissues, may appear more prominent (or brighter) in the coherence image compared to the ultrasound image itself. This provides additional information about whether the ultrasound image is suitable for measurement, such as whether features are imaged with appropriate quality and at the appropriate angle / through-the-hip slice. Therefore, an improved system exists for obtaining images suitable for hip measurements.

[0011] According to a second aspect, a method for measuring the hip in an ultrasound image is provided. The method includes obtaining an ultrasound image of the hip; obtaining one or more hysteresis-associated spatial coherence maps of ultrasound waves associated with the ultrasound image; determining a quality metric based on the ultrasound image of the hip and the spatial coherence maps, wherein the quality metric indicates the suitability of the ultrasound image for hip measurement; and indicating that the ultrasound image is suitable for hip measurement if the quality metric indicates that the ultrasound image is above a threshold quality.

[0012] In a third aspect, a computer program product including a computer-readable medium is provided, the computer-readable medium having computer-readable code contained therein, the computer-readable code being configured such that, when executed by a suitable computer or processor, it causes the computer or processor to perform the methods described in the second aspect. Attached Figure Description

[0013] To better understand the embodiments and to more clearly illustrate how they are implemented and effective, reference will now be made to the accompanying drawings only by way of example, wherein,

[0014] Figure 1 illustrates the existing technical measurement scheme for determining DDH;

[0015] Figure 2 Example systems according to some embodiments herein are illustrated;

[0016] Figure 3 Figure a shows an ultrasound image of the hip;

[0017] Figure 3 Figures b, 3c, and 3d illustrate the... Figure 3 Coherence plots of ultrasound images with different hysteresis correlations in a;

[0018] Figure 4 e, 4f, 4g and 4h illustrate different average spatial coherence values associated with different lags and positions within the anatomical structure; Figure 3 a different lags of the hip ultrasound image;

[0019] Figure 5 is a plot showing different average spatial coherence values for different lags and positions within the anatomical structure;

[0020] Figure 6 shows an example ultrasound system according to some embodiments herein; and

[0021] Figure 7 illustrates a method according to some embodiments herein. DETAILED DESCRIPTION

[0022] As mentioned above, one application of ultrasound imaging is to image the hips of infants to diagnose and monitor developmental dysplasia of the hip (DDH). Compared to MRI, taking measurements in ultrasound images has various advantages, for example, but the accuracy of the measurements is sensitive to the quality of the ultrasound images obtained, both in terms of image quality and whether a proper slice through the hip has been obtained.

[0023] Figure 1 shows a prior art example of measurements that can be made in ultrasound images of the hip in order to diagnose or monitor DDH according to the Graf measurement system. In this system, the positions of the iliac bone 102 and the femoral head 104 are used to determine the values of the angles a and b and to diagnose DDH. See the article by R. Graf, M. Mohajer and F. Plattner entitled ‘Hip sonography update. Quality-management, carastrophes-tips and tricks’ (Med Ultrason, 2013, Vol. 15: 299-303) for further details.

[0024] In order to standardize the measurement and diagnosis of DDH, various guidelines recommend the use of the Graf method. However, the Graf measurement system still requires a high level of user (e.g. operator or sonographer) experience to measure the angles a and b correctly. For example, it can take 50-100 examinations before one is able to adapt to the ultrasound scanning procedure required by Graf. Limitations of the Graf method include relatively high inter-observer and inter-scan variability, which can change the final diagnosis of 50% to 70% of infants if scanned by non-experts. Over-diagnosis of DDH leads to increased health care costs and unnecessary treatment, particularly for patients with borderline dysplasia. Furthermore, the repeatability or reproducibility of these measurements from 2D ultrasound is actually dependent on the operator / examiner for the following reasons:

[0025] 1) 2D ultrasound examination is highly dependent on the operator.

[0026] 2) Probe or transducer orientation determines the image US quality of the standardized plane and influences the measurement results, as 2D ultrasound imaging only shows a partial view of the complex 3D acetabular shape.

[0027] For example, the above difficulties related to the Graf method are discussed in the article by D. Zonoobi et al. entitled ‘Developmental hip dysplasia diagnosis at Three-dimensional US: A multicenter study’ (Radiology, 2018, Vol. 287: 1003-1015).

[0028] 3D ultrasound for DDH was introduced and investigated in the 1990s and showed some benefits in providing a more complete view of the hip geometry and obtaining high-fidelity 3D MRI. Some 3D measures (similar to Graf angles) have been introduced and proven to have higher reliability than developmental dysplasia measures measured from 2D ultrasound images.

[0029] Some research has proposed to automatically measure the hip (e.g. Graf or 3D equivalent) in ultrasound imaging, for example, by training a machine learning model to predict angles a and b from an input image. However, such approaches can be complex and require high computational load, which can not be suitable for bedside ultrasound clinical applications in DDH diagnosis and screening and follow-up examinations. Therefore, there is a need for reliable, lightweight methods to measure the hip in ultrasound images.

[0030] To this end, some embodiments herein propose systems and methods of evaluating an ultrasound image of a hip during an ultrasound examination and determining whether a suitable image for taking measurements has been obtained. This ensures that measurements are taken on images of appropriate quality to improve accuracy. Furthermore, the quality metric can be used in real-time to provide feedback and / or guidance to a user performing the ultrasound examination (e.g. an ultrasound physician) to let them know whether they have captured an image of sufficient quality to take hip measurements. The proposed quality metric is based on the ultrasound image of the hip and a spatial coherence map, which has the advantage of providing an objective metric to determine whether the image is of suitable quality. Without such a method, the quality of the ultrasound image obtained can depend on the operator’s ability or intuitive “feeling” of whether the image is suitable for real-time examination.

[0031] Figure 2A system (e.g. apparatus) 200 for recording ultrasound images is illustrated in accordance with some embodiments herein. The system 200 can be used to record (e.g. acquire or take) ultrasound images. In some embodiments, the system 200 can comprise or form part of a medical device (e.g. an ultrasound system).

[0032] With reference to Figure 2 The system 200 comprises a processor 202 which controls the operation of the system 200 and can implement the methods described herein. The processor 202 can comprise one or more processors, processing units, multi-core processors or modules configured or programmed to control the system 200 in the manner described herein. In particular implementations, the processor 202 can comprise a plurality of software and / or hardware modules each configured to perform or for performing individual or multiple steps of the methods described herein.

[0033] Briefly, the processor 202 of the system 200 is configured to obtain an ultrasound image of a hip; obtain a spatial coherence map associated with one or more lags of ultrasound waves associated with the ultrasound image; determine a quality metric based on the hip ultrasound image and the spatial coherence map, wherein the quality metric is indicative of a suitability of the ultrasound image for making measurements of the hip; and indicate that the ultrasound image is suitable for hip measurements if the quality metric indicates that the ultrasound image is above a threshold quality.

[0034] Technically, this can provide an improved way of obtaining ultrasound images of sufficient quality that are suitable for making hip measurements that can be used to make a DDH diagnosis. Without such a method, an ultrasound practitioner can effectively be “blindly” imaging without knowing whether they are capturing images that can be used to accurately determine the required measurements (e.g. the Graf measurements described above).

[0035] In some embodiments, as illustrated in Figure 2 The system 200 can further comprise a memory 204 configured to store program code which can be run by the processor 202 to perform the methods described herein. Alternatively or additionally, the one or more memories 204 can be external to (i.e. separate or remote from) the system 200. For example, the one or more memories 204 can be part of another device. The memory 206 can be used to store images, information, data, signals and measurements acquired or generated by the processor 202 of the system 200 or any interface, memory or storage device external to the system 200.

[0036] In some embodiments, as illustrated in Figure 2As shown in FIG. 2, the system 200 can also include a transducer 208 for capturing ultrasound images. Alternatively or additionally, the system 200 can receive (e.g., through a wired or wireless connection) a data stream of two-dimensional images taken using an ultrasound transducer external to the system 200.

[0037] The transducer 208 can be formed of a plurality of transducer elements. Such transducer elements can be arranged to form an array of transducer elements. The transducer can be included in a probe such as a handheld probe that can be held by a user (e.g., an ultrasound physician, radiologist, or other clinician) and moved over a patient’s skin. Those skilled in the art will be familiar with the principles of ultrasound imaging, but briefly, an ultrasound transducer includes piezoelectric crystals that can be used to both generate and detect / receive sound waves. Ultrasound produced by the ultrasound transducer enters the patient’s body and reflects from underlying tissue structures. The reflected waves (e.g., echoes) are detected by the transducer and compiled (processed) by a computer to produce an ultrasound image of the underlying anatomy, also known as an echogram.

[0038] In some embodiments, the transducer 208 can include a matrix transducer that can interrogate a volume space. In some embodiments, the transducer can include a high frequency linear ultrasound probe. In some embodiments, the transducer can include a one-dimensional transducer formed, for example, by a tightly packed and lightweight array or a wireless array of about 6 MHz (or higher). In some embodiments, the transducer can include a high frequency 2D array probe, such as a capacitive micromachined ultrasonic transducer (CMUT).

[0039] In some embodiments, as Figure 2 As shown in FIG. 2, the system 200 can also include at least one user interface, such as a user display 206. The processor 202 can be configured to control the user display 206 to display or present to a user, for example, a received data stream or portions of an ultrasound image. The user display can further display an indication of whether an ultrasound image captured using the transducer is suitable for hip measurements (e.g., the user display 206 can display an output of the methods described herein). The user display 206 can include a touchscreen or application (e.g., on a tablet or smartphone), a display screen, a graphical user interface (GUI), or other visual presentation component.

[0040] Alternatively or additionally, the at least one user display 206 can be external (i.e. separate or remote) from the system 200. For example, the at least one user display 206 can be part of another device. In such embodiments, the processor 202 can be configured to send instructions to the user display 206 external to the system 200 (e.g. over a wireless or wired connection) in order to trigger (e.g. cause or initiate) the external user display to display an indication of whether the ultrasound image captured using the transducer is suitable for a hip measurement. For example, the user display 206 can display the output of the method described herein.

[0041] It will be appreciated that, Figure 2 Only the components necessary to illustrate this aspect of the disclosure are shown and in a practical implementation the system 200 can comprise additional components to those shown. For example, the system 200 can comprise a battery or other means for connecting the system 200 to a mains power supply. In some embodiments, as shown in Figure 2 The system 200 can also comprise a communications interface (or circuitry) for enabling the system 200 to communicate with any interfaces, memories and devices internal or external to the system 200, for example over a wired or wireless network, as shown in

[0042] In more detail, the system 200 can be adapted for making a measurement of a hip (e.g. of a newborn or infant) in an ultrasound image. The measurement can comprise, for example, a measurement suitable for diagnosing DDH. The measurement can thus comprise a developmental dysplasia measurement. For example, in some embodiments the measurement can comprise a Graf measurement as described above. However, the skilled person will appreciate that a Graf measurement is merely an example and that other measurements of the hip can equally be made, for example the three-dimensional hip measurements described in the paper by C. Diederichs et al. entitled ‘Cross-modality validation of acetabular surface models using 3-D ultrasound versus magnetic resonance imaging in normal and dysplastic infant hips’ (Ultrasound in Med. & Biol., 2016, Vol. 42: 2308-2314).

[0043] The ultrasound image can be captured as part of an ultrasound examination of a hip of, for example, an infant, baby or young child for the purpose of making a DDH diagnosis.

[0044] Typically, the ultrasound image can comprise a two-dimensional image or a three-dimensional image. The image can comprise different modalities, for example the image can comprise a B-mode ultrasound image, a Doppler ultrasound image or an elastography mode image.

[0045] In some embodiments, ultrasound images can be acquired in real time. For example, ultrasound images may be acquired as part of an ongoing or “live” ultrasound examination. In other embodiments, ultrasound images may be obtained from historical ultrasound examinations (e.g., retrieved from a database).

[0046] In embodiments where system 200 includes transducer 208, ultrasound images can be obtained from data captured using transducer 208 and processed by a processor to produce ultrasound images. Alternatively, ultrasound images can be received from a remote transducer (e.g., transmitted via the Internet).

[0047] In some embodiments, a display 208 may be used to show a standardized 2D ultrasound image, including the iliac bone and femoral head, to the user as a reference for performing an ultrasound examination. Using such an example 2D image and its landmarks as a guide, the user may be able to move the ultrasound probe until a properly high-quality ultrasound image is obtained, as described below.

[0048] Typically, during an ultrasound examination, the user (the sonographer or the person performing the examination) places the probe in the hip area at the appropriate position and angle, and then acquires 2D or 3D images.

[0049] During a 2D US examination, the user can scan the infant's hip area with an ultrasound probe at a fixed / variable speed. A sequence of 2D US images is collected during this dynamic scan and saved to memory, for example, in DICOM image format or as raw US data. A smart AI algorithm may select the best representative frame (called the optimal image for later measurement).

[0050] In other words, in some embodiments, enabling the processor to acquire an ultrasound image of the hip may include enabling the processor to: receive multiple two-dimensional ultrasound images; use a model trained using a machine learning process to determine, based on the multiple two-dimensional ultrasound images, the optimal image for subsequent measurements; and select the optimal image for subsequent measurements as the acquired ultrasound image. The following quality measurement methods may be performed on the acquired ultrasound image (“the optimal image for subsequent measurements”) according to the methods described below.

[0051] As part of an ultrasound examination, the methods described herein can also be performed on any (or all) 2D frames imaged by an ultrasound physician.

[0052] In some embodiments, 3D US volume can be obtained from the acquired 2D US image sequence. For example, a 2D probe can be moved around the hip region until a suitable 3D US volume or volume sequence is acquired. In these embodiments, quality metrics can be calculated based on several slices selected at the median line for this purpose.

[0053] In other words, in some embodiments, causing the processor to obtain an ultrasound image can comprise causing the processor to: receive a plurality of two-dimensional ultrasound images; use a model trained using a machine learning process to reconstruct a three-dimensional volume from the plurality of two-dimensional ultrasound images, and select a slice from the three-dimensional volume as the obtained ultrasound image.

[0054] The machine learning process can be trained to reconstruct a 3D volume from a series of 2D ultrasound images using a process such as that proposed by R. Prevost et al in the article entitled ‘3D freehand ultrasound without external tracking using deep learning’, Medical Image Analysis, 2018, Vol. 48: 187-202.

[0055] Turning now to coherence maps, the skilled person will be familiar with spatial coherence maps, but briefly, a spatial coherence map describes the spatial coherence properties of the tissue backscattered signal of an object (e.g. a baby / newborn) being examined and measured by an ultrasound transducer array. In general, the spatial coherence of a wave is a measure of how it varies with distance. For example, a spatial coherence map can produce bright pixel values in regions with partial or high coherence, and dark pixel values in regions where there is low or no coherence in the backscattered wavefront (or vice versa). Coherence can be influenced by the properties of the tissue over which the wave reflects or scatters, and so coherence maps can be used to image the underlying tissue.

[0056] In more detail, the spatial coherence function is a measure of the correlation function between signals with a given lag or separation of m elements. The function is defined at each field point x in the ultrasound image. Thus, Rm(x) is the correlation coefficient of transducer element signals measured by a transducer element from field point x for a relative lag of m. For a two-dimensional transducer array, m refers to a two-dimensional lag, with components in both dimensions of the array. In some cases, these measured correlation coefficients are computed by averaging over a correlation window (e.g. a one-wavelength axial signal window centred on x). In other cases, no averaging is performed to provide these correlation coefficients, resulting in so-called ‘single-pixel’ calculations. Here the field point x can be a pair of two coordinates (x = (x1, x2)) or a triple (x = x1, x2, x3) of three coordinates for 2D and 3D cases respectively.

[0057] The integration of the spatial coherence function can be performed by an analytical integration over the appropriate range, but in practice this integration is usually performed by a discrete summation over indices in the appropriate range.

[0058] For a one-dimensional transducer array, the ultrasound transducer array can have N elements, and the predetermined integration window is preferably a lag range from 1 to N / 2, and more preferably a lag range from 1 to N / 4. For a two-dimensional transducer array, the ultrasound transducer array can have Nx x Ny elements, and the predetermined integration window is preferably a lag range from 1 to Nx / 2 and from 1 to Ny / 2, and more preferably a lag range from 1 to Nx / 4 and from 1 to Ny / 4 and more preferably the set of all two-dimensional lags (mx, my) such that (1 - mx / Nx)(1 - my / Ny) > 0.75. Further details of beamforming methods suitable for spatially coherent imaging are given in the article by Hyun et al. (IEEE Transactions on Ultrasonics, Ferroelectrics and Frequency Control, v61 n7, pages 1101-1112, 2014).

[0059] Figure 3 An example ultrasound image of an infant's hip is provided in a, showing a B-mode image of the iliac bone (linear structure 302) and the near-spherical femoral head 304. Figure 3 b, 3c and 3d show spatially coherent images corresponding to the image in a with lag times of 2, 3 and 4 respectively. Figure 3 The images in a correspond to the spatially coherent images in b, 3c and 3d with lag times of 2, 3 and 4 respectively. Figure 4 e, 4f, 4g and 4h illustrate spatially coherent images with lags of 5, 6, 7 and 8 respectively. In this example, the iliac bone lies within a depth range of pixels [360 410] and a width range of pixels [1 400], and the femoral head appears quite spherical but lacks sharpness. However, in Figure 3 The circle corresponding to the femoral head is clearly visible in the spatially coherent images in b to 4h. This is because the ultrasound signals from the femoral head boundary are highly coherent, whereas noise or other artefacts are less coherent.

[0060] In the example shown in Figure 3 and Figure 4 the optimal sliding window length was determined by varying the length from 10 pixels to 80 pixels, and ranged between 30 pixels to 50 pixels; thus a length of 41 pixels was used in calculating these spatially coherent images. However, it will be appreciated that this is merely an example, and different sliding window lengths can be suitable for different embodiments of the herein.

[0061] In general, a spatial coherence map can be obtained using a coherence function Rm(x) which is computed using cross-correlation performed between pairs of transducer elements separated by a distance m(c). The coherence function R is evaluated as a function of distance in element number m (or lag) by establishing an autocorrelation between all pairs of receiver elements. For example, the coherence for lag m can be computed according to the following equation:

[0062]

[0063] The processor is then caused to determine a quality metric based on the ultrasound image of the hip and the spatial coherence map. In this sense, the quality metric is indicative of the suitability of the ultrasound image for taking a hip measurement. For example, the quality metric can generally represent an estimate of the suitability of the ultrasound image for taking a measurement of a baby's hip. In some embodiments, the quality metric can represent the sharpness or image quality of the ultrasound image. In other embodiments, the quality metric can represent the suitability of the angle, or the suitability of the two-dimensional slice through the hip represented by the ultrasound image. In some embodiments, the quality metric can combine these aspects.

[0064] According to Figure 3 b-d and Figure 4 e-h, the inventors have noticed that the coherence (e.g. relatively high gradient in the coherence map) drops for the region under the iliac bone, where there can be noise due to the ultrasound's inability to penetrate bone-like tissue like the iliac bone, whereas the coherence reduces less for the soft tissue above the iliac bone. Thus, the average spatial coherence values for the iliac bone, the noise region under the iliac bone and the soft tissue above the iliac bone are computed for a lag range of 2 to 8, and these curves are shown in Figure 5 Fig. 5, which shows the average spatial coherence values for a lag of 2 to 8 for the iliac bone region (curve 502), the noise region under the iliac bone (curve 504) and the soft tissue above the iliac bone (curve 506).

[0065] From the above studies, the inventors have drawn the following conclusions:

[0066] (1) there is (high or relatively high, e.g. above a threshold) spatial coherence on the ultrasound image of the iliac bone;

[0067] (2) the shadowing due to the iliac bone is highly incoherent on the US image;

[0068] (3) human soft tissue is partially spatially coherent, but the boundaries between different soft tissues show relatively high coherence on the US image.

[0069] The spatial coherence map can also be used to estimate the size or other properties of the circle / loop of the femoral head, as the appearance of the femoral head in the spatial coherence map is strongly enhanced. Thus, the contrast between the iliac bone (or soft tissue above the iliac bone) and / or the noise region between the iliac bones from the B-mode US image with the spatial coherence map can be used as an indicator of whether the B-mode image is of sufficient quality to provide a later development dysplasia measurement.

[0070] For example, if the contrast and coherence values of the iliac bone and soft tissue are both high, then the quality of the US image can be sufficient to correctly measure the dysplasia measure, otherwise the US image can not be sufficient for such a measurement.

[0071] Thus, in general, the ultrasound image can be of sufficient quality if some or all of the following criteria are met:

[0072] • there is a high contrast resolution between the key structures (the iliac bone and soft tissue above the iliac bone) and the noise region below the iliac bone (acoustic shadow)

[0073] • the iliac bone contains a horizontal line (or close to horizontal line) in the middle of the 2D ultrasound image

[0074] • the femoral head comprises a circle / loop close to the centre of the 2D ultrasound image

[0075] • the US image shows a plane through the femoral head that bisects the femoral head at its widest point.

[0076] Thus, in some embodiments, the quality measure can be based on a measure of the spatial coherence at the iliac bone or femoral head; a measure of the contrast between the iliac bone and noise in the region adjacent to the iliac bone; and / or a measure of the diameter of the femoral head.

[0077] In more detail, in some embodiments, the quality measure can indicate that the ultrasound image is above a threshold quality if the coherence value associated with the iliac bone is above a first threshold value. For example, if the average coherence value of the iliac bone is closer to the (normalised) value 1.0, then the ultrasound image can be of sufficient quality. The first threshold value can thus be set to a value of, for example, 0.8, or a value of 0.9 or a value of 0.95.

[0078] In some embodiments, the quality measure can indicate that the ultrasound image is above a threshold quality if a measure of the contrast between the iliac bone and noise in the region adjacent to the iliac bone is above a second threshold value. For example, if the contrast between the iliac bone (or soft tissue near it) and the noise region below the iliac bone is relatively high, then the second threshold value can be set such that the ultrasound image is considered to be of sufficient quality. For example, if the contrast is above a value of about, for example, 3dB. The ratio is a parameter that is calibrated, whereas the absolute parameter can be calibrated.

[0079] In some embodiments, the quality metric can indicate that the ultrasound image is above a threshold quality if the ultrasound image includes a plane that passes through the femoral head, where the diameter of the femoral head is at (or approximately at, e.g. within a threshold tolerance of) its maximum value.

[0080] In general, the femoral head is not a perfect sphere, and in embodiments, e.g. where the measurement comprises a Graf measurement, the measurement should be made in an ultrasound image that includes a plane (e.g. cross-section) that bisects the femoral head at its widest point through the hip joint. In other words, the measurement should be made on a plane that passes through the femoral head at the maximum diameter and / or maximum cross-sectional area. This ensures that all operators / sonographers use a consistent reference plane.

[0081] Imagine the femoral head as a sphere, there is a plane that bisects the femoral head at its maximum diameter, and the task of the operator can be to determine this plane. Thus, in some embodiments, the processor 202 can be configured to determine the diameter of the femoral head in real-time during the ultrasound examination and send instructions to the display 206 to instruct the display to display the determined diameter to the user (e.g. operator or sonographer) performing the ultrasound examination. This can help the user to obtain an image that includes a plane close to or passing through the maximum diameter of the femoral head.

[0082] The diameter of the femoral head can be determined from the ultrasound image by...

[0083] In some embodiments, the quality metric can indicate that the ultrasound image is above a threshold quality if a coherence value associated with the femoral head is above a third threshold value. The coherence value can comprise an average coherence value.

[0084] In some embodiments, the above measurements can be combined into a single number or score. For example, if the US image score is above a fourth threshold value (e.g. if the US score is calibrated so that the maximum score is 100, in one example the threshold value can be 90), then the ultrasound image quality is classified as good enough for later measurements.

[0085] An appropriate value for the fourth threshold can be determined based on characteristics of the ultrasound scanner (e.g., how it is calibrated) and / or characteristics of the patient population. The combined score (e.g., combining the average coherence of the ilium, the contrast between the ilium and nearby soft tissue at similar imaging depths, and / or the noise level of the 2D ultrasound image; and / or whether a measure including a plane through the maximum diameter of the femoral head has been reached) can be determined from a dataset collected (including normal subjects and patients with DDH) and the ground truth of the dataset determined by clinical diagnosis. An experienced sonographer or ultrasound specialist typically has a strong“feeling pattern” (targetedness) for high quality ultrasound images with all attributes within the image. Thus, rules or patterns can be identified based on these annotations.

[0086] In general, as described above, the quality metric can be used to determine whether the US image is sufficient for measurements to be made. If the US image quality is not sufficient, the processor can be caused to indicate to the user (e.g., by sending instructions to the display) that the obtained image is not suitable for measurements to be made. The processor can be caused to indicate to the user to move the probe and / or change the probe angle until an appropriate high quality ultrasound image is acquired for which the quality metric indicates that the ultrasound image is above a threshold quality.

[0087] Once an image is obtained for which the quality metric indicates that the ultrasound image is above a threshold quality, the processor can be configured to provide an indication to the user that the ultrasound image is suitable for measurements to be made of the hip (e.g., by sending instructions to the display 208). For example, the indication can include an audio or visual cue.

[0088] In some embodiments, the processor can also be caused to determine a measurement of the hip using the ultrasound image and / or the coherence maps. To this end, in some examples, the processor can be caused to segment the ultrasound image and / or one or more of the coherence maps associated with the ultrasound image. Returning to Figure 3 In a–4h, the strong appearance of the femoral head and ilium in the coherence maps makes it possible that improved iliac and / or femoral head segmentation can be produced by combining one or more spatial coherence maps with the 2D mode US image. Thus, in some embodiments, one or more spatial coherence maps and / or the ultrasound image can be combined prior to performing the segmentation.

[0089] In some embodiments, spatial coherence maps having different lags can be combined into a“combined spatial coherence map,” and the processor can be caused to determine a measurement of the hip using the combined spatial coherence map. In other words, the combination (e.g., normalization and averaging, or some other manner of combining) of the spatial coherence maps can be segmented.

[0090] In some embodiments, the ultrasound image and the spatial coherence map can be combined into a single image, which is referred to herein as a "combined image". The processor can then be caused to segment the ilium and the femoral head in the combined image and use the segmentation of the ilium and the femoral head to perform the measurements of the hip. For example, the coherence map can be used as a mask. The combined image can be determined according to the following manner: mask x original US image. In another example, a threshold can be applied to the coherence map to convert the coherence map to a binary image. For example, the threshold can be 0.6. In such an example, the combined image can then be determined as binary image x original US image. In another example, a linear combination of the two methods described above can be employed.

[0091] The skilled person will be familiar with image segmentation, but briefly, image segmentation involves extracting shape / form information about objects or shapes captured in an image. This can be achieved by converting the image into constituent blocks or "segments", each of which has a common attribute. In some approaches, image segmentation can involve fitting a model to one or more features in the image.

[0092] Examples of image segmentation that can be performed by the processor 202 on the combined image include, but are not limited to, model-based segmentation (MBS), in which a triangular mesh of a target structure (e.g. heart, brain, lung, etc.) is adapted to features in the image in an iterative manner, and machine learning (ML) segmentation, in which an image is converted into a plurality of constituent shapes (e.g. block shapes or block volumes) using an ML model based on similar pixel / voxel values and image gradients. Further information on MBS applied to medical images can be found in, for example, the article by Ecabert, O. et al. in 2008 entitled "Automatic Model-Based Segmentation of the Heart in CT Images" (IEEE Trans. Med. Imaging 27(9), 1189-1201).

[0093] A sphere can be fitted to the femoral head in order to determine, for example, its position and diameter. Thus, in some embodiments, causing the processor to use the segmentation of the ilium and the femoral head to perform the measurements of the hip can comprise causing the processor to fit an ellipse to the segment of the segmentation corresponding to the femoral head, for example using a Hough transform.

[0094] In embodiments in which segmentation is performed on the coherence map, the combined coherence map or the combined image, the use of the coherence can provide an improved fitted Hough transform, as the femoral head is brighter in the spatial coherence map compared to the ultrasound image (see Figure 3 b-d and Figure 4 e-h and Figure 3 a comparison of the USB mode image in a).

[0095] The processor can then be caused to send instructions to the display to cause the display to mark the fitted ellipse and / or the segments of the segmentation result corresponding to the iliac bone and the femoral head onto the ultrasound image, and / or to make measurements of the iliac bone relative to the marked fitted ellipse and / or the segmentation result corresponding to the iliac bone and the femoral head.

[0096] The processor can also be caused to use the measurements of the hip to diagnose developmental dysplasia of the hip, DDH. This can be performed, for example, by comparing the measurements to well-known scales associated with the particular type of measurement performed. For example, there are threshold values associated with the Graf values of the angles a and b shown in Fig. 1, which can be used to diagnose DDH.

[0097] In addition to or instead of the segmented iliac bone and femoral head, the determined measurements (e.g. dysplasia metrics) can be displayed on the screen.

[0098] The processing performed by the processor 202 described above can be divided into modules. For example, some embodiments can comprise one or more of the following modules: i) a module to perform a freehand sweep / scan of the surface of an infant's hip using a 1D high frequency linear ultrasound probe to obtain a sequence of 2D ultrasound images ii) a module to use a deep learning algorithm to reconstruct a 3D ultrasound volume image from the sequence of 2D ultrasound images (as described above) iii) a module to perform automatic key structure (e.g. iliac bone and femoral head (circle)) detection and measurement of dysplasia metrics. This module can exploit the spatial coherence properties of US images as described above iv) a module for determining the quality metrics described above, which can determine for a given transducer position whether the acquired 2D ultrasound image is suitable for use in later metric measurements, e.g. checking that the key structures (ilium and femoral head) are displayed well enough in the 2D ultrasound image or 3D ultrasound volume, e.g. (a) detected iliac bone (b) detected femoral head or circle; (c) an indication that the infant is normal (metric above threshold) or (d) an indication that the infant is DDH (metric is small), and vi) a module to provide a graphical representation of the 2D US images or 3D US (or sliced 2D image) volume with detected typical landmarks of the hip.

[0099] Embodiments of the system 200 can be used for ultrasound imaging for infant DDH diagnosis and screening. The system can be used (a) in a pre-hospital setting (small clinics or at home), (b) for initial assessment in a hospital, and (c) for follow-up after appropriate DDH treatment. The teachings herein are applicable to many ultrasound imaging systems and are therefore not limited to any particular ultrasound imaging device. The teachings are also not limited to any particular DDH study (diagnosis and screening).

[0100] Therefore, System 200 can provide inexperienced users with increased confidence in ultrasound-based DDH diagnosis and screening; improve workflow by reducing examination and interpretation time; reduce user reliance and standardize reporting schemes; and facilitate easier archiving and documentation for later use, such as for billing or educational purposes.

[0101] Turn now Figure 6 , Figure 6 An example embodiment of an ultrasound system 600 constructed according to the principles described herein is shown. Figure 6 One or more components shown may be included within a system configured to perform any of the methods described herein. For example, one or more processing components may be configured to acquire an ultrasound image of the hip, acquire a spatial coherence map associated with one or more hysteresis of ultrasound waves associated with the ultrasound image, determine a quality metric based on the ultrasound image of the hip and the spatial coherence map, wherein the quality metric indicates the suitability of the ultrasound image for hip measurement, and indicates that the ultrasound image is suitable for hip measurement if the quality metric indicates that the ultrasound image is above a threshold quality.

[0102] For example, any of the aforementioned functions of processor 202 can be programmed into the processor of system 600, for example, via computer-executable instructions. In some examples, the functions of processor 202 can be provided by... Figure 6 One or more processing components are shown that implement and / or control, including, for example, an image processor 636.

[0103] exist Figure 6 In the ultrasound imaging system, the ultrasound probe 612 includes a transducer array 614 for transmitting ultrasound waves into a body region and receiving echo information in response to the transmitted waves. The transducer array 614 may be a matrix array comprising multiple transducer elements configured to be individually activated. In other embodiments, the transducer array 614 may comprise a one-dimensional linear array. The transducer array 614 is coupled to a microwave beamformer 616 within the probe 612, which controls the transmission and reception of signals by the transducer elements in the array. In the illustrated example, the microwave beamformer 616 is connected by a probe cable to a transmit / receive (T / R) switch 618, which switches between transmission and reception and protects the main beamformer 622 from high-energy transmitted signals. In some embodiments, the T / R switch 618 and other components of the system may be included in the transducer probe rather than in a separate ultrasound system base.

[0104] Transmission of ultrasound beams from the transducer array 614 under the control of the microwave beamformer 616 is directed by a transmit controller 620 coupled to the T / R switches 618 and the beamformer 622, which receives input from, for example, user operation of a user interface or control panel 624. One of the functions controlled by the transmit controller 620 is the direction in which the beam is steered. The beam can be steered straight ahead (perpendicular to the transducer array), or at different angles for a wider field of view. The partially beamformed signals produced by the microwave beamformer 616 are coupled to the beamformer 622, where the partially beamformed signals from individual facets of the transducer elements are combined into fully beamformed signals.

[0105] The beamformed signals are coupled to a signal processor 626. The signal processor 626 can process the received echo signals in various ways, such as bandpass filtering, decimation, I and Q component separation, and harmonic signal separation. The data generated by the different processing techniques employed by the signal processor 626 can be used by a data processor to identify internal structures, such as ribs, or anatomical features of a neonate, and their parameters.

[0106] The processor 626 can also perform signal enhancements, such as speckle reduction, signal compounding, and noise elimination. The processed signals can be coupled to a B-mode processor 628, which can employ amplitude detection to image structures in the body, including, for example, ribs, heart, and / or pleural interfaces. The signals produced by the B-mode processor are coupled to a scan converter 630 and a multiple planar reformatter 632. The scan converter 630 arranges the echo signals in a desired image format from the spatial relationships in which the echo signals were received. For example, the scan converter 630 can arrange the echo signals into a two-dimensional (2D) sector-shaped format. The multiple planar reformatter 632 is capable of converting echoes received from points in a common plane in a volumetric region of the body into an ultrasound image of that plane, as described in U.S. Patent 6,663,896 (Detmer). A volume Tenderer 634 converts echo signals of a 3D data set into a projected 3D image as seen from a given reference point, for example, as described in U.S. Patent 6,530,885 (Entrekin et al.).

[0107] 2D or 3D images are coupled from the scan converter 630, the multiple planar reformatter 632, and the volume Tenderer 634 to an image processor 636 for further enhancement, buffering, and temporary storage for display on an image display 638.

[0108] The graphics processor 640 can generate graphical overlays for display with the ultrasound images. These graphical overlays can include, for example, segmentation results produced as a result of performing segmentation on the ultrasound images or spatial coherence maps of the ultrasound images and / or locations of the iliac bone or femoral head as determined according to the methods herein.

[0109] The graphical overlays can also include other information, such as standard identifying information, e.g., patient name, date and time of the images, imaging parameters, etc. The graphical overlays can also include one or more signals indicating that a target image frame has been obtained and / or that the system 600 is in the process of identifying a target image frame. The graphics processor 640 can receive input from the user interface 624, such as a typed patient name. The user interface 624 can also receive input prompting adjustment of settings and / or parameters used by the system 600. The user interface can also be coupled to the multiplanar reformatter 632 for selecting and controlling the display of multiple multiplanar reformatted (MPR) images.

[0110] Those skilled in the art will appreciate that Figure 6 The embodiments shown in FIG. 6 are merely examples, and the ultrasound system 600 can also include additional components to those shown in FIG. 6, such as a power source or battery. Figure 6

[0111] Figure 7 A method 700 for making measurements of a hip in an ultrasound image is illustrated. The method 700 can be performed, for example, by any of the systems 200 or 600 as described above. Briefly, in a first block 702, the method 700 includes obtaining an ultrasound image of a hip. In a second block 704, the method includes obtaining a spatial coherence map associated with one or more lags of ultrasound waves associated with the ultrasound image. In a third block 706, the method includes determining a quality metric based on the ultrasound image of the hip and the spatial coherence map, wherein the quality metric is indicative of a suitability of the ultrasound image for making measurements of the hip. In a fourth block 708, it is indicated that the ultrasound image is suitable for making measurements of the hip if the quality metric indicates that the ultrasound image is above a threshold quality.

[0112] The functionality of the blocks 702, 704, 606, and 708 is described in detail above with respect to the functionality of the processor 202, and the details therein will be understood to apply equally to embodiments of the method 700.

[0113] In another embodiment, a computer program product comprising a computer- readable medium having computer-readable code embodied therein is provided, the computer-readable code being configured such that, on execution by a suitable computer or processor, the computer or processor is caused to perform the method or methods described herein. ​

[0114] Thus, it will be appreciated that the present disclosure also applies to computer programs, particularly computer programs on or in a carrier, adapted to put the embodiments into practice. The program can be in the form of a source code, an object code, a code intermediate source and an object code (such as in partially compiled form) or in any other form suitable for a used language to implement the methods according to the embodiments described herein.

[0115] It will further be appreciated that such a program can be stored on a carrier in propria lese or on or in equipment that is suitable for the use of the software. The carrier can comprise a storage medium or a medium that is suitable for the transmission of the program element by means of radio or internet. The carrier can be a non-transitory carrier such as a semiconductor (ROM, for example) or a disk (such as a CD or a DVD) or a cassette. The carrier can also be an ephemeral carrier such as an electrical or optical signal, electrical or optical cable, or the air between two electronic devices that exchange information by radio or infrared transmission.

[0116] The carrier of a computer program can be any entity or device that is capable of carrying the program. For example, the carrier can comprise a storage medium, such as a ROM, for example, a CD-ROM or a semiconductor ROM, or a magnetic recording medium, for example, a hard disk. Further, the carrier can be a transmissible carrier such as an electrical or optical signal, electrical or optical cable, or the air between two electronic devices that exchange information by radio or infrared transmission. When the program is embodied in such a signal, the carrier can be constituted by the electrical or optical cable or by the air between the two electronic devices that exchange information by radio or infrared transmission. Alternatively, the carrier can be an integrated circuit in which the program is embedded, the integrated circuit being suitable for performing the relevant method or for use in the implementation of the relevant method.

[0117] Variations to the disclosed embodiments can become apparent to those of ordinary skill in the art from a reading of the drawings, the disclosure, and the claims. In the claims, the term comprising does not exclude the presence of other elements or steps than those claimed. The word a or an preceding an element does not exclude the presence of a plurality of such elements. A processor or other unit performing an action can do so individually or in conjunction with another processor or unit. Although a particular measure is described in a particular dependent claim, this does not exclude the measure from being used in another dependent claim. A computer program can be stored / distributed on a suitable medium, such as an optical storage medium or a solid-state storage medium supplied together with or as part of other hardware, but can also be distributed in other forms such as via the internet or other wired or wireless telecommunication systems. Any reference signs in the claims should not be construed as limiting the scope.

Claims

1. A system for measuring a hip in an ultrasound image, the system comprising: a memory comprising instruction data representing a set of instructions; and a processor configured to communicate with the memory and execute the set of instructions, wherein the set of instructions, when executed by the processor, cause the processor to: obtain an ultrasound image of the hip; obtain a spatial coherence map associated with one or more lags of an ultrasound wave associated with the ultrasound image; determine a quality metric based on the ultrasound image of the hip and the spatial coherence map, wherein the quality metric indicates suitability of the ultrasound image for the measurement of the hip; and if the quality metric indicates that the ultrasound image is above a threshold quality, indicate that the ultrasound image is suitable for the measurement of the hip.

2. The system of claim 1, wherein, the quality metric is based on: a measure of spatial coherence at the iliac bone or femoral head; a measure of contrast between the iliac bone and noise in a region adjacent to the iliac bone; and / or a measure of diameter of the femoral head.

3. The system of claim 2, wherein, the quality metric indicates that the ultrasound image is above a threshold quality in the event that: a coherence value associated with the iliac bone is above a first threshold.

4. The system of claim 2 or 3, wherein, the quality metric indicates that the ultrasound image is above a threshold quality in the event that: the measure of contrast between the iliac bone and noise in a region adjacent to the iliac bone is above a second threshold.

5. The system of claim 2 or 3, wherein, the quality metric indicates that the ultrasound image is above the threshold quality in the event that: the ultrasound image comprises a plane through the femoral head, wherein the diameter of the femoral head is at its maximum.

6. The system of claim 2 or 3, wherein, the quality metric indicates that the ultrasound image is above the threshold quality in the event that: a coherence value associated with the femoral head is above a third threshold.

7. The system of claim 2 or 3, wherein, the processor is further caused to determine a measurement of the hip using the ultrasound image.

8. The system of claim 7, wherein, causing the processor to determine the measurement of the hip using the ultrasound image comprises causing the processor to: combine the ultrasound image and the spatial coherence map into a combined image; segment the iliac bone and femoral head in the combined image; and use the segmentation of the iliac bone and femoral head to make the measurement of the hip.

9. The system of claim 8, wherein, causing the processor to use the segmentation of the iliac bone and the femoral head to make the measurement of the hip comprises causing the processor to: use a Hough transform to fit an ellipse to a segment of the segmentation corresponding to the femoral head.

10. The system of claim 9, wherein, the processor is further caused to: send instructions to a display to cause the display to mark the fitted ellipse and / or segments of the segmentation corresponding to the iliac bone and femoral head onto the ultrasound image; and make the measurement of the hip with respect to the marked fitted ellipse and / or the segmentation corresponding to the iliac bone and femoral head.

11. The system of any one of claims 1 to 3, wherein, causing the processor to obtain an ultrasound image of the hip comprises causing the processor to: receive a plurality of two-dimensional ultrasound images; determine an optimal image for later measurement from the plurality of two-dimensional ultrasound images using a model trained using a machine learning process; and selecting the best image for later measurements as the obtained ultrasound image.

12. The system of any one of claims 1 to 3, wherein, obtaining an ultrasound image of the hip joint; receiving a plurality of two-dimensional ultrasound images; reconstructing a three-dimensional volume from the plurality of two-dimensional ultrasound images using a model trained using a machine learning process; and selecting a slice through the three-dimensional volume as the obtained ultrasound image.

13. The system of any one of claims 1 to 3, wherein, The processor is further caused to use the measurements of the hip joint to diagnose dysplasia, DDH, of the hip joint.

14. A method for measuring a hip joint in an ultrasound image, the method comprising: obtaining an ultrasound image of the hip joint; obtaining a spatial coherence map associated with one or more lags of an ultrasound wave associated with the ultrasound image; determining a quality metric based on the ultrasound image of the hip joint and the spatial coherence map, wherein the quality metric is indicative of a suitability of the ultrasound image for a measurement of the hip joint; and if the quality metric indicates that the ultrasound image is above a threshold quality, indicating that the ultrasound image is suitable for the measurement of the hip joint.

15. A computer program product comprising a computer readable medium having computer readable code embodied therein, the computer readable code being configured such that, on execution by a suitable computer or processor, the computer or processor is caused to perform the method of claim 14.

Images