Spectroscopic dark field imaging

Abstract

The invention relates to an image processing device (1) comprising an input (2) for receiving image data representing a region of interest in a patient's body from a medical X-ray imaging apparatus (100). The image data comprises a first dark-field image obtained for a first X-ray energy spectrum and a second dark-field image obtained for a second, different X-ray energy spectrum. A combination unit (3) provides a combined image representing a medical condition map (e.g. a lung condition map) by combining the first dark-field image and the second dark-field image.

Description

Technical Field

[0001] This invention relates to the field of dark-field imaging (e.g., in medical diagnostic applications). Specifically, this invention relates to image processing apparatus, image processing methods, imaging systems, and related computer program products for medical diagnostic imaging (e.g., for assisting in the detection and diagnosis of different lung conditions). Background Technology

[0002] As is known in the art, X-ray dark-field imaging is a useful and very versatile mode of ionizing radiation imaging (e.g., for medical diagnostic applications). Particularly advantageous is the ability to extract information about structures at scales below the spatial resolution limits of the imaging system from X-ray dark-field image data.

[0003] It is well known that this imaging modality can be used, for example, in the detection of lung diseases, to achieve highly sensitive detection. Considering representative examples of lung imaging, it is noted that the alveoli in the lungs, responsible for gas exchange between blood and inhaled air, are particularly small and delicate structures. While it is difficult or even impossible to directly image these alveoli (at least by non-invasive imaging), useful information about the alveoli and their potential diseases can be conveyed through the ionizing radiation scattering properties of the alveoli, which scatter at small angles. These small-angle scattering properties can be characterized by X-ray dark-field imaging, making this modality particularly useful for the accurate detection of lung diseases. Unfortunately, changes in the observed signal detected by X-ray dark-field imaging can lack specificity. For example, it is not easy to determine the type of lung disease based solely on the dark-field signal. Therefore, there is a need in the art to provide good means and methods to allow the differentiation of different lung diseases, while preferably maintaining high detection sensitivity of dark-field imaging for lung diseases (e.g., sensitivity at least comparable to prior art dark-field imaging methods).

[0004] Dark-field radiography images obtained using methods known in the art allow clinicians to detect deviations from expected lung health, but they cannot differentiate between different types of lung diseases because many such diseases cause very similar changes in the dark-field signal (e.g., localized reduction). For example, emphysema may cause a significant increase in alveolar size, while acute inflammation may cause affected alveoli to fill with fluid or cellular material. Nevertheless, both conditions show similar reductions in the dark-field signal, often requiring additional data for differential diagnosis and thus allowing for the determination of appropriate treatment, management, or disease control for the patient.

[0005] US 2018 / 271465 discloses a prior art method in which X-ray dark-field imaging information is used in a patient's lung examination. In this method, a lung depth map is calculated based on attenuated image data after applying lung segmentation and bone suppression algorithms. The X-ray dark-field image is then normalized using spatially corresponding estimated lung thickness values ​​to obtain a map indicative of lung condition (e.g., chronic obstructive pulmonary disease (COPD)). While COPD is typically not visible on conventional attenuated X-ray images, the normalized dark-field image is particularly sensitive to conditions affecting alveolar microstructure.

[0006] While this specification focuses on lung imaging as a representative example and application, those skilled in the art will understand that the methods and apparatus according to the embodiments can be applied more broadly to various applications. For example, skeletal structures share many similarities with lung structures, where the small-angle scattering behavior of alveoli can be considered analogous to scattering on trabeculae. Furthermore, the complex structures of the human or animal body contain many structures with similar properties, to which the teachings of this disclosure can be applied. Another example of a microstructure that may exhibit similar scattering behavior is the nephron. Summary of the Invention

[0007] The purpose of embodiments of the present invention is to provide easy, effective, efficient and / or good means and / or methods for processing diagnostic images to allow differentiation of different diseases and / or medical conditions (e.g., to assist in the differential diagnosis of lung diseases).

[0008] The advantage of embodiments of the present invention is that the high detection sensitivity of dark-field imaging for lung or other diseases, as is currently achievable in the art, can be advantageously combined with the methods and / or devices according to embodiments of the present invention to achieve good detection specificity (and / or selectivity) for a variety of different conditions or symptoms (e.g., different lung diseases).

[0009] The advantage of embodiments of the present invention is that the methods and / or apparatus according to the embodiments can be easily adapted to work with readily available X-ray imaging devices and or can be easily included in established medical imaging methods and / or workflows.

[0010] The advantage of embodiments of the present invention is that multi-energy (e.g., dual-energy) X-ray attenuation information can be used for image segmentation and / or bone suppression and / or pixel value normalization, while multi-energy X-ray dark-field information (typically acquired in conjunction with the acquisition) can be used to detect and differentiate different types of lung diseases.

[0011] The advantage of embodiments of the present invention is that it is possible to detect and identify lung conditions associated with normal or abnormal alveolar function from different but similar situations.

[0012] The advantage of embodiments of the present invention is that it is able to detect and identify different bone conditions (e.g., osteoporosis and bone edema).

[0013] The advantage of embodiments of the present invention is that, according to embodiments of the present invention, it is easier to distinguish different conditions that present similar or nearly identical images in conventional dark-field X-ray images acquired using single-peak photon energy (i.e., single-energy spectrum), or at least that these different conditions are difficult to distinguish in such images.

[0014] An advantage of embodiments of the present invention is that it is possible to collect information about the normal or abnormal function of alveoli and associated lung conditions from images obtained by an imaging modality that has (for example, although it has) a spatial detection threshold (e.g., spatial resolution) that is insufficient to independently image such alveoli or even to image the microstructure of such alveoli in detail.

[0015] The apparatus, method, system, and computer program product according to embodiments of the present invention achieve the above-mentioned objectives.

[0016] In a first aspect, the present invention relates to an image processing apparatus comprising an input unit for receiving image data representing a region of interest within a patient from a medical X-ray imaging device. The image data includes a first dark-field image obtained for a first X-ray energy spectrum and a second dark-field image obtained for a second X-ray energy spectrum, wherein the first X-ray energy spectrum and the second X-ray energy spectrum are (substantially) different. The image processing apparatus includes a combining unit adapted to provide a combined image representing a condition map by combining the first dark-field image and the second dark-field image. For example, the condition map may be a map of anatomical conditions, physiological conditions, and / or disease conditions, e.g., which conditions may (but are not necessarily) specific to the region of interest.

[0017] In an image processing device according to an embodiment of the present invention, the region of interest may be a chest region, wherein the condition map is a lung condition map.

[0018] In the image processing apparatus according to an embodiment of the present invention, the region of interest may be a skeletal region, and the condition map may be a skeletal condition map, such as a map of osteoporosis and bone edema.

[0019] In an image processing apparatus according to an embodiment of the present invention, the combining unit may be adapted to construct the combined image into a vector value image, wherein the components of the vector value image correspond to the first dark field image and the second dark field image or are calculated based on the first dark field image and the second dark field image.

[0020] In an image processing apparatus according to an embodiment of the present invention, the combining unit may be adapted to determine a dark field signal as a function of photon energy or as a function of correlation length for pixel positions in the combined image.

[0021] In an image processing apparatus according to an embodiment of the present invention, the combining unit may be adapted to calculate a deviation metric between the first dark field image and the second dark field image, wherein the pixel values ​​in the combined image correspond to the deviation metric.

[0022] In an image processing apparatus according to an embodiment of the present invention, the input unit is adapted to receive the image data, wherein the image data includes at least three dark-field images obtained for at least three corresponding different X-ray energy spectra.

[0023] An image processing apparatus according to an embodiment of the present invention may include a controller for controlling the operation of the X-ray imaging apparatus when connected to the X-ray imaging apparatus in order to acquire the image data.

[0024] In an image processing apparatus according to an embodiment of the present invention, the controller may be adapted to: step through a plurality of phase steps by controlling the phase stepping mechanism of the X-ray imaging apparatus; acquire image data from the image detector of the X-ray imaging apparatus for each of the plurality of phase steps; and control the X-ray source of the X-ray imaging apparatus to switch between a first X-ray energy spectrum for at least one of the plurality of phase steps and a second X-ray energy spectrum for at least another of the plurality of phase steps.

[0025] An image processing apparatus according to an embodiment of the present invention may include a segmenter for segmenting a first dark-field image, a second dark-field image, the combined image, and / or another image received via the input unit to identify structures of interest (e.g., lung structures).

[0026] An image processing apparatus according to an embodiment of the present invention may include a classifier (i.e., a classification unit) for labeling each identified structure of interest using classifier labels, the classifier labels being selected by the classifier from a plurality of classifier labels based on the combined image, the plurality of classifier labels corresponding to different conditions (e.g., different lung conditions).

[0027] In a second aspect, the present invention relates to an X-ray imaging system comprising an image processing apparatus and an X-ray imaging device according to an embodiment of a first aspect of the present invention, the X-ray imaging device being configured to: acquire multiple X-ray dark-field images of a patient (e.g., chest or skeletal region...) of a corresponding plurality of X-ray energy spectra, and provide the X-ray dark-field images to the input unit of the image processing apparatus.

[0028] In an imaging system according to an embodiment of the present invention, the X-ray imaging apparatus may include an X-ray source and an X-ray (e.g., image) detector.

[0029] In an image processing apparatus according to an embodiment of the present invention, the X-ray detector may be an energy-resolved photon counting detector.

[0030] In an image processing apparatus according to an embodiment of the present invention, the X-ray imaging device may include a dual-energy imaging device or an energy spectrum imaging device.

[0031] An image processing apparatus according to an embodiment of the present invention may include a registration unit for spatially registering the plurality of X-ray dark-field images in order to compensate for patient movement between image acquisitions.

[0032] In an image processing apparatus according to an embodiment of the present invention, the X-ray imaging apparatus may include a grating interferometer apparatus for differential phase contrast and dark field imaging.

[0033] In a third aspect, the present invention relates to a clinical workstation comprising an image processing device according to an embodiment of the first aspect of the present invention.

[0034] In a fourth aspect, the present invention relates to a method for processing image data, such as a computer-implemented method. The method includes: acquiring image data representing a region of interest within a patient's body from a medical X-ray imaging apparatus. The image data includes a first dark-field image obtained for a first X-ray energy spectrum and a second dark-field image obtained for a second X-ray energy spectrum, wherein the first X-ray energy spectrum and the second X-ray energy spectrum are significantly different. The method includes providing a combined image representing a condition map by combining the first dark-field image and the second dark-field image.

[0035] In the method according to an embodiment of the present invention, the region of interest may be the chest region. The condition map may be a lung condition map.

[0036] In a fifth aspect, the present invention relates to a computer program product, which, when run on a suitable processor, is used to perform a method according to a fourth aspect of the invention (e.g., computer-implemented).

[0037] In a sixth aspect, the present invention relates to a medical diagnostic image of a region of interest within a patient's body (e.g., a chest region, a skeletal region, etc.), comprising a combination of a first dark-field image obtained for a first X-ray energy spectrum and a second dark-field image obtained for a second X-ray energy spectrum, wherein the first X-ray energy spectrum and the second X-ray energy spectrum are (sufficiently) different.

[0038] When it is mentioned that the first and second energy spectra are significantly different (i.e., sufficiently different), those skilled in the art will understand that this means the difference between the first and second X-ray energy spectra is sufficient to obtain supplementary information from images obtained for these different energy spectra. Therefore, the energy spectra may differ by at least 1 kV, preferably at least 5 kV, and preferably by at least 10 kV (e.g., at least 20 kV), or even more (e.g., at least 50 kV), in terms of the average or peak energy (kVp) of the spectra. It should be noted that even for the same peak photon energy (kVp), the energy spectra can still be significantly different, for example, by filtering with different beam-tuning filters (i.e., different beam filter selections).

[0039] It should also be understood that embodiments of the present invention can also relate to image data being acquired and / or used, including multiple dark-field images obtained for a first X-ray energy spectrum, a second X-ray energy spectrum, an optional third X-ray energy spectrum, an optional fourth X-ray energy spectrum, etc. In other words, the number of dark-field images obtained for different energy spectra can be extended to any number (greater than one), for example, as a particular application might require or benefit from. Those skilled in the art will understand that the number of energy spectra used can be selected based on a trade-off between achieving accuracy and / or further identification information by combining a large number of images and the increased complexity and cost (e.g., costs in terms of operation time, processing resources, energy consumption, and / or clinician evaluation time).

[0040] The independent and dependent claims describe the specific and preferred features of the invention. Features of the dependent claims can be combined with features of the independent and other dependent claims, as long as deemed appropriate, and not necessarily only as expressly stated in the claims. Attached Figure Description

[0041] Figure 1 The apparatus according to an embodiment of the present invention is illustrated schematically.

[0042] Figure 2 A system according to an embodiment of the present invention is shown.

[0043] Figure 3 Relevant functions for different types of microstructures are shown to illustrate various aspects of embodiments of the present invention.

[0044] Figure 4 A method according to an embodiment of the present invention is shown.

[0045] The accompanying drawings are illustrative and not restrictive. Elements in the drawings are not necessarily presented to scale. The invention is not necessarily limited to the specific embodiments shown in the drawings. Detailed Implementation

[0046] While exemplary embodiments are described below, the invention is limited only by the claims. The claims are thus expressly incorporated into this specific embodiment, wherein each claim, and each combination of claims permitted by the dependent structures defined by the claims, forms a separate embodiment of the invention.

[0047] The word "comprising" as used in the claims is not limited to the features, elements, or steps described herein, and does not exclude additional features, elements, or steps. Therefore, this specifies the presence of the mentioned features, but does not exclude the further presence or addition of one or more features.

[0048] In this specific embodiment, various specific details are presented. Embodiments of the present invention can be implemented without these specific details. Furthermore, for the sake of clarity and brevity, it is not necessary to describe well-known features, elements, and / or steps in detail.

[0049] In a first aspect, the present invention relates to a medical image processing apparatus. This medical image processing apparatus can (e.g., by using a dual-energy or spectral X-ray imaging system) advantageously utilize dark-field image information acquired against two or more different X-ray energy spectra to improve the specificity of X-ray dark-field imaging for diagnostic purposes in evaluating the condition of a predetermined region of interest (e.g., a lobe or multiple lobes of the lung or another organ or body structure). For example, different lung conditions may affect the alveoli of the lung differently, and these different effects can be detected and identified relative to each other after processing by the medical image processing apparatus according to embodiments of the invention.

[0050] refer to Figure 1An illustrative image processing apparatus 1 according to an embodiment of the present invention is shown. The image processing apparatus 1 may include a computer, for example, a computer specifically programmed to perform the functions of the apparatus. Such a computer may include input and output units, for example, communication interfaces(one or more) for receiving and transmitting data, for example, via a data carrier and / or a communication network interface. Such input and output units may also include user interface hardware, for example, human-computer interaction devices (e.g., keyboard, mouse, voice interpreter, touch interface, gyroscope, or accelerometer, etc.) for receiving input from a human user, a monitor for presenting information to a user, a speaker, a printer for drawing information onto a physical carrier (e.g., paper), a 3D printer for generating a 3D physical model of the data, and other such elements known in the art.

[0051] A computer may include a general-purpose processor for executing instructions (e.g., computer code) and memory for storing such instructions. A computer may include memory for storing data, such as memory for manipulating data according to instructions. The device is not necessarily limited to a general-purpose computer and may also include application-specific integrated circuits (ASICs) and / or configurable processing hardware, such as field-programmable gate arrays (FPGAs). Furthermore, device 1 may be included in a single processing device (e.g., a computer), but may also be distributed across multiple such devices that are operatively interconnected, for example, such that the processing described herein is performed through the combined action of a server and(one or more) client devices or a parallel processing system (e.g., a computing cluster).

[0052] In a second aspect, the present invention also relates to a medical imaging system. Figure 2A system 10 according to an embodiment of the present invention is schematically illustrated. System 10 includes an image processing apparatus 1 according to an embodiment of the first aspect of the present invention. System 10 also includes an X-ray imaging apparatus 100 configured to supply image data to an input section 2 of the image processing apparatus 1. It should be understood that the "input section" can be a physical connection (e.g., using a signal line, connector, etc.), but only an operational connection is required (e.g., it can refer to a connection via a data communication network, a wireless connection, or even a connection via a removable data carrier that can be connected to the imaging apparatus 100 to receive and store image data, and can (e.g., subsequently) be connected to the processing apparatus 1 to provide access to the stored data). Furthermore, the image processing apparatus 1 can be integrated with the processing and / or control components of the imaging apparatus 100, for example as a software component configured to run on a computing device, which can also be used to run other functions of the imaging apparatus. For example, “input unit” 2 may refer only to shared memory (volatile or non-volatile memory), through which a part of the imaging system makes image data available to processing device 1 (e.g., in the form of another part of the imaging system, possibly including only software).

[0053] X-ray imaging apparatus 100 (e.g., X-ray image acquisition device) may include an X-ray source and an X-ray detector. The X-ray imaging apparatus can be configured to acquire X-ray dark-field information of a patient's chest at different representative energies (e.g., for different X-ray energy spectra). While most X-ray imaging systems have adjustable photon energy settings (e.g., kVp), it should be noted that the X-ray imaging apparatus can be particularly well-suited to acquire X-ray dark-field information concurrently (i.e., simultaneously or at least rapidly sequentially (e.g., with a delay of less than 10 seconds, preferably less than 1 second, even more preferably less than 100 milliseconds, even more preferably less than 10 milliseconds) for different energy spectra. The X-ray imaging apparatus can also be adapted to concurrently acquire X-ray attenuation information and / or X-ray phase contrast information. It should be noted that the X-ray imaging apparatus can also be adapted to sequentially acquire X-ray dark-field image information of different representative energies at relatively low rates (e.g., with a delay of more than 10 seconds between acquisitions of different energy spectra). X-ray imaging apparatus 100 or device 1 may include a registration unit to compensate for patient movement between acquisitions (i.e., to correct the first dark-field image and / or the second dark-field image to ensure good alignment of corresponding image features in different images). The latter situation may be particularly useful if the X-ray imaging apparatus is not particularly well-suited for simultaneously acquiring images for different energy spectra (e.g., energy ranges or energy sub-boxes) or for switching between energy spectra at high switching rates.

[0054] Device 1 includes an input unit 2 for receiving image data from a medical X-ray imaging apparatus 100 representing a region of interest (e.g., the patient's chest region, such as the patient's chest or a portion thereof) within a patient's body. The image data includes a first dark-field image obtained for a first X-ray energy spectrum and a second dark-field image obtained for a second X-ray energy spectrum. The second X-ray energy spectrum is substantially different from the first X-ray energy spectrum. "Image data" can refer to any type of information, such as typical digital data derived from an X-ray imaging apparatus and suitable for constructing a visual representation of the region of interest (e.g., the patient's chest region (or an imaged portion thereof)). In other words, the term should not be narrowly interpreted as referring only to a matrix of pixel data, but can also be, for example, compressed, encrypted, and / or encoded image data, which can be in processed formats (e.g., segmented image data, annotated image data), raw formats (e.g., data signals directly recorded by an image sensor or detector), vector-value formats (e.g., a matrix containing multiple values ​​for each pixel), and / or parametric representations (e.g., iso-intensity (iso-value) curves, segmented contours in an image, and / or parametric descriptions of surfaces).

[0055] A first dark-field image and a second dark-field image can be obtained by collecting small-angle X-ray scattering information for different source energy spectra (e.g., generating the first and second dark-field images based on different peak energies and / or material filtering parameters), or by collecting small-angle X-ray scattering information by radiation probing with detectors of different configurations to be sensitive to different X-ray energy spectra. Therefore, the first and second dark-field images can correspond to dark-field images obtained for different source energy spectra, or can be obtained from raw detector data acquired based on different detector response functions (e.g., detector sensitivity as a function of photon energy), or can correspond to combinations thereof (e.g., different radiation source energy spectra and collecting detector response functions). For example, a dark-field image can correspond to different energy channels of an energy-resolved detector (e.g., an energy-resolved photon counting detector), or to (substantially) different combinations of these energy channels.

[0056] As will be clearly understood by those skilled in the art, the first dark-field image and the second dark-field image both represent the same region (e.g., the chest region) of the same patient at substantially the same time (e.g., in a single chest imaging examination), and are preferably naturally co-registered, for example, by using substantially the same imaging geometry and being acquired at substantially the same time (even if, as is generally understood in the art, small and inconspicuous differences in imaging geometry and / or small time intervals between image acquisitions for different energy spectra are obviously produced due to some system design choices).

[0057] For example, the X-ray imaging device 100 may be a dual-energy imaging device configured to image a patient (or at least the patient's chest region) with two different (e.g., different mean or different peak) photon energies.

[0058] The X-ray imaging apparatus 100 can also be adapted to acquire two or more different images with corresponding representative photon energies, for example, it can be an energy spectrum imaging device capable of acquiring two or more images with different energies. Therefore, the input unit can be adapted to receive image data, which includes two or more dark-field images obtained for two or more different X-ray energy spectra respectively.

[0059] This imaging apparatus is suitable for dark-field imaging and / or phase-contrast imaging. For example, the system may include a grating interferometer, such as comprising two or more gratings. For example, a three-grating interferometer apparatus may be used, as is known in the art. This approach is particularly advantageous because its implementation does not impose any troublesome limitations on the X-ray source 101 and detector 102; for example, it can, in principle, be used with conventional X-ray sources 101 and detector 102 suitable for conventional attenuated X-ray imaging. For example, the imaging apparatus may be adapted (e.g., using a Talbot-Lau type interferometer) for differential phase-contrast imaging. For example, the X-ray imaging apparatus 100 may include a source grating 103, which is typically located between the X-ray source 101 and the target 106 to be imaged (e.g., a patient's chest), and the X-ray imaging apparatus 100 may include a phase grating 104 and an absorption (or analyzer) grating 105, which are typically located between the target 106 and detector 102. Based on established knowledge in the art, gratings can typically have well-defined grating line periods and can be positioned at well-defined distances relative to each other, to the source, and to the detector to achieve the desired effect. Source gratings can create independent, coherent, but uncoherent, virtual source arrays. However, it should be understood that alternatives exist for achieving the same or similar coherence in X-ray beams. Phase gratings can be used to split the beam into its first diffraction order. Absorption gratings can act as transmission masks to transform localized fringe variations into detectable intensity values, where these fringe patterns can be generated by the Talbot effect, particularly when the distance between the phase grating and the analyzer grating corresponds to the Talbot distance.

[0060] X-ray imaging apparatus 100 may include a phase stepping mechanism to translate a phase grating about an analyzer grating (e.g., in small steps) (or vice versa). This allows raw image detector data to be collected for each phase step, from which X-ray attenuated images, phase-contrast images, and dark-field images can be reconstructed using methods known in the art.

[0061] Dark-field imaging does not necessarily require a phase-stepping mechanism. For example, fringe analysis methods, such as spatial fringe scanning, can be used. Furthermore, different dark-field images for different energy spectra are not necessarily acquired using the same technique. For instance, phase stepping may be used for one energy spectrum, while spatial fringe scanning may be used for another. Even if one image in the dataset has a lower spatial resolution, this can reduce acquisition time. For example, an image acquired at the highest spatial resolution can be used to detect features of interest (e.g., lung abnormalities), while other images or images with lower spatial resolution can be used to determine the type of abnormality (once the abnormality is detected).

[0062] Acquisition time can also be reduced by altering the energy spectrum (e.g., switching source energies) while collecting the phase step curve, rather than collecting the entire phase step curve for each energy spectrum individually (i.e., a fixed number of phase step points). Therefore, for example, two energy spectra (energys) can be used alternately while undergoing phase stepping. It's also important to note that the number of phase step samples for each energy spectrum does not need to be equal. For example, a large number of phase steps can be determined for one energy spectrum (e.g., a reference energy spectrum), while only a small number of phase steps are used for another energy spectrum (and possibly other spectra). For example, the number of phase step samples required to determine a dark field signal can be less than the number required to determine a phase contrast signal. Therefore, a reference energy spectrum can be used to obtain high-quality attenuation, phase contrast, and dark field images, while another(one or more) energy spectra can be used to obtain only one or more additional dark field images, which may have lower spatial resolution or potentially more noise.

[0063] Device 1 may include a controller 7 for controlling the operation of the X-ray imaging apparatus 100 (i.e., when the device is operatively connected to the X-ray imaging apparatus) to acquire image data. For example, controller 7 may be adapted to control an X-ray source and / or an X-ray image detector. Furthermore, controller 7 may be adapted to control the phase-stepping mechanism of the X-ray imaging apparatus. For example, controller 7 may be adapted to acquire (raw) image data from the image detector for each of a plurality of phase steps, such as controlling the phase-stepping mechanism to perform a phase step before acquiring each image. As is known in the art, dark-field images, as well as attenuation and phase-contrast images, can be calculated from such raw image data collected for multiple phase steps. Therefore, device 1 may include a pixel value calculator to determine a dark-field signal value (and possibly attenuation and / or phase-contrast value) for each image pixel based on the raw image data (e.g., different detection signal values ​​corresponding to each phase step in the phase steps). The controller 7 may be adapted to control the X-ray source to switch, for example, between a first X-ray energy spectrum for at least one phase step (e.g., a first set) of multiple phase steps and a second X-ray energy spectrum for at least another phase step (e.g., a second set) of multiple phase steps. For example, the controller may alternate between two X-ray energy spectra as the phase steps are iteratively performed.

[0064] However, other methods for obtaining phase-contrast and / or dark-field image information (e.g., different grating interferometer designs or phase-contrast / dark-field imaging systems different from grating interferometer designs) are known in the art and are not necessarily excluded from embodiments of the present invention. Dark-field and phase-contrast imaging systems are considered well-known in the art and will therefore not be discussed in detail.

[0065] It should be noted that, typically, phase-contrast imaging information and dark-field imaging information can be obtained simultaneously by the same system through specific but different processing of the acquired raw image data. Similarly, attenuation image data can typically be obtained from the same image data. However, for the purposes of this application, the X-ray imaging apparatus 100 should be able to generate dark-field image information or be able to derive raw data of such dark-field images from it. Nevertheless, it will be appreciated that phase-contrast information and attenuation information can be obtained (substantially) simultaneously, even if this is not strictly necessary, and that both phase-contrast and attenuation information can be presented as useful information to the user (e.g., a clinician using the device according to an embodiment of the invention). As is known in the art, dark-field image information is generated by ultra-small angle scattering (e.g., multiple refractions on microstructures), while attenuation image contrast is caused by local differences in photon absorption, and phase contrast is sensitive to changes in electron density, which may be caused, for example, by local changes in refractive properties (e.g., refractive index).

[0066] It will be recognized that these three distinct complementary imaging modalities (which can be advantageously acquired in conjunction with each other) provide insights into different properties of the imaged target. The dark-field imaging component provides insights into the structural properties of the imaged target (e.g., lung tissue), structures that appear at scales significantly smaller than the spatial resolution of the imaging system (i.e., smaller than the system's spatial detection threshold).

[0067] The dark field signal V (e.g., for a given geometric ray or pixel location) is the so-called correlation length ξ. corr Functions:

[0068] y(ξ corr )=exp(σt(G(ξ corr )-1))

[0069] Where t is the sample thickness (as experienced by the ray at hand), and σ is the total scattering probability per unit length, which is given by the formula Given. Furthermore, the differential cross section can be expressed as:

[0070]

[0071] The correlation function can be calculated as And typically, the relevant length can be calculated as Where λ is the wavelength, p2 is the grating period of the absorption grating, and It is the distance between the sample and the absorption grating.

[0072] The grating period is typically predetermined and fixed. Furthermore, the distance between the sample and the absorption (analyzer) grating can also preferably be fixed to maintain, for example, a constant amplification factor. However, spectral imaging systems such as dual-energy imaging devices (but not limited to) allow for the independent detection of different wavelengths (or spectral components) of the X-ray beam. Therefore, it is possible to determine the dark field signal simultaneously (or at least concurrently) for different correlation lengths.

[0073] If X-ray radiographs are obtained at (at least) two different X-ray energy spectra, the dark-field signal will change due to different correlation lengths, implied by the different X-ray energies. Fitting a function to the measured dark-field signal produces additional information about the sample's microstructure; that is, it complements the information about the microstructure already provided by the dark-field signal obtained at single-photon energy or single-energy spectra, allowing, for example, the microstructure to be better characterized by this additional information. For example, the correlation functions for shell-like spheres (healthy alveoli) and solid spheres (fibrotic tissue, etc.) are very different and are shown below. Figure 3The correlation function G(z) for sphere 31 and shell 32 is illustrated. The x-axis has been normalized to units z / R of the radius R of the sphere (and shell). In this example, the inner diameter of the shell (plotted line 32) corresponds to 0.9 times the outer diameter R. It can be seen that some conditions of the lungs can well (or better) conform to the example of a solid sphere (e.g., when scattering is caused by fluid-filled alveoli or fibrotic tissue), while other conditions of the lungs can well (or better) conform to the example of an empty shell (e.g., healthy alveoli). However, it should be noted that this is merely one example illustrating that different microstructures can correspond to substantially different correlation functions, and therefore different microstructures can be identified based on dark-field images for different correlation lengths (i.e., different energy spectra, e.g., different energies).

[0074] X-ray imaging apparatus 100 may include an X-ray source 101 and an X-ray detector 102. The X-ray source and / or X-ray detector may be adapted to perform dual-energy and / or spectral imaging, i.e., adapted to image the same target (e.g., a patient's chest) at different representative photon energies and substantially simultaneously (or to acquire images at negligible time intervals to obtain naturally aligned and registered images at different energies). For example, the X-ray source may be adapted to rapidly switch between at least two different photon spectral ranges (e.g., by allowing the peak energy kVp to vary between different settings at a high switching rate). The different spectral ranges may also be characterized (alternatively or additionally) by other characteristics besides differing peak energies alone (e.g., different X-ray filters, anode target properties, spot size, etc.). The X-ray detector may be adapted to switch between different response characteristics to selectively sensitize different X-ray spectral ranges, or may be adapted to simultaneously acquire (raw) image data for different spectral ranges (e.g., by resolving multiple different energy windows for the same spatial point). For example, the X-ray detector may have different sensor elements that have different sensitivities to the incident X-ray radiation (inherently or according to the filter design), or be able to directly determine the energy of the incident photons. It should be noted that, as is known in the art, for this purpose, a dual-energy or spectral imaging device can be implemented by specifically adjusting the X-ray source 101 or the X-ray detector 102, or both the source and the detector. Dual-energy and spectral imaging systems are well known in the art and therefore will not be described in detail herein.

[0075] X-ray imaging apparatus 100 may be a computed tomography system, for example, comprising a gantry that rotates around a source, detector, and additional image-forming elements (e.g., interferometer gratings) around a target to be imaged (e.g., a patient). Such a system may also include an image reconstructor, for example, a processor or computer device adapted to reconstruct tomographic images from image data acquired by the detector from multiple different projection angles (e.g., corresponding to different orientations of the gantry). However, X-ray imaging apparatus 100 may also be a digital projection radiography system, which may not necessarily have tomographic reconstruction capabilities. For example, X-ray imaging apparatus 100 may be adapted to acquire a first dark-field image and a second dark-field image in a single projection view of a patient's chest region (e.g., anterior-posterior (AP) or posterior-anterior (PA) projection view image acquisition). Clearly, the user (e.g., a radiologist) may decide to capture multiple views from different angles, but the X-ray imaging apparatus 100 according to embodiments of the invention does not necessarily require the ability to run coordinated (e.g., automated) tomographic scans (e.g., spiral scans). Therefore, the dark-field images (first, second, other, etc.) mentioned in this disclosure may refer to projection images, but may also refer to tomographic reconstruction images.

[0076] Device 1 includes a combination unit 3 for calculating a combined image (e.g., a lung condition map) of a first dark-field image and a second dark-field image, such that, for example, a user can easily identify different lung conditions when evaluating the combined image. For example, the combined image may show different lung conditions, such as emphysema, COPD, and / or acute inflammation (but is not limited to these).

[0077] Combining unit 3 can construct a vector-valued image based on the first dark-field image and the second dark-field image (and optionally, one or more additional dark-field images). For example, such a vector-valued image can be a color image, where different color components correspond to pixels in different dark-field images. However, the vector values ​​in this image can also be constructed using more complex algorithms, such as displaying the average value of the dark-field images as pixel intensity (“value”) and displaying differences between dark-field images (e.g., in a hue-saturation-value representation) as color hue or color saturation. If additional images are obtained, different combinations can be considered to allow for visualization of the data. For example, a contrasting image or attenuation image can be used as one color component (e.g., “value” or “red”), while the dark-field images can be encoded in other color components (e.g., “hue” and “saturation” or “green” and “blue”).

[0078] For pixel locations in the combined image, the combining unit 3 can be adapted to determine a representation of the dark field signal as a function of photon energy (e.g., mean or peak energy) or as a function of correlation length. For example, while it is possible to combine a limited number of dark field images in a vector-valued image for visual evaluation, this combination may be unsuitable for a larger number of dark field images, for example, due to the limitations of human vision requiring the selection of a limited number of features. However, the combined image can include (e.g., at least for pixels or regions of interest) a detailed functional representation, allowing the user to select points in the image to view the corresponding functional map.

[0079] For example, input unit 2 may be adapted to receive image data, wherein the image data includes multiple dark-field images (e.g., more than two, at least four, at least eight, at least 16, at least 32, at least 64, at least 128, at least 256, etc.) obtained for corresponding multiple (different) X-ray energy spectra. For example, image data may be obtained from an energy-resolved detector, and each image may include a dark-field component calculated based on different energy bins (or sets of bins) collected by the energy-resolved detector. Therefore, the representation of the dark-field signal as a function of energy may include a function specified by the dark-field values ​​obtained for each energy bin, but may also include a parameter definition of the function determined based on that data. For example, such a parameter definition may include spline representation, Fourier representation, linear, quadratic or cubic interpolation, Gaussian mixture representation, or another suitable form. The combining unit may, for example, fit a parameter model to the data obtained for each pixel and store the obtained parameters in the combined image.

[0080] Combining unit 3 can calculate a deviation metric between the first dark-field image and the second dark-field image, and store this deviation metric as a pixel value (or a component of a pixel value) in the combined image. For example, this deviation metric can be a difference, an absolute difference, a ratio, and the logarithm of the difference, and / or another suitable operation for comparing two values ​​(in absolute or relative terms). It can be seen that a ratio can be a useful metric, for example, dividing a dark-field signal obtained for relatively low photon energy (e.g., longer wavelength, longer correlation length) by a dark-field signal obtained for relatively high photon energy (e.g., shorter wavelength, shorter correlation length) (for the corresponding pixel position). This ratio can be used as an approximation or substitute for the slope of the correlation function, which can encode useful information about the nature of the underlying condition (see, for example...). Figure 3 (significantly different slopes near z / R = 0).

[0081] Composite images can be suitable for presentation to users (e.g., radiologists), but can also be technical images for further image processing (e.g., image segmentation).

[0082] Device 1 may also include a segmenter 4 for segmenting an image to identify lung structures of interest, such as regions affected by abnormal lung conditions. The image used by the segmenter may be at least one of a dark-field image, a composite image, or another image received via an input (e.g., an attenuated image or a phase-contrast image). The segmenter may also use different images (e.g., in a stepwise segmentation method or as a vector-value composite image). For example, X-ray attenuation information may be used to determine a lung mask image, for example, as a mask (e.g., based on a composite image) to limit further segmentation of different lung structures of interest.

[0083] Referring to US 2018 / 271465, the device may further include a normalization unit adapted to normalize the first dark-field image, the second dark-field image, and / or any additional dark-field image using a lung depth map for normalization. As described in detail in the aforementioned patent application, such a lung depth map can be calculated by applying a bone suppression algorithm (which can be advantageously performed based on spectral attenuation information (e.g., spectral attenuation information already available in the same raw detector data acquired for calculating the first and second dark-field images)) and applying a radiation attenuation model to boneless lung image data.

[0084] Device 1 may also include a classifier 5, which is used to label segmented lung structures based on a composite image using multiple classifier labels. The classifier labels correspond to different lung conditions, such as (selected from two or more of the following): inflamed lung tissue, emphysema, lung cancer nodules, COPD, cystic fibrosis, bronchiectasis, pleural effusion, etc. The classifier may be adapted to annotate the composite image using the labeled segmented lung structures to obtain an annotated image.

[0085] Therefore, classifier 5 can select one of several possible labels for each detected lung structure based on the combined image. However, classifier 5 can also consider other image information, such as one or more attenuation images (e.g., for different energy spectra), one or more phase-contrast images (e.g., for different energy spectra), or even images obtained through different modalities (e.g., preferably images co-registered with other images used).

[0086] Device 1 may include an output unit 6 for outputting combined and / or segmented and / or annotated images. The output unit may include a data storage device, a network connection, a data bus, or other types of digital communication and / or storage interface. The output unit may also include a user interface device (e.g., a computer monitor, printer, etc.) for displaying the combined images. As is known in the art, such a user interface may also include a module for receiving (e.g., interactive) commands from a user. Thus, the combined images can be presented to a user (e.g., a radiologist) for review. One advantage is that by combining dark-field images targeting different X-ray energies (different spectra), users can more easily distinguish different lung conditions without requiring additional imaging (e.g., further examination using different imaging modalities) or further examinations (e.g., functional lung tests, biopsies, etc.).

[0087] In a third aspect, the present invention relates to a clinical workstation comprising the image processing device 1 described above. The clinical workstation may be adapted to present visual information, for example, an imaging visualization workstation. The workstation may include one or more graphical display devices (e.g., a monitor, one or more user interface devices (e.g., a keyboard and / or mouse) and / or other human-machine interface devices known in the art) and a processor. The workstation may be implemented by a computer, smartphone, tablet computer, web server computer, and / or a combination thereof. Such a clinical workstation may be suitable for use in or integrated into radiology suites, biopsy laboratory suites, operating room suites, radiotherapy planning and / or operating systems, etc.

[0088] refer to Figure 4 In a fourth aspect, the present invention relates to a method 20 for processing image data, for example, a computer-implemented method. Method 20 includes obtaining 21 image data representing a region of interest of a patient from a medical X-ray imaging apparatus. The image data includes a first dark-field image obtained for a first X-ray energy spectrum and a second dark-field image obtained for a second X-ray energy spectrum, wherein the first X-ray energy spectrum and the second X-ray energy spectrum are substantially different. The method includes providing 22 a combined image representing a lung condition map by combining the first dark-field image and the second dark-field image.

[0089] In view of the above description relating to the device and / or system according to embodiments of the present invention, the details of other features of the method according to embodiments of the present invention or the features described above should be clear.

[0090] In a fifth aspect, the present invention relates to a computer program product, which, when run on a suitable processor, is used to perform a method (e.g., computer-implemented) according to an embodiment of the invention.

[0091] In a sixth aspect, the present invention relates to a medical diagnostic image of a patient’s chest region, the medical diagnostic image comprising a combination of a first dark-field image obtained for a first X-ray energy spectrum and a second dark-field image obtained for a second X-ray energy spectrum, wherein the first X-ray energy spectrum and the second X-ray energy spectrum are different.

Claims

1. An image processing device (1), comprising: An input unit (2) is configured to receive image data representing a region of interest within a patient's body from a medical X-ray imaging apparatus (100), wherein the image data includes a first dark-field image obtained for a first X-ray energy spectrum and a second dark-field image obtained for a second X-ray energy spectrum, wherein the first X-ray energy spectrum and the second X-ray energy spectrum are different. Combination unit (3) is used to provide a combined image representing a condition map by combining the first dark field image and the second dark field image.

2. The image processing apparatus according to claim 1, wherein, The region of interest is the chest region, and the condition diagram is a lung condition diagram.

3. The image processing apparatus according to claim 1 or 2, wherein, The combining unit (3) is adapted to construct the combined image into a vector value image, wherein the components of the vector value image correspond to the first dark field image and the second dark field image or are calculated based on the first dark field image and the second dark field image.

4. The image processing apparatus according to claim 1 or 2, wherein, The combining unit (3) is adapted to determine a dark field signal as a function of photon energy or as a function of correlation length for the pixel position in the combined image.

5. The image processing apparatus according to claim 1 or 2, wherein, The combining unit (3) is adapted to calculate a deviation metric between the first dark field image and the second dark field image, wherein the pixel values ​​in the combined image correspond to the deviation metric.

6. The image processing apparatus according to claim 1 or 2 further includes a controller (7) for controlling the operation of the X-ray imaging apparatus (100) when connected to the X-ray imaging apparatus (100) in order to acquire the image data.

7. The image processing apparatus according to claim 6, wherein, The controller (7) is adapted to: step through a plurality of phase steps by controlling the phase stepping mechanism of the X-ray imaging apparatus; acquire image data from the image detector of the X-ray imaging apparatus for each of the plurality of phase steps; and control the X-ray source of the X-ray imaging apparatus to switch between a first X-ray energy spectrum for at least one of the plurality of phase steps and a second X-ray energy spectrum for at least another of the plurality of phase steps.

8. The image processing apparatus according to claim 1 or 2, comprising a segmenter (4) for segmenting the first dark field image, the second dark field image, the combined image and / or another image received via the input unit to identify a structure of interest.

9. The image processing apparatus of claim 8, comprising a classifier (5) for labeling each identified structure of interest using classifier labels, the classifier labels being selected by the classifier from a plurality of classifier labels based on the combined image, the plurality of classifier labels corresponding to different conditions.

10. An imaging system (10) comprising an image processing device (1) according to any one of the preceding claims and an X-ray imaging apparatus (100), the X-ray imaging apparatus being configured to: acquire multiple X-ray dark-field images of regions of interest within a patient for corresponding multiple X-ray energy spectra, and provide the X-ray dark-field images to the input unit (2) of the image processing device.

11. The imaging system according to claim 10, wherein, The X-ray imaging device (100) includes an X-ray source (101) and an X-ray detector (102).

12. The imaging system according to claim 11, wherein, The X-ray detector is an energy-resolved photon counting detector.

13. The imaging system according to any one of claims 10 to 12, wherein, The X-ray imaging apparatus (100) includes grating interferometer devices (103, 104, 105) for differential phase contrast and dark field imaging.

14. A method (20) for processing image data, the method comprising: Image data representing regions of interest within a patient's body is obtained from a medical X-ray imaging device (21), wherein the image data includes a first dark-field image obtained for a first X-ray energy spectrum and a second dark-field image obtained for a second X-ray energy spectrum, wherein the first X-ray energy spectrum and the second X-ray energy spectrum are different, and (22) A combined image representing a condition map is provided by combining the first dark field image with the second dark field image.

15. A computer program product, when run on a processor, wherein the computer program product is configured to perform the method according to claim 14.

Images