METHOD FOR PRODUCEING A SYNTHETIC MAMMOGRAM BASED ON A DUAL-ENERGY TOMOSYNTHESE RECORDING
Patent Information
- Application Number
- DE502019012428
- Authority / Receiving Office
- DE · DE
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2019-09-30
- Publication Date
- 2024-11-21
- Estimated Expiration
- 2039-09-30
AI Technical Summary
Conventional methods for generating synthetic mammograms from tomosynthesis images often result in insufficient clarity of tumor-relevant areas due to superimposition from dense tissue, which hinders tumor visibility and malignancy assessment.
A method utilizing dual-energy tomosynthesis images, where low-energy and high-energy X-ray spectra are used to create a three-dimensional probability map, allowing for weighted projection and subtraction to generate a synthetic mammogram with reduced overlay of tumor-bearing areas, leveraging iodine contrast agent distribution for enhanced lesion visualization.
This approach provides a clearer representation of tumor structures, improving diagnostic accuracy by reducing overlays and enhancing the visibility of tumor margins, thus aiding in malignancy assessment.
Description
[0001] The invention relates to a method for generating a synthetic mammogram based on a dual-energy tomosynthesis image of an examination area and a mammography system therefor, which enable the calculation of a synthetic mammogram with a reduced overlay of tumor-affected areas.
[0002] Contrast-enhanced dual-energy mammography of the breast (CEDEM / Contrast Enhanced Dual Energy Mammography, CEDM / Contrast Enhanced Digital Mammography) is a diagnostic X-ray technique in which high-energy (HE) X-ray images are created after contrast medium administration and low-energy (LE) images are subtracted from them in a weighted manner. The goal is to improve the detection of lesions by visualizing contrast enhancement within the blood vessels supplying the tumors. Dual-energy imaging in digital mammography (DM) using iodine as a contrast agent, also known as the dual-spectrum method, is expected to improve diagnosis for radiologists as well as improve sensitivity and specificity. CEDEM can, for example, be performed in the form of a so-called titanium contrast-enhanced mammography.
[0003] The adaptation of the above procedure from 2D to 3D is called CEDET (Contrast Enhanced Dual Energy Tomosynthesis). Two tomosynthesis scans are acquired with different X-ray spectra or different tube voltages, which, analogous to CEDEM, visualize iodine contrast enhancement. The resulting subtraction volume based on the two tomosynthesis scans corresponds to a 3D map or 3D distribution of the tumor-relevant regions. In a digital tomosynthesis scan (Digital Breast Tomosynthesis, DBT), a synthetic mammogram is usually calculated in addition to the (three-dimensional) tomographic images. This synthetic mammogram corresponds to a two-dimensional image of the existing data set and is reconstructed from the layers.
[0004] For example, a method for calculating a synthetic mammogram is known from US Pat. No. 9,808,215 B2. A 2D mammogram image is synthesized from at least one of the tomosynthesis projection images and / or the image data reconstructed by tomosynthesis. In its simplest form, the mammogram can be synthesized by selecting one of the tomosynthesis projection images for display as a synthesized mammogram. Other methods for synthesizing a mammogram include reprojecting and filtering projection data and / or reconstructed data. The synthesized mammogram is displayed together with at least a portion of the reconstructed data to facilitate review of the reconstructed data. A familiar image is generated that can be used to facilitate review of a tomosynthesis dataset.
[0005] From the publication HAILIANG HUANG ET AL: "Comparison of contrast-enhanced digital mammography and contrast-enhanced digital breast tomosynthesis for lesion assessment", JOURNAL OF MEDICAL IMAGING, Vol. 6, No. 03, February 13, 2019 (2019-02-13), page 1, XP055671269, 1000 20th St. Bellingham WA 98225-6705 USA ISSN: 2329-4302, DOI: 10.1117 / 1.JMl.6.3.031407 ; Contrast-enhanced digital mammography (CEDM) is known to reveal the neovasculature of breast lesions in a two-dimensional contrast-enhancement map. Contrast-enhanced digital breast tomosynthesis (CEDBT) provides three-dimensional contrast enhancement, which can improve lesion characterization and localization. Synthetic CEDM can be generated from CEDBT data and provides lesion contrast enhancement comparable to CEDM.
[0006] US 2017 / 011534 A1 discloses a method for generating a synthetic two-dimensional mammogram with enhanced contrast for structures of interest. This method comprises acquiring a three-dimensional digital breast tomosynthesis volume with a plurality of voxels. A three-dimensional relevance map is generated that encodes the relevance of the underlying structure for a diagnosis for the voxels. A synthetic two-dimensional mammogram is calculated based on the three-dimensional digital breast tomosynthesis volume and the three-dimensional relevance map.
[0007] In a typical calculation of a synthetic mammogram based on an average intensity projection (AIP) and a maximum intensity projection (MIP) based on the tomosynthesis slices, not all tumor-relevant areas are sufficiently clearly displayed. For example, the superimposition of dense tissue can lead to a loss of visibility of the structures bordering a tumor, such as spiculi, and the margins. However, the visualization of such structures is particularly relevant and interesting for diagnosis, as it facilitates tumor detection for the radiologist and provides an indicator of tumor malignancy.
[0008] In conventional tomosynthesis, for example, in the context of screening, artificial intelligence (AI)-based CAD methods are used to detect lesions within the 3D slices. For example, EP 3 326 535 A1 discloses a display system for displaying digital breast tomosynthesis (DBT) data. First and second DBT volume images of a woman's left breast and first and second DBT volume images of the woman's right breast are provided by an image generation unit.In addition, a two-dimensional navigation image is provided for each DBT volume image by a navigation image provision unit. A user is allowed to specify a position in the navigation image using a user interface. A CAD marker associated with the location is then determined in a DBT volume image of a breast, and a corresponding CAD marker, if present, is determined in another DBT volume image of the breast. Sections of these DBT volume images associated with the CAD markers are shown on a display. These CAD methods are based on deep learning techniques, in which neural networks such as the U-Net are trained with a large number of data sets with known lesion positions, in particular with image annotations. The trained networks are then able to generate a probability distribution of tumor-relevant regions based on unknown data sets.
[0009] The resulting probability maps can then be used to visualize the detected areas in the synthetic two-dimensional image. During generation, the individual layers are weighted differently based on the probability distributions or probability maps, thus preventing or reducing disruptive overlaps of tumor-affected areas. It is known from US Patent No. 10,010,302 B2 that a 2D mammogram image is synthesized from at least one of the tomosynthesis projection images and / or the image data reconstructed by tomosynthesis. In its simplest form, the mammogram can be synthesized by selecting one of the tomosynthesis projection images for display as a synthesized mammogram. Other methods for synthesizing a mammogram include reprojecting and filtering projection data and / or reconstructed data.The synthesized mammogram is displayed along with at least a portion of the reconstructed data to facilitate review of the reconstructed data. This creates a familiar image that can be used to facilitate review of a tomosynthesis dataset.
[0010] It is an object of the invention to provide a method for generating a synthetic mammogram, a mammography system, a computer program product and a computer-readable medium which enable a simplified and alternative calculation of a synthetic mammogram with a reduced overlay of tumor-affected areas.
[0011] The object is achieved according to the invention by a method for generating a synthetic mammogram based on a dual-energy tomosynthesis image according to claim 1, a mammography system according to claim 12, a computer program product according to claim 13 and a computer-readable medium according to claim 14.
[0012] The invention relates to a method for generating a synthetic mammogram, in particular with a reduced overlay of tumor-affected areas, based on a dual-energy tomosynthesis image of an examination region. The method comprises the steps of acquiring a low-energy tomosynthesis image, acquiring a high-energy tomosynthesis image, determining, creating, generating, and displaying. In the step of acquiring a low-energy tomosynthesis image, a low-energy tomosynthesis image is acquired with a first X-ray energy spectrum. In the step of acquiring a high-energy tomosynthesis image, a high-energy tomosynthesis image is acquired with a second X-ray energy spectrum that is higher in energy than the first X-ray energy spectrum, wherein the examination region has a contrast agent distribution.In the determination step, a subtraction volume is determined based on the high-energy tomosynthesis image and the low-energy tomosynthesis image. In the creation step, a three-dimensional probability map is created with a weighting factor per voxel based on the subtraction volume. In the generation step, a synthetic mammogram is generated based on the three-dimensional probability map, and in particular the low-energy tomosynthesis image. The synthetic mammogram can be referred to as a contrast-enhanced synthetic mammogram.
[0013] To prepare for the scan, the patient is administered a contrast agent. During a possible waiting period, for example between 1.5 and 2 minutes, the contrast agent begins to distribute, particularly in the examination area. The patient can now be positioned and the breast compressed. First, a low-energy tomosynthesis image can be acquired, for example using a tungsten anode with a rhodium filter at a tube voltage between 24 and 32 kV. In the low-energy tomosynthesis image, the contrast agent is already at least partially distributed in the tissue, but the influence of the contrast agent on the low-energy tomosynthesis image is minimal. Subsequently, the high-energy tomosynthesis image can be acquired, for example using a titanium filter and a tube voltage of 49 kV.In the high-energy tomosynthesis image, the contrast agent is distributed in the tissue; the influence of the contrast agent on the high-energy tomosynthesis image is clearly visible.
[0014] The high-energy tomosynthesis image is acquired after a contrast agent has been introduced into the examination region. During the high-energy tomosynthesis image, contrast agent enhancement, in particular iodine enhancement, is present in the examination region. During the low-energy tomosynthesis image, however, contrast agent enhancement or no contrast agent enhancement may be present in the examination region. Typically, particularly to increase patient comfort, contrast agent enhancement can be present in both the low-energy tomosynthesis image and the high-energy tomosynthesis image. The examination region may exhibit contrast agent distribution during the acquisition of the low-energy tomosynthesis image. The high-energy tomosynthesis image and the low-energy tomosynthesis image each comprise a data set with a plurality of projection data sets acquired at a plurality of projection angles.Based on this, a tomosynthesis volume can be reconstructed for the high-energy tomosynthesis acquisition and the low-energy tomosynthesis acquisition.
[0015] During generation, different weightings of individual slices or voxels can be applied based on the probability distribution of the three-dimensional probability map. In particular, the low-energy tomosynthesis image can be combined with the three-dimensional probability map to create a contrast-enhanced synthetic mammogram.
[0016] The subtraction volume can be calculated based on the reconstructed tomosynthesis volume of the high-energy tomosynthesis image and the low-energy tomosynthesis image. The image values of corresponding voxels of the tomosynthesis volume of the high-energy tomosynthesis (HE) image and the tomosynthesis volume of the low-energy tomosynthesis (LE) image are subtracted in a weighted manner. For example, the subtraction can be performed according to the following equation: ln x = ln HE − w ⋅ ln LE .
[0017] The factor w can be selected depending on the breast thickness / density, for example, based on the compression thickness. The weighting factor of the three-dimensional probability map can be based on the value x. The value range of the weighting factors of the three-dimensional probability map can preferably extend from 0 to 200 or from 0 to 2. The weighting factor specifies a probability or weight for a voxel. The weighting factor of the three-dimensional probability map can include a contrast agent concentration, which can be determined based on the value x. Alternatively, the weighting factor can assume discrete values, for example, 0 for a background, 1 for breast tissue, and 2 for a lesion.
[0018] The three-dimensional probability map is created in the creation step based on the subtraction volume. The weighting factors of the three-dimensional probability map can be a measure of the contrast agent concentration, particularly iodine concentration.
[0019] The inventors propose a new, alternative approach for a probability map for contrast-enhanced tomosynthesis imaging. Instead of a deep learning-based method with corresponding training data, the probability map is based on the 3D map or 3D distribution of the suction volume obtained by iodine enrichment. The strength of the enriched iodine portion corresponds to the probability or degree of malignancy of the lesion that caused iodine enrichment. In one embodiment, the data from a quantitative analysis of the iodine concentration from the CEDET data (in mg iodine / cm³) can be used for this purpose. Subsequently, an overlay-free or at least overlay-reduced representation of the regions within the synthetic mammogram that are particularly relevant for diagnosis can also be achieved.
[0020] In contrast-enhanced breast tomosynthesis, the corresponding existing contrast-enhanced image, particularly the iodine image or subtraction image, can be advantageously used in addition to the creation of a three-dimensional probability map. This advantageously avoids the laborious collection of annotated data for a CAD-based procedure. Collecting annotated data would be necessary for an approach using neural networks. Since contrast-enhanced dual-energy tomosynthesis represents a new type of X-ray diagnostics for tumor detection, a sufficient database is currently not available. However, the output quality of such trained networks largely depends on the quantity and quality of the annotated training data.Advantageously, a more readily available method for calculating a three-dimensional probability map can be enabled for contrast-enhanced dual-energy tomosynthesis imaging.
[0021] According to one aspect of the invention, in the generating step, a first synthetic mammogram is created based on the low-energy tomosynthesis image. The first synthetic mammogram can be generated based on the low-energy tomosynthesis image, for example, using known methods for generating a synthetic mammogram.
[0022] The first synthetic mammogram or a synthetic mammogram can generally comprise the following steps and can be based in particular on the low-energy tomosynthesis image with a plurality of projection data sets. Preferably, at least one (first) synthetic mammogram is generated corresponding to the mean projection data set, in particular at a projection angle of 0 degrees. The mean projection angle can in particular refer to the angle at which the central beam of the X-ray source is incident substantially perpendicularly on the compressed breast or perpendicularly on the upper compression element. A (first) synthetic mammogram can be assigned to a respective projection angle. A (first) synthetic mammogram can be generated for a projection angle.The (first) synthetic mammogram can in particular be based on the projection image of the associated projection angle, for example using this projection image in the form of an average intensity projection.
[0023] In one embodiment, multiple (first) synthetic mammograms can be generated. The maximum number of (first) synthetic mammograms generated can, for example, correspond to the number of projection angles. Preferably, a (first) synthetic mammogram is generated for the projection angle 0 degrees. Furthermore, 2 to the maximum number of recorded projection angles, preferably 10 to 20, particularly preferably 17, (first) synthetic mammograms are generated. For example, 17 (first) synthetic mammograms are generated based on 25 projection data sets.
[0024] An average intensity projection (AIP) can be determined as the first image component based on the plurality of projection data sets. Preferably, a projection data set of a projection angle, in particular corresponding to the assignment of a projection angle to the (first) synthetic mammogram, can be used as the average intensity projection or as the basis for the average intensity projection. In particular, a limited-angle AIP (LARIP) can be used as the average intensity projection. One projection data set or multiple projection data sets can be selected, on the basis of which the average intensity projection is determined. Advantageously, the average intensity projection, as the first image component, comprises essential information about the examination subject in the two-dimensional plan view corresponding to the projection angle.However, the average intensity projection alone can hardly meet the requirements for a comprehensive assessment of the breast, as it has a relatively high noise level due to the only proportional dose used.
[0025] At least one of the following steps can be applied to the average intensity projection: intensity adjustment, grayscale distribution adjustment, scatter correction, and calcification-preserving noise filtering. The average intensity projection can be processed such that the processed average intensity projection can be used as a noisy base or first image component for the (first) synthetic mammogram. Advantageously, the information from this projection can serve as the basis for the (first) synthetic mammogram. Building on the first image component, a (first) synthetic mammogram that is essentially qualitatively equivalent to a conventional digital full-field mammography image can be generated by adding edge information and contrast information from a second image component.
[0026] A maximum intensity projection (MIP) can be determined as a second image component based on the plurality of projection data sets. In particular, a plurality of projection data sets can be used to determine the maximum intensity projection. In particular, projection data sets that are adjacent to the projection data set assigned to the (first) synthetic mammogram, or their projection angle, can be used. For example, projection data sets from a suitable angular range around the assigned projection angle can be used. In particular, all projection data sets can be used to determine the maximum intensity projection. Edge information and contrast information from a plurality of projection data sets can advantageously be used as a second image component to add them to the first image component. The second image component can advantageously be used for denoising.
[0027] From the tomosynthesis volume of the low-energy tomosynthesis image, a weighted intensity projection, for example based on the three-dimensional probability map, can be used as a second image component instead of an MIP.
[0028] The first image component and the second image component can be recombined to form the (first) synthetic mammogram. The (first) synthetic mammogram can be generated by combining or recombining the first image component and the second image component. Advantageously, the information from the associated projection data set in the form of the first image component is used together with additional information from several or further projection data sets in the form of the second image component. This allows the information based on the patient dose used to be advantageously used to generate the (first) synthetic mammogram.
[0029] According to one aspect of the invention, in the step of generating, a two-dimensional probability map is created by forward projection of the three-dimensional probability map. The three-dimensional probability map with the weighting factors as values for the voxels of the three-dimensional probability map can be forward projected. This generates a two-dimensional probability map. The two-dimensional probability map can preferably have weighting factors in the value range between 0 and 2. The value 1 means no enhancement. The value 0 means complete suppression. The value 2 means maximum enhancement.
[0030] The forward projection of the three-dimensional probability map can be advantageously used for synthetic, particularly contrast-enhanced, mammograms. The weighting factors in the two- or three-dimensional probability map can depend on breast thickness or the tube voltage(s).
[0031] According to one aspect of the invention, the synthetic mammogram is an overlay of the first synthetic mammogram with the two-dimensional probability map. In the synthetic mammogram, the two-dimensional probability map can be displayed, for example, as a colored overlay. Areas with tumor-affected regions, particularly overlaid regions, can advantageously be highlighted. This can advantageously improve the diagnosis.
[0032] According to one aspect of the invention, the synthetic mammogram is based on weighting the first synthetic mammogram with the two-dimensional probability map. The first synthetic mammogram can be generated based on the average intensity projection and the maximum intensity projection. The first synthetic mammogram can be weighted with the two-dimensional probability map. This allows a weighted synthetic mammogram to be obtained.
[0033] The two-dimensional probability map, preferably with values between 0 and 2, can be combined with the first synthetic mammogram to create the synthetic mammogram. The weighting factors can be applied pixel by pixel to the image values of the first synthetic mammogram. This performs a weighting in the two-dimensional image space. This can advantageously enable an optimized synthetic mammogram. This can advantageously enable a reduced overlay of tumor-bearing areas in the synthetic mammogram. This can advantageously be achieved by evaluating only the data pertaining to the examination or image.
[0034] According to one aspect of the invention, in the generating step, the three-dimensional probability map is combined with the low-energy tomosynthesis image to form a weighted low-energy tomosynthesis image, and the weighted low-energy tomosynthesis image is forward-projected. The voxels, in particular each voxel, of the tomosynthesis volume based on the low-energy tomosynthesis image can be multiplied or weighted by the weighting factor or probability factor of the three-dimensional probability map. The value range of the three-dimensional probability map can preferably correspond to the range between 0 and 2. Advantageously, the weighting can be performed in three-dimensional image space. This weighted three-dimensional image information can be transferred into two-dimensional image space by means of forward projection.
[0035] According to one aspect of the invention, the forward-projected weighted low-energy tomosynthesis image is displayed as a synthetic mammogram. According to one aspect of the invention, the forward-projected weighted low-energy tomosynthesis image is the synthetic mammogram. A clear representation in two-dimensional image space can advantageously be used for diagnosis, particularly as an overview image.
[0036] According to one aspect of the invention, the forward-projected weighted low-energy tomosynthesis image is combined with an average intensity projection to create the synthetic mammogram. The average intensity projection can, for example, be a mean projection, e.g., 0 degrees, of the low-energy tomosynthesis image. The forward-projected weighted low-energy tomosynthesis image includes a weighted intensity per pixel. A so-called calcium maximum intensity projection can also be taken into account when combining the images to create the synthetic mammogram. Advantageously, a weighted synthetic mammogram can be enabled. Advantageously, weighting in the three-dimensional image space can lead to a particularly strong reduction of overlaps of tumor-affected areas.
[0037] According to one aspect of the invention, the synthetic mammogram is an overlay of the forward-projected weighted low-energy tomosynthesis image with a first synthetic mammogram based on the low-energy tomosynthesis image. In the synthetic mammogram, the two-dimensional forward-projected weighted low-energy tomosynthesis image can be displayed, for example, as a colored overlay over the first synthetic mammogram. Regions with, in particular superimposed, tumor-affected regions can advantageously be highlighted. The diagnosis can advantageously be improved. A conventional first synthetic mammogram can advantageously be displayed with the overlay based on the three-dimensional weighting.
[0038] According to one aspect of the invention, the sum of the weighting factors of the three-dimensional probability map for a pixel of the synthetic mammogram, in particular all slices, is normalized to a predetermined value. The normalization of associated voxels of the multiple slices can be 1 or 100. The associated voxels can be defined such that these voxels contribute to a common pixel in the synthetic mammogram. This advantageously enables a uniform representation of different synthetic mammograms, for example, based on different examinations.
[0039] According to the invention, the weighting factors of the three-dimensional probability map are based on a quantitative analysis of the contrast agent concentration. The contrast agent can, in particular, be iodine. The contrast agent concentration can be specified in mg iodine / cm3.
[0040] According to one aspect of the invention, the three-dimensional probability map is displayed superimposed in a volume image, in particular in addition to the synthetic mammogram, based on the low-energy tomosynthesis image and / or the high-energy tomosynthesis image. For example, scrolling or moving the view of the volume image in the form of the subtraction volume can be enabled. Preferably, a region of interest (ROI) can be selected and, if necessary, marked in the synthetic mammogram. The volume image or the corresponding slice of the volume image can be automatically displayed based on the selection of the region of interest. Advantageously, the clarity of the two-dimensional synthetic mammogram can be combined with the depth resolution of the three-dimensional volume image. Advantageously, an improved diagnosis can be enabled.
[0041] The invention further relates to a mammography system comprising means for carrying out a method according to the invention. The advantages of the method according to the invention can be advantageously transferred to the mammography system. The mammography system can comprise the following units: a recording unit for recording a low-energy tomosynthesis image with a first X-ray energy spectrum and for recording a high-energy tomosynthesis image with a second X-ray energy spectrum that is higher in energy than the first X-ray energy spectrum, wherein the examination region has a contrast agent distribution, a determination unit for determining a subtraction volume based on the high-energy tomosynthesis image and the low-energy tomosynthesis image, a creation unit for creating a three-dimensional probability map with a weighting factor per voxel based on the subtraction volume, a generation unit for generating a synthetic mammogram based on the three-dimensional probability map, and a display unit, for example in the form of a screen, for displaying the synthetic mammogram. The determination unit, the creation unit, and the generation unit can be comprised of a computer unit of the mammography system. The acquisition unit can, in particular, comprise the X-ray source and the X-ray detector.
[0042] The invention further relates to a computer program product comprising a computer program which can be loaded directly into a memory device of a control device of a mammography system, comprising program sections for executing all steps of a method according to the invention when the computer program is executed in the control device of the mammography system.
[0043] The invention further relates to a computer-readable medium on which program sections are stored that can be read and executed by a computer unit in order to carry out all steps of a method according to the invention when the program sections are executed by the mammography system or its computer unit.
[0044] In the following, exemplary embodiments of the invention are explained in more detail with reference to the drawings. Herein: FIG 1 a schematic representation of the method according to the invention; FIG 2 a schematic representation of the mammography system according to the invention in a first embodiment; and FIG 3 a schematic representation of the mammography system according to the invention in a second embodiment.
[0045] The Fig. 1 shows an exemplary embodiment of the method 20 according to the invention for generating a synthetic mammogram based on a dual-energy tomosynthesis image of an examination region. The method 20 has the following steps: In the recording step 21, a low-energy tomosynthesis image is acquired with a first X-ray energy spectrum. In the recording step 23, a high-energy tomosynthesis image is acquired with a second X-ray energy spectrum that is higher in energy than the first X-ray energy spectrum, wherein the examination region has a contrast agent distribution. In the determining step 25, a subtraction volume is determined based on the high-energy tomosynthesis image and the low-energy tomosynthesis image. In the creating step 27, a three-dimensional probability map with a weighting factor per voxel is created based on the subtraction volume.In the generation step 29, a synthetic mammogram is generated based on the three-dimensional probability map. This may be followed by a display step 31 of the synthetic mammogram.
[0046] In the generating step 29, a first synthetic mammogram can be generated based on the low-energy tomosynthesis image. In the generating step 29, a two-dimensional probability map can also be generated by forward projection of the three-dimensional probability map. The synthetic mammogram can be an overlay of the first synthetic mammogram with the two-dimensional probability map. Alternatively, the synthetic mammogram can be based on a weighting of the first synthetic mammogram with the two-dimensional probability map.
[0047] In an alternative embodiment, in the generating step 29, the three-dimensional probability map can be combined with the low-energy tomosynthesis image to form a weighted low-energy tomosynthesis image, and the weighted low-energy tomosynthesis image can be forward-projected. The forward-projected weighted low-energy tomosynthesis image can be the synthetic mammogram. Alternatively, the forward-projected weighted low-energy tomosynthesis image can be combined with an average intensity projection to form the synthetic mammogram. Alternatively, the synthetic mammogram can be an overlay of the forward-projected weighted low-energy tomosynthesis image with a first synthetic mammogram based on the low-energy tomosynthesis image.
[0048] In the displaying step 31, the three-dimensional probability map can further be displayed superimposed in a volume image, in particular the subtraction volume image, based on the low-energy tomosynthesis image and / or the high-energy tomosynthesis image.
[0049] The Fig. 2 shows an exemplary embodiment of the mammography system according to the invention in a first embodiment. For a low-energy tomosynthesis image and a high-energy tomosynthesis image, a plurality of projection data sets are acquired at a plurality of projection angles PI -1, 0, 1, 2,..., 12. The X-ray source 2.1 is moved, in particular, along a radius around a point in the breast 8, with a projection data set being acquired at each of the projection angles PI -1, 0, 1, 2,..., 12. During the image acquisition, the patient's breast 8 is arranged as the examination object between an upper compression element 3.1 and a lower compression element 3.2.
[0050] The Fig. 3shows an exemplary embodiment of the mammography system 1 according to the invention in a second embodiment. The mammography system 1 comprises a stand 1.1, on which the X-ray housing 2 comprising the X-ray source 2.1 and the X-ray detector 5 are arranged together with a compression unit 3. In particular, the X-ray housing 2 is rotatably mounted on the stand 1.1 with respect to the stand 1.1 and the X-ray detector 5 as well as the compression unit 3. The compression unit 3 comprises an upper compression element 3.1 and a lower compression element 3.2, between which the breast 8 of a patient is arranged. The mammography system 1 is connected to a data processing unit 10. The data processing unit 10 comprises at least one processor unit or computer unit 10.1, a display unit 10.2, and an input unit 10.3. The region of interest can be marked via the input unit. The computer unit 10.1 can comprise the determination unit, the creation unit, and the generation unit. The mammography system 1 can be controlled at least partially via a foot switch 1.2.
[0051] Although the invention has been illustrated in detail by the preferred embodiment, the invention is not limited by the disclosed examples and other variations may be derived therefrom by those skilled in the art without departing from the scope of the invention.
Claims
1. Method (20) for creating a synthetic mammogram on the basis of a dual energy tomosynthesis recording of an examination region, having the steps: a. making (21) a low energy tomosynthesis recording with a first X-ray energy spectrum, b. making (23) a high energy tomosynthesis recording with a second X-ray energy spectrum of higher energy compared with the first X-ray energy spectrum, wherein the examination region has a contrast medium distribution, c. determining (25) a subtraction volume on the basis of the high energy tomosynthesis recording and the low energy tomosynthesis recording, d. generating (27) a three-dimensional probability map with a weighting factor per voxel on the basis of the subtraction volume, e. creating (29) a synthetic mammogram on the basis of the three-dimensional probability map, characterised in that the weighting factors of the three-dimensional probability map are based upon a quantitative analysis of the contrast medium concentration.
2. Method according to claim 1, wherein in the step of creating, a first synthetic mammogram is generated on the basis of the low energy tomosynthesis recording.
3. Method according to claim 2, wherein in the step of creating, by means of forward projection of the three-dimensional probability map, a two-dimensional probability map is generated.
4. Method according to claim 3, wherein the synthetic mammogram is a superposition of the first synthetic mammogram with the two-dimensional probability map.
5. Method according to claim 3, wherein the synthetic mammogram is based upon a weighting of the first synthetic mammogram with the two-dimensional probability map.
6. Method according to claim 1, wherein in the step of creating, the three-dimensional probability map is combined with the low energy tomosynthesis recording to a weighted low energy tomosynthesis recording and the weighted low energy tomosynthesis recording is forward projected.
7. Method according to claim 6, wherein the forward-projected weighted low energy tomosynthesis recording is the synthetic mammogram.
8. Method according to claim 6, wherein the forward-projected weighted low energy tomosynthesis recording is combined with an average intensity projection to the synthetic mammogram.
9. Method according to claims 2 and 6, wherein the synthetic mammogram is a superposition of the forward-projected weighted low energy tomosynthesis recording with a first synthetic mammogram on the basis of the low energy tomosynthesis recording.
10. Method according to one of the preceding claims, wherein the total of the weighting factors of the three-dimensional probability map is normalised to a predetermined value for slices contributing one pixel of the synthetic mammogram.
11. Method according to one of the preceding claims, wherein the three-dimensional probability map is displayed in a volume image on the basis of the low energy tomosynthesis recording and / or superposed on the high energy tomosynthesis recording.
12. Mammography system (1) having means for carrying out a method according to one of claims 1 to 11, having the units: a. a recording unit for making a low energy tomosynthesis recording with a first X-ray energy spectrum and for making a high energy tomosynthesis recording with a second X-ray energy spectrum of higher energy compared with the first X-ray energy spectrum, wherein the examination region has a contrast medium distribution, b. a determining unit for determining a subtraction volume on the basis of the high energy tomosynthesis recording and the low energy tomosynthesis recording, c. a generating unit for generating a three-dimensional probability map with a weighting factor per voxel on the basis of the subtraction volume, d. a creating unit for creating a synthetic mammogram on the basis of the three-dimensional probability map, and e. a display unit, for example in the form of a screen, for displaying the synthetic mammogram, characterised in that the weighting factors of the three-dimensional probability map are based on a quantitative analysis of the contrast agent concentration.
13. Computer program product having a computer program which can be directly loaded into a memory store of a control apparatus of a mammography system, having program portions in order to carry out all the steps of a method according to one of claims 1 to 11 when the computer program is executed in the control apparatus of the mammography system.
14. Computer-readable medium on which program portions that are configured to be read in and executed by a computer unit are stored, in order to carry out all the steps of a method according to one of claims 1 to 11 when the program portions are executed by the mammography system.
