Medical image processing apparatus and medical image processing method

Abstract

To assess bone density. [Solution] The medical image processing apparatus according to the embodiment includes a processing circuit. The processing circuit receives a medical image dataset including intensity values ​​as a position function, obtains segmentation of bone regions in the medical image dataset, processes the medical image dataset to identify a plurality of subregions within the bone region where the intensity values ​​are within an intensity threshold range relative to the calibration intensity values, and determines the density and / or distribution of the subregions within the bone region.

Description

[[Technical Field]]

[0001] Embodiments disclosed herein generally relate to a medical image processing apparatus and a medical image processing method, for example, to a medical image processing apparatus and a medical image processing method for analyzing bone quality and evaluating osteoporosis and fracture risk. [[Background Art]]

[0002] Existing methods for estimating bone mineral density (BMD) include estimating BMD based on attenuation values, for example measured in Hounsfield units (HU). The average value of HU within a specific vertebral region or segmentation is used. The problem with this approach is that it is specific to scanners and / or protocols. Even though this method works well for one protocol on one scanner, since scanners are not normally calibrated for high-density materials such as bone, HU values vary significantly between different scanners and protocols, for example with respect to different kilovolt peak (kVp) values applied to an X-ray tube.

[0003] Osteoporosis, also referred to as "porous bone disease", is a pathological decrease in bone density and bone strength caused by reduced formation of new bone tissue, increased resorption of old bone tissue, or both. Osteoporosis is a serious and widespread health problem. It is said that one in three women and one in five men over 50 years of age will experience a secondary fracture due to osteoporosis, and furthermore, 19% of osteoporotic women with a vertebral fracture will experience a new vertebral fracture within one year after the initial fracture.

[0004] A hip fracture increases the probability of early death for both men and women, particularly when the fracture occurs at an older age. This elevated mortality risk persists for more than 10 years after the fracture. While the prevalence of osteoporosis is higher in women than in men, approximately 30 to 40% of osteoporotic fractures occur in men, and men also have an increased mortality risk regardless of the fracture site. For women over 45 years of age, fractures resulting from osteoporosis result in longer hospital stays than diabetes, heart attack, and breast cancer.

[0005] BMD is an important clinical factor and one of several factors that increase the risk of fracture. The trabecular bone score (TBS) or T-score is a measure of the structural quality of bone tissue calculated from dual-energy X-ray absorptiometry (DEXA) images. DEXA is a low-dose dual-energy X-ray projection imaging technique commonly used for routine quantitative measurement of BMD.

[0006] Computed tomography (CT) scans use a higher radiation dose than DEXA, and therefore are not expected to replace DEXA in routine osteoporosis screenings. Regarding the evaluation of cancellous bone, CT imaging has advantages such as excluding cortical bone, clearly capturing osteophytes, and providing a true 3D distribution that allows for the detection of smaller holes and cavities in the cancellous bone structure compared to DEXA.

[0007] Early detection and intervention of bone degeneration are desirable. Radiologists do not constantly re-examine / report on the spine, and 34-55% of spinal fractures are not reported in scan results. Unfortunately, standard single-energy CT images do not reliably represent bone mineral density (BMD).

[0008] CT scanners are generally calibrated so that the Hounsfield units (HU) value for water at standard temperature and pressure is zero, and the value for air is -1000 HU. Outside of these two calibration points, HU measurements depend on the energy spectrum of the X-ray beam and the characteristics of the X-ray detector, and the difference becomes larger for materials with high attenuation values, such as bone. Furthermore, the HU measurement value of a specific voxel in a CT image is also affected by the body composition of the subject, due to factors such as changes in beam hardening in subjects with different body types. The zero value of HU is particularly important because conditions that cause bone tissue degeneration, such as osteoporosis, may appear as low-density regions with HU close to or below zero in images obtained from a CT scanner.

[0009] In quantitative CT (QCT) technology, it is known that a calibration phantom is placed after each scan. However, this is not particularly useful for scans that are not intended for bone density evaluation. [Prior art documents] [Patent Documents]

[0010] [Patent Document 1] Chinese Patent Application Publication No. 117710288 Specification [Overview of the project] [Problems that the invention aims to solve]

[0011] One of the problems that the embodiments disclosed herein and in the drawings aim to solve is the evaluation of bone density. However, the problems that the embodiments disclosed herein and in the drawings aim to solve are not limited to the above problem. Problems corresponding to the effects of each configuration shown in the embodiments described later can also be positioned as other problems. [Means for solving the problem]

[0012] A medical image processing apparatus according to an embodiment of the embodiment includes a processing circuit. The processing circuit receives a medical image dataset including intensity values ​​as a position function, obtains segmentation of bone regions in the medical image dataset, processes the medical image dataset to identify a plurality of subregions within the bone region where the intensity values ​​are within an intensity threshold range relative to a calibration intensity value, and determines the density and / or distribution of the subregions within the bone region. [Brief explanation of the drawing]

[0013] An embodiment will be described as a non-limiting example with reference to the following drawings. [Figure 1] Figure 1 is a schematic diagram showing an example of a data processing device according to the embodiment. [Figure 2] Figure 2 is a flowchart showing an example of a method for processing medical image data according to the embodiment. [Figure 3] Figure 3 shows three medical images in which low-attenuation voxels have been labeled. [Figure 4] Figure 4 is a graph showing the T-value and the proportion of low-attenuation voxels. [Figure 5] Figure 5 is a graph showing bone density and the percentage of low-attenuation voxels. [Modes for carrying out the invention]

[0014] A medical image processing apparatus is provided that includes a processing circuit based on an embodiment. The processing circuit receives a medical image dataset including intensity values ​​as a position function, obtains segmentation of bone regions in the medical image dataset, processes the medical image dataset to identify a plurality of subregions within the bone region where the intensity values ​​are within an intensity threshold range relative to a calibration intensity value, and determines the density and / or distribution of the subregions within the bone region.

[0015] A medical image processing method is provided based on an embodiment. The method comprises receiving a medical image dataset including intensity values ​​as a position function, obtaining segmentation of bone regions in the medical image dataset, processing the medical image dataset to identify a plurality of subregions within the bone region in which the intensity values ​​are within an intensity threshold range relative to a calibration intensity value, and determining the density and / or distribution of the subregions within the bone region.

[0016] Chest and abdominal CT scans, performed for various purposes, typically include images of the spine. While the spine is not the primary focus of CT scans such as chest and abdominal scans, they provide an opportunity to obtain information about spinal bone degeneration.

[0017] Figure 1 is a schematic diagram showing an example of a data processing device 10 according to an embodiment. In this embodiment, the data processing device 10 processes medical image data such as imaging data from a CT scan. In other embodiments, the data processing device 10 may process other medical image data as appropriate.

[0018] The data processing apparatus 10 comprises an arithmetic device 12. In this example, the arithmetic device 12 is a personal computer (PC) or a workstation. The arithmetic device 12 is connected to one or more output devices 16 such as a screen or other display devices, and one or more input devices 18 such as a computer keyboard and a mouse.

[0019] The arithmetic device 12 acquires a data set from a data storage unit 22. At least a part of the data acquired from the data storage unit 22 includes medical image data, such as data acquired by a scanner 20, for example. The medical image data may include two-dimensional, three-dimensional or four-dimensional data of any imaging modality. For example, the scanner 20 may include a CT (computed tomography) scanner, a micro CT scanner, a cone beam CT scanner, a photon counting CT scanner, or any three-dimensional X-ray tomography scanner, and the like.

[0020] The arithmetic device 12 may receive data from one or more other data storage units (not shown) instead of or in addition to the data storage unit 22. For example, the arithmetic device 12 may receive medical image data from one or more remote data storage units (not shown) that form part of a picture archiving and communication system (PACS) or other information systems.

[0021] The arithmetic device 12 provides processing resources for automatically or semi-automatically processing data. The arithmetic device 12 comprises a processing apparatus 14. The processing apparatus 14 comprises a data processing circuit 24 and an interface circuit 26. The interface circuit 26 acquires user input and / or other inputs and / or outputs results of data processing. The data processing circuit 24 executes one or more of processing tasks, for example, image segmentation, classification, image caption generation, display and / or automatic reporting, or other tasks as appropriate.

[0022] In this embodiment, the data processing circuit 24 and the interface circuit 26 are each implemented in the arithmetic unit 12 by a computer program that includes computer-readable instructions for performing the method of this embodiment. However, in another embodiment, these circuits may be implemented as one or more application-specific integrated circuits (ASICs) or field-programmable gate arrays (FPGAs).

[0023] Furthermore, the arithmetic unit 12 includes other PC components such as a hard drive, RAM, ROM, a data bus, an operating system with various device drivers, and a graphics card. For clarity, these components are not shown in Figure 1.

[0024] The data processing device 10 in Figure 1 is configured to perform the methods illustrated and / or described below.

[0025] Figure 2 is a flowchart illustrating an example of a method for processing medical image data according to an embodiment. The medical image data is processed for bone quality evaluation.

[0026] In step 42, medical image data is collected and provided to the processing unit 14. That is, the data processing circuit 24 receives the medical image dataset.

[0027] In this embodiment, the medical image data includes data representing bone tissue. The medical image data is collected using one of many imaging techniques. In this embodiment, the medical image data includes computed tomography (CT) images. The medical image data may be collected using scanner 20 or any other scanner. The medical image data is provided directly from scanner 20 to processing unit 14 or received from data storage unit 22 or other sources.

[0028] Medical image data can be used in any format. For example, medical image data may include intensity values ​​and / or attenuation values ​​as position functions. In one embodiment, medical image data may include semantic data related to features within the image data.

[0029] The medical image data used in the processing example described with reference to Figure 2 includes three-dimensional image data composed of voxels. In other embodiments, the medical image data may be two-dimensional data composed of pixels. The two-dimensional image data may include, for example, two-dimensional slices of a tomographic three-dimensional CT image.

[0030] In step 44, segmentation processing is performed on the medical image data to segment the bone regions within the image data. This allows the data processing circuit 24 to obtain segmentation of bone regions within the medical image dataset. Segmentation of medical image data involves selecting pixels / voxels associated with one or more different bone tissues. For example, segmentation processing according to known techniques may be used as appropriate. Segmentation processing may include, for example, labeling different bones and / or defining the boundaries of specific bones. Rigid body segmentation, non-rigid body segmentation, or other segmentation processing based on any segmentation technique may be used as appropriate.

[0031] In other embodiments, medical image data is preprocessed and bone tissue is segmented. The segmented and medical image data are provided, for example, from the data storage unit 22 to the processing unit 14.

[0032] In the example illustrated with reference to Figure 2, the bone region includes one or more vertebrae that constitute part or all of the vertebral column. The segmentation in this example includes labels for the vertebral segments, selected from a vertebral labeling scheme.

[0033] In step 46, the processing device 14 processes the medical image dataset to identify multiple subregions within the bone region whose intensity values ​​are within an intensity threshold range relative to the calibration intensity value. The intensity threshold range is, in one example, between zero HU and 100 HU, but in other embodiments, other intensity ranges are selected as appropriate. For example, depending on the various embodiments, it may be within 1 HU, 10 HU, 50 HU, or 100 HU of the calibration value.

[0034] In this example, each sub-region is a single voxel. The calibration intensity value is zero HU, which is the intensity value when the scanner device (in this case, a CT scanner) was last calibrated. In other embodiments, the calibration intensity value may be the intensity value when the scanner performed an intensity measurement that is known or likely to be accurate. Each sub-region does not have to consist of a single voxel or pixel; instead, each sub-region of interest may, as desired, consist of, for example, multiple voxels or pixels.

[0035] One embodiment is characterized by using voxels or pixels within the bone region that have been identified as having an intensity value close to zero HU (for example, within the intensity threshold range) to determine whether or not a pathological condition such as low bone quality is observed in the bone region.

[0036] Bone regions containing low-quality skeletons typically have larger cavities and lower density than healthy bone, making them less permeable to radiation. A decrease in solid bone tissue can be detected by imaging techniques highly sensitive to bone density or the ratio of solid to non-solid material within the bone.

[0037] X-rays are particularly sensitive to high-density structures such as bone. CT scans, consisting of a series of X-ray images, are a suitable candidate for evaluating bone quality due to the high attenuation caused by X-rays within bone. CT scan images are rendered to include attenuation values ​​as a function of position, or parameters derived from attenuation values ​​that are a function of position. The Houndsfield (HU) unit is a measure of the radiation density of a substance, related to the attenuation exhibited by the substance. Generally, the radiation density of distilled water at standard temperature and pressure is defined as zero Houndsfield units (HU), and the radiation density of air at standard temperature and pressure is defined as -1000 HU.

[0038] CT scanners are typically calibrated against both air and distilled water at standard temperatures and pressures. Standard pressure and temperature values ​​for CT scanner calibration are defined as 20 degrees Celsius and 1 atmosphere. Attenuation measurements at zero HU or -1000 HU are robust demonstrative results because they do not depend on the transmission spectrum of the X-ray beam source or the characteristics of the detector. Attenuation measurements at zero HU are far below those of normal, healthy bone. Attenuation measurements at zero HU are particularly important in bone pathologies such as osteoporosis and osteopenia, leading to low-density areas within bone tissue that are close to and / or below zero HU. This is due to the breakdown and demineralization of the bone's cancellous structure.

[0039] Based on the voxels or pixels within the bone region identified in step 46, it is determined whether the bone quality is low or whether there is a diseased condition.

[0040] The relevant pixels, voxels, or other subregions are identified using appropriate threshold range values, and then the pixels, voxels, or other subregions are used to identify bone regions with low bone quality.

[0041] By operating at low attenuation values, variations in attenuation measurements between different scanners and / or between scanners with different kilovoltage peak (kVp) values ​​are avoided when imaging materials with high attenuation values. It is possible to avoid the range of improperly calibrated CT scanners by measuring the absence of bone material or the presence of holes, rather than attempting to quantify the density of the bone material being examined. In step 46, the entire medical image, including the medical image dataset, or only the segment of bone tissue is analyzed.

[0042] The output of step 46 is a dataset containing the locations of subregions (e.g., voxels or pixels) in segmented medical image data. The intensity and / or attenuation of a region / pixel / voxel is below one or more reference intensity and / or reference attenuation values ​​corresponding to one or more reference intensity and / or reference attenuation values. The reference attenuation and / or reference intensity values ​​include calibration attenuation and / or calibration intensity values ​​associated with the equipment used for medical imaging.

[0043] In step 48, the number and / or percentage and / or density of low-quality bone tissue in individual bones, for example, in each vertebra, is calculated. This can be done by defining the total area or volume within an image or region of an image that contains bone tissue with low-intensity and / or low-attenuation (low-HU) areas.

[0044] In one embodiment, the number of low-HU pixels or voxels in each vertebra is counted for each 2D or 3D medical image. The 2D image data includes 2D slices from tomographic 3D CT images. The area / volume or number of pixels / voxels can be expressed as a ratio or percentage to the total area / volume or number of pixels / voxels constituting each bone. The density of subregions containing low-quality bone tissue is determined for all bone regions in the image data, or for individual bones such as each vertebra. Alternatively, the density of subregions may be determined for the entire spine or for a subset of the entire spine (cervical, thoracic, lumbar, etc.). For the convenience of the user / operator, the density of subregions within one or more bone regions may be displayed on the output device 16.

[0045] In step 50, a dataset is generated showing the spatial distribution of subregions containing low-quality bone tissue within the medical image data. This dataset is represented in the form of one or more medical images. These medical images are overlaid with visual indicators of the medical image segmentation for bone tissue and visual indicators of subregions evaluated as having low bone quality based on their respective intensity and / or attenuation. This dataset is stored for subsequent processing and analysis, and is also displayed on the output device 16 for the convenience of the user / operator.

[0046] In step 50, the results from step 48 are compared with the predicted results to determine the health status of individual bones, such as each vertebra. The processing device 14 determines the presence or absence of conditions such as osteoporosis, osteopenia, and vertebral compression fractures, as well as the risk of future fractures / re-fractures.

[0047] One or more conditions are determined according to the concentration and / or distribution of subregions within one or more bone regions. The processing device 14 is given data indicating general values ​​and / or ranges of the number or percentage of low-attenuation regions in individual bones, meaning low bone quality, healthy bone quality, and other possible classifications of bone quality.

[0048] Furthermore, the processing device 14 has a concentration threshold for one or more subregions of a bone region. The concentration thresholds for different bone regions, such as different vertebrae, are different from each other. The concentration threshold varies depending on at least one of the following: the sex, weight, body mass index, and race of the subject in the medical image data. The concentration threshold may be compared with the concentration of the subregion determined in step 48. Alternatively, the concentration threshold may be a ratio, such as the ratio of the low-attenuation subregion to the total area / volume. The concentration threshold may also be a single number representing the prevalence of low-quality bone tissue, such as the number of applicable pixels, the number of voxels or units of area and volume.

[0049] In one embodiment, bone tissue quality is graded into three or more classifications. There may be further bone quality classifications corresponding to other attenuation thresholds of the Hounsfield criteria. For example, areas corresponding to attenuation below zero HU in medical image data represent degraded bone tissue, areas corresponding to attenuation between 10 HU and zero HU are considered to be of moderate quality, and areas corresponding to attenuation above 10 HU are considered to be healthy bone tissue. In other examples, there may be more or fewer classifications at the same or other threshold attenuation points. Step 52 includes providing output to the user. The output of step 52 may include a bone label table and an indication of whether the bone tissue quality of the labeled bones falls within one or more attenuation threshold ranges or values ​​from the criterion attenuation. The output of step 52 may also include any combination of parameters / distributions determined in steps 46 and 50.

[0050] Figure 3 shows three medical images obtained by X-ray imaging, showing all or part of the subject's spine. The ground truth regarding the prevalence of fractures in the vertebrae of each subject's spine is publicly known and is represented by labeled text for each vertebra on the left side of Figures 3(a), 3(b), and 3(c). To avoid spurious density effects due to incomplete segmentation, the images are segmented based on the ground truth.

[0051] Segmented bone regions are highlighted as well as subregions containing pixels showing attenuation below zero HU. The subjects' bone mineral density (BMD) and T-score have been previously measured using a DEXA imaging device. The subjects' bone quality generally decreases from Figure 3(a) to Figure 3(c). The subject in Figure 3(a) has no vertebral fractures, but the number and severity of vertebral fractures in the subject in Figures 3(b) and 3(c) have increased. The images are processed using the method in Figure 2, with the display of regions / pixels showing intensity corresponding to attenuation below zero Houndsfield units superimposed. Figure 3(a) shows a subject with no vertebral fractures, BMD 176.6, T-score 1.4, and 0.13% of the image pixels in the vertebral images being fracture-free (below zero HU). Figure 3(b) shows a subject with multiple vertebral fractures, BMD 35.4, T-score -3.4, and 3.1% of the image pixels in the vertebral images being fracture-free (below zero HU). Figure 3(c) shows a subject with multiple severe vertebral fractures, a BMD of zero, a T-score of -2.1, and 7.9% of the pixels in the image of the fractured vertebrae were less than zero HU. For the subjects in Figures 3(a), (b), and (c), the pixels showing less than zero HU attenuation in the fractured vertebrae were 0.08%, 1.5%, and 3.9%, respectively. Subjects with low BMD and fractures in other vertebrae had a high density of pixels less than zero HU in both fractured and fractured vertebrae, and this was particularly high in the lumbar region. Fractured vertebrae often show a higher density rather than a lower density due to compression of the remaining bone.

[0052] Figure 4 is a graph showing the T-value and the percentage of voxels with less than zero HU, representing the results of a study conducted to understand the relationship between the two parameters. Bone mineral density is measured as a T-score. Bone mineral density is obtained using DEXA scans and compared to the “mean” or normal range of bone mineral density in healthy young women. A T-score greater than -1 means that bone mineral density is within the normal range. A T-score between -1 and -2.5 indicates that bone mineral density is below average, and the subject may have osteopenia, but not low enough to meet the criteria for osteoporosis. A T-score less than -2.5 means that bone mineral density is in the low range, and the subject may have osteoporosis. In Figure 4, the T-score or T-value measured using DEXA is plotted against the percentage of vertebral voxels showing attenuation below zero HU as a result of CT scans. The horizontal dashed line represents T-values ​​of -1 and -2.5, separating the areas of normal bone mineral density, osteopenia, and osteoporosis. This graph does not take fractured vertebrae into consideration. This is because fractures can sometimes result in higher-than-normal bone density, for example, due to compression. In the graph, vertebrae without fractures are represented by a cross shape. Also, in the graph, vertebrae with fractures are represented by another cross shape.

[0053] It can be seen that many data points from vertebrae without fractures appear above the horizontal line indicating a T value of -1. These data points were obtained from multiple subjects without fractures, with high T values ​​but attenuation below zero HU. Possible explanations for this result include the following: (a) Subjects who are at risk of vertebral fracture (b) Low-resolution or noisy images (c) Failure of spinal segmentation

[0054] In practice, noisy images are automatically filtered out. The failure of spinal segmentation is unlikely to be the cause of the false positive detection in Figure 4, as this segmentation has been validated against ground truth in the study. The DEXA T-score is above zero, which is good, but it can be seen that there are many subjects with fractures.

[0055] Figure 5 is a graph showing the BMD (grams / cm²) and voxel ratio of fracture-free vertebrae for subjects with attenuation values ​​less than zero HU, illustrating the results of a study conducted to understand the relationship between these two parameters. BMD is obtained by DEXA scan. In Figure 5, BMD is plotted against the percentage of voxels in vertebrae showing attenuation lower than zero HU as a result of CT scan. Fracture-free vertebrae are shown as points on the graph, while fractured vertebrae are shown as different points. It can be seen that most fractures occur when the BMD value is low. Furthermore, most fractures have a high percentage of low-density voxels. There are some cases where fractures are present even with high BMD, and many of these also have a high percentage of low-density voxels.

[0056] In one embodiment, the processing unit 14 reduces noise in the image dataset before processing in step 42 of the method shown in Figure 2, for example. The processing unit 14 reduces noise in the image using a low-pass filter, anisotropic smoothing, or other similar techniques. Images containing noise exceeding a predetermined threshold are identified by checking imaging parameters related to the measuring instrument. In the case of a CT scan, these imaging parameters include parameters of the X-ray dose used during the scan, such as X-ray tube voltage and current. High levels of X-ray radiation can generate a lot of noise in the image. The amount of noise in the image is determined by noise analysis in homogeneous regions of the image, such as regions containing the liver or aorta. The homogeneity of tissue in such regions implies that sub-regions detected within these regions or similar regions have a high noise level in the image data or in specific images containing these regions. If a noisy image is detected, the image is processed to reduce or compensate for the noise in the image, or the image is excluded from processing.

[0057] In other embodiments, the processing unit 14 may detect common artifacts in the image data. The presence of artifacts is determined based on density or radiation density measured by the imaging device. Artifacts include external objects such as metal implants, pacemakers, wire leads, and other similar objects. These artifacts are limited to the range of image slices containing high-density objects, and outliers are detected from all vertebrae based, for example, on intensity-based statistics. Affected vertebrae are excluded from subsequent analysis.

[0058] In one embodiment, the processing device 14 has vertebral shape measurement criteria and detects vertebral compression fractures (VCFs). Vertebrals with VCFs may have an abnormally small number of low-density voxels or subregions due to compression of the cancellous skeleton, which results in increased density. Based on the spatial distribution of bone regions or subregions within individual vertebrae, skeletons with a high risk of future fractures are detected, for example, by a Gray-level co-occurrence matrix, Gray-level run-length analysis, or similar method.

[0059] The medical conditions described herein or other conditions related to bone health may be identified as incidental findings in the CT imaging workflow. Chest and abdominal X-rays may unintentionally include images of bone regions, such as part or all of the spine. Such X-ray images can be analyzed using the methods and apparatus disclosed herein to detect bone health from scans where bone health is not the medical condition being examined. The methods and apparatus described herein provide additional input for fracture / re-fracture risk calculation (see FRAX score), particularly when DEXA measurement is unavailable. For example, large longitudinal changes in the skeleton can be measured in patients undergoing regular CT scans or treatments affecting bone strength for cancer treatment, or in patients after VCF correction surgery.

[0060] Based on the embodiment, a method is provided for utilizing the calibration points of a CT scanner by identifying regions of low bone attenuation and low density in the range of attenuation values ​​near zero HU. In this case, the results are more consistent across scanners than quantifying high HU values ​​that depend on the scanner / energy. The distribution of low-density voxels in the CT image contains structural information about bone strength. This information is provided to the user and / or stored in a memory device. The spatial distribution of low-density voxels provides more robust data than the average HU values ​​of individual bones, such as each vertebra. Because the amount of contrast agent absorbed by bone is limited, such a method is suitable for both non-contrast and contrast-enhanced CT scan images. This method is applicable to CT images, including photon-counting CT images. It is easily automated and can evaluate the secondary skeleton in scans targeting different primary skeletons. It can also be used as a method specifically for, for example, spinal column scans and bone quality analysis.

[0061] A medical image processing apparatus comprising a processing circuit is provided based on various embodiments. The processing circuit receives a medical image dataset including intensity values ​​as a position function, obtains segmentation of bone regions in the medical image dataset, processes the medical image dataset to identify, for example, a plurality of subregions within the bone region whose intensity values ​​are within an intensity threshold range relative to a calibration intensity value, and determines the density and / or distribution of the subregions within the bone region.

[0062] The processing circuit may further determine the presence or absence of a disease depending on the concentration and / or distribution of a specific subregion within the bone region. The disease may be osteopenia or osteoporosis, for example.

[0063] For user convenience, the processing circuit may output an image representing the density and / or distribution of a specific subregion within the bone region, or store this data in memory. A bone region in a medical image dataset may contain one or more vertebrae.

[0064] The segmentation received or determined by the aforementioned device may include one or more labels that identify bone regions. These labels may be labels associated with individual vertebrae, for example. Each identified sub-region may be a pixel or a voxel.

[0065] The processing circuit may compare the determined density of a sub-region with one or more density thresholds. The density thresholds may be associated with individual vertebrae. The density thresholds may differ for each vertebra based on the characteristics of the bone region. The density thresholds may depend on one or more characteristics of the subject in the medical image dataset. These characteristics may include, for example, the subject's age, sex, weight, height, body mass index, race, and other characteristics. The calibration intensity values ​​may include calibration values ​​in zero Houndsfield units and / or water. The water temperature and pressure are standard temperatures and pressures for CT scanner calibration. Standard temperatures and pressures for CT scanner calibration are 20 degrees Celsius and 1 atmosphere or 293.15 Kelvin and 1.013 × 10⁻¹⁰, respectively. 5 The unit is Pascal. Subregions within the bone region may be identified based on an intensity lower than the calibration value indicating zero Houndsfield units and / or water. Alternatively, subregions may be identified based on other calibration values ​​or other attenuation values.

[0066] The medical image dataset is acquired, at least partially, from computed tomography (CT) scans. CT scans may include micro-CT scans. The processing circuit reduces noise in the medical image dataset. Noise may be reduced before processing using one or more of the following methods: low-pass filtering, anisotropic smoothing, or other similar methods. Images containing noise exceeding a threshold may be excluded from processing. The processing circuit may detect images containing external structures and exclude such images from processing. External structures include high-density objects such as metal implants, pacemakers, wire leads, and similar objects.

[0067] The processing circuit may further determine the presence of a fracture based on the concentration and / or distribution of identified subregions within the bone region. The processing circuit may analyze the distribution of subregions within the bone region and assess the risk of future fracture / re-fracture of the bone region.

[0068] Furthermore, the processing circuit may determine the density and / or distribution of subregions within the bone region based on medical image data acquired during imaging for purposes other than detecting pathological conditions in the bone region. Thus, the assessment of bone health and fracture risk may be incidental findings in scans intended to detect related or unrelated pathological conditions.

[0069] The processing circuit may detect changes in bone region dimensions based on medical image data obtained from the subject at two or more different time points. In this way, the circuit can measure large longitudinal changes in the skeleton caused by vertical compression fractures, etc. Furthermore, this configuration can be used for patients undergoing regular CT scans or treatments that affect bone strength for cancer treatment. It may also be used to monitor the effects of surgery and to correct vertical compression fractures.

[0070] While specific circuits have been described herein, in other embodiments, one or more functions of these circuits can be implemented by a single processing resource or other component. Alternatively, functions implemented by a single circuit can be implemented by combining two or more processing resources or other components. A single circuit includes multiple components, whether spaced apart or not, that implement the functions of that circuit. Multiple circuits include single components that implement the functions of those circuits.

[0071] Although one embodiment has been described, these embodiments are presented as examples and are not intended to limit the scope of the invention. The novel methods, apparatus, and systems can be implemented in a variety of other forms, and various omissions, substitutions, and modifications can be made without departing from the spirit of the invention. These embodiments and their variations are included in the scope and spirit of the invention, as well as in the claims and their equivalents. [Explanation of Symbols]

[0072] 10 Data Processing Devices 12 Arithmetic unit 14 Processing Unit 16 Output device 18 Input device 20 Scanners 22 Data Storage Unit 24 Data Processing Circuits 26 Interface Circuit

Claims

1. A medical image dataset containing intensity values ​​as a position function is received. Obtain segmentation of bone regions within the aforementioned medical image dataset, The medical image dataset is processed to identify multiple subregions within the bone region whose intensity values ​​are within the intensity threshold range relative to the calibration intensity value. A processing circuit that determines the concentration and / or distribution of the sub-region within the bone region. A medical image processing device equipped with [a specific feature].

2. The processing circuit further determines the presence of a disease condition according to the concentration and / or distribution of the identified sub-region within the bone region. The medical image processing apparatus according to claim 1.

3. The aforementioned medical conditions include osteopenia and / or osteoporosis. The medical image processing apparatus according to claim 2.

4. The processing circuit further outputs an image representing the density and / or distribution of a specified subregion within the bone region. The medical image processing apparatus according to claim 1.

5. The aforementioned bone region includes one or more vertebrae. The medical image processing apparatus according to claim 1.

6. The segmentation includes one or more labels that identify the bone region. The medical image processing apparatus according to claim 1.

7. Each of the aforementioned sub-regions is a pixel or a voxel. The medical image processing apparatus according to claim 1.

8. The processing circuit compares the determined concentration with one or more concentration thresholds. The medical image processing apparatus according to claim 1.

9. The concentration threshold is based on the characteristics of the bone region and / or at least one characteristic of the subject in the medical image dataset. The medical image processing apparatus according to claim 8.

10. The characteristics of the subject include one or more of the following: age, sex, weight, height, body mass index, race, or other characteristics. The medical image processing apparatus according to claim 9.

11. The aforementioned calibration strength values ​​include calibration values ​​representing zero Houndsfield units and / or water. The medical image processing apparatus according to claim 1.

12. The sub-regions within the bone region are identified based on calibration values ​​indicating their respective strengths and / or water content, which are lower than zero Houndsfield units. The medical image processing apparatus according to claim 1.

13. The aforementioned medical image dataset is a CT dataset. The medical image processing apparatus according to claim 1.

14. The processing circuit reduces noise in the medical image dataset. The medical image processing apparatus according to claim 1.

15. The aforementioned noise is reduced using a low-pass filter or anisotropic smoothing. The medical image processing apparatus according to claim 14.

16. The processing circuit detects an image containing an external structure and excludes the image from processing. The medical image processing apparatus according to claim 1.

17. The processing circuit further determines the presence of a fracture according to the concentration and / or distribution of the identified sub-region within the bone region. The medical image processing apparatus according to claim 1.

18. The processing circuit analyzes the distribution of the sub-regions within the bone region and evaluates the future fracture / re-fracture risk of the bone region. The medical image processing apparatus according to claim 1.

19. The processing circuit detects changes in the dimensions of the bone region based on medical image data obtained from the subject at two or more different time points. The medical image processing apparatus according to claim 1.

20. Receiving a medical image dataset that includes intensity values ​​as a position function, To obtain segmentation of bone regions within the aforementioned medical image dataset, The medical image dataset is processed to identify multiple subregions within the bone region whose intensity values ​​are within the intensity threshold range relative to the calibration intensity value, and To determine the concentration and / or distribution of the sub-region within the bone region, A medical image processing method including [a specific term].

Images