DE102023200942A1 - Estimating the weight of a subject in a medical imaging procedure - Google Patents

Abstract

Method for estimating the weight of a subject (4), in particular a patient, in the context of a medical imaging method, comprising the following steps: (a) acquiring image data of at least a section of the subject (4), (b) estimating the mass and/or the mass distribution of the subject (4) based on the image data.

Description

The invention relates to a method for estimating the weight of a subject in the context of a medical imaging procedure, a computer program product and a medical imaging system.

Regardless of the grammatical gender of a particular term, persons with male, female or other gender identity are included.

When registering a patient, it is common practice to query or enter the patient’s weight, among other things. Weight is required for and influences several medical and technical aspects. For example, weight plays an important role in calculating a specific absorption rate (SAR), which is required for magnetic resonance imaging (MRI), and for the SUV (“standard uptake value”), which is required in positron emission tomography (PET). Weight is therefore an important parameter, especially for combined PET/MR hybrid procedures. In practice, however, it is often not easy to determine weight during an imaging procedure. Weight is sometimes forgotten or entered incorrectly, or medical or technical staff make a rough estimate based on the patient’s stature.

Furthermore, the weight of a patient can also have an influence on other components, e.g. a patient bed. For example, the precise movement of the patient bed on the x, y and z axes can depend on the weight or weight distribution of the patient.

The current state of the art involves manually entering the patient's weight during patient registration. However, this requires knowledge of the patient's weight. Entries can therefore often be incorrect or out of date, particularly since a body scale is not always available on site or there is not enough time to check the weight using a body scale at the time. For example, a patient can lose a significant amount of weight during the course of a treatment, e.g. chemotherapy.

Furthermore, patient beds and their suspension or fastening are usually built so robustly that the patient's weight has only a small influence on the movement and shape of the patient bed. However, this has the disadvantage that the bed is relatively more expensive due to the more solid construction than a less solid bed would be. Furthermore, such a more solid patient bed can also have an influence on the acquisition of image data. For example, attenuation during a PET measurement can be increased as a result.

It is therefore an object of the present invention to provide a new, preferably improved, possibility with which a patient's weight can be determined relatively easily and reliably within the framework of a medical imaging method. In addition, it is desirable to obtain weight information that is as precise as possible.

This object is achieved by a method according to claim 1, a computer program product according to claim 14 and a medical imaging system according to claim 15.

1. (a) Akquirieren von Bilddaten zumindest eines Ausschnitts des Subjekts,
2. (b) Abschätzen der Masse und/oder der Masseverteilung des Subjekts anhand der Bilddaten.

According to a first aspect of the invention, a method is provided for estimating the weight of a subject, in particular a patient, in the context of a medical imaging method. The method comprises the following steps:

1. (a) acquiring image data of at least a portion of the subject,
2. (b) Estimating the mass and/or mass distribution of the subject from the image data.

The image data are preferably acquired by a medical imaging system, in particular by the imaging system with which the medical imaging method is carried out. It can preferably be provided that after the acquisition of image data in step (a), further image data, in particular the image data intended for the actual main measurement, are recorded or acquired in a further step. This can be provided before, during or after step (b). The subject can in particular be a human or an animal. It can be provided that the subject or the patient is not repositioned between and during the acquisition of the image data and the further image data. This can, for example, have the purpose of ensuring that the position of the subject in the image data and the further image data is essentially the same. Alternatively, it can be provided that the patient is moved or repositioned between the acquisition of different image data. This can be made possible in a particularly advantageous manner by determining the mass distribution, whereby, for example, a changed attenuation and its influence on a measurement can be traced based on the mass distribution of the patient.

The imaging method can be, for example, a positron emission tomography/magnetic resonance imaging method (PET/MR imaging method). The image data can preferably be MRI image data. However, the method according to the invention can also be carried out with other imaging methods and image data. The image data can preferably be three-dimensional (3D) image data. Alternatively, the image data can be two-dimensional (2D) image data and/or a stack of 2D image data, eg of slices of the subject.

The estimation of the mass and/or the mass distribution is preferably carried out by a computer unit, e.g. a CPU of a computer, a server or an imaging system, e.g. an MR scanner. The mass or the mass distribution of the subject corresponds in particular to a weight or a weight distribution of the subject. Accordingly, the estimation of a weight or a weight distribution of the subject is to be understood as included in the context of this invention and the terms mass/weight can be regarded as interchangeable. It can be provided that a user enters the part, in particular the body part, of the subject and/or the section of the subject from which image data is acquired, wherein the user's information can be taken into account in particular when estimating the mass and/or the mass distribution. Entering the body part or section can enable a more precise or reliable estimation of the respective weight, for example by using a density or mass model suitable for the respective body part or section. The body part can be, for example, an organ, in particular an organ that is primarily to be examined in the medical imaging method. The method is in particular computer-implemented.

The method according to the invention can advantageously make it possible to estimate a patient's weight by using acquired image data. This can in particular be image data that is recorded anyway as part of the imaging procedure. Advantageously, the mass can thus be determined automatically. It is therefore possible, for example, to dispense with requesting information from a patient. A corresponding workflow can thus be accelerated by omitting requesting the patient's weight or weighing the patient before an examination. It can also thus make it possible to rule out errors due to misunderstandings or typos/spelling errors.

As it turns out, a weight distribution or mass distribution can also be estimated using the image data. This can be advantageous, for example, because it allows not only the total weight of the subject but also a distribution of the mass to be taken into account. If the mass distribution is estimated, even more detailed weight-based information can be determined, e.g. the distribution of the weight on a patient bed or other component and the associated effect on the shape and behavior when moving the patient bed or the other component. The mass distribution can be used to determine which forces or pressures act on the patient bed through the subject. It can be provided (e.g. also in an additional process step) that the mass distribution is used to determine an influence of the subject lying at least partially on a component on at least one property, e.g. a shape or density, and/or a behavior of the component. The component can be a component of the imaging system, for example a patient bed or a cushion. The property or behavior of the component can be estimated using a known stiffness of the material. In particular, it can be provided (e.g. also in an additional method step) that the mass distribution is used to determine an influence of the subject lying on the patient bed on at least one property and/or behavior of the patient bed. It can be provided that the mass distribution is used to calculate a deformation of a patient bed when the subject is lying on it. Alternatively or additionally, it can be provided to use the mass distribution of the subject to determine or correct the movement of the patient bed when the subject is lying on it. Advantageously, this can make it possible to use lighter and less stable components, in particular patient beds, because deformations that can influence a measurement can be calculated via the mass distribution and corrected (for the evaluation of the data or the setting of the measurement parameters). Advantageously, lighter components can have a smaller influence on specific measurements, which can lead, for example, to less PET attenuation and/or less PET scattering and/or less PET losses. Alternatively or additionally, knowledge of the mass distribution can be used to determine an influence of the mass distribution on an imaging measurement. For example, a locally larger mass or density can result in a greater attenuation of the measurement signal there. Bending of the patient bed due to the weight of the subject can also cause a slight dislocation of the subject to This can result in distortions that can affect the accuracy of the measurement. This can be counteracted, for example, by adjusting the measurement parameters according to the expected bending. This can potentially increase the measurement accuracy.

• - Zuteilen der Bildpunkte in verschiedene Kategorien von Bestandteilsarten, wobei den einzelnen Kategorien von Bestandteilsarten jeweils eine vorbestimmte Massendichte zugeordnet ist;
• - Zählen der Bildpunkte jeder Kategorie und Bestimmen von Massebereichen jeweils einer Kategorie; und
• - Abschätzen der Gesamtmasse und/oder der Masseverteilung des Subjekts anhand der Massebereiche, der Zahl der Bildpunkte in den jeweiligen Massebereichen und der Massendichte der Massebereiche.

According to one embodiment, the image data comprises pixels representing components of the subject, wherein estimating the mass and/or mass distribution based on the image data comprises the following substeps:

• - allocating the pixels into different categories of component types, each category of component type being assigned a predetermined mass density;
• - Counting the pixels of each category and determining mass ranges of each category; and
• - Estimating the total mass and/or mass distribution of the subject based on the mass regions, the number of pixels in each mass region, and the mass density of the mass regions.

The image points can in particular be pixels (of two-dimensional image data) or voxels (of three-dimensional image data). The component types can in particular include tissue types and/or implants or types of implants. Different component types can include, for example, bone, soft tissue, air, metal or plastic (of implants).

Soft tissue can in particular include fatty tissue, muscle tissue, skin tissue, etc. The categories can in particular include categories of tissue types and/or at least one category of implants. Different categories or component types can differ in terms of material properties, in particular composition. The different categories can preferably differ in terms of the material density of the component types. Different tissue categories can include, for example, bone, blood, lung tissue and/or fatty tissue. Different categories of implants can, for example, differ in terms of the material of the respective implant. In the image data, the different categories of component types can be differentiated in particular by the contrast values of the image data (e.g. the MR image data). Different imaging methods or different sequence protocols of a specific method (e.g. MR) can each have different contrast values for the different component types. The allocation can preferably be coordinated with the imaging sequence used in each case. The image data can in particular be segmented into the different mass ranges of different categories. The total mass and/or the mass distribution can be determined, for example, based on the gray values of the pixels, with different gray values representing different types of components. The pixels counted in each mass range can be multiplied by previously known typical or average mass densities for the respective category or component type in order to determine the mass or mass distribution. The previously known mass densities can, for example, be mass densities standardized in the prior art. The mass densities can, for example, be taken from a database of mass densities. The mass ranges or the pixels of the mass ranges can be assigned to their local position. The mass distribution can be determined based on the allocation to the local position. It can be provided that a user is enabled or that the user is prompted to specify a body part or the section of the subject from which the image data is acquired. The user input can be used to make an allocation to the categories more reliable. For example, by selecting a body part, the categories in question can be restricted or defined more precisely.

According to one embodiment, the step of estimating the weight of the subject comprises: if the image data does not comprise the entire subject but only a section of the subject, extrapolating from the specific mass ranges of the section to the entire mass and/or the entire mass distribution of the subject using a predetermined mass model. The mass model can be, for example, a model known from the prior art for a typical mass distribution of a corresponding subject, in particular a human. Advantageously, this makes it possible to draw conclusions about the entire mass or mass distribution of the subject from a section of the subject.

According to one embodiment, the categories of component types comprise cortical bone and bone marrow, wherein the mass ranges of these tissue areas are determined by means of an algorithm comprising bone segmentation. The bone segmentation may in particular comprise a division of the bone into its various components, in particular various components which produce different contrast levels during measurement and have different mass densities. It may happen that different components have the same contrast levels (eg bone and air). It may be possible to assign data to different categories based on a distribution model of the respective body part from which data is acquired, even if the contrast levels are the same. The distribution model can include a typical distribution of the various types of components in the body part.

According to one embodiment, the image data are acquired as part of a preparatory measurement before a main measurement of the medical imaging method. The preparatory measurement can preferably be an ultra-fast measurement with low resolution. Ultra-fast and low resolution refer in particular to a main measurement provided as part of the imaging method. Ultra-fast or low resolution can preferably mean that the preparatory measurement is at least an order of magnitude, i.e. a factor of 10, faster or has a number of pixels that is at least a fifth lower. For example, a spatial resolution of 4 to 50 mm ³ can be provided. The preparatory measurement can preferably be a measurement of three-dimensional image data. The preparatory measurement can be the recording of a scout image. A low-resolution scout image can in particular have a spatial resolution of 2×2mm ² to 8×8mm ² , preferably 3×3 to 5×5mm ² . Preferably, by using the preparatory measurement, which is anyway useful or necessary for the actual main measurement, it can be achieved that no additional measurement is required to determine the weight.

According to one embodiment, the preparatory measurement is a Dixon sequence, in particular for determining an attenuation map for a magnetic resonance imaging/positron emission tomography measurement. For PET/MR measurements, an attenuation map is required to enable quantitative PET. It is common practice in the prior art to use an MR Dixon sequence for this purpose. Advantageously, a Dixon sequence that is required anyway can also be used for weight estimation. It has been found that a Dixon sequence is also particularly well suited to determining the various categories of component types for weight determination. In particular, it can be used particularly well to distinguish between water and fatty tissue or between fatty tissue and other types of tissue.

According to one embodiment, the mass distribution is estimated by using the respective associated mass density for the mass regions to produce a mass map. The mass map can advantageously be used to represent the mass distribution. The mass map can, for example, assign a specific mass or mass density to different sections. The sections can in particular be sections of the subject. The mass map can be produced three-dimensionally or two-dimensionally. For example, the mass map can represent a two-dimensional mass distribution along a patient bed. The mass map can be used, for example, for a SAR model, determining the influence of the subject on the patient bed, an SUV distribution, an elastography or determining an influence on measuring coils or an arrangement of measuring coils. The attenuation map can in particular be produced analogously to the production of a PET attenuation map known from the prior art.

According to one embodiment, forces exerted locally on at least one component of the imaging system are determined based on the determined mass distribution and the position of the subject. The at least one component can be, for example, a patient bed, a patient cushion, e.g. for supporting body parts, in particular the head, and/or a measuring component, e.g. an MR coil. It can be provided to use the determined forces to determine how a geometric deformation of the component caused by the weight of the subject turns out. It can be provided to determine a characteristic curve or curve of the component based on the forces exerted. For example, a patient bed in the method can have a characteristic curve or curve that is influenced by the weight of the subject. The characteristic curve or curve can be designed in such a way that it sets a relationship between the weight distribution and a dependent variable. The dependent variable can be a parameter relating to the component. For example, the dependent variable can be a table position of the patient table or the patient couch after a travel path of the patient table or the patient couch. According to one embodiment, the characteristic curve or curve can be a hysteresis characteristic. Advantageously, by determining these forces, the components, e.g. the patient couch, or their behavior can be better characterized. In this sense, a characterization can include a movement behavior or a deformation of the component. For example, the weight can generate increased friction (possibly dependent on the mass distribution), which counteracts a controlled movement. Material properties of the component, e.g. a stiffness of the material, can be used to calculate the deformation. The precise Determining deformations can be useful when a precise correction of an attenuation is required, such as in modern, high-resolution PET/MR systems. Preferably, this embodiment can use a lighter component and/or a less stable attachment of the component. This can, for example, save material or use a less expensive material. Additionally or alternatively, an attenuation of the measurement signal can be reduced by choosing a lighter material, e.g. the attenuation of a PET signal.

According to one embodiment, an attenuation map for a positron emission tomography measurement is determined based on the estimated mass distribution. For PET/MR measurements, an attenuation map must be created to enable quantitative PET imaging. This is required in particular to take into account an attenuation of the signal by the anatomical structure of the subject around the area being examined as well as by components of the measurement system, e.g. MR body coils, when evaluating the image data. The attenuation would otherwise distort the measurement results or make a quantitative evaluation impossible. The quality of this attenuation map can advantageously be improved by the information about the mass distribution, in particular by determining a more precise attenuation map in which the current position and shape of the components at the time of the measurement can be determined more precisely.

According to one embodiment, the estimated mass and/or the estimated mass distribution is used for patient registration. Patient registration can thus advantageously be simplified in that a direct query (e.g. verbally or by questionnaire) of the patient's weight is not required. In this context, patient registration can be understood in particular as patient admission or anamnesis.

According to one embodiment, a specific absorption rate is determined based on the estimated mass distribution. In particular, a local specific absorption rate can be determined. For patient safety and to comply with safety regulations, it is necessary to limit the specific absorption rate (SAR) in MR examinations. The SAR is used to measure energy absorption in tissue, which is caused in particular by high-frequency excitation pulses. High-frequency excitation pulses are typically used in MR imaging, for example. The energy absorption heats the tissue and the SAR must be limited to avoid tissue damage due to overheating. Not only the total absorption of energy across the entire body plays a role, but local absorption at different locations on the body is also important. Depending on the local thermal conductivity and also the local heat absorption rate, radiation can also lead to locally excessive heat. This in turn depends on the local tissue type and also on the mass in the irradiated area (e.g. abdomen, diaphragm-pelvis). If the mass distribution is known, knowledge of the local mass can be used to estimate how much mass the radiated energy is distributed over. Advantageously, this makes it possible to determine the maximum possible energy absorption or heating limit for the individual local areas of the subject more precisely. In the prior art, for example, a generic model of the mass distribution is used for this purpose. This enables greater accuracy by using the actual mass distribution of the respective individual subject. In order not to exceed the SAR, low flip angles, low high-frequency excitation energies and low repetition times often have to be used in MR measurements. This limits the achievable image contrast and also the scan time. By being able to determine the SAR more precisely locally by knowing the mass distribution, it is possible to exploit these scan parameters further because there is less uncertainty regarding the SAR caused by them. This enables better measurement results, i.e. with improved contrast, for example, to be achieved.

According to one embodiment, a standardized uptake value, SUV, is determined based on the estimated mass and/or mass distribution, preferably at least also on the basis of the estimated mass distribution. The SUV can preferably be determined in a PET imaging. The SUV is defined in particular as the ratio of the measured activity and the injected activity per kilogram of body weight as a unitless quantity. It can be provided to derive a density distribution (in particular a density distribution of the body tissue) from the mass distribution in order to determine the SUV from this. The SUV can be required in particular for PET measurements in order to determine a quantification of (local) concentrations of radioactivity. This can be used, for example, to compare uptakes from tumors from different measurements and patients. Tumors often have a significantly increased SUV. The SUV can be used in particular to track a response to tumor therapy. To standardize the values obtained from the As described, the patient's weight is typically used to determine the radioactivity concentration determined from the PET image data. It should be noted that a patient's weight can change, for example during the course of therapy. For example, a patient can lose weight during chemotherapy. A weight loss usually leads to a lower SUV value, although the actual uptake in the tumor can still be identical at the same time, since the value is standardized using a lower weight. If the weight in the patient files is therefore not current, this can lead to an incorrect determination. For example, it can happen that the patient appears to be responding to the tumor therapy, but the changed SUV is exclusively or largely due to the patient's (not taken into account) weight loss. The possibility of determining the mass directly before a corresponding PET measurement (e.g. using a preparatory measurement) can therefore be advantageous in order to have current weight data available without having to weigh the patient separately. Knowledge of the current mass distribution can also enable even more precise monitoring of the course of the disease. In particular, it can be provided to determine a corrected SUV based on the mass distribution. For example, the activity concentration required for the SUV can be standardized to the fat-free body mass. It can be provided to determine the fat-free body mass based on the mass distribution. This can, for example, compensate for the effect of a changed body mass during therapy, e.g. due to the reduction of body fat, in order to limit the changes in the SUV value purely to a change in the uptake of the tumor during therapy.

According to one embodiment, in order to estimate the mass and/or the mass distribution of the subject, the image data are input into a trained neural network, wherein the neural network is trained to determine the mass and/or the mass distribution based on the image data, wherein the neural network outputs the mass or the weight and/or the mass distribution. The output of the mass can be provided as a scalar value. For example, the mass can be output as a weight value. The neural network can be a fully connected network or comprise fully connected layers. The network can preferably have a U-Net structure. In particular for determining a mass distribution, it can be provided that the neural network is a convolutional neural network (CNN). The neural network, in particular the CNN, can preferably comprise 2 to 5, particularly preferably 2 to 3 layers. The network can preferably have a U-Net structure or a variant of a U-Net with an additional layer at the end of the network. By providing relatively few layers, in particular 2 to 3 layers, the computational effort can be reduced, whereby 2 to 3 layers can be sufficient to achieve reliable results. The result or output of the U-Net can be a mass distribution with the same number of voxels and/or pixels as the image data used for input into the U-Net. A scalar weight value or the weight or total weight can be determined via a summation layer of all output voxels or pixels of the U-Net. Existing image data, in particular 3D image data, can be used to train the neural network. A weight or weight distribution (or mass or mass distribution) can be assigned to the 3D image data as a reference. The weight or weight distribution can be annotated manually for this purpose. Alternatively or additionally, the weight distribution used for training purposes can be determined using a point scale.

According to one embodiment, a user input relating to the mass is received from a user, wherein the estimation of the mass and/or the mass distribution is also carried out based on the user input. The user input can include, for example, the age and/or a fitness level of the subject. For example, the bones of the subject can vary depending on age and fitness. Corresponding information can be used, for example, to be able to more precisely assume bone density and to be able to determine the mass distribution more precisely. According to a preferred embodiment, the user input is a weight of the subject, wherein the mass distribution is estimated taking into account the weight of the subject, in particular such that the mass distribution is estimated such that a resulting total mass corresponds to the weight of the subject. In other words, it can be provided that the determination of the mass distribution or weight distribution is set or calibrated to the total weight. This can advantageously enable a more precise estimation of the mass distribution by ensuring that the total mass resulting from the estimated total mass distribution corresponds to the actual mass.

According to one embodiment, at least one optical camera image of the subject is recorded and/or received, wherein the estimation of the mass and/or the mass distribution is also carried out based on the at least one camera image. In particular, it can be provided be that a total mass is estimated based on the body shape of the subject resulting from the camera image and/or information on a mass distribution is derived. Provision can be made to adapt the mass distribution to the body shape resulting from the camera image. Advantageously, this embodiment can enable a more precise estimation of the mass distribution by ensuring that the estimated mass distribution conforms to the body shape of the subject.

According to one embodiment, a mass of the subject is estimated based on the at least one camera image, wherein the mass distribution is estimated taking into account the mass estimated by the camera image, in particular such that the mass distribution is estimated such that a total mass resulting from the mass distribution corresponds to the mass of the subject estimated from the at least one camera image.

A further aspect of the invention is a computer program product comprising instructions which, when the program is executed by a computer, cause the computer to carry out the steps of the method as described herein. All advantages and features of the method can be transferred analogously to the computer program product and vice versa. The computer program product can be stored, for example, on a computer-readable storage medium, in particular a non-volatile storage medium. The storage medium can be, for example, a hard disk, an SSD, a flash memory, an online server, etc.

A further aspect of the invention is a medical imaging system, wherein the imaging system comprises a measuring system for generating image data of a subject, wherein the imaging system is configured to acquire image data of the subject with the measuring system, wherein the imaging system is configured to estimate a mass and/or a mass distribution of the subject based on the image data. The imaging system can in particular be configured to carry out the method as described herein. All advantages and features of the method and the computer program product can be transferred analogously to the system and vice versa. The medical imaging system can in particular be a PET/MR system (positron emission tomography/magnetic resonance system). The imaging system can comprise a computer unit that is configured to estimate the mass and/or the mass distribution of the subject based on the image data. In other words, the imaging system can be configured to estimate the mass/mass distribution using the computer unit. The computer unit can further be configured to control the image acquisition of the imaging system. The computing unit can be a local computing unit or alternatively a remote computing unit. For example, the imaging system can be connected to the computing unit via a network, via the Internet or via radio. In other words, the medical imaging system does not have to be completely localized at the location of the measurement, but at least the estimation of the mass/mass distribution can also be carried out on a component located remotely from the rest of the system.

The imaging system may comprise a user input device, e.g. a computer mouse or keyboard, and/or a user interface for entering user inputs. The imaging system may be configured to receive a user input relating to the mass from a user via the user input device and/or the user interface. The imaging system may comprise an interface configured to receive an optical camera image, in particular of the subject. Optionally, the imaging system may comprise an optical camera configured to record an optical camera image of the subject. The camera may be a 2D camera or a 3D camera. The camera, in particular a 3D camera, may optionally be configured to record the optical camera image with additional depth information. The optical camera may be configured to forward the optical camera image to the computer unit. The imaging system and/or the computer unit may be configured to carry out the estimation of the mass and/or the mass distribution based on the optical camera image and/or the user input.

According to one embodiment, the imaging system comprises a patient bed, wherein the imaging system is configured to estimate a deformation of the patient bed caused by the subject and/or a driving behavior of the patient bed influenced by the subject based on the estimated mass distribution of the subject. Advantageously, the imaging system can thus be configured to estimate the distribution of the weight of the subject on the patient bed and the associated effect on the shape and behavior when moving the patient bed. The mass distribution can be used to determine which forces or pressures act on the patient bed through the subject. The imaging system can be configured to estimate the deformation or behavior of the patient bed using a known stiffness of the material of the patient bed. Advantageously For example, it may thus be possible to use a lighter and less stable patient bed because deformations that may influence a measurement can be calculated and corrected via the mass distribution. Advantageously, a lighter patient bed can have a smaller influence on specific measurements, which can lead to, for example, less PET attenuation and/or less PET scattering and/or lower PET losses. Additionally or alternatively, the imaging system can comprise at least one further component, e.g. a patient cushion or an MR coil, and can be configured to estimate a deformation of the at least one further component caused by the subject based on the estimated mass distribution of the subject. The deformation of the patient bed and/or of the at least one further component can be estimated in particular according to the method described herein relating to estimating the deformation of a component.

All embodiments described herein can be combined with each other, unless explicitly stated otherwise.

• 1 zeigt ein Flussdiagramm eines Verfahrens gemäß einer Ausführungsform der Erfindung,
• 2 zeigt ein Flussdiagramm eines Verfahrens gemäß einer weiteren Ausführungsform der Erfindung, und
• 3 zeigt ein medizinisches Bildgebungssystem gemäß einer Ausführungsform der Erfindung.

In the following, embodiments are described with reference to the accompanying figures.

• 1 shows a flow chart of a method according to an embodiment of the invention,
• 2 shows a flow chart of a method according to another embodiment of the invention, and
• 3 shows a medical imaging system according to an embodiment of the invention.

1 shows a flow chart of a method according to an embodiment of the invention. The method is used to estimate the weight of a subject 4, in particular a patient, and is carried out as part of a medical imaging method. In a first step 10, image data of at least a section of the subject 4 is acquired. The image data is preferably acquired as part of a preparatory measurement, e.g. when taking a scout image or when carrying out a Dixon sequence before a main measurement of the medical imaging method. In a subsequent step 20, the mass and/or mass distribution of the subject is estimated using the image data. Optionally, to estimate the mass and/or mass distribution, the image data can be entered into a trained neural network, wherein the network outputs the mass and/or mass distribution based on the entered image data in accordance with its training.

2 shows a flow chart of a method according to a further embodiment of the invention. In a first step 10, image data of at least one section of the subject 4 is also acquired here. The subsequent step of estimating the mass and/or the mass distribution of the subject is divided into several steps. First, in a step 21, the image points are allocated to different categories of component types, with a predetermined mass density being assigned to each individual category. The categories of component types can in particular include categories of tissue types and/or categories of implant types. For example, the categories of tissue types can include cortical bone and bone marrow. In a subsequent step 22, the image points of each category are counted and mass ranges of each category are determined. If the tissue types also include bone, these can be determined in particular by means of an algorithm that includes bone segmentation. In a further step 23, the total mass and/or the total mass distribution of the subject 4 is estimated based on the mass ranges, the number of pixels in the respective mass ranges and the mass density of the mass ranges. Optionally, it can be provided that the image data does not include the entire subject 4, but only a section of the subject 4, which is extrapolated from the specific mass ranges of the section to the total mass and/or the total mass distribution of the subject 4 based on a predetermined mass model. Additionally or alternatively, it can optionally be provided that the mass distribution is estimated by using the respective associated mass density for the mass ranges in order to produce a mass map. According to an optional method step 24, a user input relating to the mass can be received from a user and the estimation of the mass and/or the mass distribution can also be carried out based on the user input. The user input can preferably be a weight of the subject 4, in which case the mass distribution is estimated taking into account the weight of the subject 4. This can be provided in particular in such a way that the mass distribution is estimated in such a way that a resulting total mass corresponds to the weight of the subject 4. According to a further optional additional step 25, at least one optical camera image of the subject 4 can be recorded and/or received and used to determine the mass and/or the mass distribution. In this case, the optical image is used in particular to of the subject to the mass of the subject. Finally, it can be provided (optionally) in a further step 30 that an attenuation map for a positron emission tomography measurement is determined based on the estimated mass distribution. Alternatively or additionally, in this step 30 a specific absorption rate and/or an SUV can also be determined based on the estimated mass distribution.

3 shows a medical imaging system according to an embodiment of the invention with a measuring system 2 for generating image data of a subject 4. The imaging system is configured to acquire image data of the subject with the measuring system 2 and to estimate a mass and/or a mass distribution of the subject based on the image data. To estimate the mass or the mass distribution, the imaging system comprises a computer unit 8 which is connected to the measuring system 2. In particular, the computer unit 8 can be configured to initiate the acquisition of the image data by the measuring system 2. A user can enter user inputs such as commands and/or information into the computer unit 8 via at least one user input device 7. The user inputs can in particular relate to the mass of the subject 5. The computer unit 8 can be configured to carry out the estimation of the mass and/or the mass distribution taking the user inputs into account. Furthermore, the imaging system can optionally comprise an optical camera 6, which is configured to record an optical camera image of the subject and forward it to the computer unit 8. The computer unit 8 can in this case be configured to carry out the estimation of the mass and/or the mass distribution based on the camera image, i.e. in addition to the acquired image data. The imaging system further comprises a patient bed 5 and is configured to use the estimated mass distribution of the subject 4 to estimate a deformation of the patient bed 5 caused by the subject 4 or a movement behavior of the patient bed 5 changed by the subject 4. Analogously, the imaging system can also be configured to estimate the influence of the subject 4 on other components on which the subject 4 lies in whole or in part.

Claims (15)

Method for estimating the weight of a subject (4), in particular a patient, in the context of a medical imaging method, comprising the following steps: (a) acquiring image data of at least a section of the subject (4), (b) estimating the mass and/or the mass distribution of the subject (4) based on the image data. Procedure according to Claim 1 , wherein the image data comprise pixels representing components of the subject (4), wherein the estimation of the mass and/or the mass distribution based on the image data comprises the following sub-steps: - allocating the pixels into different categories of component types, wherein the individual categories of component types are each assigned a predetermined mass density; - counting the pixels of each category and determining mass ranges of each category; and - estimating the total mass and/or the mass distribution of the subject (4) based on the mass ranges, the number of pixels in the respective mass ranges and the mass density of the mass ranges. Procedure according to Claim 2 , wherein the step of estimating the weight of the subject (4) comprises: if the image data does not comprise the entire subject (4) but only a section of the subject (4), extrapolating from the determined mass regions of the section to the entire mass and/or the entire mass distribution of the subject (4) using a predetermined mass model. Procedure according to Claim 2 or 3 , wherein the categories of component types comprise cortical bone and bone marrow, the mass ranges of these tissue regions being determined by means of an algorithm comprising bone segmentation. Method according to one of the preceding claims, wherein the image data are acquired as part of a preparatory measurement before a main measurement of the medical imaging method. Procedure according to Claim 5 , wherein the preparatory measurement is a Dixon sequence, in particular for determining an attenuation map for a magnetic resonance imaging/positron emission tomography measurement. Procedure according to one of the Claims 2 until 6 , where the mass distribution is estimated by using the corresponding mass density for each mass range to produce a mass map. Procedure according to Claim 7 , wherein forces exerted locally on at least one component of the imaging system are determined based on the determined mass distribution and the position of the subject (4). Method according to one of the preceding claims, wherein an attenuation map for a positron emission tomography measurement, a specific absorption rate and/or a standardized uptake value, SUV, is determined based on the estimated mass and/or mass distribution. Method according to one of the preceding claims, wherein, to estimate the mass and/or the mass distribution of the subject (4), the image data are input into a trained neural network, wherein the neural network is trained to determine the mass and/or the mass distribution based on the image data, wherein the neural network outputs the mass and/or the mass distribution. Method according to one of the preceding claims, wherein a user input concerning the mass is received from a user, wherein the estimation of the mass and/or the mass distribution is also carried out based on the user input. Procedure according to Claim 11 , wherein the user input is a weight of the subject (4), wherein the mass distribution is estimated taking into account the weight of the subject (4), in particular such that the mass distribution is estimated such that a resulting total mass corresponds to the weight of the subject (4). Method according to one of the preceding claims, wherein at least one optical camera image of the subject (4) is recorded and/or received, wherein the estimation of the mass and/or the mass distribution is also carried out based on the at least one camera image. Computer program product comprising instructions which, when the program is executed by a computer, cause the computer to carry out the steps of the method according to any one of the preceding claims. Medical imaging system, wherein the imaging system comprises a measuring system for generating image data of a subject (4), wherein the imaging system is configured to acquire image data of the subject (4) with the measuring system (2), wherein the imaging system is configured to estimate a mass and/or a mass distribution of the subject (4) based on the image data, wherein the imaging system is in particular configured to carry out the method according to one of the Claims 1 until 13 to carry out.

Images