Method for evaluating availability of 4D tomographic image data, computer program product and scanner device

Abstract

A method, a computer program product, and a scanner apparatus for evaluating availability of 4D tomographic image data are provided. The method includes receiving 4D tomographic image data including a plurality of pieces of 3D tomographic image data of an examination object, the plurality of pieces of 3D tomographic image data corresponding to a plurality of time points, applying a segmentation algorithm to the plurality of pieces of 3D tomographic image data, wherein the segmentation algorithm is configured to segment at least one organ in the 3D tomographic image data to which the algorithm is applied, applying a scoring function to the segmented organ of the 3D tomographic image data, wherein the scoring function is configured to determine a score value of the segmented organ to which the scoring function is applied, the score values include and / or correspond to metrics that quantify the extent to which a vicinity of voxels at the surface of the segmented organ in the 3D tomographic image data contains image artifacts, comparing the score values to a threshold value, and providing a user notification when at least one score value exceeds the threshold value.

Description

Technical Field

[0001] The invention relates to a method for assessing the usability of 4D tomographic image data. The method is in the field of medical technology, in particular in the field of medical imaging and the use of medical images in radiation therapy. Background Art

[0002] Time-resolved 4D computed tomography (4DCT) is essential for radiotherapy treatment planning of mobile tumors. It generates 3D tomographic images at multiple time points throughout the patient's respiratory cycle. However, technical challenges often plague the process. Irregularities in the patient's breathing pattern during 4DCT acquisition, or mismatches between the scanner's rotation time and the patient's breathing rate, can lead to severe image artifacts. These artifacts (including stack transition artifacts, interpolation artifacts, and motion artifacts) are likely to render the acquired images unsuitable for treatment planning. The consequences of these problems are far-reaching. If these artifacts are not detected during the initial 4DCT scan and are only discovered during the treatment planning phase, a complete repetition of the scan is required, resulting in a significant increase in both cost and time, patient discomfort, and potentially compromised treatment results due to delays. Therefore, it is extremely important to develop an automatic algorithm for detecting and locating such image artifacts. The algorithm can pinpoint the exact location of the artifact, thereby providing the treating physician with the necessary information to make an informed decision about whether the scan needs to be repeated.

[0003] In the current scenario of clinical practice, the evaluation of 4DCT images occurs after the scan, usually by a therapist or physician. During this evaluation, critical decisions must be made quickly - whether to proceed with the acquired data or to initiate a rescan of the patient in the presence of artifacts that could significantly affect subsequent treatment plans. Given the inherent complexity of 4DCT images and their high-dimensional nature, a comprehensive manual review of the entire dataset is not practically feasible. Instead, health care professionals resort to examining only a portion of the image data, usually focusing on a subset of key slices and specific time points.

[0004] This selective review approach, while efficient in terms of time, is not without its disadvantages. An inherent risk is the potential oversight of image artifacts, which may remain undetected during the initial assessment, but surface later during the treatment planning phase. This delay in the discovery of these artifacts may result in several disadvantages. It amplifies the necessity for repeat scans, thereby imposing additional financial costs and extended treatment timelines on the patient. Furthermore, it may cause increased levels of patient discomfort due to the repeated procedures, and may compromise the efficacy of the treatment itself. The limitations of this current practice highlight the urgent need for more effective and comprehensive technical solutions in the evaluation of 4DCT images to enhance the overall quality and efficiency of patient care.

[0005] Document US10803587B2 describes a method comprising: capturing respiratory movement of a patient; determining respiration-related parameters from the respiratory movement of the patient; specifying a measurement area for an imaging examination, the measurement area comprising at least one z position; automatically calculating at least one measurement parameter based on the respiratory movement using the respiration-related parameters as input parameters; and performing an imaging examination of the patient via computed tomography based on at least one measurement parameter in the measurement area to capture projection data, wherein the projection data, when captured, depicts the respiratory cycle of the patient at at least one z position over the complete duration of the respiratory cycle. Summary of the invention

[0006] The present invention effectively solves a series of key challenges in the field of tomographic imaging. First, it ensures comprehensive detection of artifacts within time-resolved tomographic data sets without leaving room for negligence. In addition, the innovative technical solution facilitates real-time artifact detection, thereby ensuring that potential problems are identified when they occur rather than after the fact. In addition, it achieves these goals with excellent efficiency, so that it can be performed on a conventional computer or via a cloud-based system using standard system resources. Crucially, the system enables artifact detection while the patient remains in the scanner, thereby allowing instant data acquisition and eliminating the need for additional scanning sessions. In addition, it enables the scanner operator to view only those data sets or 3D data that exhibit artifacts, thereby simplifying the decision-making process regarding the necessity of further imaging. Importantly, the present invention ensures that all scanned artifacts are carefully scrutinized, leaving no room for negligence or error in the evaluation process. Ultimately, it eliminates the possibility of artifacts escaping the operator's attention, thereby enhancing the overall quality and reliability of the tomographic image data.

[0007] Independent of the grammatical use of the term, individuals with male, female, or other gender identities are included within the term.

[0008] The present invention relates to a method for assessing the availability of 4D tomographic image data, the method comprising the following steps:

[0009] - receiving 4D tomographic image data, wherein the 4D tomographic image data comprises a plurality of 3D tomographic image data of the examination object, wherein the plurality of 3D tomographic image data corresponds to a plurality of time points,

[0010] - applying a segmentation algorithm to the plurality of 3D tomographic image data, wherein the segmentation algorithm is configured to segment at least one organ in the 3D tomographic image data to which the algorithm is applied,

[0011] - applying a scoring function to the segmented organ of the 3D tomographic image data, wherein the scoring function is configured to determine a scoring value for the segmented organ to which the scoring function is applied, wherein the scoring value comprises and / or corresponds to a metric quantifying the extent to which a vicinity of a voxel at a surface of the segmented organ in the 3D tomographic image data contains image artifacts,

[0012] - compare the score value to a threshold,

[0013] - Provide user notification when at least one rating value exceeds a threshold.

[0014] 4D tomographic image data preferably refers to a dynamic data set generated by combining 3D tomographic images acquired at various time points. This technology is particularly important in medical imaging, which includes modalities such as computed tomography (CT), magnetic resonance imaging (MRI), positron emission tomography (PET), etc. The term "4D" means adding time as a fourth dimension, allowing visualization of anatomical structures or pathological processes over a specified time period. In the context of medical imaging, 4D imaging (especially time-resolved 3D tomography) is essential for understanding organ or tumor motion, which is often crucial for accurate diagnosis and treatment planning. 4D tomographic image data is preferably configured as 4D-CT image data or 4D-MR image data.

[0015] In medicine, 4D tomographic image data plays a key role in radiation therapy planning and gating. It is particularly beneficial in managing moving organs and tumors, a common challenge in oncology. By capturing the real-time movement of internal structures, clinicians can optimize radiation therapy, ensuring that the tumor receives the indicated dose while minimizing damage to healthy tissue. This is achieved through gating, a technique that synchronizes radiation delivery with the patient's breathing or cardiac cycle. For example, in lung cancer treatment, 4D imaging helps create a treatment plan that adjusts for lung tumor movement during breathing.

[0016] Artifacts in 3D and 4D tomographic image data can take various forms, with particular concern for motion-based artifacts. In 4D tomographic image data, only some of the 3D tomographic image data may have artifacts, especially only at some points in time. Artifacts include blurring, misalignment, and distortion of images caused by patient motion during image acquisition. For example, respiratory motion can cause image smearing, where structures appear elongated or distorted. Artifacts compromise the accuracy and reliability of the image. They can lead to incorrect tumor localization, making it challenging to accurately plan and administer radiation therapy. As a result, tomographic image data with these artifacts cannot be used for radiotherapy, or may require extensive post-processing, which can be time-consuming and is often insufficient to mitigate the effects of motion artifacts.

[0017] 4D tomographic image data collected by scanners such as CT, MRI and PET can be provided by the scanner and / or can be received by various devices. 4D tomographic image data can be transmitted via a physical cable connection such as USB, Ethernet or a dedicated connector. The method ensures a direct and reliable link between the scanner and the receiving device, which makes the method of the present invention particularly fast and reliable. Alternatively, wireless data transmission can be used. Wi-Fi or Bluetooth connectivity allows cable-free data transmission, which can be particularly convenient for mobile or portable scanners. Network-based data transmission is common, where the scanner is connected to a local area network (LAN) or a wide area network (WAN). The method enables remote access, storage and retrieval of image data.

[0018] In terms of data format, 4D tomographic image data is preferably provided and / or received in a standard format such as DICOM (Digital Imaging and Communications in Medicine) or FHIR. This format ensures compatibility with medical imaging systems and allows seamless integration with picture archiving and communication systems (PACS).

[0019] 4D tomographic image data can be provided and / or received as raw data or as pre-processed data. Raw data can be configured as raw projection data. Raw data needs to be further processed using software to generate tomographic images. Although this method provides flexibility, it requires a lot of post-processing. Pre-processed data preferably includes reconstructed 2D or 3D images. These images have undergone initial processing, including filtering and reconstruction. Although more user-friendly and suitable for clinical use, they may be less flexible for certain research applications.

[0020] The step of receiving 4D tomographic image data may include extracting individual 3D tomographic image data from the 4D tomographic image data for different time points. This extraction may be necessary to separate the 4D data into its constituent 3D components, thereby enabling analysis and / or segmentation of the anatomical structure or pathological process at a specific moment. The extraction of 3D tomographic image data may include data decomposition. Data decomposition involves acquiring 4D tomographic image data representing a dynamic data set and decomposing it into individual frames corresponding to different time points. These frames are essentially 3D tomographic image data slices, each of which captures the state of the subject being examined at a specific moment during the imaging process. Extracting 3D tomographic image data may be based on time information. The 4D data contains time information, such as the time interval between each 3D image acquisition. This information is critical for associating each 3D image with the precise time point it represents. Extracting 3D tomographic image data may include synchronization. The extraction process may ensure that the individual 3D images are synchronized with the exact moment in the imaging sequence. This synchronization is crucial for accurate time assessment, especially in applications such as cardiac imaging, where the phase of the heartbeat must be precisely determined.

[0021] The extracted 3D tomographic image data may be in the same format as the original 4D data, typically DICOM or another medical imaging format. Alternatively, it may be saved in a format optimized for temporal analysis and display.

[0022] Segmentation algorithms are computational methods designed to separate or delineate specific structures or regions of interest within medical images, such as 3D tomographic image data. The goal is to identify and delineate the boundaries of these structures, thereby distinguishing them from surrounding tissues or regions. The segmentation algorithm is preferably configured to operate by analyzing pixel or voxel values ​​in the 3D tomographic image data to distinguish one type of tissue or organ from another type of tissue or organ.

[0023] For example, the segmentation algorithm can be based on thresholding, where an intensity threshold is set and all pixels or voxels above or below the threshold are assigned to the segmented organ. More complex segmentation algorithms can include region growing, where a seed point is selected and the algorithm iteratively adds neighboring pixels or voxels with similar intensities to the segmented region. Another approach for segmentation algorithms, known as active contour models or snakes, employs a deformable shape that evolves to fit the boundaries of the organ by minimizing an energy function based on image features.

[0024] In particular, the segmentation algorithm may include machine learning-based methods, such as deep learning convolutional neural networks (CNNs). For example, the segmentation algorithm is trained on labeled medical images, and by learning patterns and features within the data, the segmentation algorithm can perform the segmentation task efficiently.

[0025] The input data of the segmentation algorithm are 3D tomographic image data containing voxel values ​​representing anatomical structures. In particular, for each time point, the 3D tomographic image data associated with that time point are the input data of the segmentation algorithm. The output can be a binary mask or a label map indicating the organ of interest within the original image. Pixels or voxels belonging to the segmented organ are labeled as "1", while those outside are labeled as "0". The output can also be a 3D model, 3D shape or 3D contour of the segmented organ.

[0026] Implementation of the segmentation algorithm may be accomplished by specialized medical imaging software, automated scripts, or custom software developed in a programming language such as Python or MATLAB.Preferably, the segmentation algorithm is configured to segment the organ without manual intervention.

[0027] The scoring function is applied to the segmented organs in the 3D tomographic image data. In other words, the input of the scoring function is the segmented organs and / or the output from the segmentation algorithm. Preferably, the input of the scoring function is the 3D tomographic image data, including the segmented organs or attached to the segmented organs. The scoring function is configured to determine a score value for the segmented organ to which the scoring function is applied, wherein the score value includes and / or corresponds to a measure of the extent to which image artifacts are contained near a voxel at the surface of the segmented organ in the 3D tomographic image data. The segmented organs may include lungs, heart, liver, kidneys or other internal organs. The segmented organs may include segmented tumors.

[0028] The scoring function is an algorithm or method applied to the segmented organ. It evaluates the extent to which image artifacts exist near voxels on the surface of the organ. The scoring function can detect and evaluate image artifacts based on various techniques and algorithms. The scoring function is preferably configured as a machine learning and / or training function. Preferably, the scoring function is applied to each segmented organ individually. Alternatively, the scoring function is applied to multiple or all segmented organs in the 3D tomographic image data. The metric is preferably a mathematical formula or process for quantifying the extent of image artifacts. The metric can be based on various parameters, such as artifact intensity, its spatial distribution or other characteristics.

[0029] The scoring value generated by the scoring function is preferably a numeric indicator. A higher value may indicate that there are more image artifacts near the surface of the organ, while a lower value will suggest fewer image artifacts. Alternatively, the scoring value is a string value or a mixed character value.

[0030] Image artifacts are unwanted perturbations or irregularities in medical images. They can be caused by factors such as patient motion during acquisition, image noise, metal implants, or other issues. These image artifacts can compromise the accuracy of diagnosis and analysis.

[0031] In the step of comparing the score value to a threshold, the score value obtained from the previously described scoring function is compared to a predefined threshold. The threshold is a predetermined limit or criterion that is used as a reference point for evaluating the quality of the segmented organ in the 3D tomographic image. If the score value exceeds the threshold, this indicates that the image contains a significant level of image artifacts and further action may be required. The threshold may be an organ-specific value, e.g., the lungs and the heart have different thresholds.

[0032] For example, the threshold is set to 0.7, and the score value for a specific organ is calculated to be 0.8. This indicates that the image of this particular organ has a high degree of image artifacts exceeding the predefined threshold.

[0033] Once the comparison is performed, and it is determined that one or more of the score values ​​have exceeded a threshold, a user notification is generated. The notification is used to alert the user or operator of the 3D tomographic imaging system that significant image artifacts are present in the segmented organs, thereby requiring attention or further investigation. For example, when the score value of the patient's lung segmentation exceeds a threshold of 0.7, the radiologist or technician operating the medical imaging device receives an immediate notification on their workstation. The notification prompts the user to review the image for potential image artifacts or rescan for a more accurate diagnosis.

[0034] In summary, the process ensures that when the quality of the segmented organs and / or the quality of the segmented sub-parts of the organs in the 3D tomographic images drops below an acceptable threshold due to image artifacts, the system generates notifications to alert the relevant users, thereby enabling them to take appropriate actions for a more accurate and reliable evaluation.

[0035] In particular, the 4D tomographic image data comprises a number N of 3D tomographic image data, wherein a segmentation algorithm is applied to the N 3D tomographic image data, wherein in the step of applying a scoring function, at least N scoring values ​​are determined. In particular, for each time point and / or for each 3D tomographic image data, at least one scoring value is determined. In particular, the process involves analyzing the 4D tomographic image data by applying the segmentation algorithm to a sequence of 3D images and then using a scoring function to evaluate the quality of each segmentation. By generating one or more scoring values ​​for each image, changes in image quality over time or across different 3D images in a data set can be tracked, which may be crucial for accurate diagnosis or research purposes.

[0036] Preferably, 4D tomographic image data refers to a collection of sequential 3D tomographic images captured over time. For example, in medical imaging, a 4D data set may represent a series of CT scans acquired at various phases of a patient's respiratory cycle. Each 3D tomographic image data in the data set represents the same anatomical region, but at different time points. For example, in cardiac imaging, 4D tomographic image data may include 20 3D tomographic image data of the heart, each of which is collected at a different phase of the cardiac cycle. N represents the total number of individual 3D tomographic image data in the 4D data set, in particular the total number of 3D tomographic images in the 4D data set. It means the range of data that can be used for analysis and segmentation.

[0037] In particular, a segmentation algorithm is applied to N 3D tomographic image data, where N is an integer. The number N is preferably between 5 and 100, and in particular between 10 and 25. The segmentation algorithm is used to identify and mark a region of interest (e.g., an organ or structure) within each of the N 3D tomographic images. The process results in a segmentation specific to each individual image within the data set. For example, in the context of lung imaging, a segmentation algorithm is applied to each of 100 3D CT scans within a 4D data set to delineate the lungs within a respiratory cycle. After segmentation, a scoring function is applied to each of the N segmented 3D tomographic images. The scoring function quantifies the extent of image artifacts near the surface of the segmented organ in each image. For example, after segmenting the lungs in each of the 100 3D CT scans of a patient, a scoring function is applied to evaluate the presence and extent of image artifacts at the surface of the liver in each image. For each time point or for each of the N 3D tomographic images, one or more scoring values ​​are determined. The number of scoring values ​​corresponds to the number of segmented images and represents the quality of each segmentation. For example, in imaging of the lungs, a scoring value is determined for each of 100 3D images acquired at different time points, allowing evaluation of image quality and potential motion artifacts at each phase.

[0038] In particular, the segmentation algorithm is configured to segment M organs in the 3D tomographic image data to which the segmentation algorithm is applied, wherein M is an integer and / or M is equal to or greater than 2, preferably equal to or greater than 5. Preferably, the segmentation algorithm is configured to segment all organs in the 3D tomographic image data. Optionally, the segmentation algorithm is configured to segment the M organs with the most contrast. In the step of applying the scoring function, at least M scoring values ​​are determined for each time point and / or each 3D tomographic image data. In other words, for each of the M segmented organs at each time point and / or each 3D tomographic image data, a scoring value is determined. In particular, for a segmentation algorithm applied to N 3D tomographic image data and configured to segment M organs for each 3D tomographic image data or each time point, N*M scoring values ​​are determined.

[0039] Preferably, the segmentation algorithm is configured to generate a 3D contour and / or a 3D volume of the segmented organ, wherein the step of applying the segmentation algorithm comprises providing the generated 3D contour and / or 3D volume to the step of applying a scoring function. A 3D contour (short for three-dimensional contour) refers in particular to a three-dimensional representation of the outer boundary or surface of the segmented object. This is similar to outlining the edge of an object in a 3D space, outlining its shape in a way that can be easily visualized and measured. A 3D volume (also referred to as a three-dimensional volume) in particular represents the interior of the segmented object, effectively filling the outlined space with volume data. In the context of medical imaging, it can provide a detailed representation of the internal structure of the object and can be used for various quantitative analyses. Once the segmentation algorithm has generated a 3D contour and / or a 3D volume, the data is then forwarded or input into a scoring function. The scoring function processes this information.

[0040] According to a preferred embodiment of the present invention, the 4D tomographic image data includes a plurality of segmentable organs, wherein the segmentation algorithm is configured to segment a plurality of sub-portions of the segmentable organs of the 4D tomographic image data, wherein the sub-portions of the segmentable organs include organs with the highest contrast and / or the most prone to motion artifact errors. Multiple segmentable organs particularly mean that the 4D tomographic image data covers a plurality of organs or anatomical structures that can be isolated and analyzed separately. In medical imaging, various organs within the field of view can be segmented or separated for detailed examination. The sub-portion in this context represents a portion or selection from a larger set of segmentable organs in the 4D tomographic image data. In medical imaging, high contrast refers to a clear and obvious distinction between two or more tissues or structures in an image. Organs or structures with high contrast can be easily distinguished from their surroundings. For example, in CT scans, bones usually have high contrast because they appear bright white, making them stand out relative to soft tissues. In tomographic imaging, motion artifacts occur when there is undesirable movement during image acquisition. Certain organs or structures are more susceptible to these artifacts due to their location and inherent mobility. For example, structures near areas of the body with significant motion, such as the diaphragm, may also be affected by motion artifacts.

[0041] Preferably, the scoring function comprises a local Hough transform, and / or the scoring function is configured as a local Hough transform. The local Hough transform is configured to detect lines and / or the direction of lines, wherein the scoring factor determined in the step of applying the scoring function is based on a number of detected lines and / or the direction of the detected lines. The local Hough transform is a variation of the traditional Hough transform, which is a mathematical technique for detecting geometric shapes, patterns or specific features within an image in computer vision and image processing. The local Hough transform, as the name implies, focuses on a local or specific area within an image and is generally used to find patterns or structures that are not necessarily global in nature. The Hough transform process begins with an input image (e.g., 3D tomographic image data) containing an area of ​​interest or an area where it is desired to identify a specific pattern or structure. The area may contain lines, curves, circles or other geometric shapes of interest. Unlike the standard Hough transform that considers the entire image, the local Hough transform focuses on a specific area or area of ​​interest within the image. The local area is selected to limit the computational workload and to target the detection of patterns in a specific area. The local region may be a segmented organ, a volume of a segmented organ, or a contour. In particular, the local region is the surface of the segmented organ and / or the region around the surface. In the selected local region, the local Hough transform accumulates information about the presence of certain patterns or structures. For example, the transform accumulates evidence of potential lines in the region. The local Hough transform may create a cumulative space, which is a data structure that stores information about patterns and their parameters. In the case of line detection, this may represent different angles and distances of lines. Each pixel in the local region votes for the pattern it believes it has found. After the accumulation step, the local Hough transform may identify a peak in the cumulative space. The peak corresponds to the pattern or structure that has received the most votes from the pixels in the local region. The peak in the cumulative space corresponds to the parameters of the detected pattern. In the case of line detection, these parameters may include the angle and distance of the detected line. The output of the local Hough transform is a detection of the pattern in the selected region. The output typically includes information about the parameters of the detected pattern, such as their position, orientation, and size. These parameters may then be used for further analysis or processing, such as object recognition, feature extraction, or other computer vision tasks.

[0042] In particular, the scoring function comprises a local Hough transform, wherein the local Hough transform is configured to detect lines and / or the direction of lines, wherein in the step of applying the scoring function, the Hough transform is applied to the segmented organ and the vicinity of the segmented organ, wherein the determined scoring factor is based on a comparison of the result of applying the Hough transform to the segmented organ and the result of applying the Hough transform to the vicinity. The local Hough transform is preferably designed to find lines in the image and potentially determine their direction. When applying the scoring function, they use the Hough transform not only on the segmented organ itself but also in the area around the segmented organ (which is called the "neighborhood" of the segmented organ). This implies that they are not only interested in things inside the organ, but also in things around the organ. The final score is preferably calculated based on a comparison between what the Hough transform finds inside the segmented organ and what it finds in the vicinity. In summary, this embodiment describes a method that involves using a specific technique (local Hough transform) to analyze images of the segmented organ and the area around it. The scoring function calculates a score based on the result of the analysis.

[0043] In particular, the scoring function is configured to evaluate the local image contrast along the organ boundary of the segmented organ, wherein by applying the scoring function to the segmented organ, when the local image contrast has a sharp transition, a scoring value corresponding to a stacking transition artifact is determined, and / or when the local image contrast has a blurred transition, a scoring value corresponding to a motion artifact is determined. Preferably, the scoring function is set to analyze the local image contrast specifically along the boundary of the segmented organ. Image contrast refers to the intensity difference between adjacent pixels in an image. When the scoring function detects a sharp transition of the local image contrast along the organ boundary, it assigns a scoring value corresponding to a "stacking transition artifact". A stacking transition artifact can be something like a sudden change in a layer or slice of an image stack, which can indicate a problem in the imaging process or the structure being imaged. On the other hand, when the scoring function identifies a blurred transition of the local image contrast along the organ boundary, it assigns a scoring value corresponding to a "motion artifact". Motion artifacts in medical imaging may be caused by patient movement during scanning, resulting in blurring or distortion in the image.

[0044] Optionally, the step of comparing the scoring values ​​comprises generating a user notification when at least one scoring value exceeds a threshold value. The user notification may be an optical signal, an acoustic signal and / or a tactile signal. Preferably, the user notification comprises a time point associated with the 3D tomographic image data having image artifacts, the associated 3D tomographic image data and / or an overlay image, wherein the overlay image comprises the associated 3D tomographic image, an indication of the image artifact and / or the location of the image artifact. In other words, the user notification preferably indicates to the user the 3D tomographic image data associated with the scoring value. This provides a very efficient way to show the user the problematic 3D tomographic image data without having to scan through all the 3D tomographic image data of the 4D tomographic image data.

[0045] The threshold value may be predefined or user definable, in particular the threshold value is an organ-specific value.If at least one of the score values ​​is above the threshold value, a user notification is triggered.

[0046] The notification may include the original 3D tomographic image data, which allows the user to see the unaltered image. It may also include an overlay image, which is a modified version of the image that highlights or indicates the presence and location of image artifacts. The overlay image is designed to help the user identify the image artifact or its location. It may include visual cues, such as markers or labels that make the artifact more visible.

[0047] Preferably, the segmentation algorithm and / or the scoring function are configured as a machine learning algorithm and / or a machine learning function. The segmentation algorithm and / or the scoring function can be based on machine learning techniques, where the algorithm or function is trained on data to improve their performance. The segmentation algorithm can be a deep learning model, such as a convolutional neural network (CNN), which has been trained on a large data set of medical images to accurately identify and delineate specific organs within the image. The scoring function can be a machine learning function that evaluates image quality and / or segmentation based on learned patterns and features. A machine learning algorithm refers to an algorithm that has been trained on data to perform a specific task.

[0048] In particular, the step of providing a user notification includes providing the user notification to a scanner console of a scanner used to acquire 4D tomographic image data and / or displaying the user notification on the scanner console. This step involves ensuring that the user notification (which may include important information about the quality of the data or the imaging process) is delivered directly to the scanner console where the operator or user is interacting with the imaging equipment.

[0049] Furthermore, the present invention relates to a computer program product comprising a computer readable medium storing a computer program code, which, when executed by a computer processor, configures the computer processor to perform the method as claimed in any one of the preceding claims. The computer program product is in particular configured to perform the following steps when executed by a computer processor:

[0050] - receiving 4D tomographic image data, wherein the 4D tomographic image data comprises a plurality of 3D tomographic image data of the examination object, wherein the plurality of 3D tomographic image data corresponds to a plurality of time points,

[0051] - applying a segmentation algorithm to the plurality of 3D tomographic image data, wherein the segmentation algorithm is configured to segment at least one organ in the 3D tomographic image data to which the algorithm is applied,

[0052] - applying a scoring function to the segmented organ of the 3D tomographic image data, wherein the scoring function is configured to determine a scoring value for the segmented organ to which the scoring function is applied, wherein the scoring value comprises and / or corresponds to a metric quantifying the extent to which a vicinity of a voxel at a surface of the segmented organ in the 3D tomographic image data contains image artifacts,

[0053] - compare the score value to a threshold,

[0054] - Provide user notification when at least one rating value exceeds a threshold.

[0055] The present invention further relates to a scanner device. The scanner device is configured to perform the method according to the present invention. In addition, the scanner device is configured to execute a computer program and / or computer program code of a computer program product according to the present invention. The scanner device comprises an interface module and a processor module. The interface module can be a hardware or software module. The processor module can also be a hardware module or a software module (e.g. in cloud computing).

[0056] The interface module is configured to receive 4D tomographic image data, wherein the 4D tomographic image data includes a plurality of 3D tomographic image data of an examination object, wherein the plurality of 3D tomographic image data corresponds to a plurality of time points.

[0057] The processor module is configured to apply a segmentation algorithm to the plurality of 3D tomographic image data, wherein the segmentation algorithm is configured to segment at least one organ in the 3D tomographic image data to which the algorithm is applied.

[0058] In addition, the processor module is configured to apply the scoring function to the segmented organ of the 3D tomographic image data, wherein the scoring function is configured to determine a scoring value for the segmented organ to which the scoring function is applied, wherein the scoring value includes and / or corresponds to a metric quantifying the extent to which a vicinity of a voxel at a surface of the segmented organ in the 3D tomographic image data contains image artifacts.

[0059] The processor module is configured to compare the score value with a threshold value. The processor module and / or the interface module is configured to provide a user notification when at least one score value exceeds the threshold value. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Further embodiments and / or advantages are described and / or illustrated in the drawings. The drawings show:

[0061] Figure 1 is a flow chart of a method for evaluating the usability of 4D tomographic image data;

[0062] Figure 2 is a scanner device for performing the method;

[0063] Figure 3 A system that is part of a scanner device. DETAILED DESCRIPTION

[0064] Figure 1 A flow chart depicting an example of a method for evaluating the availability of 4D tomographic image data (with reference mark 4D) is shown. The method can be performed by various devices, including a scanner device 1 (e.g., computed tomography), a computer, a processor, or a cloud-based system. The purpose is to determine the quality of the acquired tomographic image data (especially 4D tomographic data), and in particular to determine their availability for radiotherapy planning and delivery. Ideally, the method is performed while the examination object (usually a patient) is still in the scanner device 1, or shortly after the 4D tomographic image data 4D is acquired. This allows the user to evaluate the suitability of the recently acquired data and decide whether to repeat the scan or data acquisition.

[0065] In step 100, 4D tomographic image data 4D are received. These data may be acquired using various medical imaging devices such as computed tomography or magnetic resonance tomography and are associated with an examination object (usually a human patient). The 4D tomographic image data 4D provide a three-dimensional representation of a region of interest within the examination object over time, wherein each set of 3D tomographic image data corresponds to a specific point in time. In this example, the 4D tomographic image data 4D are collected over at least one complete cycle of the patient.

[0066] The method may optionally include step 200, which involves extracting 3D tomographic image data from the 4D tomographic image data. For each time point, the extracted relevant 3D tomographic image data is extracted and subsequently processed. These 3D data (particularly the extracted 3D data) are then passed to step 300.

[0067] In step 300, a segmentation algorithm is applied to each of the 3D tomographic image data extracted from the 4D tomographic image data. The segmentation algorithm, which may be a machine learning algorithm, is designed to identify and delineate one or more organs within the extracted 3D tomographic image data 3D. Ideally, the segmentation algorithm targets specific regions, such as user-defined regions of interest or regions relevant to the examination. Organs with high contrast in the 3D tomographic image data are preferably segmented. After applying the segmentation algorithm to the plurality of 3D data sets, segmented organ data, typically represented as a 3D model, is obtained.

[0068] In step 400, a scoring function (which may be a machine-learned scoring function) is applied to the segmented organs from step 300. The scoring function is applied to each of the 3D tomographic image datasets, in particular the segmented organs within these datasets. The scoring function evaluates the extent to which the vicinity of a voxel contains image artifacts and provides a score value that quantifies the extent. The scoring function may include a Hough transform. The scoring function determines and / or takes into account factors such as the presence, number, and orientation of lines on or near the surface of the organ.

[0069] The score values ​​obtained from step 400 are then used in step 500 where they are compared to at least one threshold value. The comparison may involve summing, weighting or averaging the score values ​​for each organ or each 3D tomographic image data set. If the score value exceeds the threshold value, this indicates a high probability of image artifacts associated with the segmented organ and, by extension, with the associated 3D tomographic image data.

[0070] In step 600, if at least one of the scoring values ​​exceeds a threshold, a user notification is generated. The notification should include the time point corresponding to the 3D tomographic image data containing the segmented organ with a high scoring value. Optionally, the notification may display the 3D tomographic image data with an overlay indicating the region with image artifacts and the segmented organ. This information helps the user (typically a scanner operator or a medical professional) decide whether to repeat the acquisition of 4D tomographic image data for that specific time point, or whether the identified image artifact is not critical for further use of the data.

[0071] Figure 2An example of a scanner device 1 is provided, which consists of a tomography unit 2 and a scanner console 3 connected for data exchange. The scanner console 3 can also be linked to a cloud 4 for data storage or processing outsourcing. A patient or examination subject is scanned by the tomography unit 2, wherein 4D tomographic image data 4D is acquired and transmitted to the scanner console 3 and / or the cloud 4.

[0072] Figure 3 A system 5 is depicted, which is part of a scanner device 1, in particular part of a scanner unit 2, a scanner console 3 and / or a cloud 4. The system 5 performs a method for evaluating the availability of 4D tomographic image data. The system 5 comprises an interface unit 6, a processor unit 7 and a storage unit 8. The interface unit 6 is connected to the tomographic unit 2 and the scanner console 3, thereby facilitating data exchange. The processor unit 7 is responsible for executing steps 200, 300, 400 and / or 400 of the method. The storage unit 8 stores received 4D tomographic image data 4D, extracted 3D data 3D, segmented organs, scoring values ​​and user notifications.

[0073] When the score value exceeds the threshold, the processor unit 7 sends a user notification to the scanner console 3 using the interface unit 6. The user (typically a radiologist, medical physicist or radiation therapist) can then make an informed decision as to whether to repeat the acquisition of the 4D tomographic image data 4D.

Claims

1. A method for assessing the usability of 4D tomographic image data (4D), comprising the following steps: - receiving (100) 4D tomographic image data (4D), wherein the 4D tomographic image data (4D) comprises a plurality of 3D tomographic image data (3D) of an examination object, wherein the plurality of 3D tomographic image data (3D) corresponds to a plurality of time points, - applying (300) a segmentation algorithm to the plurality of 3D tomographic image data (3D), wherein the segmentation algorithm is configured to segment at least one organ in the 3D tomographic image data (3D) to which the algorithm is applied, - applying (400) a scoring function to the segmented organ of the 3D tomographic image data (3D), wherein the scoring function is configured to determine a scoring value for the segmented organ to which the scoring function is applied, wherein the scoring value comprises and / or corresponds to a metric quantifying the extent to which a vicinity of a voxel at a surface of the segmented organ in the 3D tomographic image data (3D) contains image artifacts, - comparing the score value with a threshold value (500), - When at least one rating value exceeds said threshold, providing (600) a user notification.

2. The method according to claim 1, wherein the 4D tomographic image data (4D) comprises a number N of 3D tomographic image data (3D), wherein the segmentation algorithm is applied to the N 3D tomographic image data (3D), wherein in the step of applying (400) the scoring function, at least N scoring values ​​are determined.

3. A method according to claim 1 or 2, wherein the segmentation algorithm is configured to segment M organs in the 3D tomographic image data (3D) to which the segmentation algorithm is applied, wherein in the step of applying (400) the scoring function, at least M scoring values ​​are determined for each time point and / or each 3D tomographic image data (3D).

4. A method according to one of the preceding claims, wherein the segmentation algorithm is configured to generate a 3D contour and / or a 3D volume of the segmented organ, wherein the step of applying the segmentation algorithm comprises providing the generated 3D contour and / or 3D volume to the step of applying (400) the scoring function.

5. A method according to one of the preceding claims, wherein the 4D tomographic image data (4D) comprises a plurality of segmentable organs, wherein the segmentation algorithm is configured to segment sub-portions of the plurality of segmentable organs of the 4D tomographic image data (4D), wherein the sub-portions of the segmentable organs comprise organs having the highest contrast and / or being most susceptible to motion artifact errors.

6. A method according to one of the preceding claims, wherein the scoring function comprises a local Hough transform, wherein the local Hough transform is configured to detect lines and / or directions of lines, wherein the scoring factor determined in the step of applying the scoring function is based on the number of detected lines and / or the directions of the detected lines.

7. A method according to one of the preceding claims, wherein the scoring function comprises a local Hough transform, wherein the local Hough transform is configured to detect lines and / or the direction of lines, wherein in the step of applying the scoring function, the Hough transform is applied to the segmented organ and the vicinity of the segmented organ, wherein the determined scoring factor is based on a comparison of the result of applying the Hough transform to the segmented organ with the result of applying the Hough transform to the vicinity.

8. A method according to one of the preceding claims, wherein the scoring function is configured to evaluate local image contrast along an organ boundary of the segmented organ, wherein by applying the scoring function to the segmented organ, when the local image contrast has sharp transitions, a scoring value corresponding to a stacking transition artifact is determined, and / or when the local image contrast has blurry transitions, a scoring value corresponding to a motion artifact is determined.

9. A method according to one of the preceding claims, wherein the step of comparing (600) the score values ​​comprises generating the user notification when at least one score value exceeds the threshold value, wherein the user notification comprises the time point associated with the 3D tomographic image data having the image artifact, the associated 3D tomographic image data (3D) and / or an overlay image, wherein the overlay image comprises the associated 3D tomographic image (3D), an indication of the image artifact and / or a location of the image artifact.

10. The method according to one of the preceding claims, wherein the segmentation algorithm and / or the scoring function are configured as a machine-learned algorithm and / or a machine-learned function.

11. The method according to one of the preceding claims, wherein the step of providing (600) the user notification comprises: The user notification is provided to and / or displayed on a scanner console of a scanner used to acquire the 4D tomographic image data.

12. A computer program product comprising a computer readable medium storing computer program code which, when executed by a computer processor, configures the computer processor to perform the method as claimed in any one of the preceding claims.

13. A scanner device comprising a processor unit (7) configured to perform the method according to one of claims 1 to 11.

Images