Quality Control in Medical Imaging

Abstract

The system provides quality control in medical imaging. Noise sources that may adversely affect image quality are evaluated. The evaluated noise sources include at least two of device noise sources, operator noise sources, and subject (patient being imaged) noise sources, all of which create noise sources during acquisition of medical images. Based on the evaluated noise sources, a quality control indicator is determined and output to an operator.

Description

Technical Field

[0001] The present invention relates to automatic quality evaluation and quality control of medical images.

Background Art

[0002] To increase the possibility of making an accurate diagnosis from an image, it is clearly desirable to ensure that the highest quality medical images are obtained. For this purpose, it is necessary to take measures for quality evaluation and quality control.

[0003] Conventionally, this has been a manual process where the operator of an MRI system determines whether the captured image has sufficient quality or whether a rescan is necessary based on a visual inspection of the image. However, quality control procedures are difficult to standardize and automate. For example, quality control of magnetic resonance imaging (MRI) has been demonstrated to be difficult to standardize because there is no simple objective quality control metric for capturing image quality in various MRI sequences.

[0004] Furthermore, quality control depends not only on the objective characteristics of the image but also on the clinical goals of physicians and radiologists. Ideally, when the clinical goal is clear but the diagnosis has not yet been determined, quality control of the MRI should be performed during acquisition.

[0005] There is a need for a system that provides at least partial automation of image quality control and thus consistency.

[0006] International Patent Publication WO2017 / 102860 discloses a scan monitoring device for medical imaging. This scan monitoring device includes a controller unit and a display. The controller unit acquires tracking data of a subject in a medical scanner and scanner data indicating the operating parameters of the medical scanner during a scan session, determines the output of a verification function based on the tracking data and the scanner data, and controls the scan monitoring device according to the output of the verification function.

[0007] U.S. Patent Application Publication No. 2016 / 259898 discloses an apparatus for providing reliability in computer-aided diagnosis (CAD). This includes a raw data collector that collects raw data including images acquired by a probe, an image reliability determiner that determines a reliability level of an image using the collected raw data, and a reliability provider that provides the reliability of the determined image to a user.

[0008] U.S. Patent No. 9,836,840 discloses a method for evaluating the quality of a medical image of at least a part of a patient's anatomical structure. One method includes receiving one or more images of at least a part of a patient's anatomical structure, determining one or more image characteristics of the received images, performing anatomical localization or modeling of at least a part of the patient's anatomical structure based on the received images, obtaining an identification of one or more image features associated with anatomical features of the patient's anatomical structure based on the anatomical localization or modeling, and calculating an image quality score based on the one or more image characteristics and the one or more image features.

Summary of the Invention

[0009] The present invention is defined by the claims.

[0010] According to an example of an aspect of the present invention, a system for quality management in medical imaging is provided. The system includes a processor, and the processor is configured to evaluate noise sources that may adversely affect image quality, the noise sources including a device noise source caused by a medical imaging device during acquisition of a medical image of a subject, an operator noise source caused by an operator during acquisition of a medical image, and at least two of a subject noise source caused by the subject during acquisition of a medical image, Based on the evaluated noise sources, determine a quality control indicator and output it to the operator.

[0011] An "operator" is a person who operates medical imaging equipment. This is usually different from a person who provides a medical interpretation of an image, i.e., a radiologist.

[0012] This system can generate a quality control indicator in real time during imaging to assist the operator or enable automatic evaluation of image quality. Furthermore, by identifying the causes of image artifacts, corrective measures can be taken to improve image quality. Three different categories of noise sources may each cause multiple possible noise causes. Therefore, it is necessary to understand that a "noise source" has factors that may contribute to some noise. Evaluate each noise source to identify one or more possible causes of noise that can result in artifacts in the acquired images.

[0013] Similarly, a "quality control indicator" may include multiple information units such as multiple quality evaluation criteria.

[0014] Therefore, the present invention provides a real-time multimodal quality control system that can be easily interpreted according to the context and uses metrics that provide insights into the causes of insufficient signals or MRI image quality.

[0015] Preferably, the processor evaluates noise sources including all three enumerated categories.

[0016] The system may include a first arrangement unit that determines parameters contributing to the device noise source, and the first arrangement unit means for determining the uniformity of the imaging field, a monitoring circuit for monitoring the resonance frequency, means for detecting the normal function of the imaging transmitter and / or receiver, The operating temperature sensor of the imaging device, and a circuit for determining the energy consumption of the imaging device, include one or more of the following.

[0017] These are part of how to determine the factors contributing to the noise introduced by the device. Some of these parameters can be obtained by dedicated sensors, and other parameters can be obtained based on, for example, the analysis of medical images.

[0018] The resonance frequency defines the frequency at which atoms / molecules in the subject resonate based on the applied imaging field. The resonance frequency can vary, for example, based on the temperature of the subject. For example, in MRI imaging, the proton resonance frequency can be used to measure the temperature of the subject.

[0019] The system may include a second arrangement unit for determining the parameters contributing to the operator noise source, and the second arrangement unit a system for collecting data regarding the positioning, field of view, slice direction, timing, and number of scans set by the operator, a timer for recording the time the operator was working, and a display of the operator's qualification or level of expertise, include one or more of the following.

[0020] These are part of how to determine the factors contributing to the noise introduced by the operator's actions.

[0021] The slice direction defines the direction of the imaging plane with respect to the subject. For example, in MR imaging of the head of a subject, there are three basic orthogonal slice directions, namely, the transverse direction, the sagittal direction, and the coronal direction.

[0022] The system may include a third arrangement unit for determining the parameters contributing to the subject noise source, and the third arrangement unit A sensor that senses the movement of the subject's body and / or eyes, A respiration sensor that senses the respiration of the subject, A skin conductance sensor, A body temperature sensor, A heart rate sensor that senses the heart rate of the subject, and A display of the duration of the imaging process, include one or more of the above.

[0023] These are part of how to determine factors contributing to noise caused by subject behavior or characteristics.

[0024] For example, the processor may further control a display device to present a medical image to the operator along with a quality control indicator.

[0025] The quality control indicator may be a general indicator of image quality, but more preferably, it interprets the medical image, for example, by flagging image artifacts and indicating noise sources considered to be the cause of the image artifacts, to provide additional context to the user interpreting the medical image. Thereafter, a proposal for obtaining a higher quality image may be provided.

[0026] The processor may further determine an image quality metric of the medical image, identify image artifacts in the medical image, and be able to identify the image artifacts to the operator.

[0027] Image quality metrics (such as signal-to-noise ratio, contrast-to-noise ratio, etc.) are related to general evaluation criteria for image quality, and image artifacts are related to specific image problems. These can be associated with various noise sources. When determining the quality control indicator, both of these evaluation criteria can be considered.

[0028] The processor can identify image artifacts using machine learning algorithms. For example, the machine learning algorithm can further output a corrected medical image with the image artifacts removed. The machine learning may be based on deep learning or a neural network.

[0029] In this way, the system can improve image quality and identify the noise source and / or the image artifact source.

[0030] The processor can further determine the most likely noise source of the image quality problem based on the evaluated noise source and present the most likely noise source to the operator. Then, the operator can perform a repair procedure and determine whether the problem is solved, for example, by rescan.

[0031] The processor can also highlight the image artifacts identified on the scan to inform the operator where in the image they are visible. For example, some operators may not know or be unsure where to look for the artifacts.

[0032] The processor can further determine whether the medical image should be reacquired based on the quality control indicator and present a proposal to the operator on whether to reacquire the medical image. This helps the operator to determine whether a rescan is appropriate.

[0033] This rescan can also be performed semi-autonomously, i.e., without obtaining the operator's consent and without requiring any operator action, or fully autonomously with only notification to the operator.

[0034] The processor can further track the operator's judgments during and after the acquisition of the medical image, detect non-general operator judgments from the operator's judgments, and request the operator for the basis of the non-general operator judgment.

[0035] Such non - general operator judgments may indicate operator errors. Therefore, the system can further track and trace medical judgments. This may include, for example, a radiologist acting as a further agent to make a judgment. For example, the radiologist may request the operator to perform a rescan (or a follow - up scan based on diagnostic information requiring further scans) based on the quality of the image. Therefore, the system can implement a decision - making support system, especially regarding quality control.

[0036] The processor can further send the medical image to a radiologist for delegation of judgment, and can track the radiologist's judgment.

[0037] For example, the system can use the radiologist's judgment to improve image analysis.

[0038] The processor can further create a judgment map based on the operator's judgment, the basis for the judgment, and the radiologist's judgment. With this judgment map, the effectiveness of the procedure can be evaluated.

[0039] For example, the processor can control the display to show the operator past imaging data and / or demographic imaging data. If there are differences between the medical image and the corresponding past imaging data and / or demographic imaging data, these differences can be displayed to the operator.

[0040] The processor can further generate a quality control report indicating how the medical device, the operator, and the subject are affecting the quality of the medical imaging process. Using this, local quality control over time becomes possible, and comparisons between different devices, operators, and patient groups become possible.

[0041] The present invention also provides a medical imaging system. This system For example, a medical imaging device that is an MRI scanner, and a system for the aforementioned quality control, wherein the processor of the system for quality control further executes an analysis of an image captured by the medical imaging device, determines a quality control indicator based on the image analysis, and outputs it to an operator.

[0042] The present invention also provides a computer-implemented method for providing quality control in medical imaging. The method includes evaluating a noise source that may adversely affect the quality of a medical image, wherein the noise source includes a device noise source caused by a medical imaging device during acquisition of a medical image of a subject, an operator noise source caused by an operator during acquisition of the medical image, and a subject noise source caused by the subject during acquisition of the medical image, evaluating, including at least two of the above; determining a quality control indicator based on the evaluated noise source and outputting it to an operator.

[0043] The present invention also provides a computer program including computer program code means. When the computer program is executed on a computer, the aforementioned method is implemented.

[0044] These and other aspects of the present invention will become apparent from the embodiments described below and will be described with reference to the embodiments.

Brief Description of the Drawings

[0045] To understand the present invention more deeply and to more clearly show how it is implemented, reference is made to the accompanying drawings by way of example only.

[0046]

Figure 1

Figure 2

Mode for Carrying Out the Invention

[0047] The present invention will be described with reference to the drawings.

[0048] The detailed description and specific examples show exemplary embodiments of the apparatus, system, and method, but are for illustrative purposes only and are not intended to limit the scope of the invention. It should be understood that these and other features, aspects, and advantages of the apparatus, system, and method of the present invention will become better understood from the following description, the appended claims, and the accompanying drawings. It should be understood that the drawings are only schematic and are not drawn to scale. Also, it should be understood that the same reference numerals are used throughout the drawings to indicate the same or similar parts.

[0049] The present invention provides a system for providing quality control in medical imaging. Noise sources that may adversely affect image quality are evaluated. The noise sources to be evaluated include at least two of a device noise source, an operator noise source, and a subject (patient being imaged) noise source, all of which create noise sources during the acquisition of medical images. Based on the evaluated noise sources, a quality control indicator is determined and output to the operator.

[0050] To address the problem of the lack of objective quality control specific to medical imaging, the present invention is based on the concept of estimating noise sources known to degrade image quality, such as the quality of MRI images. Next, context is provided to the user about the artifacts most likely or potentially causing low-quality images.

[0051] In the example of MR imaging, noise can originate from the MRI device, the operator, or the patient being scanned.

[0052] Regarding MRI devices, fluctuations in the stability of the MRI device may cause low-quality control via magnetic field inhomogeneity, radio frequency noise, and malfunction of either the RF transmitter / receiver. Regarding the operator, the MRI operator may cause low-quality control due to inadequate positioning of the field of view. Patients may cause low-quality management due to movement within the scanner or other physiological fluctuations (such as respiration, heart rate, etc.). Such patient-specific noise sources are frequently seen in psychiatric patients, tremor patients, back pain patients, children, and the elderly, resulting in noisier images and lower standards for normal quality control.

[0053] The present invention is based on the recognition that the interpretation of quality for providing a quality control indicator (i.e., a set of one or more quality control metrics) is facilitated by evaluating these noise sources. Accordingly, the present invention integrates the automatic estimation of noise from MRI devices, operators, and patients with the quality control estimated from the MRI images themselves.

[0054] When implemented on an MRI console, the estimated value of noise and quality control metrics can be evaluated in the context of clinical goals. The integration of context-dependent quality control can also be standardized across scanners employed in clinical protocols and workflows. When appropriately implemented in the workflow by a human operator, quality control metrics can be developed to aid in the training of machine learning algorithms to automatically detect artifacts during MRI acquisition.

[0055] FIG. 1 shows a system according to an example of the present invention.

[0056] FIG. 1 shows a system for quality control in medical imaging, which in this example is combined with a medical imaging device 10, which is an MRI scanner. A subject 12 is shown within the scanner bore, and an operator 14 of the device is also shown.

[0057] The imaging unit 16 generates medical images in a conventional manner.

[0058] The processor 18 evaluates noise sources that may have an adverse effect on image quality.

[0059] The noise sources include (i) device noise sources caused by the medical imaging device 10 during acquisition of medical images of the subject, (ii) operator noise sources caused by the operator 14 during acquisition of medical images, and (iii) subject noise sources caused by the subject 12 during acquisition of medical images.

[0060] The system evaluates any two, preferably all three, of these noise sources.

[0061] To evaluate these noise sources, the processor receives various sensor inputs as described below and medical images from the imaging unit 16.

[0062] Next, the processor determines a quality control indicator from these noise sources. The quality control indicator is used to adapt the image displayed on the display 20 and / or to provide additional output information. This provides the operator with additional context.

[0063] This system can generate a quality control indicator in real time during imaging to assist the operator or enable automatic evaluation of image quality. By identifying and displaying the causes of image artifacts, corrective measures can be taken to improve image quality. The three different categories of noise sources can each be the cause of multiple possible noises. Therefore, each noise source is evaluated to identify one or more possible causes of the noise that can introduce artifacts in the acquired image.

[0064] Generally, the first arrangement part is used to determine parameters that contribute to device noise sources, such as the uniformity of the imaging field, resonance frequency, normal function of the imaging transmitter and / or receiver, operating temperature of the imaging device, and energy consumption of the imaging device.

[0065] The second arrangement part is used to determine parameters that contribute to operator noise sources, such as the positioning, field of view, slice direction, timing, and number of scans set by the operator, the time the operator was working, and the level of the operator's qualifications or expertise.

[0066] The third arrangement part is used to determine parameters that contribute to subject noise sources, such as the movement of the subject's body and / or eyes, the subject's breathing, sweating, body temperature, and heart rate. Other physiological information about the subject can also be monitored during the imaging process.

[0067] The parameters can be monitored by dedicated sensors or based on the analysis of medical images or existing outputs from the imaging device.

[0068] For example, there are internal MRI sensors (such as log files), and these can be used to measure the noise caused by the MRI device. For example, real-time access to the MRI console enables access to the scan planner (data related to positioning, field of view, timing, and rescan, etc.) set by the operator to detect errors by the operator.

[0069] By reconstructing the MRI image, various metrics related to quality control can be collected, such as magnetic field uniformity, foreground / background energy, relative signal of each RF coil, sudden spikes in the RF signal, realignment in the case of time series, and estimated signal-to-noise for each MRI sequence.

[0070] The present invention provides real-time analysis (and potentially display) of quality control metrics based on MRI reconstruction and further integrates an estimated value of noise.

[0071] Also, external sensors such as psychometric measurements (respiration, heart rate, skin conductance), a peripheral pulse unit for movement, and an in-bore camera can be used.

[0072] For three noise sources, some specific examples are shown below regarding the specific situations to be monitored, the ways in which they can be monitored, and the possible impacts on image quality.

[0073] 1. Device noise source 1.1 Magnetic field inhomogeneity due to poor (active) shimming or incorrect shim current setting. Detection: Automatic analysis during image reconstruction of magnetic field inhomogeneity based on geometric distortion, etc. Alternatively, fat signal intensity can be monitored in an MRI scan using fat suppression. Image artifacts: Image inhomogeneity at the image boundary and incomplete suppression of fat signal when fat suppression is used.

[0074] 1.2 Gradient coil heating / cooling Detection: Automatic analysis during resonance frequency scanning. Image artifacts: Signal drift and fluctuations in the time series. Geometric distortions (such as sheering, scaling, and shifting related to eddy current artifacts) may be observed.

[0075] 1.3 RF shield violation by other peripheral devices Detection: Automatic analysis during RF noise reconstruction using spikes detected in the image data in the Fourier domain, for example, combined with values from during acquisition or from the MRI preparation stage including shimming. Image artifacts: Signal spikes, noisy images, reduction in foreground / background ratio, zipper artifacts.

[0076] 1.4 Dropout or disconnection between MRI and peripheral devices Detection: Use software applications to check device connectivity, measure transmission delays, and log system offsets of various devices. If MRI acquisition is synchronized with the cardiac or respiratory cycle, there are solutions for gated / triggered MRI. In these scans, usually, acquisition is temporarily stopped or connectivity is completely lost. This can lead to longer acquisition times. It is also possible to report to the operator the reasons for the longer scan time. Image artifacts: In triggered and gated MR, the acquisition time becomes longer. If peripheral devices are not linked to the MRI, it may lead to data timing or phase offsets.

[0077] 1.5 Active remote array coils outside the field of view Detection: Detection of coil movement relative to the scanner, especially the anterior body coil, by including sensors on the coil. Also, automatic analysis during image data reconstruction can be used for SENSE reference scans. Image artifacts: Signal dropout, Annefact cusp artifact, star artifact.

[0078] 1.6 Reconstruction (runtime) errors Detection: Software applications that perform automatic analysis during the reconstruction of data error artifacts. Image artifacts: Data error artifacts, arcs or crosshatch / herringbone patterns.

[0079] 2. Operator noise source 2.1 Installation of RF coils Detection: Sensors within the RF coil for detecting patient-based positioning, combined with automatic analysis during image reconstruction of MRI scout view images. Image artifacts: Signals related to the target (such as the liver) are lost or absent.

[0080] 2.2 Field of view setting / size Detection: Automatic analysis during image reconstruction of automatic alignment and smart survey. Image artifact: Phase wrap-around artifact or clipping.

[0081] 2.3 Saturated slab setting and orientation Detection: Automatic analysis during image reconstruction of the saturated slab based on the field of view and image target. Image artifact: Insufficient suppression of the selected spectrum.

[0082] 2.4 Slice direction setting Detection: Automatic analysis during image reconstruction of fold-over artifact or phase shift artifact related to slice setting. Image artifact: Propagation of fold-over artifact or phase shift artifact from tissues outside the field of view into the image.

[0083] 2.5 Signal dropout due to metallic objects (hairpins, rings, dental objects, etc.) Detection: Automatic analysis during image reconstruction to detect signal dropout artifact. Image artifact: Local signal dropout artifact.

[0084] 2.6 RF shield violation due to the door of the shielded room having an MRI magnet being open Detection: Sensors on the door to detect the open / closed state. Automatic analysis during image reconstruction of image noise, signal ratio, and zipper artifact. Image artifact: Noisy image, decrease in foreground / background ratio, and zipper artifact.

[0085] 2.7 Timing misalignment between contrast agent and MRI scan Detection: Registration of the timing of gadolinium contrast bolus injection, very fast low-resolution preparatory MRI scan, start of high-resolution MRI acquisition, and automatic analysis during image reconstruction to track the bolus signal using software for automatic detection of Maki artifacts. Image artifact: Maki artifact (lack of signal at the center of blood vessels compared to the edges).

[0086] 2.8 Incorrect follow-up or diagnostic scan Detection: Use a dictionary to check for uncommon / improbable combinations of scans (subsequently, suggestions for contacting a radiologist may be generated). Image artifact: Not visible in the image (except when contrary to the operator's expectations).

[0087] 3. Patient noise source 3.1 Spontaneous patient movement. Detection: Camera system, pressure sensors in the mattress, head coil, headphones, surface coil, or other objects in contact with the patient. Analysis of the raw signal from the RF coil, and automatic analysis during image reconstruction of the k-space trajectory scan and navigator scan. Image artifact: Blurring, ghosting from moving structures, signal loss due to spin dephasing.

[0088] 3.2 Respiratory motion Detection: Respiratory belt, pressure sensor, or camera. Image artifact: Blurring due to spin dephasing, ghosting, and signal loss.

[0089] 3.3 Respiration-dependent blood oxygenation level Detection: Respiratory belt, peripheral pulse unit (PPU), respiratory mask (end-tidal oxygen or carbon dioxide). Image artifact: T2 * Confounding of estimates of blood oxygenation level-dependent imaging (= fMRI).

[0090] 3.4 Eye Movements and Blinking Detection: Eye tracking camera and automatic analysis during image reconstruction of phase shift ghosting artifacts. Depending on the phase encoding direction, for example, these artifacts can occur outside the brain and may reduce the foreground / background signal ratio. Image Artifact: Phase shift artifact that depends on the reduced foreground / background signal.

[0091] 3.5 Changes in Blood Flow and Blood Pressure Detection: Heart rate monitor (volume clamp, PPU), inflatable cuff, and arterial pressure applanation tonometry. Image Artifact: Changes in cerebral blood flow bias MRI acquisition (T2 * , perfusion, fMRI).

[0092] 3.7 Heart Operation and Heart Rate Variation Detection: Heart rate monitor (e.g., ECG, EKG, ultrasound), PPU, and cardiac MRI flow measurement. Image Artifact: Variation in echo planar imaging time series.

[0093] 3.7 Patient Movements Related to the Patient's Sleep / Wake State or Anesthesia Level Detection: Through EEG, PPG, contact electrodes such as skin conductance, anesthesia device, or automatic analysis of fMRI time series. When combined with the estimated value of spontaneous movement, it may indicate an insufficient anesthesia level. Image Artifact: Confusion of the estimated value of the fMRI signal related to cognitive ability.

[0094] The present invention is based on using one or more such monitoring options from at least two of the categories. Thereby, the interaction between at least two noise sources related to the patient, the operator, and the medical imaging device is considered, which makes it possible to identify the cause of the degradation of the image quality and generate an image quality indicator.

[0095] The image quality indicator is integrated into an easy-to-use dashboard and presented to the operator along with the acquired MRI image. Further, for post-analysis of data quality, quality control metrics are exported (e.g., within the DICOM header image).

[0096] The information provided (quality control indicator) may be a general indicator of image quality, but more preferably provides additional context to the user interpreting the medical image, such as flagging image artifacts and indicating the source of noise considered to be the cause of the image artifacts. Subsequently, proposals for obtaining higher quality images may be provided.

[0097] Real-time quality control evaluations from multiple sources support the detection and interpretation of MRI artifacts, which are currently performed by human visual inspection. Implementing this real-time quality control framework significantly facilitates the development of automated artifact detection and quality control using machine learning.

[0098] In the most basic example, the system can provide context to the operator by providing a quality control indicator, for example, along with an estimate of the noise, during visual inspection of the MRI image.

[0099] In a more advanced case, the system can flag image artifacts, identify the most likely source of noise, and provide a proposal for rescan.

[0100] When a large amount of quality data can be utilized at one site, the system can detect outliers. These can be used for predictive maintenance and improvement regarding the local site (i.e., a specific MRI device).

[0101] When aggregated quality data becomes available across multiple sites, scanners, and (clinical) populations, quality control metrics can be compared across distributed and local systems to further evolve the operation.

[0102] When user feedback is safely collected in the MRI console (or post - processing), machine - learning algorithms for artifact detection can be developed and optimized using data from image reconstruction and quality - management metrics. When such machine - learning tools are implemented in image reconstruction and, for example, image artifacts caused by motion are automatically removed, context - dependent quality - management metrics are used to show the "amount" and "reason" for which the data is interpolated. Such context - dependent quality - assessment criteria also help to further improve machine - learning - based image reconstruction.

[0103] Also, user feedback can be used to validate machine - learning algorithms at various sites, so they can be generalized across different patient groups, scanners, and hospitals. Thus, the present invention can be used for testing (i.e., validating) solutions.

[0104] The operator, rather than the radiologist, is usually the first person to evaluate the scan and make an autonomous on - site judgment. The operator can decide to perform a rescan based on image quality, perform an immediate follow - up scan based on visible pathology, or, if in doubt or in case of legal liability, contact the radiologist and defer the judgment.

[0105] Such on - site judgments have medical consequences and are costly. This is because about 5 - 10% of patients require additional scans at a new visit after examination by a radiologist, but unknown scans have become unnecessary. Unless the radiologist requests an additional scan, most of the operator's judgments are made informally, logged only partially, and are difficult to retrospectively evaluate.

[0106] To address this problem, the system is further used as a medical - imaging judgment assistant that tracks, traces, and supports autonomous and human judgments during medical imaging.

[0107] Thus, the system can show the operator, almost in real-time, the scan and image analysis results as described above, along with proposals regarding the need for rescan or the need to contact a radiologist.

[0108] The operator makes autonomous workflow decisions to end the examination, perform a rescan, obtain a follow-up scan, or contact a radiologist. These workflow decisions are recorded by the system, and if a deviation from the expected workflow occurs, the system can request the operator for the basis of a particular decision. This applies, for example, when the system detects an operator decision that is not common.

[0109] When deciding to refer to a radiologist, the system may transfer information remotely to the radiologist via a push-type application. The remote decision by the radiologist is also recorded by the system.

[0110] Thus, in addition to providing support functions, the system acts as a medical imaging flight recorder that tracks human decisions during the medical imaging process. Then, the system data can be periodically analyzed to evaluate the effectiveness of the procedure.

[0111] The system can create aggregated regular reports (weekly, monthly, yearly) to provide insights into how medical devices, operators, and patients are affecting the quality of the medical imaging process. Using these aggregated reports enables local quality management over time and comparison between different device, operator, and patient groups. These reports help identify problems or bottlenecks in the medical imaging process at the local site. By being in an aggregated form, these reports allow comparison between devices, hospitals, and regions. This report helps determine the priority of solution development to improve the quality of medical imaging.

[0112] For example, the system autonomously determines whether a medical image should be reacquired based on a quality control indicator, and presents a proposal to the operator as to whether to reacquire the medical image. This helps the operator to determine whether a rescan is appropriate.

[0113] The system can create a decision map based on operator judgment, basis for judgment, and radiologist judgment. This decision map can be used to evaluate the effectiveness of the procedure.

[0114] For example, the processor controls the display to show the operator past imaging data and / or demographic imaging data. If there are differences between the medical image and the corresponding past imaging data and / or demographic imaging data, these differences are displayed to the operator.

[0115] The system integrates information from a Picture Archiving and Communication System (PACS), thereby highlighting differences with previously acquired medical images and supporting workflow decisions. The system is also used to provide individual training for medical imaging technicians. The system may also be extended to assist in making decisions during image-guided radiation therapy, HIFU, or other image-guided interventions.

[0116] Figure 2 shows a computer-implemented method for providing quality control in medical imaging. This method includes the following steps: In step 30, a noise source that may affect image quality is evaluated. This includes at least two, preferably all three, of the following substeps: Step 30a of evaluating a device noise source caused by a medical imaging device during acquisition of a medical image of a subject, Step 30b of evaluating an operator noise source caused by an operator during acquisition of the medical image, and Step 30c of evaluating the subject noise sources caused by the subject during acquisition of medical images.

[0117] In step 32, based on the evaluated noise sources, a quality control indicator is determined and output to the operator.

[0118] The present invention has been described in relation to an MRI system. However, the present invention can be more generally applied to medical imaging.

[0119] Modifications of the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claimed invention, from a study of the drawings, the disclosure, and the appended claims. In the claims, the term "comprising" does not exclude other elements or steps, and the singular does not exclude the plural.

[0120] A single processor or other unit can perform the functions of several items recited in the claims.

[0121] The mere fact that certain means are recited in mutually different dependent claims does not mean that these means cannot be used advantageously in combination.

[0122] The computer program can be stored / distributed on any suitable medium, such as an optical storage medium or a solid-state medium, supplied together with or as part of other hardware, but can also be distributed in other forms, such as via the Internet or other wired or wireless communication systems.

[0123] It should be noted that when the term "adapted to" is used in the claims or the description, the term "adapted to" is intended to be equivalent to the term "configured to".

[0124] Any reference signs in the claims shall not be construed as limiting the scope.

Claims

1. A system for quality control during acquisition of medical images, said system including a processor, said processor being to evaluate noise sources that may have an adverse effect on the quality of said medical images, said noise sources being a device noise source caused by a medical imaging device during acquisition of said medical image of a subject, an operator noise source caused by an operator during acquisition of said medical image, and a subject noise source caused by said subject during acquisition of said medical image, including at least two of them, to evaluate; to determine a quality control indicator based on the evaluated noise sources and output it to said operator; In a system that performs, said processor further to identify one or more possible causes of noise that may introduce artifacts into said medical image based on the evaluated noise sources, A system, characterized by performing.

2. including a first arrangement unit for determining parameters contributing to said device noise source, said first arrangement unit being means for determining the uniformity of an imaging field when said medical image is acquired using a magnetic resonance imaging device, a monitoring circuit for monitoring a resonance frequency when said medical image is acquired using a magnetic resonance imaging device, means for detecting the normal function of an imaging transmitter and / or receiver, an operating temperature sensor of the imaging device, and a circuit for determining the energy consumption of said imaging device, The system according to claim 1, including one or more of them.

3. including a second arrangement unit that determines parameters contributing to the operator noise source, the second arrangement unit a system that collects data on positioning, field of view, slice direction, timing, and number of scans set by the operator, a timer that records the time the operator was working, and a display of the operator's qualification or level of expertise, The system according to claim 1 or 2, including one or more of the above.

4. including a third arrangement unit that determines parameters contributing to the subject noise source, the third arrangement unit a sensor that senses movement of the body and / or eyes of the subject, a respiration sensor that senses the respiration of the subject, a skin conductance sensor, a body temperature sensor, a heart rate sensor that senses the heart rate of the subject, and a display unit for the duration of the imaging process, The system according to any one of claims 1 to 3, including one or more of the above.

5. The processor further controls a display device to present a medical image to the operator together with the quality management indicator, the system according to any one of claims 1 to 4.

6. The processor further determining an image quality metric of the medical image, and further determining the quality management indicator based on the image quality metric, identifying an image artifact in the medical image, and further determining the quality management indicator based on the image artifact, identifying the image artifact to the operator, The system according to any one of claims 1 to 5, which executes

7. The system according to claim 6, wherein the processor identifies the image artifact using a machine learning algorithm.

8. The system according to any one of claims 1 to 7, wherein the processor further determines the most likely noise source of the image quality problem based on the evaluated noise source and presents the most likely noise source to the operator.

9. The system according to any one of claims 1 to 8, wherein the processor further determines whether the medical image should be reacquired based on the quality management indicator and presents a proposal to the operator on whether to reacquire the medical image.

10. The processor further tracks the operator's judgment during and after the acquisition of the medical image, detects non - general operator judgments from the operator's judgment, and requests the basis for the non - general operator judgment from the operator. The system according to any one of claims 1 to 9.

11. The processor further sends the medical image to a radiologist for judgment commission, and tracks the radiologist's judgment. The system according to claim 10.

12. The processor further generates a quality management report indicating how the medical imaging device, the operator, and the subject affect the quality of the acquisition of the medical image, and generating the quality management report is based on the quality management indicator. The system according to any one of claims 1 to 11.

13. A medical imaging device, and a system for quality management according to any one of claims 1 to 12, A medical imaging system including , wherein the processor of the system for quality control further executes analysis of an image captured by the medical imaging device, determines the quality control indicator based on the image analysis, and outputs it to the operator.

14. A computer-implemented method for providing quality control during acquisition of a medical image, the computer-implemented method comprising: Evaluating a noise source that may adversely affect the quality of the medical image, the noise source including: A device noise source caused by a medical imaging device during acquisition of a medical image of a subject; An operator noise source caused by an operator during acquisition of the medical image; and A subject noise source caused by the subject during acquisition of the medical image, wherein the evaluating step includes at least two of them; Determining a quality control indicator based on the evaluated noise source and outputting it to the operator; Identifying one or more possible causes of noise that may introduce artifacts in the medical image based on the evaluated noise source; A computer-implemented method including the above steps.

15. A computer program including computer program code means which, when the computer program is executed on a computer, implements the method according to Claim 14.

Images