System and method for monitoring and suppressing radio frequency (RF) noise

Abstract

In the environment of a medical imaging or treatment system (1), radio frequency (RF) noise should be suppressed and / or avoided. This is accomplished by a system for monitoring radio frequency (RF) noise and a computer-implemented method for suppressing and / or avoiding radio frequency (RF) noise. A medical imaging or treatment system (1), whereby the medical imaging or treatment system (1) is located inside an operating chamber (2), the system comprising: a plurality of sensors (4, 5, 7), where the sensors (4, 5, 7) are configured for continuous broadband detection of radio frequency (RF) signals, where a first sensor (4) is located inside the operating chamber (2) and a second sensor (5) is located outside the operating chamber (2), where the sensors (4, 5, 7) are configured to provide sensor data, the system further comprises: a processor unit (8) wherein the sensors (4, 5, 7) are connected to the processor unit (8), the processor unit comprising: at least one memory (9) for storing instructions; at least one processor (10) executing the instructions to cause the following to be performed: calculating a noise level and a spectrum based on sensor data in the processor unit (8); a procedure is performed for suppressing and / or avoiding RF noise and spurious signals based on noise levels.

Description

Technical Field

[0001] The present invention relates to the field of monitoring radio frequency (RF) noise in the environment of a medical imaging or treatment system. The present invention also relates to a method for suppressing and / or avoiding radio frequency (RF) noise in the environment of a medical imaging or treatment system. Background Art

[0002] In the field of medical imaging or treatment systems, a positive diagnostic experience and simultaneously good image quality are crucial. Medical imaging or treatment systems will become increasingly autonomous, with fewer local operators, relying on actions and automated remotely controlled workflow steps. Imaging or treatment guidance requires a seamless workflow. For this and other reasons, new features will be introduced, among which features should be used in close proximity to the medical imaging or treatment system. For example, compatibility with image quality across numerous devices is crucial. External and additional equipment such as portable monitoring equipment, sensors for vital signs, and mobile anesthesia systems are controlled by local hospital services and clinical operators. This additional equipment is located in, for example, the operating room of the medical imaging or treatment system and connected to an AC power source and local communication channels. Positioning and placement are often performed relative to clinical requirements and constraints, such as limited space and accessibility. Until now, signal integrity has not been controlled. For example, reports from MRI sites have shown image quality issues caused by spurious signal injection of digital noise in the MRI image band. Spurious signals from digital components (such as local switching boost converters) are unstable in frequency, amplitude, and phase and are often related to the operating mode, positioning, and orientation of the equipment. Summary of the Invention

[0003] An object of the present invention is to provide a system and method for monitoring radio frequency (RF) noise in an environment of a medical imaging or treatment system to suppress and / or avoid the radio frequency (RF) noise.

[0004] According to the invention, this object is solved by the subject matter of the independent claims. Preferred embodiments of the invention are described in the dependent claims.

[0005] Thus, according to the present invention, a system for monitoring radio frequency (RF) noise in the environment of a medical imaging or treatment system is foreseen, wherein the medical imaging or treatment system is located inside an operating room. The system comprises: a plurality of sensors, wherein the sensors are configured for continuous broadband detection of radio frequency (RF) signals, wherein a first set of sensors is located inside the operating room and a second set of sensors is located outside the operating room, wherein the sensors are configured to provide sensor data. The system further comprises: a processor unit, wherein the sensors are connected to the processor unit, the processor unit comprising: at least one memory for storing instructions; and at least one processor that executes the instructions to: calculate a noise level and spectrum based on the sensor data in the processor unit; and execute a procedure for suppressing and / or avoiding RF noise and spurious signals based on the noise level. In this application, an operating room is understood to be a room shielded from the outside by an RF cage. This distinguishes the terms "inside" from "outside," defined by the RF-shielded cabin and the outside, unshielded world with all types of RF radiation. Continuous broadband sampling contrasts with known intermittent sampling in the Larmor band. In particular, provision may be made for the continuous wideband sampling of the RF signal to be at least wider than the MR bandwidth.RF noise and spurious signals are typically signals in the RF spectrum that may adversely affect the functioning of a medical imaging or therapy system.

[0006] Very-broadband sampling of RF signals enables the capture of RF signals that exceed the typical bandwidth of imaging data (e.g., magnetic resonance signals) acquired by medical imaging or therapy systems. A data compression module on the sensor assembly enables efficient transmission of the acquired data, thereby avoiding the transmission of large amounts of uncompressed data output by the sensor.

[0007] In a technically advantageous embodiment of the system, at least one processor is arranged to execute a machine learning algorithm. Machine learning can include neural networks, autoencoders, decision trees, support vector machines, and other kinds of machine learning. Along with the various advantages and possible applications of neural networks and other machine learning techniques, the most common is that they help us perform classification and clustering. Such machine learning models learn (or train) by processing examples, each of which contains a known "input" and "result", forming a probabilistic weighted association between the two, which is stored in a data structure of the model itself. Training from a given example is typically performed by determining the difference between the processed output of the model (usually a prediction) and the target output.

[0008] Learning is the process of adapting a machine learning model to better perform a task by considering sample observations. Learning involves adjusting the model's parameters to improve the accuracy of the results. This is accomplished by minimizing the observed error. Learning is complete when examining additional observations does not significantly reduce the error rate.

[0009] A machine learning algorithm is trained to return characterizations of (imminent) hardware failure and / or artifactual operation of a medical imaging or therapy system. The returned characterizations can be employed to control settings of an imaging sequence in order to reduce sensitivity to the characterized (imminent) hardware failure or artifactual operation. For example, the settings can be adjusted to move anticipated image artifacts outside the field of view of images produced by the imaging sequence. Furthermore, artifactual operation or hardware failure may be attributable to an RF-active object or metallic device inadvertently placed in or near an examination region. The artifactual operation will cause artifactual image artifacts when controlling the medical imaging or therapy system, which adversely affects image quality for efficacy or accuracy of the therapy.

[0010] The present invention enables automatic assessment of the technical status of a medical imaging or treatment system based on acquired extremely wideband RF signals, particularly automatic assessment of whether the medical imaging or treatment system is in a sufficient and safe state for executing a planned imaging or treatment protocol. This assessment can also be performed during installation of the medical imaging or treatment system.

[0011] The extremely broadband spectrum of the sensor assembly can be divided into several narrower frequency ranges. The divided ranges can be selected by switchable filter bands.

[0012] In another technically advantageous embodiment of the system, at least one of the sensors is a portable sensor device for on-site RF signal measurements.Portable sensor devices have the advantage that they can be used to take measurements at particularly important locations.

[0013] In a technically advantageous embodiment of the system, the sensor is coupled to the processor unit via a non-galvanic signal path. In particular, in a technically advantageous embodiment of the system, the sensor is connected to the processor unit via an optical and / or shielded wire-based and / or wireless signal path. This has the advantage that the non-galvanic signal path prevents common-mode coupling.

[0014] In a technically advantageous embodiment of the system, the sensor includes a dual-mode antenna for WLAN control and RF signal detection, and / or the sensor includes separate antennas for WLAN control and RF signal detection using a duplexer. Antennas can be implemented and configured and mixed for sensitivity with respect to electromagnetic E and H field components.

[0015] In another technically advantageous embodiment of the system, the medical imaging or therapy system is a magnetic resonance imaging (MRI) system, wherein the sensor is temporally aligned with an MRI multi-receiver channel of the MRI system.

[0016] In a technically advantageous embodiment of the system, the medical imaging or treatment system is a magnetic resonance imaging (MRI) system, wherein at least one sensor for detecting radio frequency (RF) signals is arranged around the MRI system. Using the sensors arranged around the MRI system, the RF noise field within the MRI system can be calculated. Before a scan begins, the RF noise spectrum in the MRI system from sources external to the scanner (e.g., surrounding equipment) can be determined and used as a baseline noise floor estimate. During the scan, the field outside the scanner is monitored by the sensor, and based on real-time measurements and calibration from before the scan, the RF noise inside the scanner can be retrospectively removed from the acquired k-space signal.

[0017] In another aspect of the invention, a medical imaging or treatment system is envisioned comprising a system for monitoring radio frequency (RF) operation of a magnetic resonance imaging system as described above.

[0018] In another aspect of the present invention, the object is achieved by a computer-implemented method for suppressing and / or avoiding radio frequency (RF) noise in the environment of a medical imaging or treatment system by a system for monitoring radio frequency (RF) noise in the environment of the medical imaging or treatment system, wherein the medical imaging or treatment system is located inside an operating room, the system comprising: a plurality of sensors, wherein the sensors are configured to detect radio frequency (RF) signals, wherein a first sensor is located inside the operating room and a second sensor is located outside the operating room, wherein the sensors are configured to provide sensor data, the system further comprising: a processor unit, wherein the sensors are connected to the processor unit, the processor unit comprising: at least one memory for storing instructions; at least one processor for executing the instructions, the method comprising the steps of: measuring RF signals by the sensors inside and outside the operating room and providing sensor data; calculating a noise level in the processor unit based on the sensor data; and executing a process for suppressing and / or avoiding RF noise based on the noise level.

[0019] The term medical imaging or therapy system may be used to refer to different systems, such as magnetic resonance imaging (MRI) systems, X-ray systems, computed tomography (CT) systems, or therapy systems, like linear accelerator-based therapy systems or proton beam therapy systems.

[0020] In a technically advantageous embodiment of the method, the step of performing a procedure based on the noise level is based on a thresholding technique with adaptive window selection. An advantage of this technique is that it provides for the effectiveness of an appropriate procedure.

[0021] In another technically advantageous embodiment of the method, the step of performing a procedure for suppressing and / or avoiding RF noise based on the noise level includes the step of modifying the operation of a component of the medical imaging or therapy system based on the noise level. Modifying the operation of a component of the medical imaging or therapy system can achieve an improvement in the noise level, provided the noise is related to the operating mode.

[0022] In a technically advantageous embodiment of the method, the step of measuring the RF signal by means of the sensor and providing sensor data comprises the step of sampling signals from the sensor inside the operating chamber and from the sensor outside the operating chamber, and the step of processing the sensor data in the processor unit comprises the step of subtracting or dividing the sensor data from inside and outside the operating chamber to calculate a corresponding noise level.

[0023] In another technically advantageous embodiment of the method, at least one processor is arranged to execute a neural network machine learning algorithm, and the step of processing the sensor data in the processor unit comprises the following steps: feeding the sensor data to the neural network machine learning algorithm, wherein the neural network machine learning algorithm is trained to identify the device by a fingerprint of the device's spurious signal frequency band.

[0024] In a technically advantageous embodiment of the method, the step of performing a procedure based on the noise level comprises the step of proposing to an operator an adapted operating mode for the medical imaging system to achieve clinical images with better image quality.

[0025] In another technically advantageous embodiment of the method, the medical imaging or treatment system is a magnetic resonance imaging (MRI) system, wherein sensors inside and outside the operating room are temporally aligned with an MRI digital multi-receiver of the MRI system, and the step of performing a procedure based on noise levels includes the steps of removing noise from the clinical image in k-space or image space as much as possible. In an embodiment of the present invention, noise removal is performed with the aid of a trained machine learning model. In cases where the signal is too strong (the MRI receive signal exceeds the dynamic range), machine learning can help estimate the true signal level without additive RF noise. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] These and other aspects of the invention will be apparent from and elucidated with reference to the embodiments described hereinafter. However, such embodiments do not necessarily represent the full scope of the invention, and reference is made therefore to the claims and their interpretation herein.

[0027] In the attached figure:

[0028] Figure 1 schematically depicts a system for monitoring radio frequency (RF) noise in the environment of a medical imaging or treatment system according to an embodiment of the present invention,

[0029] Figure 2 Depicted is a flow chart of a computer-implemented method for suppressing and / or avoiding radio frequency (RF) noise in the environment of a medical imaging or treatment system, in accordance with an embodiment of the present invention.

[0030] List of reference numerals:

[0031] Medical imaging or treatment systems1

[0032] RF shielded operation room 2

[0033] Door 3

[0034] Sensor 4 inside the operating room

[0035] Sensor 5 outside the operating room

[0036] Signal Path 6

[0037] Sensors around medical imaging systems7

[0038] Processor unit 8

[0039] Memory 9

[0040] Processor 10

[0041] Collection Module 11

[0042] Service Module 12

[0043] Wall bracket 13 DETAILED DESCRIPTION

[0044] Figure 1 A system for monitoring radio frequency (RF) noise in the environment of a medical imaging or treatment system 1 according to an embodiment of the invention is schematically depicted.

[0045] exist Figure 1 In FIG. 1 , a medical imaging or treatment system 1 is shown which is located inside an operating room 2, wherein the medical imaging or treatment system 1 can be Figure 1The door 3 shown is used to enter the room. The medical imaging or treatment system 1 can be, for example, a magnetic resonance imaging (MRI) system, an X-ray system or a computed tomography (CT) system or a treatment system, such as a linear accelerator-based treatment system or a proton beam therapy system. Another field of application of the invention is application in interventional radiology suites, where biopsies, diagnoses or treatments are precisely guided using real-time fluoroscopy and / or MRI. Here, many individual devices are located in the operating room 2 and signal integrity is very important for patient safety. The operating room is understood to be a room that is shielded from the outside by an RF cage. This distinguishes the terms "inside" from "outside" as being defined by the RF-shielded cabin and the outside unshielded world with all kinds of RF radiation.

[0046] A system for monitoring radio frequency (RF) noise includes multiple sensors 4, 5, and 7, each configured to detect radio frequency (RF) signals. A first sensor 4 is located inside operating room 2, and a second sensor 5 is located outside operating room 2. Sensors 4, 5, and 7 are configured to provide sensor data. Specifically, the sensors may be broadband sensors for measuring RF noise. In particular, broadband sampling of the RF signal is continuous, wherein, in embodiments, the sampling is at least wider than the MR bandwidth. Continuous broadband sampling contrasts with known intermittent sampling in the Larmor band. Sensors 4, 5, and 7 are connected to a processor unit 8, which includes a memory 9 for storing instructions and a processor 10 for executing the instructions. For example, the sensors may be coupled to processor unit 8 via a non-current signal path 6. In one embodiment, optical and / or wireless signal paths 6 may be used. In one embodiment, continuous broadband sampling of the RF signal, wherein the sampling is at least wider than the MR bandwidth, is contemplated. Sensor data from sensor 4 located inside operating room 2 can be compared with sensor data from second sensor 5 located outside operating room 2 to monitor the integrity of the RF cage. Therefore, predictive maintenance can be requested.

[0047] Processor unit 8 can act as a sensor hub, where the sensor signal paths 6 converge. To this end, processor unit 8 can also include a software-defined radio (SDR), which allows sensors 4, 5, and 7 to be remotely controlled and configured. Processor 10 can, for example, be a field-programmable gate array (FPGA), onto which logic circuits suitable for evaluation can be loaded. In particular, provision can be made for processor 10 to execute a neural network machine learning algorithm. This neural network machine learning algorithm allows analysis of individual noise and signal sources. For this purpose, the neural network is previously trained accordingly. The neural network determines the information flow, which is used, for example, for remote operators and system operators for predictive maintenance. The trained network identifies devices by fingerprinting spurious signal frequency bands. In addition, the signals of sensors 4, 5, and 7 can be digitized by processor unit 10. For this purpose, processor unit 10 includes corresponding equipment, such as an ADC converter. The digitized data and preselected information are forwarded, collected, and stored in collection module 11. Local data processing before remote submission is advantageous to avoid large amounts of data resulting from continuous broadband sampling. The collection module 11 can be added to the processor unit 8 as a standalone module, but it can also be integrated into the processor unit 8. For example, it can be a cloud-based collection module 11 to store data outside the system. The collection module 11 can be connected to the service module 12, which is intended for example for service coordination and fault prediction. In addition, the service module 12 can provide additional customer guidance features included in the app, or it can display guidance data on an output unit such as a monitor.

[0048] Sensors 4, 5, and 7 are equipped with dual-mode antennas or separate antennas using a duplexer for WLAN control and separate MR band antennas. Antennas can be implemented, configured, and mixed based on their sensitivity to the E-field and H-field electromagnetic components. Signals from sensor 4 inside operating room 2 and from sensor 5 outside operating room 2 are sampled in a narrowband within the MRI channel and subtracted or divided. Signals are monitored over time, with significant changes and signal events processed using timestamps and stored, for example, in collection module 11. Furthermore, these events are reported, for example, via an output unit of service module 12. This reduces the amount of data to a reasonable level. Measurements inside operating room 2 help locate problems and categorize them into categories such as "internal noise source," "RF cage damage," "RF gate damage," or "external noise source." Defined threshold alerts inform service personnel and operators about image quality. In another embodiment, based on noise level analysis, a more rugged MRI sequence is recommended to the operator to achieve clinical images with acceptable image quality.

[0049] Furthermore, it can be provided that sensors 4, 5, and 7 are partially portable sensor devices. Here, it can be provided that the portable sensor device is placed in a wall mount 13. For analysis, the portable sensor device can be removed from the wall mount 13 and moved through the environment of the medical imaging or treatment system 1 for on-site noise source analysis. To achieve this, the portable sensor device is configured to operate, optionally with battery power, and to transmit data wirelessly to, for example, the wall mount 13. The wall mount 13 serves as a power supply / charger and wirelessly collects the data and forwards it to the processor unit 8.

[0050] When using a system for monitoring radio frequency (RF) noise in the environment of an MRI system, sensor 4 inside operating room 2 and sensor 5 outside operating room 2 are temporally aligned with an MRI digital multi-receiver. The signals from sensors 4, 5, and 7 are fed to a processor unit 8, on which a neural network machine learning algorithm is run. The neural network machine learning algorithm is trained to remove noise from clinical images in k-space or image space to exclude possible cases with too strong a signal because the spurious signal saturates the MR RX chain, causing nonlinearity.

[0051] In another embodiment, the signal chain of sensors 4, 5, and 7 can be switched between different modes. For example, modes for narrowband or broadband signal detection can be provided. Furthermore, different modes can be provided, for example, for magnetic resonance imaging. These can be, for example, modes for MR imaging or service or installation monitoring modes.

[0052] During the installation of the medical imaging or treatment system 1, continuous monitoring of signal integrity by sensors 4, 5, 7 can be provided. The results (e.g., a representation of the signal and noise levels) can be presented on a display unit (e.g., a monitor or 3D glasses). By doing so, guidance and transparency for field service can be provided during installation. This also facilitates the perception of noise generation based on external X-ray scanners and accessory equipment (e.g., anesthesia, monitors, surgical robots, etc.), as well as mitigation through guidance on positioning and orientation. The software subroutine, for example, communicates with the MRI system software and monitors noise analysis for the user and remote operator. At the same time, the operational phase of the MRI system is added to the noise sensors 4, 5, 7 for data acquisition.

[0053] In another embodiment, sensors 7 can be placed directly around the MRI system, for example, mounted to an external bore cover, such as a magnetic field camera. In addition to receiving from or pairing with sensors around or within the MRI system, sensors 7 around the MRI system can also transmit. This enables active noise cancellation during acquisition by sending a signal via a transmitter that cancels unwanted RF noise through destructive interference within the MRI system bore. To process RF noise in real time during scanning, fast processing hardware such as an FPGA can be used to determine the transmit field required for active cancellation. This hardware can be programmed to handle the field propagation equations from sensors 4, 5, and 7, and receive coils programmed for efficient processing. Alternatively, to add a dedicated transmitter for active noise cancellation, the RF pulses sent by the main transmit coil in the MRI system can be designed so that they are RF noise-aware and issue cancellation terms to suppress RF noise at the receive coil during acquisition. For example, active sampling techniques with automatic real-time sequence adjustment can be used. To calculate the noise level based on the sensor data, Maxwell's equations can be used to calculate the RF noise field within the MRI system from sensors outside the MRI system.

[0054] Sensor data from the first sensor 4 inside the operating room and the sensors 7 placed directly around the MRI system 1 can typically be used to correct the acquired MRI data in post-processing / reconstruction. This enables image denoising / enhancement. The sensor data from the sensor 7 and the MRI receive coils can be compared to rate the fidelity of the acquired MRI data (or data to be acquired in the planning sequence). If RF noise dominates the MRI signal, embodiments of the present invention can perform the following: estimate the effective SNR of the planning image and / or warn the operator and / or recommend a more noise-robust MRI sequence (e.g., with more signal averaging).

[0055] Before the scan begins, the spectrum of RF noise in the MRI system from sources external to the MRI system (e.g., surrounding equipment) can be determined and used as a baseline noise floor estimate. During the scan, the field outside the MRI system is monitored by sensors 4, 5, 7, and based on real-time measurements and calibration from before the scan, the RF noise inside the MRI system can be retrospectively removed from the acquired k-space signal. In embodiments, additional RF noise can be removed from the acquired k-space data during post-processing.

[0056] Figure 2A flow chart depicts a computer-implemented method for suppressing and / or avoiding radio frequency (RF) noise in the environment of a medical imaging or treatment system, according to an embodiment of the present invention. In step S1, a system for monitoring radio frequency (RF) noise in the environment of a medical imaging or treatment system 1 is provided. The medical imaging or treatment system 1 is located within an operating room 2. The system includes a plurality of sensors 4, 5, and 7, wherein the sensors 4, 5, and 7 are configured to detect RF signals, wherein a first sensor 4 is located within the operating room 2 and a second sensor 5 is located outside the operating room 5, wherein the sensors 4, 5, and 7 are configured to provide sensor data. The system also includes a processor unit 8, wherein the sensors 4, 5, and 7 are connected to the processor unit 8. The processor unit 8 includes at least one memory 9 for storing instructions and at least one processor 10 for executing the instructions. In step S2, RF signals are measured by the sensors 4, 5, and 7 within and outside the operating room 2, wherein sensor data is provided. In step S3, a noise level and spectrum based on the sensor data are calculated in the processor unit 8. In an embodiment of the present invention, the processor 10 is configured to execute a neural network machine learning algorithm. Thus, the step of processing sensor data in processor unit 8 may include the additional step of feeding the sensor data to a neural network machine learning algorithm, wherein the neural network machine learning algorithm is trained to identify the device by the fingerprint of the device's spurious signal frequency band. In step S4, a process for suppressing and / or avoiding RF noise based on noise level is executed. In one embodiment, to suppress and / or avoid RF noise, provision may be made to change the operation of components of the medical imaging or treatment system 1 based on the noise level. In another embodiment, an adjusted operating mode for the medical imaging or treatment system 1 is proposed to the operator to achieve clinical images with better image quality. Thus, the process for suppressing and / or avoiding RF noise involves guiding and alerting the operator of the medical imaging or treatment system 1. If a noise issue is detected, this adds a layer of interaction with the operator of the medical imaging or treatment system 1. Only certain MRI sequences may be affected, or increased scan times may be required to maintain the scanner until service arrives.

[0057] In an embodiment, it is envisioned that clinical and interventional workflows can be controlled and guided based on continuous monitoring of RF noise inside and outside the operating room 2 of a medical imaging or treatment system 1 using real-time broadband signal and noise monitoring. Furthermore, threshold and window-based actions for predictive maintenance and customer guidance are provided. A local processor unit 8 (e.g., comprising an FPGA) performs analysis using machine learning algorithms to generate and prepare data for operator guidance and predictive maintenance. Thus, among other things, operator guidance and monitoring are provided to optimize positioning and alignment of equipment for maximum signal integrity. This improves workflows for acquiring MR images, for example.

[0058] In another embodiment, sensors 4, 5, and 7 inside and outside the operating room 2 are temporally aligned with the MRI digital multi-receiver of the MRI system. Knowledge of the noise spectrum of the sensor signals from sensors 4 and 7 (in the operating room / RF cabin) is used to remove as much noise as possible from the clinical image in k-space or image space. This can be achieved, for example, by feeding the signals to a neural network machine learning algorithm running on a processing unit, which removes noise from the clinical image in k-space or image space.

[0059] Although the invention has been illustrated and described in detail in the drawings and the foregoing description, such illustration and description are to be regarded as illustrative or exemplary and not restrictive; the invention is not limited to the disclosed embodiments. By studying the drawings, the disclosure and the claims, a person skilled in the art will be able to understand and implement other variations to the disclosed embodiments when practicing the claimed invention. In the claims, the word "comprising" does not exclude other elements or steps, and the word "a" or "an" does not exclude a plurality. Although certain measures are recited in mutually different dependent claims, this does not indicate that a combination of these measures cannot be used to advantage. Any reference signs in the claims should not be interpreted as limiting the scope. In addition, for the sake of clarity, not all elements in the drawings may have been provided with reference signs.

Claims

1. A system for monitoring radio frequency (RF) noise and spurious signals in the environment of a medical imaging or treatment system (1), wherein: The medical imaging or treatment system (1) is located inside an operating room (2), and the system comprises: A sensor assembly comprising a plurality of sensors (4, 5, 7), wherein the sensors (4, 5, 7) are configured for continuous very wideband detection of radio frequency (RF) signals with a bandwidth of at least 2 MHz or 5 MHz or 10 MHz, wherein a first sensor (4) is located inside the operating room (2) and a second sensor (5) is located outside the operating room (2), wherein the sensors (4, 5, 7) are configured to provide sensor data, the sensor assembly comprising a data compression module for compressing the detected RF signals, and the system further comprising: a processor unit (8), wherein the sensors (4, 5, 7) are connected to the processor unit (8), the processor unit comprising: at least one memory (9) for storing instructions comprising a machine learning algorithm, At least one processor (10) that executes the instructions to cause the following operations to be performed: calculating a noise level and a frequency spectrum in the processor unit (8) based on the sensor data, controlling the machine learning module to return a characterization of an (imminent) malfunction or artifactual operation of the medical imaging or therapy system, A process for controlling the medical imaging or treatment system is executed to avoid or reduce the characterized malfunction or artifactual operation.

2. The system according to claim 1, wherein: The process for controlling the medical imaging or treatment system involves resetting one or more imaging sequences to reduce sensitivity to characterized malfunctioning or artifactual operation.

3. A system according to any preceding claim, wherein: At least one of the sensors (4, 5, 7) is a portable sensor device for on-site RF signal measurement.

4. A system according to any preceding claim, wherein: The sensors (4, 5, 7) are coupled to the processor unit (8) via a non-current signal path (6).

5. The system according to claim 4, wherein: The sensors (4, 5, 7) are connected to the processor unit (8) via optical signal paths and / or shielded wire-based signal paths and / or wireless signal paths (6).

6. A system according to any preceding claim, wherein: The sensors (4, 5, 7) include dual-mode antennas for WLAN control and RF signal detection, and / or the sensors (4, 5, 7) include separate antennas for WLAN control and RF signal detection using a duplexer.

7. A system according to any preceding claim, wherein: The medical imaging or treatment system (1) is a magnetic resonance imaging (MRI) system, wherein the sensors (4, 5, 7) are temporally aligned with MRI multi-receiver channels of the MRI system.

8. A system according to any preceding claim, wherein: The medical imaging or treatment system (1) is a magnetic resonance imaging (MRI) system, wherein at least one sensor (7) for detecting radio frequency (RF) signals is arranged around the MRI system.

9. A medical imaging or treatment system (1) comprising a system for monitoring radio frequency (RF) operation of a magnetic resonance imaging system according to any one of claims 1 to 8.

10. A computer-implemented method for suppressing and / or avoiding radio frequency (RF) noise in an environment of a medical imaging or treatment system (1) by a system for monitoring radio frequency (RF) noise in an environment of a medical imaging or treatment system (1), wherein: The medical imaging or treatment system (1) is located inside an operating room (2), and the system comprises: A plurality of sensors (4, 5, 7), wherein the sensors (4, 5, 7) are configured for continuous broadband detection of radio frequency (RF) signals, wherein a first sensor (4) is located inside the operating room (2) and a second sensor (5) is located outside the operating room (5), wherein the sensors (4, 5, 7) are configured to provide sensor data, the system further comprising: A processor unit (8), wherein the sensors (4, 5, 7) are connected to the processor unit (8), the processor unit (8) comprising: at least one memory (9) for storing instructions, At least one processor (10) executes the instructions, the method comprising the steps of: Measuring RF signals by the sensors (4, 5, 7) inside the operating room (2) and outside the operating room (2) and providing sensor data, calculating a noise level in the processor unit (8) based on the sensor data, A procedure for suppressing and / or avoiding RF noise is performed based on the noise level.

11. The method according to claim 10, wherein: The steps of performing a procedure for suppressing and / or avoiding RF noise based on the noise level include the following steps: The operation of a component of the medical imaging or treatment system (1) is altered based on the noise level.

12. The method according to any one of claims 10 or 11, wherein: The step of measuring RF signals by means of the sensors (4, 5, 7) and providing sensor data comprises the following steps: The step of sampling the signals from the sensor (4) inside the operating room (2) and the signals from the sensor (5) outside the operating room (2), and processing the sensor data in the processor unit (8) comprises the following steps: The sensor data from inside and outside the operating room (2) are subtracted or divided to calculate the corresponding noise level.

13. The method according to any one of claims 10 to 12, wherein: The at least one processor (10) is arranged to execute a neural network machine learning algorithm, and the step of processing the sensor data in the processor unit (8) comprises the following steps: The sensor data is fed to the neural network machine learning algorithm, wherein the neural network machine learning algorithm is trained to identify the device by a fingerprint of the device's spurious signal frequency band.

14. The method according to any one of claims 10 to 13, wherein: The step of executing a process based on the noise level comprises the following steps: An adjusted operating mode for the medical imaging system (1) is proposed to an operator to achieve clinical images with better image quality.

15. The method according to any one of claims 10 to 14, wherein The medical imaging or treatment system (1) is a magnetic resonance imaging (MRI) system, wherein the sensors (4, 5, 7) inside and outside the operating room (2) are time-aligned with an MRI digital multi-receiver of the MRI system, and the step of performing a procedure based on the noise level comprises the following steps: Noise is removed from clinical images in k-space or image space as much as possible.

Images