Memory-efficient monitoring of a magnetic resonance measurement using a prefix-free code

Abstract

The invention relates to a method for monitoring a magnetic resonance measurement, a magnetic resonance device, and a computer program. During the magnetic resonance measurement, monitoring values ​​are acquired. These monitoring values ​​are quantized to generate monitoring memory values. The monitoring memory values ​​are stored as a prefix-free code in a memory unit. The magnetic resonance measurement is then monitored for compliance with a monitoring limit value based on these stored monitoring values.

Description

[0001] The invention relates to a method for monitoring a magnetic resonance measurement, a magnetic resonance device and a computer program product.

[0002] In medical technology, magnetic resonance imaging (MRI), also known as magnetic resonance tomography (MRI), is characterized by high soft tissue contrast. A magnetic resonance measurement is performed using a magnetic resonance scanner. Typically, a main magnet in the scanner generates a primary magnetic field, while a gradient coil unit generates a gradient magnetic field within a defined area of ​​the scanner. During the MRI scan, a patient is positioned within this area. To generate magnetic resonance signals, radio-frequency (RF) pulses are emitted into the scan area according to a magnetic resonance sequence. These signals are received by the scanner as measurement data and used to reconstruct magnetic resonance images.Magnetic resonance signals are usually received using MR local coils, i.e., receiving coils located locally on the patient to achieve a high signal-to-noise ratio.

[0003] The application of RF pulses delivers thermal energy to the patient, causing their body to heat up. In magnetic resonance imaging (MRI) devices with a single transmit channel, the RF heating, specifically the specific absorption rate (SAR), is proportional to the RF power applied to the patient. Regulatory standards require that the average SAR value be limited to a specific value over certain time intervals. Typically, separate SAR limits are defined for 10-second and 6-minute averages. These limits apply to each 10-second or 6-minute interval and therefore require a moving average calculation to obtain the average SAR value. For example, for a 6-minute interval, the SAR value is recorded every 100 ms, and 3600 values ​​are summed to obtain the moving average. This typically requires approximately 15 kB of storage space, primarily RAM.Documents US2014 / 0292332A1 and US2015 / 0022206A1 disclose exemplary procedures for SAR monitoring.

[0004] However, in MRI systems with multiple transmit channels, especially those with eight or more, the moving average must be calculated not just for a single parameter, but for a multitude of parameters. This often requires significantly more memory, e.g., 15 MB or more. Since SAR monitoring algorithms are typically implemented on small embedded systems, this amount of memory is often unavailable.

[0005] The object of the present invention can be considered to be to provide a storage-efficient method for monitoring a magnetic resonance measurement. This object is achieved by the features of the independent claims. Advantageous embodiments are described in the dependent claims.

[0006] Accordingly, a computer-implemented method for monitoring a magnetic resonance (MRI) measurement is proposed. This method provides monitoring values ​​acquired during the MRI measurement. Specifically, the proposed method can also include the acquisition of these monitoring values ​​during the MRI measurement itself. Monitoring memory values ​​are generated by quantizing these values. These memory values ​​are stored as a prefix-free code, particularly of variable length, in a storage unit. The stored monitoring values ​​are then used to monitor the MRI measurement for compliance with a (predefined) monitoring limit.

[0007] The monitoring threshold can, for example, be retrieved from a database and / or determined, in particular calculated, before the start of monitoring. Advantageously, the monitoring threshold is set in such a way as to ensure compliance with regulatory standards.

[0008] It is conceivable that the magnetic resonance measurement is monitored for compliance with further limit values ​​(in addition to the aforementioned monitoring limit). If several limit values ​​are monitored, the aforementioned monitoring limit value is preferably the limit value whose monitoring requires the largest storage capacity of the memory unit.

[0009] If, for example, the monitoring of the magnetic resonance measurement includes monitoring of SAR exposure in a (rolling) 10-second time window and also monitoring of SAR exposure in a (rolling) 6-minute time window, then monitoring the limit value for the 6-minute time window (assuming otherwise identical conditions such as the time resolution of the monitoring) requires a larger storage capacity of the storage unit due to the undoubtedly larger amount of monitoring memory values ​​to be stored.

[0010] Furthermore, it is conceivable that when monitoring multiple limit values, the proposed procedure could be executed several times in parallel, in particular with a separate storage unit for each instance.

[0011] In particular, the monitoring measurements are (temporally) recorded monitoring measurements taken sequentially. Specifically, the monitoring measurements are (temporally) successive monitoring measurements. Specifically, the monitoring measurements are recorded at constant time intervals. For example, the monitoring measurements can be recorded at time intervals td of 100 ms, i.e., the temporal resolution of the monitoring would be 100 ms in this case.

[0012] Preferably, the monitoring measurements characterize a quantity to be monitored. Preferably, the monitoring measurements are suitable for deriving at least one value, in particular a SAR value, which can be compared with the monitoring limit. The monitoring measurement can, for example, be a power, in particular an RF power, and / or a voltage of a transmit channel of the magnetic resonance device. The monitoring limit can, for example, include a SAR limit or be a SAR limit itself. Accordingly, the monitoring measurements and the monitoring memory values ​​derived from them can represent SAR values.

[0013] Quantizing the monitoring measurements can, in particular, include discretizing them. The monitoring measurements are, in particular, continuous and / or steady values. Specifically, each monitoring measurement is assigned a discrete monitoring memory value. Specifically, each monitoring measurement is transferred to a discrete monitoring memory value and / or mapped to a discrete monitoring memory value.

[0014] Preferably, the monitoring measurements are quantized in N quantization levels, where N ≤ 500, particularly ≤ 200. Advantageously, N is chosen to be as small as possible to save storage space. Advantageously, N is chosen to be as high as necessary to enable sufficiently accurate evaluation of the monitoring measurements, particularly to determine a patient's SAR exposure with sufficient accuracy. Advantageously, the number of quantization levels corresponds to the measurement accuracy during the acquisition of the monitoring measurements, in order to make the best possible use of the measurement accuracy. Advantageously, the storage accuracy resulting from the quantization of the monitoring measurements is not greater than the measurement accuracy.

[0015] In particular, the number of quantization levels N determines the resolution of the monitoring memory values ​​at which the monitoring is performed. Preferably, the quantization level is chosen such that the monitoring is carried out with a (still) sufficient resolution.

[0016] For example, each quantization level is assigned a specific value interval and / or a value point. For example, all monitoring measurements that fall within a specific value interval are assigned to a specific monitoring memory value. Preferably, enough value intervals are defined such that the value intervals and / or value points cover the entire range of values ​​in which the monitoring measurements lie.

[0017] Preferably, the quantization of the monitored measurements, in particular the definition of the value intervals assigned to the quantization stages, is carried out depending on the monitoring limit. Preferably, the range of values ​​covered by the value intervals assigned to the quantization stages is limited by the monitoring limit. Preferably, there is no quantization stage with a value interval whose lower limit is above the monitoring limit.

[0018] Preferably, the value ranges assigned to the quantization stages are selected such that at most one value range, and in particular no value range, exists for which the acquisition of a (single) monitoring measurement value (especially regardless of any other acquired monitoring measurements) within this value range would (in any case and / or always and / or immediately) lead to a breach of the monitoring limit, in particular to a cessation of the measurement. Preferably, there is no quantization stage for a monitoring measurement value characterized in that this monitoring measurement value alone (especially regardless of any other acquired monitoring measurements) would lead to an exceedance of the limit, in particular to a cessation of the magnetic resonance measurement. Advantageously, this allows the required storage space to be kept particularly low.

[0019] In addition to quantizing the monitoring measurements, the procedure can also include digitizing the measurement signals to capture, in particular generate, the monitoring measurements during magnetic resonance measurement.

[0020] The monitoring memory values, in particular a sequence of monitoring memory values, are stored as a prefix-free code in the memory unit. Preferably, the sequence of monitoring memory values ​​is determined from monitoring measurements acquired sequentially. Preferably, the sequence of monitoring memory values ​​comprises a maximum number of monitoring memory values. Preferably, the sequence of monitoring memory values ​​is determined from the most recently acquired monitoring measurements.

[0021] A prefix-free code (often also called a prefix code) whose codewords have the property that no codeword of the code is a prefix of another codeword. Preferably, a variable-length prefix-free code is used. In particular, the watch memory values ​​can be stored by variable-length codewords. Preferably, different watch memory values ​​are represented by different codewords. Preferably, the codewords form a code that is stored in the memory unit. Preferably, the code represents a sequence of watch memory values.

[0022] This advantageously reduces the required storage space in the memory unit. Instead of each stored value having a fixed length of, for example, 32 bits, the monitoring memory values ​​can be stored with a smaller number of bits in the memory unit.

[0023] One possible embodiment of the method involves monitoring the magnetic resonance measurement for compliance with the limit value using stored monitoring values, and stopping the magnetic resonance measurement if the limit value is exceeded. Advantageously, stopping the measurement can prevent a risk to the patient being examined, for example, from an excessively high SAR.

[0024] Stopping the magnetic resonance (MRI) measurement can include, in particular, a (permanent) termination of the MRI measurement. Stopping the MRI measurement can also include, in particular, a (temporary) pause in the MRI measurement. Stopping can also include, in particular, terminating the MRI measurement according to a current MRI sequence and resuming the MRI measurement according to a modified MRI sequence.

[0025] Monitoring the magnetic resonance measurement for compliance with the limit value based on the stored monitoring values ​​can, in particular, include determining an additive monitoring value, whereby determining the additive monitoring value involves adding the stored monitoring values. The magnetic resonance measurement is stopped, in particular, if the additive monitoring value exceeds the (predefined) limit value.

[0026] The addition monitoring memory value can be determined, for example, to calculate a, in particular a moving, mean value of the monitoring memory values ​​(and thus ultimately also of the monitoring measurements), e.g. according to equation (1) MW=AWm=1m∑k=1mSWk, where MW is the mean value, AW is the sum of monitoring memory values, and m is the number of summed monitoring memory values ​​SW. If the limit value is specified as a mean value, a mean value MW determined according to equation (1) can, for example, be compared with the specified limit value. Alternatively, in such a case, the sum of monitoring memory values ​​AW can of course also be compared with m times the specified limit value.

[0027] Advantageously, stopping the magnetic resonance measurement also limits the storage capacity of the memory unit required for monitoring the magnetic resonance measurement. Advantageously, the storage capacity of the memory unit corresponds to the predefined limit value.

[0028] Preferably, when adding the stored monitoring memory values, at most a number L of the most recently stored monitoring memory values ​​are added. Advantageously, L limits a (rolling) monitoring time window, for example a 10-second time window or a 6-minute time window.

[0029] Preferably, the monitoring measurements are acquired at constant time intervals Td, e.g., every 100 ms. Preferably, the monitoring time window Tc can be calculated using Tc = L x Td, where Td is the time interval between two successive monitoring measurements.

[0030] Preferably, the mean value according to equation (1) is formed from a maximum of L monitoring memory values, i.e., m ≤ L. For the case m <L ist insbesondere denkbar, dass der Mittelwert zusätzlich aus L-m Vorbelegungsspeicherwerten gebildet wird, die insbesondere den Wert Null aufweisen können.

[0031] One possible embodiment provides that the (new) mean value is calculated by adding a newly generated monitoring memory value to a previous mean value, and, in particular, if the previous mean value consisted of m <L Überwachungsspeicherwerten gebildet ist, von dem vorherigen Mittelwert der älteste Überwachungsspeicherwert subtrahiert wird. Vorzugsweise ist der vorherige Mittelwert ein aktueller Mittelwert. Vorzugsweise wurde der vorherige Mittelwert aus den zuletzt erzeugten Überwachungsspeicherwerten berechnet. Vorzugsweise wurde der vorherige Mittelwert aus Überwachungsspeicherwerten berechnet, die aus den zuletzt erfassten Überwachungsmesswerten erzeugt wurden.

[0032] When using the prefix-free code, the necessary storage size of the storage unit for monitoring the magnetic resonance measurement can be determined, in particular, as a function of the quantization levels N of the monitoring memory values ​​and the number L of the maximum number of monitoring memory values ​​to be added.

[0033] According to a further embodiment of the method, the method further comprises storing pre-assignment memory values ​​in the storage unit, wherein the addition of the stored monitoring memory values ​​includes an addition of K last stored monitoring memory values ​​and LK pre-assignment memory values.

[0034] The default memory values ​​can be considered, in particular, as fictitious and / or initial monitoring memory values. Preferably, the default memory values ​​are zero. Advantageously, the default memory values ​​simplify the handling of storing and / or adding the monitoring memory values.

[0035] The pre-stored values ​​can include, in particular, values ​​from a previous magnetic resonance imaging (MRI) scan. This is especially possible if the time since the end of the previous MRI scan is shorter than, for example, a monitoring window of 6 minutes.

[0036] Preferably, a maximum number M of monitoring memory values ​​is stored in the storage unit, each generated from the most recently acquired monitoring measurements. Preferably, the maximum number L of monitoring memory values ​​to be added is equal to the maximum number M of monitoring memory values ​​to be stored.

[0037] However, it is also conceivable that (initially) the available storage space is fully utilized (i.e., more than L monitoring memory values ​​are stored in the storage unit) and only when needed for the addition are superfluous and / or unnecessary monitoring memory values ​​removed from the storage unit, in particular deleted.

[0038] Preferably, only the monitoring memory values ​​are stored in the storage unit, which were each generated from the monitoring measurement values ​​that were recorded in a given, in particular immediately preceding, monitoring time window, especially most recently.

[0039] One possible embodiment of the method provides that a, in particular maximum, number M of monitoring memory values ​​is stored in the storage unit, each of which was generated from the monitoring measurement values ​​that were last recorded.

[0040] For example, only the number M of monitoring memory values ​​generated from the most recently acquired monitoring measurements are stored in the memory unit. Advantageously, the number M corresponds to the maximum number of monitoring memory values ​​required to calculate the cumulative monitoring memory value.

[0041] One possible embodiment of the method provides that the method further includes storing M pre-assignment memory values ​​in the storage unit, in particular as fictitious monitoring memory values.

[0042] Preferably, the pre-assignment memory values ​​are saved before the first of the monitoring memory values ​​is saved. Saving the pre-assignment memory values ​​can, for example, occur before the start of the magnetic resonance measurement, but it can also occur afterward.

[0043] Preferably, when a new monitoring memory value is stored, one of the pre-assigned memory values ​​is removed from the storage unit, in particular deleted, as long as at least one of the pre-assigned memory values ​​is still stored; otherwise, the oldest monitoring memory value is removed, in particular deleted. Preferably, the pre-assigned memory values ​​in the storage unit are successively replaced by the monitoring memory values.

[0044] Advantageously, the pre-stored memory values ​​supplement the monitored memory values ​​in such a way that the total number of pre-stored and monitored memory values ​​stored in the storage unit always corresponds to a time window to be monitored (monitoring time window). Advantageously, especially at the beginning of the magnetic resonance measurement, the monitoring memory values ​​covering the monitoring time window are temporarily filled by the pre-stored memory values ​​in the storage unit.

[0045] One possible embodiment of the method involves normalizing the monitoring measurements, particularly to the specified limit value, as part of generating the monitoring memory values. This advantageously simplifies the monitoring process. Furthermore, normalization avoids the need to adapt the assignment of codewords from the prefix-free code to (unnormalized) monitoring measurements; in particular, the assignment can be kept constant regardless of the specified limit value.

[0046] One possible embodiment of the method provides that the prefix-free code is a binary code, a Huffman code, or a Fibonacci code.

[0047] One possible embodiment of the method provides that the monitoring measurements describe a patient's specific absorption rate. Advantageously, medical image data, in particular magnetic resonance signals, are acquired during the magnetic resonance measurement. The medical image data are preferably suitable for generating an image of the patient.

[0048] Furthermore, a magnetic resonance device is proposed that is configured to perform a previously described method. The magnetic resonance device can, for example, comprise a directional coupler (particularly with subsequent digitization of the measured values) and / or a capacitive coupler as a unit for acquiring the monitoring measurements. The magnetic resonance device specifically includes a quantization unit for generating monitoring memory values ​​from the monitoring measurements, a storage unit for storing monitoring memory values ​​as a prefix-free code, and a monitoring unit for checking the magnetic resonance measurement for compliance with a (predefined) limit value based on the stored monitoring measurements.

[0049] Preferably, the quantization unit, the storage unit, and the monitoring unit are part of an embedded system. The embedded system can, for example, include a DSP chip (DSP: Digital Signal Processing) and / or an ARM SoC (SoC: system-on-chip) and / or an FPGA (Field Programmable Gate Array).

[0050] The advantages of the proposed magnetic resonance device essentially correspond to the advantages of the previously described method for monitoring a magnetic resonance measurement, which have been detailed below. Features, advantages, or alternative embodiments mentioned therein can also be applied to the other claimed items and vice versa.

[0051] Furthermore, a computer program product is proposed that comprises a program and is directly loadable into the memory of a programmable system control unit and / or an embedded system of a magnetic resonance device, and includes program resources, e.g., libraries and auxiliary functions, to execute a proposed method when the computer program product is executed in the system control unit and / or the embedded system of the magnetic resonance device. The computer program product may comprise software with source code that still needs to be compiled and bound or that only needs to be interpreted, or executable software code that only needs to be loaded into the system control unit for execution.

[0052] The proposed method can advantageously be executed quickly, identically, and robustly by the computer program product. The computer program product is preferably configured to execute the proposed method steps via the system control unit. The system control unit possesses the necessary prerequisites to ensure that the respective method steps can be executed efficiently.

[0053] The computer program product is stored, for example, on a computer-readable medium or on a network or server, from where it can be loaded into the processor of a local system control unit, which may be directly connected to the magnetic resonance device or be formed as part of the magnetic resonance device. Furthermore, control information of the computer program product may be stored on an electronically readable data carrier. The control information of the electronically readable data carrier may be designed such that, when the data carrier is used in a system control unit and / or an embedded system of a magnetic resonance device, it executes a proposed procedure.

[0054] Examples of electronically readable data carriers include a DVD, a magnetic tape, or a USB flash drive on which electronically readable control information, in particular software, is stored. If this control information is read from the data carrier and stored in a system control unit and / or an embedded system of the magnetic resonance device, all proposed embodiments of the methods described above can be carried out.

[0055] Further advantages, features, and details of the invention will become apparent from the exemplary embodiments described below and from the drawings. Corresponding parts are designated with the same reference numerals in all figures.

[0056] They show: Fig. 1. A magnetic resonance device in a schematic representation, Fig. 2 a diagram of a monitoring procedure in a first embodiment, Fig. 3 a diagram of a monitoring procedure in a second embodiment, Fig. 4 an exemplary table to illustrate the storage of monitoring memory values ​​as a prefix-free code according to the monitoring method of the second embodiment, Fig. 5 and Fig. 8 exemplary tables illustrating the quantization of monitoring measurement data and its assignment to a codeword of a prefix-free code, Fig. 6 a diagram of a monitoring procedure in a third embodiment, Fig. 7 An exemplary table illustrating the storage of monitoring memory values ​​as a prefix-free code according to the monitoring method of the third embodiment.

[0057] In Fig. Figure 1 schematically depicts a magnetic resonance imaging (MRI) device 10. The MRI device 10 comprises a magnetic unit 11, which includes a main magnet 12 for generating a strong and, in particular, time-constant main magnetic field 13. The MRI device 10 also includes a patient receiving area 14 for receiving a patient 15. In the present embodiment, the patient receiving area 14 is cylindrical and is cylindrically surrounded in one circumferential direction by the magnetic unit 11. However, a different configuration of the patient receiving area 14 is conceivable. The patient 15 can be moved into the patient receiving area 14 by means of a patient positioning device 16 of the MRI device 10. For this purpose, the patient positioning device 16 has a patient table 17 that is movably designed within the patient receiving area 14.

[0058] The magnet unit 11 further comprises a gradient coil unit 18 for generating magnetic field gradients, which are used for spatial encoding during imaging. The gradient coil unit 18 is controlled by a gradient control unit 19 of the magnetic resonance device 10. The magnet unit 11 also includes a high-frequency antenna unit 20, which in the present embodiment is designed as a body coil permanently integrated into the magnetic resonance device 10. The high-frequency antenna unit 20 is controlled by a high-frequency antenna control unit 21 of the magnetic resonance device 10 and emits high-frequency pulses into an examination space, which is essentially formed by a patient acquisition area 14 of the magnetic resonance device 10. This excites the atomic nuclei in the main magnetic field 13 generated by the main magnet 12.Magnetic resonance signals are generated by the relaxation of the excited atomic nuclei. The high-frequency antenna unit 20 is designed to receive these magnetic resonance signals.

[0059] The magnetic resonance imaging (MRI) device 10 includes a system control unit 22 for controlling the main magnet 12, the gradient control unit 19, and the high-frequency antenna control unit 21. The system control unit 22 centrally controls the MRI device 10, for example, by performing a predetermined imaging gradient echo sequence. The system control unit 22 also includes an evaluation unit (not shown) for evaluating the magnetic resonance signals acquired during the MRI scan. Furthermore, the MRI device 10 includes a user interface 23 connected to the system control unit 22. Control information, such as imaging parameters, and reconstructed MRI images can be displayed on a display unit 24, for example, on at least one monitor, via the user interface 23 for medical personnel.Furthermore, the user interface 23 has an input unit 25 by means of which information and / or parameters can be entered by the medical operating personnel during a measurement process.

[0060] The irradiated RF pulses are partially absorbed by the patient's tissue. The high-frequency energy absorbed per unit of time and per kilogram of body weight is commonly referred to as the specific absorption rate (SAR). The absorption of RF energy can lead to heating of the patient's body tissue. Energy absorption is an important factor in setting safety limits. At unacceptably high local concentrations of RF energy, RF burns can occur (local SAR). With a uniform distribution of RF energy throughout the entire body, the strain on the patient's thermoregulation and / or cardiovascular system is significant (whole-body SAR).

[0061] In Fig. Figure 2 describes a computer-implemented method for monitoring a magnetic resonance (MRI) measurement. Such monitoring can, in particular, include monitoring the patient's SAR exposure. In S10, an MRI measurement is started. During the MRI measurement, monitoring values ​​MW are acquired. In S30, monitoring memory values ​​are generated by quantizing the monitoring measurements. In S40, the monitoring memory values ​​SW are stored as a prefix-free code in a memory unit 26 of the MRI device 10. In S50, the MRI measurement is monitored for compliance with a (predefined) limit value based on the stored monitoring measurements.

[0062] In Fig. Figure 3 shows an extended version of a computer-implemented method for monitoring a magnetic resonance measurement. After acquiring the monitoring measurements MW in S20, the monitoring measurements are normalized in S25.

[0063] Monitoring the magnetic resonance measurement for compliance with the limit value in S50 includes determining an additive monitoring memory value in S51. Determining the additive monitoring memory value involves adding the stored monitoring memory values, whereby at most the L most recently stored monitoring memory values ​​are added. Based on the additive monitoring memory value, it is determined in S52 whether the limit value is exceeded. If the limit value is exceeded, the magnetic resonance measurement is stopped in S53. If the limit value is not exceeded, the magnetic resonance measurement continues, and further monitoring measurements are recorded and evaluated according to S20, S25, S30, S40, and S50.

[0064] Using a concrete example according to the Fig. 4 and Fig. Section 5 will explain possible aspects of the proposed procedure in more detail. According to S20, monitoring measurements MW are recorded at various times t. Preferably, these times t have constant time intervals. In this example, the times t are represented as dimensionless numbers 1, 2, ..., 10. These could, for example, be seconds.

[0065] Similarly, the monitoring measurements in MW are shown here as dimensionless numbers 260, 135, ..., 876. These monitoring measurements in MW preferably represent physical measurements such as measured power.

[0066] In this example, the limit is set so that the sum of the monitored measurements must not exceed 3000 MW within a rolling time window of five time units. This limit is equivalent to the mean of the monitored measurements within the same five-time window not exceeding 3000 / 5 = 600.

[0067] In S25, the monitoring measurement value MW is normalized to the limit value GW = 3000, and thus the normalized monitoring measurement value MW n generated; the monitoring measurement value MW is divided by the limit value GW. In S30, the normalized monitoring measurements MW are n Quantized in twenty steps; the quantization resolution is 5%. This is done in Fig. Figure 5 illustrates this. Each quantization level N is assigned a value interval INT with a width of 5%, meaning the quantization resolution is equal to the interval width. Quantization occurs in uniform steps, meaning the width of the value intervals INT is always the same for each quantization level N.

[0068] Each quantization level N is assigned a monitoring memory value SW and a corresponding codeword CW (sometimes also called code symbol). The monitoring memory value SW is the upper limit of the respective value interval INT. This results in the normalized monitoring measurement value MW being used in S30 during quantization. n (Conservatively) always rounds up. Advantageously, this makes monitoring the magnetic resonance measurement particularly reliable.

[0069] Instead of rounding up, it would also be conceivable to round down (truncate) and add the remainder to the next measurement. While this might result in slightly inaccurate timing, the sum would be correct except for a quantization inconsistency.

[0070] The codewords CW have the property of being codewords of a prefix-free code: A prefix-free code is defined as a code that satisfies the Fano condition, according to which no codeword of the code is a prefix of another codeword. Advantageously, the codewords of a prefix-free code can therefore be simply concatenated; nevertheless, the code remains uniquely decomposable and / or decodable into its codewords.

[0071] This example uses a binary code where the elements of the code symbols consist only of "0" or "1". The length of each codeword CW is equal to its quantization level N. It consists of N-1 zeros and a one at the end. However, it is also conceivable to use a different prefix-free code, such as a Huffman code or a Fibonacci code.

[0072] Now, if we apply the assignment table from Fig. 5 on the example according to Fig. 4, for example, the monitoring measurement value MW recorded and normalized at time t = 1 is used. n= 8.7% of the monitoring memory value SW = 10% and the codeword CW = 01 is assigned. The resulting code initially only includes the first codeword, i.e., 01. The additive monitoring memory value ASW determined in S51 results from the addition of the L = 5 most recently stored monitoring memory values ​​SW. At time t = 1, this only applies to the first stored monitoring memory value SW = 10%.

[0073] In the same way, the monitoring measurement value MW = 135 recorded in S20 at time t = 2 or the corresponding monitoring measurement value MW normalized in S25 is n= 135 / 3000 = 4.5%. After quantization in S30, the codeword CW = 1 is assigned. The watch memory values ​​for time points t = 1 and t = 2 are then stored in S40 as a code 101, i.e., a sequence of the respective codewords. The additive watch memory value ASW at time point t = 2 results from the addition of the watch memory values ​​SW 10% (for t = 1) and 5% (for t = 2), totaling 15%.

[0074] The monitoring in S50 shows that at both time t = 1 and t = 2, the respective addition monitoring memory values ​​ASW 10% and 15% do not exceed the (normalized) limit of 100%, so the magnetic resonance measurement is not stopped. Similarly, no limit is exceeded at the following time points up to t = 9. At time t = 10, the limit is exceeded because ASW = 110% > 100%. However, the code describing this does not need to be stored, as the magnetic resonance measurement is stopped in S53 when the limit is exceeded, thus eliminating the need for storage. Therefore, the code is shown in parentheses.

[0075] Since only the L = 5 most recently stored monitoring memory values ​​SW are used for the addition in S51, correspondingly older codewords CW can be removed from the code again from time t = 6 onwards. Preferably, a number M, in particular a maximum number, of monitoring memory values ​​is stored in the memory unit, each of which was generated from the most recently acquired monitoring measurements. Preferably, M = L.

[0076] Advantageously, the prefix-free code is a variable-length code: the code can have a different length depending on time t. For example, the length at time t = 1 is 2, at time t = 2 it is 3, at time t = 3 it is 4, at time t = 4 it is 9, and so on.

[0077] Advantageously, the code length corresponds to the addition monitoring memory value (ASW) at the respective time. The relationship is: Code length = Addition monitoring memory value (ASW) / quantization resolution. Since the addition monitoring memory value (ASW) can be a maximum of 100%, otherwise the measurement would stop, the memory size of the storage unit in this example can be limited to 20 bits.

[0078] Conventionally, each monitoring memory value would typically be stored as a 32-bit float, which, with five monitoring memory values, would require 5 x 32 bits = 160 bits of storage. For this demonstration example, this would result in a reduction factor of 8. In real-world applications, however, this reduction factor is usually significantly larger, as a much larger number of time points typically need to be considered. Monitoring a six-minute rolling time window with a time interval of 100 milliseconds results in 3600 values. Conventionally, this translates to a storage requirement of 3600 x 32 = 115200 bits.

[0079] With a practically useful quantization resolution of 0.5%, or the 200 quantization levels N required for this, the proposed method requires 200 bits of storage and thus a reduction factor of 576.

[0080] In Fig. Figure 7 shows another possible method. Here, pre-assignment memory values ​​VW are stored in memory unit 26 in S15. Advantageously, each time period of a monitoring window is assigned one of the pre-assignment memory values ​​VW. Preferably, the pre-assignment memory values ​​VW each have the value zero. Advantageously, the pre-assignment memory values ​​VW function as preliminary monitoring memory values ​​SW.

[0081] According to the table in Fig. 8 is assigned the codeword CW 1 to a monitoring memory value SW of zero. According to the example of Fig. 7 includes the monitoring window (as already shown in the example of Fig. 4 and Fig. 5) five time periods. Therefore, at time t = 0, the code consists of five codewords CW = 1, i.e., 11111.

[0082] At time t = 1, the codeword CW = 001 is determined for the recorded monitoring measurement value MW = 260 according to the previously described method. For storage of the code, one of the codewords CW = 1 then replaces the default memory value VW with this codeword CW = 001.

[0083] At time t = 2, a monitoring measurement value MW = 0 is recorded. This can occur, for example, if no RF pulse is being transmitted by the magnetic resonance device 10 at this time, meaning no RF power is being delivered to the patient 15. Consequently, a corresponding pre-assignment memory value is stored using the codeword CW = 1, and simultaneously a pre-assignment memory value VW with the codeword CW = 1 is removed, so that the code now only comprises three pre-assignment memory values ​​VW, each represented by a codeword CW = 1.

[0084] Advantageously, the pre-assigned memory values ​​in VW can simplify the addition of the memory values ​​in S51, because the number of summands is always the same, especially at the beginning.

[0085] As already explained, monitoring a six-minute rolling time window with a time interval of 100 milliseconds results in 3600 values, for which a conventional storage requirement of 115200 bits is necessary.

[0086] It has also already been explained that a quantization resolution of 0.5%, or the 200 quantization levels N required for this, results in a memory requirement of 200 bits; furthermore, according to the variant according to Fig.6. When using the pre-allocated memory values ​​VW, one additional bit of storage space is required for each time interval. This results in a total storage space of 200 + 3600 = 3800 bits. This represents a reduction of approximately a factor of 30 compared to conventional storage.

[0087] When applied to SAR monitoring, the proposed method makes particular use of the following properties: SAR values ​​have an upper and a lower limit: On the one hand, a SAR value can never be less than zero; on the other hand, it also cannot be greater than the specified limit. (The limit can, in particular, be normalized to one.) - The measurement accuracy of the monitoring values ​​is limited in practice. Therefore, it is sufficient to store the monitoring values ​​with only a certain accuracy (e.g., 0.5%; corresponding to 200 quantization levels). - In the (practically unlikely) event that the specified SAR limit is exceeded, the magnetic resonance measurement (and thus also the output of RF power) can be stopped. This reliably prevents a significant exceedance of the SAR limit.

[0088] These properties allow the SAR contributions within a (rolling) monitoring time window to be divided into discrete parts. These are advantageously stored using a prefix-free code.

[0089] Finally, it should be noted once again that the methods described in detail above, as well as the illustrated magnetic resonance device, are merely exemplary embodiments which can be modified in various ways by those skilled in the art without departing from the scope of the invention. Furthermore, the use of the indefinite articles "a" or "an" does not preclude the possibility that the features in question may be present multiple times. Likewise, the term "unit" does not preclude the possibility that the components in question consist of several interacting sub-components, which may also be spatially distributed. Regardless of the grammatical gender of a particular term, persons of male, female, or other gender identities are included.

Claims

[1] Computer-implemented method for monitoring a magnetic resonance measurement, the method comprising: - Providing monitoring measurements acquired during magnetic resonance imaging (MRI), - Generating monitoring memory values ​​by quantizing the monitoring measurements, - Storing the monitoring memory values ​​as a prefix-free code, especially of variable length, in a storage unit, - Monitoring the magnetic resonance measurement for compliance with a limit value based on the stored monitoring measurements, characterized by that the quantization of the monitoring measurements takes place depending on the limit value. [2] Method according to claim 1, wherein monitoring the magnetic resonance measurement for compliance with the predetermined limit value comprises using the stored monitoring measurement values: - Stopping the magnetic resonance measurement when the limit value is exceeded. [3] Method according to claim 1, where monitoring the magnetic resonance measurement for compliance with the limit value includes using the stored monitoring measurements: - Determining an addition monitoring memory value, which includes determining the addition monitoring memory value: Addition of the stored monitoring memory values, where at most a number L of recently stored monitoring memory values ​​are added, - in particular stopping the magnetic resonance measurement if the limit value is exceeded, especially if the addition monitoring memory value exceeds the limit value. [4] The method of claim 3, wherein the method further comprises: - Storing predefined memory values ​​in the storage unit, where the addition of the stored monitoring memory values ​​includes an addition of a number K of recently stored monitoring memory values ​​and a number LK of stored pre-assignment memory values. [5] Method according to any of the preceding claims, wherein at most a number M of monitoring memory values, each generated from the monitoring measurement values ​​that were last recorded, are stored in the storage unit. [6] The method of claim 5, wherein the method further comprises: - Storing M pre-assignment memory values ​​in the memory unit, wherein when a new watch memory value is stored, one of the pre-assignment memory values ​​is removed as long as at least one of the pre-assignment memory values ​​is still stored; otherwise, the oldest watch memory value is removed. [7] Method according to any of the preceding claims, wherein the monitoring measurements are monitoring measurements recorded at constant time intervals. [8] Method according to any of the preceding claims, wherein the monitoring measurements are quantized in N quantization levels, where each quantization level is assigned a value interval, where all monitoring measurements that lie within a respective value interval are assigned to a defined monitoring memory value. [9] Method according to one of the preceding claims, wherein the generation of the monitoring storage values ​​comprises normalizing the monitoring measurement values, in particular to the specified limit value. [10] Method according to any of the preceding claims, wherein the prefix-free code is a binary code, a Huffman code or a Fibonacci code. [11] Method according to any of the preceding claims, wherein the monitoring measurements describe a specific absorption rate of a patient. [12] Magnetic resonance device with a storage unit for storing monitoring memory values, which is configured to perform a method according to one of the preceding claims. [13] Computer program product comprising a program that can be directly loaded into a memory of a magnetic resonance device, comprising program means for executing a method according to any one of claims 1 to 11 when the program is executed by the magnetic resonance device.

Images