Magnetic resonance thermometry method, apparatus, electronic device, and storage medium

Abstract

This application relates to a magnetic resonance imaging (MRI) temperature measurement method, apparatus, electronic device, and storage medium. The method includes: acquiring MRI data, including a real-time temperature image; performing at least one of the following processes on the real-time temperature image: spatial smoothing, temporal smoothing, temperature zone-level judgment processing, and penalty processing, and obtaining at least one of the following processing results: spatial smoothing result, temporal smoothing result, temperature zone-level judgment result, and penalty result; and determining a target temperature image based on at least one processing result and / or the real-time temperature image. The solution provided by this application can improve the accuracy of MRI temperature measurement and provide reliable assurance for MRI-based surgical treatment.

Description

Technical Field

[0001] This application relates to the field of medical technology, and in particular to magnetic resonance temperature measurement methods, devices, electronic equipment and storage media. Background Technology

[0002] Magnetic resonance (MRI) signals typically contain both amplitude and phase terms. However, due to the difficulty in processing and displaying these terms, the phase term is often simply ignored. In the field of magnetic resonance thermometry, the raw data is used to reconstruct the phase image from the MRI signal. Magnetic resonance thermometry plays a crucial role in real-time temperature monitoring and control feedback in treatment scenarios such as laser ablation (LITT) and high-intensity focused ultrasound (HIFU). Therefore, accurate MRI phase reconstruction has significant medical implications.

[0003] In existing magnetic resonance thermometry techniques, the original temperature phase is typically deconvolved to the true relative temperature phase by assessing the temporal and spatial continuity of the signal. However, in practical applications, the phase difference between adjacent time points in magnetic resonance thermometry may exceed the range of [-π, π], resulting in a non-unique solution obtained from the deconvolution, thus affecting the accuracy of magnetic resonance thermometry. Summary of the Invention

[0004] To overcome the problems existing in related technologies, this application provides a magnetic resonance temperature measurement method, electronic device and storage medium. This magnetic resonance temperature measurement method can improve the accuracy of magnetic resonance temperature measurement and provide reliable protection for surgical treatment based on magnetic resonance temperature measurement.

[0005] The first aspect of this application provides a magnetic resonance temperature measurement method, comprising:

[0006] Acquire magnetic resonance data, including real-time temperature images;

[0007] Perform at least one of the following processing on the real-time temperature image: spatial smoothing, temporal smoothing, temperature zone-level judgment processing, and penalty processing, and obtain at least one of the following processing results: spatial smoothing result, temporal smoothing result, temperature zone-level judgment result, and penalty result;

[0008] The target temperature image is determined based on at least one processing result and / or a real-time temperature image.

[0009] In one implementation, the real-time temperature image is processed by at least one of the following: spatial smoothing, temporal smoothing, temperature zone-level judgment processing, and penalty processing, including:

[0010] Deconvolution processing is performed on real-time temperature images to obtain at least one preliminary temperature prediction image.

[0011] At least one of the following processes is applied to at least one preliminary temperature prediction image: spatial smoothing, temporal smoothing, temperature zone-level judgment processing, and penalty processing.

[0012] In one implementation, determining a target temperature image based on at least one processing result and / or a real-time temperature image includes:

[0013] The target temperature image is determined based on at least one processing result and / or a preliminary estimated temperature image.

[0014] In one implementation, determining a target temperature image based on at least one processing result and / or a preliminary estimated temperature image includes:

[0015] Construct a total loss function based on at least one processing result and / or a preliminary temperature prediction image;

[0016] The target temperature image is determined based on the result of the total loss function.

[0017] In one implementation, determining the target temperature image based on the result of the total loss function includes:

[0018] The preliminary estimated temperature image is iteratively updated based on the result of the total loss function until the change in the preliminary estimated temperature image reaches the preset requirement and / or the number of iterations reaches the preset number, thus obtaining the target temperature image.

[0019] In one implementation, a total loss function is constructed based on at least one processing result and / or a preliminary estimated temperature image, including:

[0020] For each preliminary estimated temperature image, a loss function is calculated based on at least one processing result and / or a preliminary estimated temperature image, and at least one of the following loss functions is obtained: spatial loss function, temporal loss function, temperature loss function, and penalty loss function.

[0021] Each preliminary temperature prediction image is used to construct a total loss function from at least one loss function.

[0022] In one implementation, a loss function is calculated for each preliminary estimated temperature image based on at least one processing result and / or a preliminary estimated temperature image, resulting in at least one of the following loss functions: a spatial loss function, a temporal loss function, a temperature loss function, and a penalty loss function, including:

[0023] Multiplying the spatial smoothing result by the first weighting coefficient yields the spatial loss function;

[0024] Multiplying the time smoothing result by the second weighting coefficient yields the time loss function;

[0025] Multiply at least one preliminary temperature prediction image by the third weighting coefficient to obtain at least one temperature loss function;

[0026] The penalty loss function is obtained by multiplying the temperature zone-level judgment result and / or penalty result by the fourth weighting coefficient.

[0027] In one implementation, determining the target temperature image based on the result of the total loss function includes:

[0028] Compare the results of multiple total loss functions;

[0029] The temperature image corresponding to the total loss function that minimizes the function result is determined as the target temperature image.

[0030] In one implementation, the preliminary estimated temperature image is iteratively updated based on the result of the total loss function, including:

[0031] The temperature image composed of the temperature values ​​of each pixel corresponding to the total loss function that minimizes the function result is used as the updated temperature image;

[0032] The process involves determining at least one updated predicted temperature image based on the updated temperature image, and then performing the step of constructing a total loss function based on at least one processing result and / or the preliminary predicted temperature image.

[0033] In one implementation, acquiring magnetic resonance data includes:

[0034] Acquire a baseline temperature image; the baseline temperature image includes at least one of the following: a baseline temperature phase image and a baseline temperature amplitude image;

[0035] Region of interest is determined based on baseline temperature images;

[0036] Real-time monitoring of magnetic resonance data in the region of interest.

[0037] In one implementation, determining the region of interest based on a baseline temperature image includes:

[0038] The region of interest is determined based on the pixel values ​​of each pixel in the base temperature phase image and / or base temperature amplitude image.

[0039] In one implementation, the real-time temperature image is deconvolved to obtain at least one preliminary estimated temperature image, including:

[0040] The real-time temperature image is deconvolved to obtain a preliminary temperature image. At least one preliminary estimated temperature image is determined based on the preliminary temperature image and a first preset bias coefficient.

[0041] In one implementation, determining at least one updated predicted temperature image based on the updated temperature image includes:

[0042] At least one updated predicted temperature image is determined based on the updated temperature image and the second preset bias coefficient.

[0043] In one implementation, deconvolution processing is performed on the real-time temperature image to obtain a preliminary temperature image, including:

[0044] The real-time temperature image is deconvolved into the sum of the differences between the preceding phase image and the real-time phase image;

[0045] The preceding phase image is the temperature image corresponding to one or more time points before the current time point;

[0046] The real-time temperature image difference is the temperature change corresponding to the phase difference obtained by performing a modulus operation on the preceding phase image and the real-time phase image.

[0047] In one implementation, spatial smoothing includes:

[0048] Spatially smooth at least one of the preliminary temperature image, the preliminary estimated temperature image, and the updated estimated temperature image using a spatial smoothing function; and

[0049] Time smoothing includes:

[0050] At least one of the preliminary temperature image, the preliminary estimated temperature image, and the updated estimated temperature image is time-smoothed using a one-dimensional time smoothing function.

[0051] A second aspect of this application provides a magnetic resonance temperature measurement device, comprising:

[0052] The data acquisition module is used to acquire magnetic resonance data, which includes real-time temperature images.

[0053] The data processing module is used to perform at least one of the following processes on the real-time temperature image: spatial smoothing, temporal smoothing, temperature zone-level judgment processing, and penalty processing, and obtain at least one of the following processing results: spatial smoothing result, temporal smoothing result, temperature zone-level judgment result, and penalty result;

[0054] An output module is used to determine a target temperature image based on at least one of the processing results and / or the real-time temperature image.

[0055] A third aspect of this application provides an electronic device, comprising:

[0056] Processor; and

[0057] A memory that stores executable code, which, when executed by the processor, causes the processor to perform the method described above.

[0058] A fourth aspect of this application provides a non-transitory machine-readable storage medium having executable code stored thereon, which, when executed by a processor of an electronic device, causes the processor to perform the method described above.

[0059] The technical solution provided in this application may include the following beneficial effects:

[0060] By acquiring magnetic resonance data, including real-time temperature images, and performing at least one of the following processing methods on the real-time temperature images: spatial smoothing, temporal smoothing, temperature zone-level judgment processing, and penalty processing, at least one of the following processing results is obtained: spatial smoothing result, temporal smoothing result, temperature zone-level judgment result, and penalty result. Then, based on at least one processing result and / or the real-time temperature image, a target temperature image is determined. This effectively handles noise generated by high temperatures, achieves the goal of fusing the joint relationship between time and space, avoids phase drift caused by changes in background pixels in a single region, and considers both spatial and temporal continuity to obtain the overall optimal solution. This improves the accuracy of magnetic resonance thermometry and provides reliable assurance for surgical treatment based on magnetic resonance thermometry.

[0061] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0062] The above and other objects, features, and advantages of exemplary embodiments of this application will become readily understood by reading the following detailed description with reference to the accompanying drawings. Several embodiments of this application are illustrated in the drawings by way of example and not limitation, and the same or corresponding reference numerals denote the same or corresponding parts.

[0063] Figure 1 This is one of the schematic flowcharts of the magnetic resonance temperature measurement method shown in the embodiments of this application;

[0064] Figure 2 This is a second schematic flowchart of the magnetic resonance temperature measurement method shown in the embodiments of this application;

[0065] Figure 3 This is the third schematic flowchart of the magnetic resonance temperature measurement method shown in the embodiments of this application;

[0066] Figure 4 This is a schematic diagram of the structure of the magnetic resonance temperature measuring device shown in the embodiments of this application;

[0067] Figure 5This is a schematic diagram of the structure of an electronic device shown in an embodiment of this application. Detailed Implementation

[0068] Embodiments will now be described with reference to the accompanying drawings. It should be understood that, for the sake of simplicity and clarity, reference numerals may be repeated in the drawings to indicate corresponding or similar elements where deemed appropriate. Furthermore, numerous specific details are set forth herein to provide a thorough understanding of the embodiments described herein. However, those skilled in the art will understand that the embodiments described herein can be practiced without these specific details. In other instances, well-known methods, processes, and components have not been described in detail so as not to obscure the embodiments described herein. Moreover, this description should not be construed as limiting the scope of the embodiments described herein.

[0069] In existing magnetic resonance thermometry techniques, the original temperature phase is typically deconvoluted to obtain the true relative temperature phase by assessing the temporal and spatial continuity of the signal. However, in practical applications, the phase difference between adjacent time points in magnetic resonance thermometry may exceed the range of [-π, π] and be mapped to the range of [-π, π]. This results in a non-unique solution obtained from the deconvolution. Furthermore, magnetic resonance phase acquisition is subject to significant noise due to the magnetic resonance equipment, coils, and the environment, thus affecting the accuracy of magnetic resonance thermometry. Moreover, current magnetic resonance thermometry techniques do not consider the impact of high temperatures on the object under test and the magnetic resonance data, nor do they consider the influence of the actual environment in which the tissue under test is located on the temperature, further impacting the accuracy of the magnetic resonance thermometry data.

[0070] To address the aforementioned issues, this application provides a magnetic resonance temperature measurement method that can improve the accuracy of magnetic resonance temperature measurement and provide reliable support for surgical treatment based on magnetic resonance temperature measurement.

[0071] The technical solutions of the embodiments of this application are described in detail below with reference to the accompanying drawings.

[0072] Figure 1 This is one of the schematic flowcharts of the magnetic resonance temperature measurement method shown in the embodiments of this application.

[0073] Please see Figure 1 The magnetic resonance temperature measurement method shown in the embodiments of this application may include:

[0074] In step 101, magnetic resonance data is acquired.

[0075] In this application embodiment, magnetic resonance data includes, but is not limited to, real-time temperature images. A real-time temperature image refers to an image that reflects temperature, obtained through real-time measurement using a magnetic resonance device.

[0076] In step 102, the real-time temperature image is processed to obtain at least one processing result.

[0077] In this embodiment of the application, the real-time temperature image needs to be processed by at least one of the following: spatial smoothing, temporal smoothing, temperature zone-level judgment, and penalty processing, and at least one of the following processing results is obtained: spatial smoothing result, temporal smoothing result, temperature zone-level judgment result, and penalty result.

[0078] Since real-time temperature images are magnetic resonance images, which are formed based on magnetic resonance signals using spatial phase coding technology, each acquired signal in the magnetic resonance image has a corresponding spatial location signal and a temporal signal. Therefore, real-time temperature images also contain spatial location signals and temporal signals. Consequently, spatial smoothing and temporal smoothing processing are required for real-time temperature images to effectively handle noise and background drift caused by high temperatures, achieving the goal of fusing the joint relationship between time and space. Adding temperature zone-level judgment processing and penalty processing can also avoid phase drift caused by changes in background pixels in a single region and avoid the problem of inaccurate temperature measurement caused by sudden high temperatures.

[0079] In step 103, a target temperature image is determined based on at least one processing result and / or a real-time temperature image.

[0080] Since the actual phase of a real-time temperature image may exceed [-π,π], it is necessary to determine a target temperature image with the true phase based on at least one processing result and / or a real-time temperature image, thereby improving the accuracy of nuclear magnetic resonance thermometry.

[0081] By acquiring magnetic resonance data, including real-time temperature images, and performing at least one of the following processing methods on the real-time temperature images: spatial smoothing, temporal smoothing, temperature zone-level judgment processing, and penalty processing, at least one of the following processing results is obtained: spatial smoothing result, temporal smoothing result, temperature zone-level judgment result, and penalty result. Then, based on at least one processing result and / or the real-time temperature image, a target temperature image is determined. This effectively handles noise generated by high temperatures, achieves the goal of fusing the joint relationship between time and space, avoids phase drift caused by changes in background pixels in a single region, and considers both spatial and temporal continuity to obtain the overall optimal solution. This improves the accuracy of magnetic resonance thermometry and provides reliable assurance for surgical treatment based on magnetic resonance thermometry.

[0082] In some embodiments, the real-time temperature image is first deconvolved to obtain at least one preliminary estimated temperature image. Then, the at least one preliminary estimated temperature image is processed to obtain at least one processing result. Finally, a loss function is constructed to determine the target temperature image. Figure 2This is the second schematic flowchart of the magnetic resonance temperature measurement method shown in the embodiments of this application. Please refer to [link / reference]. Figure 2 The magnetic resonance temperature measurement method shown in the embodiments of this application may include:

[0083] In step 201, magnetic resonance data is acquired.

[0084] In this application embodiment, magnetic resonance data includes, but is not limited to, real-time temperature images. A real-time temperature image refers to an image that reflects temperature, obtained through real-time measurement using a magnetic resonance device.

[0085] In step 202, deconvolution processing is performed based on the real-time temperature image to obtain at least one preliminary estimated temperature image.

[0086] Deconvolution is the same as anticonvolution, which can be viewed as a signal recovery process. Since the phase of a real-time temperature image may exceed the range of [-π, π], and the phase may contain noise, in this embodiment, it is necessary to deconvolve the phase of the real-time temperature image to determine the preliminary temperature image.

[0087] In this embodiment, preferably, steps 201 and 202 can be executed simultaneously, that is, the acquired real-time temperature image is deconvolved in real time, and deconvolution is performed while monitoring, thereby improving processing efficiency. This depends on the actual application and is not a unique limitation here.

[0088] Specifically, the real-time temperature image is deconvolved into the sum of the differences between the preceding phase image and the real-time phase image, thereby forming a preliminary temperature image. Wherein, the preceding phase image is the temperature image corresponding to one or more previous time points at the current time point, and the real-time temperature image difference is the temperature change corresponding to the phase difference obtained by performing a modulo operation based on the preceding phase image and the real-time phase image. For example, the deconvolution process can be represented by the following formula (1):

[0089] Formula (1)

[0090] in, This is the preceding phase image. This represents the difference between real-time temperature images; mod is the modulo operation. Ensure that it is within the range of [-π, π].

[0091] Further, at least one preliminary estimated temperature image is determined based on the preliminary temperature image and the first preset bias coefficient. Specifically, in the embodiments of this application, at least one preliminary estimated temperature image can be, for example, a first preliminary estimated temperature image, a second preliminary estimated temperature image, and a third preliminary estimated temperature image, and can be determined for example using the following formula (2):

[0092] Formula (2):

[0093] in, This is the first preliminary temperature forecast image. This is a second preliminary temperature forecast image. The third preliminary temperature prediction image is given, k is the first preset bias coefficient, and k is a natural number. In the embodiments of this application, k can preferably be set to 1, 2, 3 or other natural numbers, depending on the actual application. There is no unique limitation here, and the number of preliminary temperature prediction images can exceed 3. This application does not limit the number of them.

[0094] Understandably, since k is a natural number, the phase of the initial temperature image, after adding or subtracting an even multiple of 2kπ, does not change. However, due to spatial and temporal continuity within the preset temperature measurement range, the actual temperature changes. This allows for the prediction of other convolution temperatures that may exist under the same phase.

[0095] In this step, the possible temperature conditions of the tissue being tested are fully considered in order to obtain more comprehensive temperature data and improve the accuracy of magnetic resonance thermometry.

[0096] In step 203, at least one preliminary estimated temperature image is processed, and at least one processing result is obtained after processing each preliminary estimated temperature image.

[0097] At least one of the following processes is applied to at least one preliminary estimated temperature image: spatial smoothing, temporal smoothing, temperature zone-level judgment, and penalty processing. Spatial smoothing can be performed by applying a spatial smoothing function to at least one of the preliminary temperature image, the preliminary estimated temperature image, and the updated estimated temperature image. The spatial location signal and time signal can be represented by I(x,y,t), where x and y represent spatial distribution coordinates and t represents time. The relative temperature phase parameter is spatially smoothed using a two-dimensional Gaussian function. Specifically, I(x,y,t) can be convolved with a two-dimensional Gaussian function, as expressed by the following formula (3):

[0098] Formula (3):

[0099] Where G(σ, D) is a two-dimensional Gaussian function of size D, σ is the standard deviation, x = [-D, D], y = [-D, D], and z is the normalization factor.

[0100] It is understood that the above method of spatially smoothing at least one of the preliminary temperature image, the preliminary estimated temperature image, and the updated estimated temperature image using a two-dimensional Gaussian function is merely an example. In practical applications, there are various methods for spatial smoothing, such as Gaussian, median, and average, which need to be selected according to the actual application situation. No single limitation is made here.

[0101] Furthermore, time smoothing can be performed on at least one of the preliminary temperature image, the preliminary estimated temperature image, and the updated estimated temperature image using a one-dimensional time smoothing function. Specifically, time smoothing can be performed using quadratic spline curves or linear regression prediction methods, resulting in a fourth-order smoothed curve. It is understood that in practical applications, various time smoothing methods are used, such as Gaussian, median, and average smoothing, and the choice should be based on the specific application; no single method is specified here.

[0102] The temperature zone-level judgment and penalty processing can be specifically as follows: First, a temperature zone-level penalty function is set based on a preset temperature measurement range. In this embodiment, the preset temperature measurement range can be set to 30℃ to 90℃, or it can be 37℃ to 60℃, 45℃ to 90℃, 45℃ to 85℃, or 45℃ to 95℃. In practical applications, the preset temperature measurement range can also be set to other temperature ranges, depending on the actual application situation, and is not limited here.

[0103] Assuming the preset temperature measurement range is 30℃ to 90℃, the temperature zone-level penalty function can be set as follows (4) to limit temperature values ​​exceeding the range through the temperature zone-level penalty function:

[0104] Formula (4):

[0105] In formula (4), the temperature range determination of Tem can be regarded as a temperature zone level determination process, while 0, (30℃-Tem) 2 And (Tem-90℃) 2 This corresponds to the penalty terms for different temperature zones. Tem is the real-time temperature value after direct phase conversion of the real-time temperature image. The real-time temperature phase conversion to the real-time temperature value can be exemplarily performed using the following formula (5):

[0106] Formula (5):

[0107] in, The phase of the real-time temperature image is represented by r, a, and B0, which are all preset constants.

[0108] It is understood that the temperature zone-level penalty function and phase temperature transformation method described above are merely examples. In practical applications, the temperature zone-level penalty function and phase temperature transformation method need to be determined according to the actual application situation. There is no single limitation here.

[0109] Through this step, this application can limit the temperature of the tissue to be tested (or the region of interest for temperature measurement) to its actual possible temperature range as much as possible. That is, this application fully considers the actual temperature limitations of the tissue to be tested in real scenarios, so that magnetic resonance thermometry can more accurately obtain results with actual physical or physiological significance, including temperature or changes in blood flow velocity caused by temperature changes. The temperature measurement data will be closer to the real situation, and the temperature measurement data obtained through the solution of this application is more accurate.

[0110] In step 204, a target temperature image is determined based on at least one processing result and / or a preliminary estimated temperature image.

[0111] First, a total loss function is constructed based on at least one processing result and / or a preliminary temperature prediction image. Specifically:

[0112] First, based on at least one processing result and / or a preliminary estimated temperature image, a loss function is calculated for each preliminary estimated temperature image, resulting in at least one of the following loss functions: spatial loss function, temporal loss function, temperature loss function, and penalty loss function. Specifically, the spatial loss function is obtained by multiplying the spatial smoothing result by a first weighting coefficient; the temporal loss function is obtained by multiplying the temporal smoothing result by a second weighting coefficient; at least one temperature loss function is obtained by multiplying each preliminary estimated temperature image by a third weighting coefficient; and the penalty loss function is obtained by multiplying the temperature zone-level judgment result and / or penalty result by a fourth weighting coefficient.

[0113] Then, the total loss function is constructed by taking at least one loss function corresponding to each preliminary estimated temperature image. For example, assume that at least one preliminary estimated temperature image is from step 202. as well as The total loss function can then be determined using the following formula (6):

[0114] Formula (6):

[0115]

[0116] Where L0 is The corresponding total loss function, L -2kπ for The corresponding total loss function, L 2kπ for The corresponding total loss function, S(x,y,t), represents the spatial smoothing result, and T(x,y,t) represents the temporal smoothing result. x and y constitute the spatial distribution coordinates of each monitoring point, t represents time, and n1, n2, n3, and n4 are the first, second, third, and fourth weighting coefficients, respectively. n1, n2, n3, and n4 are all values ​​greater than or equal to 0. The values ​​of n1, n2, n3, and n4 depend on the actual application and are not uniquely defined here. In practical applications, if the influence of a certain factor does not need to be considered, such as time loss, the second weighting coefficient corresponding to the time loss factor can be set to 0. Similarly, if spatial loss does not need to be considered, the first weighting coefficient corresponding to the spatial loss factor can also be set to 0. The specific values ​​depend on the actual application and are not uniquely defined here. The determination of whether time loss, space loss, and / or temperature penalty should be considered can be based on the number of iterations, whether the temperature difference between consecutively obtained values ​​does not exceed a certain threshold, or image parameters such as grayscale values ​​and average values. For example, the number of iterations should not exceed 3, the temperature difference should not exceed 0.1℃, and the grayscale value or average value should not exceed a certain threshold or threshold range. Based on this, this application can obtain accurate temperature data in a shorter time. Compared with the prior art, this application has the characteristics of real-time dynamic judgment and rapid and accurate temperature measurement.

[0117] Understandably, in practical applications, if the difference or average variance between multiple monitored real-time temperature phases is within a preset range, the monitored real-time temperature phase can be considered relatively accurate. Therefore, the first, second, and third preliminary temperature prediction images can all use the monitored real-time temperature phase to improve data processing speed. However, this depends on the specific application and is not a single, definitive rule.

[0118] It is also understandable that in practical applications, the loss function can be constructed in various ways, depending on the specific application. For example, relevant biological, physical, and / or spatiotemporal loss function terms can be flexibly added according to the specific application. There is no single limitation here.

[0119] Next, the target temperature image is determined based on the result of the total loss function. Specifically, the results of multiple total loss functions are compared, and the temperature image composed of the pixel temperature values ​​corresponding to the total loss function with the smallest result is used as the updated temperature image (for example, the updated temperature image corresponding to the total loss function with the smallest result is...). Then, based on the updated temperature image, at least one updated predicted temperature image is determined (e.g., the updated predicted temperature image is...). as well as k' is the second preset bias coefficient, and k' is a natural number. In the embodiments of this application, k' can preferably be set to 1, 2, 3 or other natural numbers, depending on the actual application situation (not limited here), and the step of constructing a total loss function based on at least one processing result and / or updated estimated temperature image is performed.

[0120] In some embodiments, at least one loss function corresponding to each updated predicted temperature image is used to construct the total loss function. For example, suppose at least one updated predicted temperature image is as described above. as well as The total loss function for the new round of updates can then be determined using the following formula (7):

[0121] Formula (7):

[0122]

[0123] Where L′0 is The corresponding total loss function, L′ -2kπ for The corresponding total loss function, L′ 2kπ for The corresponding total loss function is as follows: S′(x,y,t) is the spatial smoothing result of the updated predicted temperature image, T′(x,y,t) is the temporal smoothing result of the updated predicted temperature image, and P′(Tem) is the temperature zone-level penalty function of the updated predicted temperature image. x and y constitute the spatial distribution coordinates of each monitoring point, t is time, and n1, n2, n3 and n4 are the first weight coefficient, the second weight coefficient, the third weight coefficient and the fourth weight coefficient, respectively. n1, n2, n3 and n4 are all values ​​greater than or equal to 0. The values ​​of n1, n2, n3 and n4 are determined according to the actual application and are not uniquely limited here.

[0124] In this method, at least one updated predicted temperature image can be determined based on the updated temperature image and a second preset bias coefficient. The second preset bias coefficient can be set to be the same as or different from the first preset bias coefficient, but it must be a natural number. The method for determining the updated predicted temperature image can be the same as the method for determining at least one preliminary predicted temperature image, and is not limited to a single method. In some embodiments, by comparing the function results of multiple total loss functions, the total loss function with the smallest function result is obtained, and the temperature image corresponding to the smallest total loss function is determined as the target temperature image. This method can quickly obtain temperature measurement data in a short time.

[0125] In other embodiments, the preliminary estimated temperature image and / or the updated estimated temperature image are iteratively updated based on the result of the total loss function until the change range of the preliminary estimated temperature image and / or the updated estimated temperature image reaches a preset requirement, and / or the number of iterations reaches a preset number, to obtain the target temperature image. The preset requirement may be that the phase of two adjacent updated temperatures no longer changes, or that the change range is within a preset range, and is not uniquely limited. Thus, while limiting the actual temperature measurement range and achieving the goal of restricting temporal and spatial continuity, a spatiotemporally smooth and temperature-reasonable optimal solution is obtained. This step can be repeated through multiple rounds of iterative updates to obtain a temperature more closely resembling reality, greatly improving the accuracy of magnetic resonance thermometry.

[0126] In some embodiments, such as in the application scenario of Laser Interstitial Therapy (LITT), LITT utilizes multimodal image fusion analysis for precise preoperative planning. Stereotactic technology is used to accurately position the laser head into the lesion tissue, and magnetic resonance imaging (MRI) and temperature measurement are employed to perform laser thermotherapy on the lesion tissue under real-time video monitoring. During the procedure, laser energy is transmitted to the lesion tissue via a treatment fiber for precise irradiation, achieving photothermal conversion. This causes the temperature of the target lesion area to rise, leading to thermal coagulation and denaturation of the lesion tissue, thus achieving the therapeutic goal. Therefore, accurately obtaining MRI temperature data is extremely important. Furthermore, since LITT uses a laser as its energy source, there will be a temperature surge during the ablation process. For patient safety, a safe operating temperature range is often set (e.g., LITT may set temperature ranges of 20℃-100℃, 30℃-100℃, 37℃-90℃, or 45℃-95℃). It is understood that cryoablation therapy, which also requires temperature control, also sets certain temperature ranges, and the method described in this application is applicable to similar operations and processes.

[0127] During treatment, factors such as background temperature drift, inaccurate temperature measurement, and noise from the MRI equipment itself can affect MRI data. Furthermore, in actual clinical practice, temperature measurement of the treatment area is subject to the following limitations: First, to ensure operational safety, temperature measurements are limited to specific ranges, such as 20℃-100℃, 30℃-100℃, 37℃-90℃, or 45℃-95℃. Second, sudden temperature spikes or abrupt changes can lead to inaccurate measurements. Third, within the aforementioned temperature ranges, temperature changes occur continuously both spatially and temporally as the treatment progresses. Therefore, obtaining more accurate and reliable MRI temperature data quickly is crucial.

[0128] Figure 3This is the third schematic flowchart of the magnetic resonance temperature measurement method shown in the embodiments of this application. Please refer to [link / reference]. Figure 3 The magnetic resonance temperature measurement method shown in the embodiments of this application may include:

[0129] In step 301, baseline temperature parameters are acquired. These parameters include at least one of the following: baseline temperature phase and baseline temperature amplitude. The temperature phase reflects the angle traversed by different protons during the relaxation process of temperature changes. Temperature amplitude reflects the magnitude of temperature change during a process of temperature variation. Base temperature phase, on the other hand, is the initial temperature phase when the temperature has not changed, and base temperature amplitude is the initial temperature magnitude when the temperature has not changed.

[0130] In step 302, the region of interest for temperature measurement is determined based on the baseline temperature parameters.

[0131] The region of interest for thermometry can be understood as the area where thermometry is planned or the area where the temperature may change, i.e., the area where the phase of the temperature may differ from the background temperature, and / or the area where the amplitude of the temperature may differ from the background temperature. For example, suppose a base temperature is increased at the site to be ablated. After the base temperature is increased, the temperature at the site to be ablated will be relatively higher than other sites that do not require ablation, thus enabling the identification of the site to be ablated as the region of interest for thermometry.

[0132] Specifically, a base temperature phase map is formed based on the base temperature phase; and / or a base temperature amplitude structure map is formed based on the base temperature amplitude. Both the phase map and the amplitude structure map belong to magnetic resonance imaging. The phase map is a magnetic resonance image that uses the phase data differences of different protons obtained by magnetic resonance imaging to form an image for image comparison, while the amplitude structure map can be formed by converting the temperature amplitude into pixel values.

[0133] The region of interest (ROI) for temperature measurement is determined based on the pixel values ​​of each pixel in the base temperature phase map and / or base temperature amplitude structure map. Specifically, the boundary of the ROI can be determined based on the pixel values ​​of each pixel in the image, i.e., by calculating the difference in pixel values ​​between adjacent pixels. In practical applications, there are various ways to determine the ROI, and the appropriate method should be determined according to the specific application. No single method is limited here.

[0134] In step 303, magnetic resonance data is monitored in the region of interest for thermometry.

[0135] In this embodiment, real-time temperature images are monitored within the region of interest (ROI). Since treatment, such as ablation, is required within the ROI, the temperature at the ablation site needs to be adjusted to a specific value during the ablation process. Therefore, close monitoring of the real-time temperature at the ablation site is necessary to prevent tissue damage due to inappropriate temperature values. Since the raw data from magnetic resonance thermometry is obtained by reconstructing phase images from magnetic resonance signals, this embodiment employs real-time temperature image monitoring to assess the real-time temperature.

[0136] In step 304, the real-time temperature phase difference is determined based on the real-time temperature image, and the phase deconvolution of the real-time temperature phase difference is performed to obtain at least one preliminary estimated temperature image.

[0137] In this embodiment, the real-time temperature phase difference can be deconvolved into the sum of the preceding temperature phase difference parameter and the phase difference parameter. The preceding temperature phase difference parameter is the temperature phase difference corresponding to the previous one or several time points of the current time point, and the phase difference parameter is the parameter obtained by performing a modulus operation based on the preceding temperature phase difference parameter and the real-time temperature phase difference.

[0138] After deconvolution, a preliminary temperature image is obtained. Further, at least one preliminary estimated temperature image is determined based on the preliminary temperature image and a first preset bias coefficient.

[0139] In step 305, at least one preliminary estimated temperature image is processed, and at least one processing result is obtained after processing each preliminary estimated temperature image.

[0140] In this embodiment of the application, the content of step 305 is the same as that of step 203, and will not be described again here.

[0141] In step 306, a target temperature image is determined based on at least one processing result and / or a preliminary estimated temperature image.

[0142] In this embodiment of the application, the content of step 306 is the same as that of step 204, and will not be described again here.

[0143] This application, taking into full account real-world clinical scenarios, incorporates temporal and spatial constraints, as well as a more reasonable phase range constraint, into its temperature measurement method. (That is, based on the actual temperature range, it fully considers the possible phase or temperature values ​​of each pixel at the current temperature and penalizes any potentially occurring phase or temperature values.) Preferably, it iteratively adjusts the phase or temperature value of each pixel and updates the phase or temperature values ​​of surrounding pixels in real time to obtain more realistic temperature measurement data and more accurate results. Compared to existing technologies, this application improves the robustness of the algorithm by using potentially occurring phase or temperature values ​​to constrain at least one of the preliminary temperature image, the preliminary estimated temperature image, and the updated estimated temperature image. In actual clinical practice, this ensures that the temperature of the treatment area (or the region of interest for temperature measurement) is within a realistic and reasonable temperature range, while maintaining good spatiotemporal continuity and avoiding inaccurate temperature measurements caused by temperature spikes. Furthermore, this application uses the prediction of loss function results to further guarantee the accuracy, effectiveness, realism, and rationality of the temperature measurement.

[0144] Corresponding to the aforementioned application function implementation method embodiments, this application also provides a magnetic resonance temperature measurement device. Please refer to [link to relevant documentation]. Figure 4 The magnetic resonance temperature measuring device shown in this application may include:

[0145] Data acquisition module 401 is used to acquire magnetic resonance data, including real-time temperature images;

[0146] The data processing module 402 is used to perform at least one of the following processes on the real-time temperature image: spatial smoothing, temporal smoothing, temperature zone-level judgment processing, and penalty processing, and obtain at least one of the following processing results: spatial smoothing result, temporal smoothing result, temperature zone-level judgment result, and penalty result;

[0147] Output module 403 is used to determine a target temperature image based on at least one of the processing results and / or the real-time temperature image.

[0148] Corresponding to the aforementioned application function implementation method embodiments, this application also provides an electronic device for performing a magnetic resonance temperature measurement method and corresponding embodiments.

[0149] Figure 5 A block diagram illustrating the hardware configuration of an electronic device 800 capable of implementing the magnetic resonance temperature measurement method of the embodiments of this application is shown. Figure 5 As shown, the electronic device 800 may include a processor 810 and a memory 820. Figure 5In the electronic device 800, only the components relevant to this embodiment are shown. Therefore, it will be apparent to those skilled in the art that the electronic device 800 may also include components related to... Figure 5 The following are common components with different constituent elements. For example, a fixed-point arithmetic unit.

[0150] Electronic device 800 can correspond to a computing device with various processing functions, such as functions for generating neural networks, training or learning neural networks, quantizing floating-point neural networks into fixed-point neural networks, or retraining neural networks. For example, electronic device 800 can be implemented as various types of devices, such as personal computers (PCs), server devices, mobile devices, etc.

[0151] The processor 810 controls all functions of the electronic device 800. For example, the processor 810 controls all functions of the electronic device 800 by executing programs stored in the memory 820 on the electronic device 800. The processor 810 can be implemented by a central processing unit (CPU), graphics processing unit (GPU), application processor (AP), artificial intelligence processor chip (IPU), etc., provided in the electronic device 800. However, this application is not limited to this.

[0152] In some embodiments, processor 810 may include an input / output (I / O) unit 811 and a computing unit 812. I / O unit 811 may be used to receive various types of data, such as magnetic resonance data. Exemplarily, computing unit 812 may be used to process, for example, a real-time temperature image received via I / O unit 811, at least one of the following: spatial smoothing, temporal smoothing, temperature zone-level judgment processing, and penalty processing, to obtain at least one of the following processing results: spatial smoothing result, temporal smoothing result, temperature zone-level judgment result, and penalty result, and then determine a target temperature image based on at least one of the processing results and / or the real-time temperature image. This target temperature image may, for example, be output by I / O unit 811. The output data may be provided to memory 820 for use by other devices (not shown), or may be directly provided to other devices.

[0153] Memory 820 is hardware used to store various data processed in electronic device 800. For example, memory 820 can store processed data and data to be processed in electronic device 800. Memory 820 can store data involved in the magnetic resonance temperature measurement method that has been processed or is to be processed by processor 810. In addition, memory 820 can store applications, drivers, etc. to be driven by electronic device 800. For example, memory 820 can store various programs related to the magnetic resonance temperature measurement method to be executed by processor 810. Memory 820 can be DRAM, but this application is not limited to it. Memory 820 can include at least one of volatile memory or non-volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), flash memory, phase-change RAM (PRAM), magnetic RAM (MRAM), resistive RAM (RRAM), ferroelectric RAM (FRAM), etc. Volatile memory may include dynamic RAM (DRAM), static RAM (SRAM), synchronous DRAM (SDRAM), PRAM, MRAM, RRAM, ferroelectric RAM (FeRAM), etc. In an embodiment, memory 820 may include at least one of hard disk drive (HDD), solid-state drive (SSD), high-density flash memory (CF), secure digital card (SD), micro-secure digital card (Micro-SD), mini-secure digital card (Mini-SD), extreme digital card (xD), cache, or memory stick.

[0154] In summary, the specific functions implemented by the memory 820 and processor 810 of the electronic device 800 provided in the embodiments of this specification can be explained in comparison with the foregoing embodiments in this specification, and can achieve the technical effects of the foregoing embodiments. Therefore, they will not be repeated here.

[0155] In this embodiment, the processor 810 can be implemented in any suitable manner. For example, the processor 810 can take the form of a microprocessor or processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) that can be executed by the (micro)processor, logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers, and embedded microcontrollers, etc.

[0156] It should be understood that the possible terms "first" or "second," etc., in the claims, specification, and drawings disclosed in this application are used to distinguish different objects, rather than to describe a specific order. The terms "comprising" and "including" used in the specification and claims disclosed in this application indicate the presence of the described features, integrals, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components, and / or collections thereof.

[0157] It should also be understood that the terminology used in this disclosure is for the purpose of describing particular embodiments only and is not intended to limit the scope of this disclosure. As used in this disclosure and claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in this disclosure and claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes such combinations.

[0158] Although the embodiments of this application are described above, the content is merely an example adopted for the purpose of facilitating understanding of this application and is not intended to limit the scope and application scenarios of this application. Any person skilled in the art described in this application may make any modifications and changes in the form and details of the implementation without departing from the spirit and scope disclosed in this application, but the scope of patent protection of this application shall still be determined by the scope defined in the appended claims.

[0159] It should also be understood that any module, unit, component, server, computer, terminal, or device that executes the instructions executorized herein may include or otherwise access computer-readable media, such as storage media, computer storage media, or data storage devices (removable) and / or non-removable) such as disks, optical discs, or magnetic tapes. Computer storage media may include volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information, such as computer-readable instructions, data structures, program modules, or other data.

Claims

1. A magnetic resonance temperature measurement method, characterized in that, include: Acquire magnetic resonance data, including real-time temperature images; Based on the real-time temperature image, deconvolution processing is performed to obtain at least one preliminary estimated temperature image; The following processing is performed on at least one preliminary temperature prediction image: spatial smoothing, temporal smoothing, temperature zone-level judgment processing, and penalty processing, and the following processing results are obtained: spatial smoothing result, temporal smoothing result, temperature zone-level judgment result, and penalty result. Based on the spatial smoothing results, a loss function is calculated to obtain a spatial loss function; based on the temporal smoothing results, a loss function is calculated to obtain a temporal loss function; based on the at least one preliminary estimated temperature image, a loss function is calculated to obtain at least one temperature loss function; based on the temperature zone-level judgment results and penalty results, a loss function is calculated to obtain a penalty loss function; the spatial loss function, temporal loss function, temperature loss function, and penalty loss function corresponding to each preliminary estimated temperature image are respectively constructed into a total loss function; The target temperature image is determined based on the result of the total loss function.

2. The magnetic resonance temperature measurement method according to claim 1, characterized in that, Determining the target temperature image based on the result of the total loss function includes: The preliminary estimated temperature image is iteratively updated based on the result of the total loss function until the change range of the preliminary estimated temperature image reaches a preset requirement and / or the number of iterations reaches a preset number, thereby obtaining the target temperature image.

3. The magnetic resonance temperature measurement method according to claim 1, characterized in that, Further includes: The spatial smoothing result is multiplied by the first weighting coefficient to obtain the spatial loss function; The time smoothing result is multiplied by the second weighting coefficient to obtain the time loss function; At least one preliminary temperature prediction image is multiplied by the third weighting coefficient to obtain at least one temperature loss function; The penalty loss function is obtained by multiplying the temperature zone-level judgment result and the penalty result by the fourth weighting coefficient.

4. The magnetic resonance temperature measurement method according to claim 2, characterized in that, Determining the target temperature image based on the result of the total loss function includes: Compare the results of multiple total loss functions; The temperature image corresponding to the total loss function that minimizes the function result is determined as the target temperature image.

5. The magnetic resonance temperature measurement method according to claim 4, characterized in that, The preliminary estimated temperature image is iteratively updated based on the result of the total loss function, including: The temperature image composed of the temperature values ​​of each pixel corresponding to the total loss function that minimizes the function result is used as the updated temperature image; Based on the updated temperature image, at least one updated predicted temperature image is determined, and the step of constructing a total loss function is performed.

6. The magnetic resonance temperature measurement method according to any one of claims 1 to 5, characterized in that, The acquisition of magnetic resonance data includes: Acquire a baseline temperature image; the baseline temperature image includes at least one of the following: a baseline temperature phase image and a baseline temperature amplitude image; The region of interest is determined based on the baseline temperature image; The magnetic resonance data is monitored in real time in the region of interest.

7. The magnetic resonance temperature measurement method according to claim 6, characterized in that, The step of determining the region of interest based on the baseline temperature image includes: The region of interest is determined based on the pixel values ​​of each pixel in the base temperature phase image and / or the base temperature amplitude image.

8. The magnetic resonance temperature measurement method according to claim 7, characterized in that, The deconvolution process based on the real-time temperature image to obtain at least one preliminary estimated temperature image includes: The real-time temperature image is deconvolved to obtain a preliminary temperature image, and at least one preliminary estimated temperature image is determined based on the preliminary temperature image and a first preset bias coefficient.

9. The magnetic resonance temperature measurement method according to claim 5, characterized in that, Determining at least one updated predicted temperature image based on the updated temperature image includes: At least one updated estimated temperature image is determined based on the updated temperature image and the second preset bias coefficient.

10. The magnetic resonance temperature measurement method according to claim 8, characterized in that, The real-time temperature image is deconvolved to obtain a preliminary temperature image, including: The real-time temperature image is deconvolved into the sum of the differences between the preceding phase image and the real-time phase image; The preceding phase image is the temperature image corresponding to one or more time points before the current time point; The real-time phase image difference is the temperature change corresponding to the phase difference obtained by performing a modulus operation on the real-time phase image and the preceding phase image.

11. The magnetic resonance temperature measurement method according to claim 8, characterized in that, The spatial smoothing process includes: At least one of the preliminary temperature image, the preliminary estimated temperature image, and the updated estimated temperature image is spatially smoothed using a spatial smoothing function; and The time smoothing process includes: At least one of the preliminary temperature image, the preliminary estimated temperature image, and the updated estimated temperature image is time-smoothed using a one-dimensional time smoothing function.

12. A magnetic resonance temperature measuring device, characterized in that, include: The data acquisition module is used to acquire magnetic resonance data, which includes real-time temperature images. The data processing module is used to perform deconvolution processing on the real-time temperature image to obtain at least one preliminary estimated temperature image; perform the following processing on the at least one preliminary estimated temperature image: spatial smoothing, temporal smoothing, temperature zone-level judgment processing, and penalty processing, and obtain the following processing results: spatial smoothing result, temporal smoothing result, temperature zone-level judgment result, and penalty result; calculate a loss function based on the spatial smoothing result to obtain a spatial loss function; calculate a loss function based on the temporal smoothing result to obtain a temporal loss function; calculate a loss function based on each of the at least one preliminary estimated temperature image to obtain at least one temperature loss function; calculate a loss function based on the temperature zone-level judgment result and the penalty result to obtain a penalty loss function; and construct a total loss function from the spatial loss function, temporal loss function, temperature loss function, and penalty loss function corresponding to each preliminary estimated temperature image. The output module is used to determine the target temperature image based on the result of the total loss function.

13. An electronic device, characterized in that, include: processor; as well as A memory having executable code stored thereon, which, when executed by the processor, causes the processor to perform the method as described in any one of claims 1-11.

14. A non-transitory machine-readable storage medium having executable code stored thereon, which, when executed by a processor of an electronic device, causes the processor to perform the method as described in any one of claims 1-11.

Images