Liver cancer immunotherapy curative effect evaluation method based on medical image

Abstract

The invention discloses a liver cancer immunotherapy curative effect evaluation method based on a medical image, and relates to the technical field of medical data processing, and the method comprises the steps: obtaining historical liver image data and current liver image data of a target patient stored in a medical system; performing image segmentation processing on the historical liver image data to obtain a liver cancer focus area before immunotherapy, and segmenting the current liver image data into a plurality of liver tissue areas; determining curative effect information indexes of different liver tissue areas based on the treatment curative effect characteristics of the liver tissue areas; based on the curative effect information index, selecting a curative effect comparison area from each liver tissue area; and based on the difference characteristics between the curative effect comparison area and the liver cancer focus area, generating a curative effect evaluation result of immunotherapy. The technical effect of accurately evaluating the immunotherapy effect is achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of medical data processing, and particularly relates to a liver cancer immunotherapy efficacy evaluation method based on medical images. BACKGROUND

[0002] In recent years, with the rapid development of tumor immunotherapy, the survival prognosis of some patients with advanced liver cancer has been significantly improved; however, the clinical response of liver cancer immunotherapy has high heterogeneity, and for some patients, the treatment may be ineffective or even immune-related adverse reactions may occur; therefore, how to timely and accurately evaluate the effect of immunotherapy to guide individualized treatment strategy adjustment in the subsequent process has become a key problem to be solved in the current field of liver cancer immunotherapy.

[0003] At present, the clinical main method is to judge the treatment response according to the change of the maximum diameter of the tumor lesion by using the modified solid tumor response evaluation standard for liver cancer; however, immunotherapy is different from chemotherapy and targeted therapy, and has atypical imaging performance such as'super progression', and the imaging performance of liver cancer is complex, the lesion boundary is fuzzy, and is easily interfered by liver background lesions, so when the efficacy comparison region of liver cancer lesion after immunotherapy is selected, the subjectivity of doctors is relied on, which leads to the inability to accurately evaluate the effect of immunotherapy. SUMMARY

[0004] The main purpose of the present application is to provide a liver cancer immunotherapy efficacy evaluation method, device, equipment and storage medium based on medical images, which aims to solve the technical problem that in the related art, it is difficult to select the efficacy comparison region of liver cancer lesion after immunotherapy in the efficacy evaluation process, the subjectivity of selecting the efficacy comparison region is strong, and the effect of immunotherapy cannot be accurately evaluated.

[0005] To achieve the above-mentioned purpose, the embodiment of the present application provides a liver cancer immunotherapy efficacy evaluation method based on medical images, which comprises: Obtaining the historical liver image data and the current liver image data of the target patient stored by the medical system; Performing image segmentation processing on the historical liver image data to obtain the liver cancer lesion region before immunotherapy, and segmenting the current liver image data into a plurality of liver tissue regions; Based on the treatment efficacy characteristics of each liver tissue region, the efficacy information index of different liver tissue regions is determined; Based on the efficacy information index, the efficacy comparison region is selected from each liver tissue region; Based on the difference characteristics between the efficacy comparison region and the liver cancer lesion region, the efficacy evaluation result of immunotherapy is generated.

[0006] In a possible implementation of the present application, the historical liver image data is subjected to image segmentation processing to obtain a liver cancer lesion region before immunotherapy, including: The historical liver image data is subjected to three-dimensional reconstruction processing to obtain first three-dimensional image data; The first three-dimensional image data is subjected to image segmentation by a preset deep learning model to obtain the liver cancer lesion region before immunotherapy and a liver region.

[0007] In a possible implementation of the present application, the current liver image data is segmented into a plurality of liver tissue regions, including: The current liver image data is subjected to three-dimensional reconstruction processing to obtain second three-dimensional image data; Based on the signal intensity features of each voxel point in the second three-dimensional image data, a curative effect evaluation feature vector of each voxel point is constructed; The two-dimensional data composed of the curative effect evaluation feature vector and the coordinates of each voxel point are subjected to clustering processing by a preset clustering algorithm to obtain the plurality of liver tissue regions.

[0008] In a possible implementation of the present application, after the current liver image data is subjected to three-dimensional reconstruction processing to obtain second three-dimensional image data, the method further includes: The liver region is taken as a rigid registration point, and the second three-dimensional image data and the liver region are subjected to multi-sequence spatial alignment processing to obtain the second three-dimensional image data after alignment.

[0009] In a possible implementation of the present application, based on the signal intensity features of each voxel point in the second three-dimensional image data, a curative effect evaluation feature vector of each voxel point is constructed, including: The voxel point region in the liver region except the liver cancer lesion region is set as a general liver tissue region; For any sequence in the voxel point, the absolute value of the difference between the signal intensity mean values between the liver cancer lesion region and the general liver tissue region in the current sequence is calculated to obtain a signal intensity mean difference; Based on the volume of the liver cancer lesion region in the current sequence and the absolute value of the difference between the volume mean values of the liver cancer lesion regions in all sequences, a volume mean difference is calculated; The ratio of the number of sequences belonging to the general liver tissue region in each voxel point to the number of all sequences is taken as a lesion morphology interference index of each voxel point; The index mean value of each lesion morphology interference index is calculated, and the index mean value is multiplied by the volume mean difference to obtain a morphology weight influence index of the current sequence; Based on the ratio between the signal intensity mean difference and the morphology weight influence index, a curative effect reference weight of each sequence is calculated. Based on the efficacy reference weight and the signal intensity feature of each voxel point in the second three-dimensional image data, an efficacy evaluation feature vector of each voxel point is constructed.

[0010] In a possible implementation of the present application, based on the treatment efficacy features of each liver tissue region, the efficacy information index of different liver tissue regions is determined, including: Based on the edema features of each liver tissue region and the signal intensity change degree of multiple sequences in each liver tissue region before and after immunotherapy, a liver cancer lesion information index is calculated; For any liver tissue region, the difference between the standard deviations of the signal intensity of each voxel point in the current liver tissue region before and after immunotherapy is calculated, and a standard deviation difference value is obtained. Based on the standard deviation difference value and the liver cancer lesion information index, the efficacy information index of different liver tissue regions is calculated.

[0011] In a possible implementation of the present application, based on the edema features of each liver tissue region and the signal intensity change degree of multiple sequences in each liver tissue region before and after immunotherapy, a liver cancer lesion information index is calculated, including: For any liver tissue region, the average of the proportion of voxel points containing the corresponding sequence of the liver cancer lesion region in the current liver tissue region and the first difference value of the signal intensity of each sequence in the voxel point before immunotherapy minus the signal intensity of the corresponding sequence after immunotherapy are calculated, and the average of each first difference value is calculated to obtain a first average difference value; The apparent diffusion coefficient and the diffusion weighted imaging signal intensity ratio of each voxel point in the liver tissue region adjacent to the current liver tissue region are calculated as the edema feature index, and the first average of the edema feature indexes of all voxel points is calculated. The adjacent liver tissue region corresponding to the maximum first average value is selected as the edema feature reference region of the current liver tissue region, and the first difference average value of the edema feature index of the voxel points in the edema feature reference region before and after immunotherapy is calculated. Based on the average proportion, the first average difference value and the first difference average value, the liver cancer lesion information index is calculated.

[0012] In a possible implementation of the present application, based on the efficacy information index, an efficacy comparison region is selected from each liver tissue region, including: The liver tissue region in which the voxel point and the liver cancer lesion region before immunotherapy have a coincident region is set as a target liver tissue region; The first distance between the centroid coordinates of each target liver tissue region and the centroid coordinates of the liver cancer lesion region is calculated. Based on the ratio of the efficacy information index of each target liver tissue region to the corresponding first distance, an efficacy comparison coefficient is obtained. Sort the therapeutic effect comparison coefficients of each target liver tissue region in descending order, and select the adjacent target liver tissue regions with the largest difference in therapeutic effect comparison coefficients; The target liver tissue region with a therapeutic effect comparison coefficient greater than or equal to the corresponding maximum therapeutic effect comparison coefficient in the adjacent target liver tissue region is taken as the therapeutic effect comparison region.

[0013] In a possible implementation of the present application, the difference features between the therapeutic effect comparison region and the liver cancer lesion region are used to generate the efficacy evaluation result of the immunotherapy, including: The multi-sequence signal intensity and the standard deviation of the multi-sequence signal intensity of the therapeutic effect comparison region and the liver cancer lesion region are calculated respectively to obtain the difference features; According to the change trend of the difference features, the efficacy evaluation result of the immunotherapy is generated.

[0014] In a possible implementation of the present application, the difference features between the therapeutic effect comparison region and the liver cancer lesion region are used to generate the efficacy evaluation result of the immunotherapy, including: If the multi-sequence signal intensity and the standard deviation of the multi-sequence signal intensity both show a decreasing trend, it is determined that the treatment effect of the current immunotherapy meets the expected standard; If the multi-sequence signal intensity shows an increasing trend, it is determined that the treatment effect of the current immunotherapy is lower than the expected standard.

[0015] The present application provides a liver cancer immunotherapy efficacy evaluation method based on medical images. In the related art, it is difficult to select the therapeutic effect comparison region of the liver cancer lesion after immunotherapy in the efficacy evaluation process, and the selection of the therapeutic effect comparison region is highly subjective, which leads to inaccurate evaluation of the immunotherapy effect. In the present application, the historical liver image data and the current liver image data of a target patient are obtained, and the historical liver image data and the current liver image data are respectively subjected to image segmentation to obtain the liver cancer lesion region before immunotherapy and a plurality of liver tissue regions after immunotherapy. Then, the therapeutic effect information index of different liver tissue regions is calculated according to the treatment efficacy features of each liver tissue region, and the therapeutic effect comparison region of the immunotherapy is selected through the therapeutic effect information index, thereby avoiding the inaccuracy caused by subjective selection. Furthermore, the difference features between the therapeutic effect comparison region and the liver cancer lesion region are used to generate the efficacy evaluation result of the immunotherapy, and the treatment effect of the immunotherapy is accurately evaluated through the efficacy evaluation result. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 The flowchart of the first embodiment of the liver cancer immunotherapy efficacy evaluation method based on medical images of the present application is shown; Figure 2 The flowchart of the second embodiment of the liver cancer immunotherapy efficacy evaluation method based on medical images of the present application is shown; Figure 3 A device structure schematic diagram of a hardware running environment involved in an embodiment of the present application. DETAILED DESCRIPTION

[0017] It should be understood that the specific embodiments described herein are merely intended to explain the present application and are not intended to limit the present application.

[0018] An embodiment of the present application provides a liver cancer immunotherapy efficacy evaluation method based on medical images. All steps of the method can be executed by a computing device such as a computer or a mobile terminal. In a first embodiment of the liver cancer immunotherapy efficacy evaluation method based on medical images, referring to Figure 1 , the method comprises the following steps. Step S10, acquiring historical liver image data and current liver image data of a target patient stored by a medical system; Step S20, performing image segmentation processing on the historical liver image data to obtain a liver cancer lesion region before immunotherapy, and segmenting the current liver image data into a plurality of liver tissue regions; Step S30, determining efficacy information indexes of different liver tissue regions based on treatment efficacy features of the liver tissue regions; Step S40, selecting an efficacy comparison region from the liver tissue regions based on the efficacy information indexes; Step S50, generating an efficacy evaluation result of the immunotherapy based on difference features between the efficacy comparison region and the liver cancer lesion region.

[0019] The embodiment aims to: select an efficacy comparison region of the immunotherapy through the efficacy information indexes, avoid inaccurate selection of the efficacy comparison region, generate the efficacy evaluation result of the immunotherapy according to the difference features between the efficacy comparison region and the liver cancer lesion region, and accurately evaluate the treatment effect of the immunotherapy through the efficacy evaluation result.

[0020] The specific steps are as follows: Step S10, acquiring historical liver image data and current liver image data of a target patient stored by a medical system; As an example, the liver cancer immunotherapy efficacy evaluation method based on medical images can be applied to a liver cancer immunotherapy efficacy evaluation device based on medical images. The liver cancer immunotherapy efficacy evaluation device based on medical images belongs to a liver cancer immunotherapy efficacy evaluation system based on medical images. The liver cancer immunotherapy efficacy evaluation system based on medical images belongs to a liver cancer immunotherapy efficacy evaluation device based on medical images.

[0021] As an example, the target patient can be a liver cancer patient after immunotherapy, wherein the historical liver image data is liver image data of the patient before immunotherapy, the current liver image data is liver image data of the patient after immunotherapy, and the liver image data can be CT image data or MRI (magnetic resonance imaging) image data.

[0022] As an example, when collecting image data of the patient, multi-sequence and multi-phase image data are collected by a high-field 3.0T MRI (magnetic resonance imaging) scanning system. When scanning, the patient adopts an inhalation breath-hold technique to reduce motion artifacts. After obtaining the medical image data, the image data is preprocessed by non-local mean filtering and noise reduction to reduce the interference of noise on the image. All image data is saved in the medical system in DICOM (Digital Imaging and Communications in Medicine) format, records the original image parameters, contrast agent dose and delay time information, and ensures the uniformity of subsequent analysis standards. When evaluating the effect of immunotherapy, the historical liver image data and the current liver image data of the target patient are extracted from the medical system.

[0023] Step S20, image segmentation processing is performed on the historical liver image data to obtain a liver cancer lesion area before immunotherapy, and the current liver image data is segmented into a plurality of liver tissue areas.

[0024] As an example, immunotherapy is different from chemotherapy / targeting, and its effectiveness is not only reflected in the change of tumor volume, but also more reflected in the heterogeneity of tumor tissue, blood perfusion mode, necrosis range, and imaging performance changes caused by immune cell infiltration. In order to further improve the accuracy of liver cancer in the evaluation of immunotherapy, the accuracy of liver analysis can be improved by three-dimensional reconstruction of the historical liver image data and the current liver image data.

[0025] As an example, after three-dimensional reconstruction of the two kinds of image data, they are respectively segmented into a liver cancer lesion area before immunotherapy and a plurality of liver tissue areas after immunotherapy, so as to perform a subsequent treatment comparison process and evaluate the actual effect of immunotherapy.

[0026] As an example, the liver cancer lesion area is a specific image area of the target patient where cancer occurs, and the liver tissue area is an image area of liver tissue formed after immunotherapy of the cancerous area.

[0027] The step S20 of performing image segmentation processing on the historical liver image data to obtain the liver cancer lesion area before immunotherapy further includes steps S21-S22, which include: Step S21, performing three-dimensional reconstruction processing on the historical liver image data to obtain first three-dimensional image data.

[0028] As an example, the first three-dimensional image data is three-dimensional body data obtained by three-dimensional conversion of historical liver image data. Before three-dimensional conversion, preprocessing is required, and DICOM format MRI images are imported into medical image processing software (such as 3D Slicer); linear normalization is performed on the signal intensity difference between different patients and different sequences to unify the signal value distribution in the 0-255 interval; all MRI images are uniformly resampled to a voxel size of 1x1x1 mm3 to ensure consistency in subsequent three-dimensional spatial registration and feature extraction.

[0029] As an example, the preprocessed two-dimensional MRI sequence is reconstructed into MRI three-dimensional body data and saved in NIfTI format (.nii) for subsequent multi-modal analysis and texture feature extraction. Three-dimensional reconstruction is a prior art and will not be described here.

[0030] In step S22, the first three-dimensional image data is segmented by a preset deep learning model to obtain a liver cancer lesion area before immunotherapy and a liver area.

[0031] As an example, after obtaining the first three-dimensional image data, the liver cancer lesion area needs to be located and segmented. Subsequently, the liver cancer characteristics in the liver cancer lesion area before and after immunotherapy are analyzed, and the liver cancer lesion area can be located and segmented in the three-dimensional body data by a deep learning model. The training process of the deep learning model is as follows: A large amount of three-dimensional liver MRI data is obtained using the Internet, wherein the sequences in the MRI data include T1WI (T1 weighted image) arterial phase, portal phase, delay phase, T2WI (T2 weighted image), and DWI (diffusion weighted imaging).

[0032] The liver area and liver cancer lesion area in the three-dimensional body data are labeled by an artificial labeling method, and the data is divided into a training set (70%), a validation set (15%), and a test set (15%).

[0033] A 3D U-Net neural network structure is adopted, wherein: the encoder is composed of 5 layers of 3D convolution layers, batch normalization layers, and ReLU activation functions, each layer has a convolution kernel size of 3x3x3, a step of 2, and a down-sampling feature map; the decoder up-samples the encoded features and jump-connects with the corresponding encoded layer feature map to restore the spatial resolution of the segmentation mask; and the loss function and the cross-entropy loss are combined as the loss function.

[0034] The training process is performed by three-dimensional data enhancement in random rotation, translation, scaling, noise, etc. to alleviate overfitting and improve model robustness.

[0035] The model training adopts an Adam optimizer to stabilize and accelerate the gradient update process, and the initial learning rate is set to 1x10-4 And automatically decrease by decay coefficient 0.95 every 10 cycles in the training process to adapt to the loss function convergence rate change; Considering that the memory occupation of three-dimensional medical image data is large, the GPU memory resource is limited, and the batch size is set to 2.

[0036] In the training process, the Dice similarity coefficient change on the validation set is continuously monitored. If the Dice similarity coefficient of the validation set does not obviously improve in the last 10 cycles, the early stopping mechanism is automatically triggered to terminate the training to prevent overfitting.

[0037] After the model training is completed, the segmentation performance of the model is comprehensively evaluated on an independent test set, and the evaluation indexes include the Dice similarity coefficient, the Jaccard coefficient, the sensitivity and the specificity.

[0038] After the model training is completed, the current patient's pre-immunotherapy input model is obtained, and the liver region and liver cancer lesion region in the first three-dimensional image data are obtained.

[0039] The step S20 of segmenting the current liver image data into a plurality of liver tissue regions further includes steps S210-S230: Step S210, performing three-dimensional reconstruction processing on the current liver image data to obtain second three-dimensional image data; As an example, the second three-dimensional image data can be three-dimensional volume data obtained by three-dimensional conversion of the current liver image data, and the three-dimensional reconstruction manner is the same as that of the first three-dimensional image data.

[0040] After the step of performing three-dimensional reconstruction processing on the current liver image data to obtain the second three-dimensional image data, the following steps are further included: Taking the liver region as a rigid registration point, the second three-dimensional image data and the liver region are subjected to multi-sequence spatial alignment processing to obtain the second three-dimensional image data after alignment.

[0041] As an example, after the patient's MRI sequences of multiple sequences and multiple phases after immunotherapy are collected, the three-dimensional data of these sequences are converted, and the three-dimensional volume data of all MRI sequences before and after immunotherapy need to be registered. Since the core mechanism of systemic immunotherapy for liver cancer does not depend on physical liver resection or local destruction, it will not directly lead to changes in liver macroscopic morphology or anatomical structure, so the liver region in the MRI image before immunotherapy can be taken as a rigid registration point to align the overall anatomical structure of the liver, so as to reduce the influence of the change of lesion morphology, volume, enhancement mode and tissue composition after immunotherapy on the registration process.

[0042] Specifically, the arterial phase T1WI before immunotherapy is taken as a main sequence, and the liver region is taken as a full liver mask of the sequence; multi-sequence spatial initial alignment is performed by rigid registration, local deformation correction is realized by B-spline non-rigid registration, and other sequences are registered to the coordinate system of the main sequence by taking the boundary of the full liver mask as a marker point to drive registration.

[0043] All images are resampled to 1x1x1 mm3 voxel size, intensity standardization is performed by Z-score method, three-dimensional body data registration after immunotherapy is completed, and the voxel points of the sequence images in the registration result are aligned with the main sequence of the arterial phase T1WI before immunotherapy. The same voxel point position corresponds to the same anatomical position, and then the second three-dimensional image data after alignment of the sequence is obtained.

[0044] In step S220, a curative effect evaluation feature vector of each voxel point is constructed based on the signal intensity features of each voxel point in the second three-dimensional image data.

[0045] As an example, in the MRI imaging performance of liver cancer, different stages of the lesion show typical but complex signal feature changes on multi-sequence MRI. These signals reflect the pathological basis changes inside the lesion. Due to the heterogeneity of the internal structure of the lesion, the reaction of different regions after immunotherapy is inconsistent, and there is microscopic spatial heterogeneity inside a single lesion region. Therefore, the macroscopic mean texture cannot be reflected, and therefore, it is necessary to quantitatively calculate the curative effect evaluation feature vector of each voxel point on different sequences and different spatial positions, which is used to describe the lesion performance characteristics of each voxel point and serves as an important basis for subsequent curative effect response evaluation and partitioning.

[0046] In step S220, a curative effect evaluation feature vector of each voxel point is constructed based on the signal intensity features of each voxel point in the second three-dimensional image data. The voxel point region in the liver region except the liver cancer lesion region is set as a general liver tissue region.

[0047] As an example, the voxel point region in the liver region except the liver cancer lesion region is a non-liver cancer lesion region, i.e., a general liver tissue region.

[0048] For any sequence in the voxel point, the absolute value of the difference between the mean signal intensity of the liver cancer lesion region and the general liver tissue region in the current sequence is calculated to obtain the mean signal intensity difference.

[0049] As an example, for each voxel point, a signal combination of multiple sequences is included, which can include signals corresponding to the liver cancer lesion area and signals corresponding to the general liver tissue area, the signal intensity difference between different sequences in the liver cancer lesion area and the general liver tissue area of the non-liver cancer lesion area is different, the greater the difference, the greater the degree to which the current sequence can distinguish the liver cancer lesion area and the general liver tissue, therefore, the absolute value of the difference between the signal intensity mean values of the liver cancer lesion area and the general liver tissue area in the current sequence is calculated to obtain the signal intensity mean difference, denoted as The greater the value, the greater the difference between the signal intensity of the general liver tissue area and the signal intensity of the liver cancer lesion area in the current sequence.

[0050] Based on the absolute value of the difference between the volume of the liver cancer lesion area in the current sequence and the volume mean value of the liver cancer lesion area of all sequences, the volume mean difference is calculated.

[0051] As an example, the volume mean difference is used to represent the volume size difference of the liver cancer lesion area in the current sequence compared with other sequences, when the volume mean difference is greater, it means that the volume size difference of the liver cancer lesion area in the current sequence compared with other sequences is greater, and the volume mean difference is denoted as .

[0052] The ratio of the number of sequences belonging to the general liver tissue area in each voxel point to the number of all sequences is taken as the lesion morphology interference index of each voxel point.

[0053] As an example, the lesion morphology interference index is used to indicate the probability of the current voxel point being affected by artifacts under the sequence, the greater the lesion morphology interference index, the higher the probability of the current voxel point being affected by artifacts under the sequence.

[0054] The index mean value of each lesion morphology interference index is calculated, and the index mean value is multiplied by the volume mean difference to obtain the morphology weight influence index of the current sequence.

[0055] As an example, the difference in morphology of the liver cancer lesion area segmented under different sequences can reflect the tissue resolution tendency of the current sequence in the efficacy evaluation, if the morphology of the liver cancer lesion area segmented under a certain sequence always deviates from other sequences, the sequence is less effective or has a greater probability of artifact interference in the tumor in this case, therefore, the product of the index mean value of each lesion morphology interference index and the volume mean difference is used to determine the morphology weight influence index of the current sequence.

[0056] Based on the ratio between the signal intensity mean difference and the morphology weight influence index, the efficacy reference weight of each sequence is calculated.

[0057] As an example, in different MRI sequences, the signal intensity of liver lesions shows differences, so it is necessary to analyze the lesion weight represented by different sequences, which should reflect the discriminant contribution of the sequence in the differentiation of different tissue components in the liver; therefore, by analyzing the difference in signal intensity between the liver lesion area before immunotherapy and the general tissue in different sequences, the reference weight of each sequence can be used as a reference when constructing the efficacy evaluation feature vector of the voxel point after immunotherapy.

[0058] As an example, after calculating the morphological weight influence index, the reference weight of each sequence is calculated by the ratio between the signal intensity mean difference and the morphological weight influence index The greater the , the higher the weight of the current sequence in the efficacy evaluation feature vector.

[0059] Based on the efficacy reference weight and the signal intensity feature of each voxel point in the second three-dimensional image data, the efficacy evaluation feature vector of each voxel point is constructed.

[0060] As an example, for any voxel point after immunotherapy in three-dimensional space, the efficacy evaluation feature vector of the liver MRI three-dimensional data after immunotherapy is constructed by combining the efficacy reference weight, and the vector is composed of the product of the signal intensity of the current voxel point in different sequences and the corresponding efficacy reference weight.

[0061] Step S230, by a preset clustering algorithm, the two-dimensional data composed of the efficacy evaluation feature vector and the coordinates of each voxel point are clustered to obtain a plurality of liver tissue regions.

[0062] As an example, the preset clustering algorithm can be a DBSCAN clustering algorithm (a density-based spatial clustering algorithm), and by the DBSCAN clustering algorithm, the two-dimensional data composed of the efficacy evaluation feature vector and the coordinates of each voxel point are clustered to obtain a plurality of clustering clusters, and each clustering cluster is taken as a liver tissue region.

[0063] Step S30, based on the treatment efficacy features of each liver tissue region, the efficacy information index of different liver tissue regions is determined.

[0064] As an example, for different liver tissue regions, the efficacy features of immunotherapy are different, and the efficacy information index of different liver tissue regions can be calculated by the specific efficacy features of different liver tissue regions, and then it is determined whether the selected liver tissue region has reference, so as to accurately select the efficacy comparison region.

[0065] As an example, the efficacy information index represents the distribution of the efficacy information carried by the liver tissue region. The greater the efficacy information index, the richer the efficacy information carried by the current liver tissue region, and the greater the reference degree of the region for analyzing the treatment effect.

[0066] Step S40, selecting the efficacy comparison region from each liver tissue region based on the efficacy information index.

[0067] As an example, the efficacy comparison region is selected from each liver tissue region according to the size distribution of the efficacy information index corresponding to different liver tissue regions.

[0068] Among them, the step S40 of selecting the efficacy comparison region from each liver tissue region based on the efficacy information index comprises: Step S41, setting the liver tissue region with overlapping region of the voxel point and the liver cancer lesion region before immunotherapy as the target liver tissue region.

[0069] As an example, for the efficacy information of different liver tissue regions, the spatial overlap degree of each liver tissue region and the liver cancer lesion region before immunotherapy can be analyzed: the greater the overlap, the more likely it is that the cluster contains lesion residues, necrosis, or immune reactive inflammation area, and has higher efficacy sensitivity. The liver tissue region with overlapping region of these voxel points and the liver cancer lesion region before immunotherapy is set as the target liver tissue region to be selected.

[0070] Step S42, calculating the first distance between the centroid coordinates of each target liver tissue region and the centroid coordinates of the liver cancer lesion region.

[0071] As an example, the first distance represents the distance value between the centroid coordinates of the target liver tissue region and the centroid coordinates of the liver cancer lesion region. The distance value is normalized to obtain the first distance, and the normalization method is maximum-minimum normalization.

[0072] Step S43, obtaining the efficacy comparison coefficient based on the ratio of the efficacy information index of each target liver tissue region to the corresponding first distance.

[0073] As an example, the ratio between the efficacy information index and the first distance of each target liver tissue region is calculated as the efficacy comparison coefficient of immunotherapy.

[0074] Step S44, sorting the efficacy comparison coefficients of each target liver tissue region from large to small, and selecting the adjacent target liver tissue regions with the largest difference in efficacy comparison coefficients.

[0075] As an example, after the therapeutic effect contrast coefficients of the target liver tissue regions are calculated, the target liver tissue regions are ranked according to the sizes of the therapeutic effect contrast coefficients.

[0076] As an example, the adjacent target liver tissue regions with the largest difference in therapeutic effect contrast coefficients satisfy two conditions: 1. the difference between the therapeutic effect contrast coefficients of the two (the difference between the former and the latter in the ranking) is greater than that of other adjacent target liver tissue regions, and 2. the two are adjacent in the ranking position.

[0077] Step S45, the target liver tissue region with a therapeutic effect contrast coefficient greater than or equal to the corresponding maximum therapeutic effect contrast coefficient of the adjacent target liver tissue region is taken as the therapeutic effect contrast region.

[0078] As an example, after the adjacent target liver tissue regions are selected, the therapeutic effect contrast coefficients of the two have a maximum value, the target liver tissue region with a therapeutic effect contrast coefficient greater than or equal to the maximum therapeutic effect contrast coefficient represents a region with obvious therapeutic effect feature contrast, and is also a turning point of the change trend of the therapeutic effect contrast coefficient. Therefore, the target liver tissue regions ranked before the maximum therapeutic effect contrast coefficient are taken as the therapeutic effect contrast regions, and these selected therapeutic effect contrast regions have obvious contrast with the liver cancer lesion region before immunotherapy.

[0079] Step S50, generating the therapeutic effect evaluation result of immunotherapy based on the difference features between the therapeutic effect contrast region and the liver cancer lesion region.

[0080] As an example, the therapeutic effect of immunotherapy is determined by comparing the difference features between the therapeutic effect contrast region and the liver cancer lesion region after treatment.

[0081] The step S50 of generating the therapeutic effect evaluation result of immunotherapy based on the difference features between the therapeutic effect contrast region and the liver cancer lesion region comprises: Step S51, respectively calculating the multi-sequence signal intensity of the therapeutic effect contrast region and the liver cancer lesion region and the standard deviation of the multi-sequence signal intensity to obtain the difference features.

[0082] As an example, the therapeutic effect of immunotherapy needs to be analyzed by features such as lesion volume (evaluating the change of lesion load after immunotherapy), MRI signal intensity (evaluating the change of tissue composition in the lesion and the characteristics of immune response), and internal composition complexity (quantifying the change of internal tissue composition complexity to reflect the sensitivity of treatment response).

[0083] As an example, the difference features can be one or more of the volume difference, the multi-sequence signal intensity, and the standard deviation of the multi-sequence signal intensity.

[0084] Step S52, generating the therapeutic effect evaluation result of immunotherapy according to the change trend of the difference features.

[0085] As an example, the change trend of the difference feature can be increasing, decreasing or unchanged, and different treatment evaluation effects can be obtained according to different change trends.

[0086] The step S52 of generating the evaluation result of the immunotherapy according to the change trend of the difference feature includes: If both the multi-sequence signal intensity and the standard deviation of the multi-sequence signal intensity show a decreasing trend, it is determined that the treatment effect of the current immunotherapy meets the expected standard.

[0087] As an example, when evaluating the treatment effect, the volume change of the liver cancer lesion area before and after treatment can also be used as a reference value. When the volume of the therapeutic effect comparison area is significantly reduced (volume shows a decreasing trend) relative to the volume of the liver cancer lesion area, the arterial enhancement subsides (signal intensity decreases), and the internal structure homogenizes (the mean of the standard deviation of the signal intensity decreases), the current immunotherapy effect is better, that is, the treatment effect of the current immunotherapy meets the expected standard, and the current regimen can be continued. It should be noted that when the treatment effect is good, the cell structure of the overall lesion area will appear homogenization, and the overall cell structure is not so chaotic. From the signal intensity, it can be seen that the standard deviation of the multi-sequence signal intensity decreases.

[0088] As an example, when the volume of the same area after immunotherapy increases relative to the volume before immunotherapy, the T2 high signal, and the ADC (apparent diffusion coefficient) increases, it may indicate a benign lesion (such as a cyst or necrotic tissue) or an improvement after treatment (such as a decrease in cells after tumor radiotherapy and chemotherapy), that is, a pseudo-progression phenomenon may occur, and follow-up observation is required. If the multi-sequence signal intensity shows an increasing trend, it is determined that the treatment effect of the current immunotherapy is lower than the expected standard.

[0089] As an example, if the volume difference shows that the volume of the therapeutic effect comparison area is greater than or equal to the volume of the liver cancer lesion area, and the multi-sequence signal intensity shows an increasing trend, indicating volume increase, arterial enhancement high signal characteristics, ADC decrease, and internal structure does not appear homogenization characteristics (the standard deviation of the multi-sequence signal intensity is basically unchanged), then residual tumor or super-progression may occur, and it is determined that the treatment effect of the current immunotherapy is lower than the expected standard, and additional intervention is required. After generating different treatment evaluation results, the doctor is provided for reference, and then the doctor makes the final intervention treatment means based on the reference results.

[0090] The application provides a liver cancer immunotherapy efficacy evaluation method based on medical images. In the related art, it is difficult to select the efficacy comparison region of the liver cancer lesion after immunotherapy in the efficacy evaluation process, and the selection of the efficacy comparison region is highly subjective, which leads to inaccurate evaluation of the immunotherapy effect. In the application, the historical liver image data and the current liver image data of the target patient are obtained, and the historical liver image data and the current liver image data are subjected to image segmentation respectively to obtain the liver cancer lesion region before immunotherapy and a plurality of liver tissue regions after immunotherapy. Then, the efficacy information index of different liver tissue regions is calculated according to the treatment efficacy characteristics of each liver tissue region, and the efficacy comparison region of the immunotherapy is selected through the efficacy information index, thereby avoiding the inaccuracy caused by subjective selection. Then, the efficacy evaluation result of the immunotherapy is generated according to the difference characteristics between the efficacy comparison region of the immunotherapy and the liver cancer lesion region, and the treatment effect of the immunotherapy is accurately evaluated through the efficacy evaluation result.

[0091] Further, with reference to Figure 2 Based on the first embodiment in the application, another embodiment of the application is provided, in which the step S30 of determining the efficacy information index of different liver tissue regions based on the treatment efficacy characteristics of each liver tissue region includes: Step S31, based on the edema characteristics of each liver tissue region and the signal intensity change degree of a plurality of sequences in each liver tissue region before and after immunotherapy, the liver cancer lesion information index is calculated. As an example, the cell density in the tumor area is high, and the diffusion of water molecules is limited, which shows the characteristics of high DWI (diffusion weighted imaging) signal and reduced ADC value. The water molecules in the edema region around the tumor diffuse freely, which shows reduced DWI signal and increased ADC. Therefore, the edema characteristics of other regions around the current region can be analyzed to further improve the accuracy of the efficacy information judgment.

[0092] As an example, liver cancer usually has an abnormally rich arterial blood supply, and the tumor neovascularization is dense. A large amount of contrast agent enters the lesion in the arterial phase, which shows high enhancement in the arterial phase. When the treatment is effective, the tumor blood vessels are necrotic and occluded, and the arterial blood supply is reduced, and the enhancement disappears. Therefore, for a certain liver tissue region, the comparison of the MRI signal intensity after immunotherapy and the MRI signal intensity before treatment can be used as one of the analysis factors of the efficacy information index, and then based on the edema characteristics of each liver tissue region and the signal intensity change degree before and after immunotherapy, the liver cancer lesion information index of each liver tissue region can be determined. The liver cancer lesion information index represents the richness of the liver cancer lesion signal characteristics. The greater the value is, the more liver cancer lesion signal characteristics are distributed in the current liver tissue region.

[0093] The step S31 of calculating the liver cancer lesion information index based on the edema characteristics of each liver tissue region and the signal intensity change degree of each sequence in each liver tissue region before and after immunotherapy comprises: For any liver tissue region, the proportion mean of the voxel points in the current liver tissue region that contain the sequence corresponding to the liver cancer lesion region, and the first difference value of the signal intensity of each sequence in the voxel points before immunotherapy minus the signal intensity of the corresponding sequence after immunotherapy are calculated, and the average of each first difference value is calculated to obtain the first average difference value; As an example, the proportion mean represents the lesion proportion of the voxel points in each sequence that belong to the liver cancer lesion region before immunotherapy, and the average of the lesion proportions of all sequences calculated is denoted as A.

[0094] As an example, the first difference value represents the signal intensity change degree of each sequence before and after treatment, and the first average difference value is obtained by calculating the average of the change degree of the signal intensity of each voxel point and normalizing the average, denoted as , wherein the normalization method for the average is maximum and minimum normalization.

[0095] The apparent diffusion coefficient to diffusion weighted imaging signal intensity ratio of each voxel point in the liver tissue region adjacent to the current liver tissue region is calculated as an edema characteristic index, and the first average of the edema characteristic indexes of all voxel points is calculated; As an example, the ratio between the apparent diffusion coefficient and the diffusion weighted imaging signal intensity can reflect the water diffusion of the tumor region, and based on the ratio between the two, the edema characteristic index is determined, and the greater the edema characteristic index, the stronger the water diffusion ability of the edema region of the tumor region.

[0096] As an example, the first average is the average of the edema characteristic indexes corresponding to each voxel point in the liver tissue region adjacent to the current liver tissue region.

[0097] The adjacent liver tissue region corresponding to the maximum first average value is selected as the edema characteristic reference region of the current liver tissue region, and the first difference value mean of the edema characteristic indexes of the voxel points in the edema characteristic reference region before and after immunotherapy is calculated; As an example, the difference value mean obtained by subtracting the edema characteristic index of the corresponding voxel point before immunotherapy from the edema characteristic index of each voxel point in the edema characteristic reference region of the current liver tissue region after immunotherapy is obtained, and after obtaining the difference value mean, the first difference value mean is obtained by normalization processing, denoted as , wherein the normalization method for the difference value mean is maximum and minimum normalization.

[0098] Based on the proportion mean, the first average difference and the first difference mean, the liver cancer lesion information index is calculated.

[0099] As an example, the calculation method of the liver cancer lesion information index may be: ; Wherein, A represents the proportion mean, represents the first average difference, represents the first difference mean, and norm() represents the normalization calculation. The greater the liver cancer lesion information index, the more liver tissue regions have liver cancer lesion signal characteristics.

[0100] Step S32, for any liver tissue region, the difference between the standard deviations of the different sequence signal intensities of each voxel point in the current liver tissue region before and after immunotherapy is calculated to obtain the standard deviation difference; As an example, in addition to the lesion signal characteristics embodied in the liver tissue region as a whole, due to the complex structure inside the lesion region, the same region after immunotherapy may be mixed with necrosis, residual tumor, fibrosis, blood supply reconstruction and other different tissues. Such heterogeneous structures lead to an increase in the difference between the signal intensities of adjacent voxels, which is manifested as an increase in the standard deviation.

[0101] As an example, for any liver tissue region, the standard deviations of the different sequence signal intensities of each voxel point in the current liver tissue region before and after immunotherapy are calculated respectively. Based on each sequence, the standard deviation between the different signal intensities of the current sequence before and after immunotherapy is calculated for each voxel point, and then the average value of the standard deviations of the different sequence signal intensities is calculated. The difference between the average value of the standard deviation after immunotherapy and the average value of the standard deviation before immunotherapy is calculated, and the value range is normalized to by the sigmoid function to obtain the standard deviation difference, denoted as .

[0102] Step S33, based on the standard deviation difference and the liver cancer lesion information index, the efficacy information index of different liver tissue regions is calculated.

[0103] As an example, the calculation method of the efficacy information index Z can be: ; Wherein, S represents the liver cancer lesion information index, represents the standard deviation difference, and the greater the efficacy information index, the richer the efficacy information carried by the current liver tissue region.

[0104] In the embodiment, by analyzing the characteristics of the immunotherapy efficacy of different regions, the efficacy information index of different liver tissue regions is evaluated, and then, according to the efficacy information index, the efficacy comparison region with more efficacy information is accurately selected.

[0105] With reference to Figure 3 , Figure 3 is a device structure schematic diagram of a hardware running environment involved in the embodiment of the application.

[0106] As Figure 3 shown, the medical image-based liver cancer immunotherapy efficacy evaluation device can include a processor 1001, a memory 1005, and a communication bus 1002. The communication bus 1002 is used to realize the connection communication between the processor 1001 and the memory 1005.

[0107] Optionally, the medical image-based liver cancer immunotherapy efficacy evaluation device can further include a user interface, a network interface, a camera, an RF (Radio Frequency, RF) circuit, a sensor, a WiFi module, and the like. The user interface can include a display screen (Display), an input sub-module such as a keyboard (Keyboard), and the optional user interface can further include a standard wired interface and a wireless interface. The network interface can include a standard wired interface and a wireless interface (such as a WI-FI interface).

[0108] Those skilled in the art can understand that Figure 3 the medical image-based liver cancer immunotherapy efficacy evaluation device structure shown in the embodiment does not constitute a limitation on the medical image-based liver cancer immunotherapy efficacy evaluation device, and can include more or fewer components than the diagram, or combine certain components, or different component arrangements.

[0109] As Figure 3 shown, the memory 1005 as a storage medium can include an operating system, a network communication module, and a medical image-based liver cancer immunotherapy efficacy evaluation program. The operating system is a program that manages and controls the hardware and software resources of the medical image-based liver cancer immunotherapy efficacy evaluation device, supports the running of the medical image-based liver cancer immunotherapy efficacy evaluation program and other software and / or programs. The network communication module is used to realize the communication between the components in the memory 1005, and the communication between other hardware and software in the medical image-based liver cancer immunotherapy efficacy evaluation system.

[0110] In Figure 3 the medical image-based liver cancer immunotherapy efficacy evaluation device, the processor 1001 is used to execute the medical image-based liver cancer immunotherapy efficacy evaluation program stored in the memory 1005, and realize the steps of the medical image-based liver cancer immunotherapy efficacy evaluation method of any one of the above.

[0111] The application based on medical image of liver cancer immunotherapy efficacy evaluation device specific implementation and the above based on medical image of liver cancer immunotherapy efficacy evaluation method each embodiment is basically same, here will not repeat.

[0112] It should be noted that in this paper, the term "includes", "contains" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, article or system including a series of elements not only includes those elements, but also includes other elements not explicitly listed, or includes elements inherent to such process, method, article or system. Without more limitations, the element defined by the sentence "includes a" does not exclude the presence of other identical elements in the process, method, article or system including the element.

[0113] The above application example serial number is only for description, not representing the pros and cons of the embodiment.

[0114] Through the description of the above embodiments, those skilled in the art can clearly understand that the above embodiment method can be realized by software and necessary general hardware platform, of course, it can also be realized by hardware, but in many cases, the former is a better embodiment. Based on such understanding, the technical solutions of the present application can be embodied in the form of software product, which is stored in the above storage medium (such as ROM / RAM, magnetic disk, optical disk), including a plurality of instructions for making a terminal device (which can be a mobile phone, computer, server, air conditioner or network device) execute the method of each embodiment of the present application.

[0115] The above is only the preferred embodiment of the present application, and does not limit the application range of the present application. Any equivalent structure or equivalent process transformation made by using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the application protection range of the present application.

[0116] It should be noted that: the above-mentioned sequence of the embodiments of the present application is only for description, not representing the pros and cons of the embodiment. The process depicted in the drawing does not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or may be advantageous.

[0117] Each embodiment in the specification is described in a progressive manner, and the same or similar parts of each embodiment refer to each other. Each embodiment focuses on the difference from other embodiments.

Claims

1. A method for evaluating the efficacy of liver cancer immunotherapy based on medical imaging, characterized in that, The method includes: Retrieve historical and current liver imaging data of the target patient stored in the medical system; The historical liver image data is segmented to obtain the liver cancer lesion area before immunotherapy, and the current liver image data is segmented into multiple liver tissue regions. Based on the therapeutic efficacy characteristics of each liver tissue region, an efficacy information index for different liver tissue regions was determined. Based on the efficacy information index, efficacy comparison areas are selected from each of the liver tissue regions; Based on the differences between the efficacy comparison area and the liver cancer lesion area, the efficacy evaluation results of immunotherapy are generated.

2. The method for evaluating the efficacy of liver cancer immunotherapy based on medical imaging as described in claim 1, characterized in that, The process of segmenting the historical liver imaging data to obtain the liver cancer lesion region before immunotherapy includes: The historical liver image data is subjected to three-dimensional reconstruction processing to obtain the first three-dimensional image data; The first three-dimensional image data is segmented using a preset deep learning model to obtain the liver cancer lesion area and liver area before immunotherapy.

3. The method for evaluating the efficacy of liver cancer immunotherapy based on medical imaging as described in claim 2, characterized in that, The step of segmenting the current liver imaging data into multiple liver tissue regions includes: The current liver imaging data is subjected to three-dimensional reconstruction processing to obtain second three-dimensional imaging data; Based on the signal intensity features of each voxel in the second three-dimensional image data, a therapeutic efficacy assessment feature vector for each voxel is constructed. By using a pre-defined clustering algorithm, the two-dimensional data composed of the efficacy evaluation feature vector and coordinates of each voxel point is clustered to obtain multiple liver tissue regions.

4. The method for evaluating the efficacy of liver cancer immunotherapy based on medical imaging as described in claim 3, characterized in that, After performing three-dimensional reconstruction processing on the current liver image data to obtain the second three-dimensional image data, the process further includes: Using the liver region as a rigid registration point, the second three-dimensional image data and the liver region are subjected to multi-sequence spatial alignment processing to obtain the aligned sequence of the second three-dimensional image data.

5. The method for evaluating the efficacy of liver cancer immunotherapy based on medical imaging as described in claim 3, characterized in that, The step of constructing a therapeutic efficacy assessment feature vector for each voxel based on the signal intensity features of each voxel in the second three-dimensional image data includes: The voxel point region within the liver area, excluding the liver cancer lesion region, is set as the general liver tissue region; For any sequence in the voxel, calculate the absolute value of the difference between the mean signal intensity of the liver cancer lesion region and the general liver tissue region in the current sequence to obtain the mean signal intensity difference; The volume mean difference is calculated based on the absolute value of the difference between the volume of the liver cancer lesion region described in the current sequence and the mean volume of the liver cancer lesion region in all sequences. The ratio of the number of sequences belonging to the general liver tissue region in each voxel point to the total number of sequences is used as the lesion morphology interference index of each voxel point. Calculate the mean index of each lesion morphology interference index, and multiply the mean index by the difference in volume to obtain the morphological weight influence index of the current sequence. Based on the ratio between the mean difference in signal intensity and the morphological weight influence index, the therapeutic reference weight of each sequence is calculated. Based on the therapeutic reference weights and the signal intensity characteristics of each voxel in the second three-dimensional image data, a therapeutic evaluation feature vector is constructed for each voxel.

6. The method for evaluating the efficacy of liver cancer immunotherapy based on medical imaging as described in claim 1, characterized in that, The determination of efficacy information indices for different liver tissue regions based on the therapeutic efficacy characteristics of each of the aforementioned liver tissue regions includes: Based on the edema characteristics of each liver tissue region and the changes in signal intensity of multiple sequences in each liver tissue region before and after immunotherapy, the liver cancer lesion information index is calculated. For any liver tissue region, calculate the difference between the standard deviations of the signal intensities of different sequences before and after immunotherapy for each voxel point within the current liver tissue region, and obtain the standard deviation difference. Based on the standard deviation difference and the liver cancer lesion information index, the efficacy information index for different liver tissue regions is calculated.

7. The method for evaluating the efficacy of liver cancer immunotherapy based on medical imaging as described in claim 6, characterized in that, The liver cancer lesion information index is calculated based on the edema characteristics of each liver tissue region and the changes in signal intensity of multiple sequences in each liver tissue region before and after immunotherapy, including: For any liver tissue region, calculate the average proportion of voxel points corresponding to the liver cancer lesion region in the current liver tissue region, and the first difference between the signal intensity of each sequence in the voxel points before immunotherapy and the signal intensity of the corresponding sequence after immunotherapy, and calculate the average of each of the first differences to obtain the first average difference. The apparent diffusion coefficient to diffusion-weighted imaging signal intensity ratio of each voxel point in the liver tissue region adjacent to the current liver tissue region is calculated as the edema characteristic index, and the first mean of the edema characteristic index of all voxel points is calculated. The adjacent liver tissue region corresponding to the maximum value of the first mean is selected as the edema feature reference region of the current liver tissue region, and the first mean difference of the edema feature index of the voxel points in the edema feature reference region before and after immunotherapy is calculated. Based on the average proportion, the first average difference, and the average of the first differences, the liver cancer lesion information index is calculated.

8. The method for evaluating the efficacy of liver cancer immunotherapy based on medical imaging as described in claim 7, characterized in that, The selection of efficacy comparison regions from each of the liver tissue regions based on the efficacy information index includes: The liver tissue region where the voxel point overlaps with the liver cancer lesion region before immunotherapy is set as the target liver tissue region; Calculate the first distance between the centroid coordinates of each target liver tissue region and the centroid coordinates of the liver cancer lesion region; The efficacy comparison coefficient is obtained based on the ratio of the efficacy information index of each target liver tissue region to its corresponding first distance; The efficacy comparison coefficients of each target liver tissue region are sorted from largest to smallest, and the adjacent target liver tissue regions with the largest difference in efficacy comparison coefficients are selected. The target liver tissue region with the efficacy comparison coefficient greater than or equal to the corresponding maximum efficacy comparison coefficient in the adjacent target liver tissue region is designated as the efficacy comparison region.

9. The method for evaluating the efficacy of liver cancer immunotherapy based on medical imaging as described in claim 1, characterized in that, The method of generating efficacy evaluation results for immunotherapy based on the differences between the efficacy comparison area and the liver cancer lesion area includes: The multi-sequence signal intensity and standard deviation of the multi-sequence signal intensity were calculated for the therapeutic effect comparison region and the liver cancer lesion region, respectively, to obtain the difference characteristics; Based on the changing trends of differential characteristics, efficacy assessment results of immunotherapy are generated.

10. The method for evaluating the efficacy of liver cancer immunotherapy based on medical imaging as described in claim 9, characterized in that, The process of generating efficacy evaluation results for immunotherapy based on the changing trends of differential characteristics includes: If both the intensity of the multi-sequence signal and the standard deviation of the multi-sequence signal intensity show a decreasing trend, then the therapeutic effect of the current immunotherapy is determined to meet the expected standard. If the intensity of the multi-sequence signal shows an increasing trend, it is determined that the therapeutic effect of the current immunotherapy is lower than the expected standard.

Images