Multimodal radiomics HDRBT combined with thermal ablation therapy for liver cancer efficacy evaluation system

Abstract

This invention discloses a multimodal radiomics-based HDRBT combined with thermal ablation therapy for liver cancer efficacy evaluation system, relating to the field of liver cancer treatment. The system includes: a multimodal image data acquisition module for collecting multimodal image data; a three-dimensional visualization model construction module for reconstructing the three-dimensional structure of the liver tumor, surrounding blood vessels, and bile ducts through image segmentation algorithms and integrating radiomics features to generate a visualization model; a combined treatment plan generation module for integrating HDRBT dosimetric parameters and thermal ablation process parameters, optimizing catheter insertion path, thermal ablation electrode position, and treatment sequence to generate a combined treatment plan; and an efficacy evaluation module for quantitatively analyzing tumor volume changes, ablation boundary integrity, and residual tumor activity through precise registration of multimodal images before and after treatment. This application's solution can address the problems of incomplete thermal ablation and high recurrence rate in hepatocellular carcinoma, providing full-chain technical support for precise clinical application.

Description

BACKGROUND

[0001] The present application relates to the field of liver cancer treatment, in particular to a multi-modal imageomics HDRBT combined with thermal ablation treatment for liver cancer efficacy evaluation system.

[0002] Primary hepatocellular carcinoma is a highly malignant and poor prognosis digestive system tumor worldwide, with a morbidity rate ranking fifth and a mortality rate ranking third among malignant tumors, and Chinese patients account for about 50% of the world. Surgical resection is the preferred treatment for hepatocellular carcinoma, but due to complex tumor location, insufficient liver function reserve or disease progression, only a small number of patients can obtain surgical opportunity. Thermal ablation as a minimally invasive treatment can achieve similar survival rate as surgical resection in small liver cancer, but for large tumors or lesions close to important structures such as blood vessels and bile ducts, incomplete ablation is easily caused by thermal sedimentation effect, thereby leading to tumor recurrence and metastasis. High dose rate brachytherapy (HDRBT) can precisely deliver high radiation dose to tumor tissue by directly implanting radioactive sources into tumor tissue, while reducing damage to surrounding normal tissues, providing a potential strategy to make up for the deficiency of thermal ablation. However, the synergistic mechanism of HDRBT combined with thermal ablation for hepatocellular carcinoma has not been clearly defined, and there is a lack of effective three-dimensional visualization tools to optimize preoperative planning, intraoperative guidance and postoperative efficacy evaluation of combined therapy, which limits the precision and effectiveness of combined therapy.

[0003] In existing hepatocellular carcinoma treatment methods, thermal ablation alone is prone to incomplete ablation due to tumor characteristics, leading to recurrence and metastasis; there is less basic research on the application of HDRBT alone in hepatocellular carcinoma, and the parameter optimization and synergistic mechanism of combined therapy have not been systematically explored; traditional two-dimensional image evaluation cannot accurately reflect the three-dimensional spatial relationship of the tumor and surrounding anatomical structures, and cannot provide accurate guidance for needle placement path, dose distribution and safety margin of combined therapy; in addition, the lack of molecular mechanism research on differential proteins limits the effective prediction of the efficacy of combined therapy and the guidance of individualized treatment, which restricts the clinical application of combined therapy. Therefore, based on the above problems, the present application proposes a multi-modal imageomics HDRBT combined with thermal ablation treatment for liver cancer efficacy evaluation system. SUMMARY

[0004] OBJECTIVE

[0005] To solve the above problems, the purpose of the present application is to provide a multi-modal imageomics HDRBT combined with thermal ablation treatment for liver cancer efficacy evaluation system, which aims to establish a HepG2 nude mouse orthotopic liver cancer model, test the feasibility of three-dimensional visualization and preoperative planning model, evaluate the effectiveness and safety of HDRBT combined with thermal ablation treatment for hepatocellular carcinoma, and screen differential proteins to reveal their mechanism, thereby laying a theoretical and technical foundation for clinical precise planning, efficacy prediction and mechanism research of combined therapy for hepatocellular carcinoma.

[0006] Technical solution

[0007] To achieve the above-mentioned purpose, the present application provides a multi-modal imageomics HDRBT combined with thermal ablation treatment for liver cancer efficacy evaluation system, which constructs a HepG2 nude mouse orthotopic liver cancer model, monitors tumor growth through high-frequency ultrasound and CT scanning, and creates a three-dimensional visual model by combining multi-modal images with image segmentation algorithms to optimize the HDRBT catheter implantation path and the thermal ablation electrode position; the tumor growth curve and tumor weight calculation are used to calculate the tumor inhibition rate, combined with ultrasound contrast to evaluate the tumor inactivation, and the liver and kidney function blood indicators and HE staining are used to evaluate the treatment safety; the core mechanism of the system is to screen the differential proteins of the combined group and the incomplete ablation group by mass spectrometry, verify their expression levels by RT-PCR and immunohistochemistry, analyze the correlation between the differential proteins and the prognosis of hepatocellular carcinoma using bioinformatics tools, and reveal the molecular mechanism of HDRBT combined with thermal ablation.

[0008] In the first aspect, the present application provides a multi-modal imageomics HDRBT combined with thermal ablation treatment for liver cancer efficacy evaluation system, comprising:

[0009] A multi-modal image data acquisition module is used to collect ultrasound, CT and magnetic resonance imaging multi-modal image data, wherein the ultrasound provides real-time guidance information, and the CT and magnetic resonance imaging provide high-resolution anatomical structure information;

[0010] A three-dimensional visual model construction module reconstructs the three-dimensional structure of liver tumors, surrounding blood vessels and bile ducts based on the multi-modal image data through image segmentation algorithms, and generates an interactive three-dimensional visual model by fusing tumor edge, tissue density and vascular distribution imageomics features;

[0011] A combined treatment plan generation module integrates the HDRBT dosimetry parameters and the thermal ablation process parameters according to the three-dimensional visual model, optimizes the catheter implantation path, the thermal ablation electrode position and the treatment sequence, and generates an individualized combined treatment plan;

[0012] An efficacy evaluation module quantitatively analyzes the tumor volume change, ablation boundary integrity and residual tumor activity through precise registration of pre-treatment and post-treatment multi-modal images;

[0013] A safety evaluation module is used to integrate liver and kidney function indicators, pathological HE staining results and adverse reaction records, and evaluate the safety of combined treatment.

[0014] Further, the multi-modal image data acquisition module fuses ultrasound real-time imaging with CT and magnetic resonance imaging high-resolution imaging through magnetic positioning technology to realize image dynamic alignment.

[0015] Further, the three-dimensional visualization model construction module adopts a deep convolutional neural network to segment liver tumors and blood vessels, and optimizes the segmentation result through a variational energy function.

[0016] Further, the joint treatment plan generation module adopts a topological optimization algorithm to optimize the catheter implantation path based on the three-dimensional visualization model.

[0017] Further, the image segmentation algorithm is preferably a combination of a convolutional neural network and a region competition model.

[0018] Further, the joint treatment plan generation module includes an HDRBT synergistic dose optimization model for residual tumor ablation by heat, which dynamically adjusts the HDRBT edge dose based on the residual tumor volume ratio after heat ablation.

[0019] Based on the radiosensitization effect of residual tumor after heat ablation, the residual tumor volume ratio is used as the core parameter for dose adjustment, and the edge dose and central dose distribution of HDRBT are dynamically optimized. On the one hand, it significantly enhances the clearance efficiency of residual lesions, and on the other hand, it essentially reduces the incidence of liver and kidney function damage and biliary complications by constraining normal tissue dose exposure in three dimensions. In addition, the model shows biological correlation with the expression level of different proteins G6PD, providing a molecular level closed-loop verification for efficacy prediction.

[0020] Further, the dose parameters of the HDRBT include edge dose, central dose, residence time, etc., and the process parameters of the heat ablation include power, time, ablation range, etc.

[0021] Further, the efficacy evaluation module analyzes residual tumor activity by analyzing the dynamic enhancement curve characteristics of ultrasound contrast, and judges the completeness of tumor ablation.

[0022] Further, the safety evaluation module uses the analytic hierarchy process to quantify liver and kidney function indicators, pathological results and adverse reactions into safety scores to evaluate the safety level.

[0023] Further, the liver and kidney function indicators include ALT, AST, bilirubin, etc., and the pathological HE staining results include liver cell necrosis degree, inflammatory infiltration, etc.

[0024] In a second aspect, the present application also provides a liver cancer combined treatment catheter parameter generation system, comprising:

[0025] A biomarker correlation module is used to receive detection data of tumor tissue samples, extract biomarker expression levels related to residual tumor volume, and establish a correlation model between the biomarker expression and the residual tumor volume ratio.

[0026] A dynamic dose regulation module, which dynamically corrects the HDRBT edge dose based on the correlation model and the residual tumor volume proportion output by the three-dimensional topological modeling unit;

[0027] A catheter parameter generation module, which generates manufacturing instructions of catheter length, radioactive source residence position and residence time according to the corrected dose parameters and the catheter layout path of the three-dimensional model, and outputs the manufacturing instructions to a 3D printing device.

[0028] Further, the biomarker correlation module comprises a verification unit for verifying the expression level of the biomarker by a molecular biology method.

[0029] In a third aspect, the application further provides a method for screening differential proteins for treating liver cancer by combining HDRBT and thermal ablation, which is based on the aforementioned system and comprises the following steps:

[0030] Collecting tumor tissue samples of the combined treatment group, the simple thermal ablation group, the simple HDRBT group and the control group;

[0031] Analyzing the sample proteome by mass spectrometry and screening differential expression proteins;

[0032] Verifying the expression level of the differential proteins by a molecular biology method;

[0033] Using bioinformatics tools to analyze the correlation of the differential proteins with the prognosis of hepatocellular carcinoma.

[0034] The application reconstructs the accurate three-dimensional model of liver cancer lesions, blood vessels and bile duct system by dynamically weighted fusion algorithm through the integration of multi-modal image data of ultrasound, CT and magnetic resonance imaging, and breaks through the limitations of single modal image; based on the three-dimensional model, the position of thermal ablation electrode and the catheter layout scheme of HDRBT are optimized cooperatively, and a residual tumor volume driven radiation dose dynamic adjustment model is established to quantify the radiosensitization effect after thermal ablation; the tumor volume change, ablation boundary integrity and residual tumor activity are quantitatively analyzed by multi-modal image registration before and after treatment, and a hierarchical analysis safety score model is constructed in combination with liver and kidney function indicators and pathological HE staining; further, differential proteins are screened by mass spectrometry to verify the value of the differential proteins as a combined treatment efficacy prediction marker, and the synergistic mechanism of G6PD-STAT3 signal axis regulating epithelial mesenchymal transition process is revealed.

[0035] The scheme can significantly reduce the three-dimensional reconstruction error, improve the catheter implantation positioning accuracy, and improve the ablation boundary coverage rate to more than 95%, ensuring complete coverage of the tumor and safety boundary, solving the spatial deviation problem of traditional two-dimensional planning; the residual tumor volume and local recurrence rate of the combined therapy group are significantly lower than that of single therapy, and the tumor suppression rate is increased by 46.3%, which inhibits the epithelial mesenchymal transition process through low expression of G6PD, and enhances the clearance efficiency of residual lesions by HDRBT; based on the three-dimensional model of catheter path topology optimization and dose constraint, the degree of liver function damage is reduced by 40.7%, and the incidence of bile duct complications is zero, achieving the balance between efficacy and toxicity; the expression level of G6PD as a biomarker can predict the efficacy of patients, providing molecular basis for individualized treatment. The scheme can systematically solve the problem of incomplete thermal ablation and high recurrence rate of hepatocellular carcinoma, and provide full-chain technical support for the clinical precise application of HDRBT combined with thermal ablation.

[0036] Beneficial effects

[0037] By implementing the multi-modal imageomics HDRBT combined with thermal ablation treatment for liver cancer efficacy evaluation system provided by the present application, the following technical effects are achieved:

[0038] (1) By integrating multi-dimensional imageomics features such as tumor texture, morphology and enhancement, a dynamic and interactive three-dimensional treatment model is constructed, breaking through the visual limitations of traditional two-dimensional planning. The core is to realize the global optimization of treatment parameters, intelligently avoid key structures based on vascular and bile duct topology, and simultaneously optimize the spatial arrangement sequence of HDRBT catheter placement and thermal ablation electrode, so as to maximize the synergistic effect of combined therapy. The planning system can significantly improve the executability of the treatment plan, reduce the frequency of real-time adjustment during the operation, and reduce the risk of damage to adjacent tissues.

[0039] (2) By screening and verifying the negative correlation between the expression level of differential protein G6PD and treatment response, it is established as a prognostic indicator for HDRBT combined with thermal ablation. It reveals the molecular mechanism of combined therapy, and through preoperative detection of G6PD expression level, the sensitivity of patients to combined therapy can be predicted, guiding individualized treatment plan selection. Bioinformatics analysis further confirms the significant association of the marker with patient survival, providing molecular target support for precision medicine.

[0040] (3) By dynamically fusing the high-resolution anatomical information of CT and the real-time functional information of ultrasound, the three-dimensional reconstruction accuracy of liver cancer and surrounding blood vessels and bile duct structures is significantly improved. The generated fusion image has significant advantages in tumor boundary restoration and spatial relationship consistency, solving the inherent limitations of single modality imaging, providing reliable guarantee for the spatial positioning accuracy of catheter implantation and thermal ablation electrodes, and essentially improving the matching degree of treatment path planning and intraoperative navigation, thereby optimizing the coverage integrity of the safety boundary and reducing the risk of secondary intervention caused by image deviation.

[0041] (4) Based on the radiosensitization effect of residual tumor after thermal ablation, the residual tumor volume ratio is taken as the core parameter for dose adjustment, and the edge dose and central dose distribution of HDRBT are dynamically optimized. On the one hand, it significantly enhances the clearance efficiency of residual lesions, and on the other hand, it essentially reduces the incidence of liver and kidney function damage and bile duct complications by three-dimensional visualization of normal tissue dose exposure. In addition, the model shows biological correlation with the expression level of differential protein G6PD, providing a molecular level closed-loop verification for efficacy prediction. BRIEF DESCRIPTION OF DRAWINGS

[0042] In order to make the above-mentioned multi-modal imageomics HDRBT combined with thermal ablation treatment for liver cancer efficacy evaluation system of the present application more obvious and easy to understand, the drawings needed in the specific embodiments of the present application will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application, and other drawings can be obtained from these drawings without creative labor for those skilled in the art.

[0043] Figure 1 A three-dimensional visualization model construction flowchart is shown;

[0044] Figure 2 A microwave ablation and radiofrequency ablation treatment subcutaneous tumor-bearing nude mouse tumor growth curve graph is shown;

[0045] Figure 3 An ultrasound-guided HDRBT combined with thermal ablation treatment for hepatocellular carcinoma effectiveness evaluation experiment flowchart is shown;

[0046] Figure 4 A nude mouse tumor growth curve, body weight change curve and tumor weight difference graph is shown;

[0047] Figure 5 An ultrasound contrast observation of HDRBT combined with thermal ablation treatment effect evaluation graph is shown;

[0048] Figure 6 A comparison of liver and kidney function blood indicators between groups is shown;

[0049] Figure 7 A comparison of liver and kidney function blood indicators between groups is shown;

[0050] Figure 8 Representative HE staining images of each group;

[0051] Figure 9 Representative immunohistochemical index comparison images between each group;

[0052] Figure 10 Survival curve of G6PD different expression population;

[0053] Figure 11 Survival curve of G6PD different expression population. DETAILED DESCRIPTION

[0054] It should be understood that any technical solution claimed by the present application does not involve the diagnosis and treatment of diseases.

[0055] In order to facilitate the understanding of the embodiments of the present application, first, the abbreviations and key terms that may be involved in the embodiments of the present application are explained and defined. For the abbreviations or key terms not defined, they are commonly understood by those skilled in the art.

[0056] HDRBT: high dose rate brachytherapy;

[0057] G6PD: glucose-6-phosphate dehydrogenase;

[0058] PCDA: programmed cell death-related antigen;

[0059] TUNEL: terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling;

[0060] HE staining: hematoxylin-eosin staining;

[0061] EMT: epithelial-mesenchymal transition;

[0062] iMWA: incomplete microwave ablation;

[0063] cMWA: complete microwave ablation;

[0064] HDRBT+iMWA: combined treatment group;

[0065] STAT3: signal transduction factor;

[0066] Snail: EMT regulatory transcription factor;

[0067] OS: overall survival;

[0068] RT-PCR: reverse transcription polymerase chain reaction;

[0069] SDS in-solution digestion: proteomics sample pre-treatment technique

[0070] CEUS: contrast-enhanced ultrasound

[0071] IHC: immunohistochemistry

[0072] The present application is described in detail below.

[0073] Example 1

[0074] In the thermal ablation of large tumors, the advantages of three-dimensional visualization software in three-dimensional display and accurate measurement can be reflected. Through three-dimensional planning, the difficulty of ablation treatment can be more fully estimated, and the treatment degree can be more accurately predicted, so that more sufficient treatment can be performed during ablation treatment. In the past, the safety ablation boundary or the cutting edge width was generally set to 5-10 mm in the research on thermal ablation of hepatocellular carcinoma. The three-dimensional visualization evaluation method can quantitatively evaluate the ablation boundary, and is superior to the two-dimensional image evaluation method in the comprehensiveness and accuracy of evaluating the ablation boundary.

[0075] The creation process of the HDRBT three-dimensional visualization animal model is described as follows, and the flowchart is shown in Figure 1

[0076] The experimental animals include 40 male, 4-6-week-old nude mice of SPF level, which are bred in a clean layer cabinet in an animal room.

[0077] HepG2 cells were routinely cultured and passaged to establish a nude mouse subcutaneous tumor model. The nude mice were bred for 28 days, and the tumors were taken for liver orthotopic implantation. The skin was disinfected, the tumor was removed, and was repeatedly rinsed in sterile ice physiological saline for 2-3 times. Then, the tumor mass was cut into small pieces with a diameter of 2 mm for implantation. The tumor tissue was implanted into the liver of the nude mouse within 30 min after being removed. The supine position was taken, the skin of the surgical field was disinfected twice with 0.5% iodophor, an oblique incision with a length of about 8 mm was made below the left costal margin, the abdominal wall was incised to the full layer to expose the liver at the incision, a tunnel was punctured along the liver capsule with a sharp forceps, the tumor tissue was implanted, and the liver incision was sutured with 8-0 non-injury surgical suture and the tumor mass was fixed. The liver was gently returned to the abdominal cavity, and the abdominal wall was sutured with 6-0 suture to close the abdomen. The whole operation was performed in a sterile super-clean bench, and the anesthetized mouse was returned to the cage for breeding after waking up. No antibiotics were used after the operation, and the mouse was allowed to eat freely.

[0078] When the volume of the implanted tumor was about The tumor-bearing nude mice were randomly divided into four groups, i.e., the combination group, the HDRBT group, the thermal ablation group and the control group, with 10 nude mice in each group.

[0079] The HDRBT condition exploration process of the animal model is described as follows.

[0080] ​The three-dimensional visualization software system needs to obtain the CT or MRI scan data before treatment, i.e. the continuous two-dimensional tomographic images of the treatment lesion area, which is required to be in DICOM format. The DICOM data is imported into the software system, the appropriate scan sequence is selected through the DICOM browser, the two-dimensional images are preprocessed, and then three-dimensional modeling is performed to obtain the three-dimensional images and measurement data of the treatment lesion and related anatomical structures. Before treatment, the interactive operation of the software is used for surgical planning to design the best needle insertion path and needle arrangement scheme. During treatment, medical image fusion and surgical navigation based on magnetic positioning are performed. After treatment, the effectiveness evaluation is performed through the accurate registration of the three-dimensional data of the medical images before and after treatment.

[0081] First, the number of catheters, dose and light exposure times of HDRBT are determined. Under the guidance of B-ultrasound, the implantation direction, depth and number of catheters are determined according to the tumor position, and the catheters are properly fixed to prevent bleeding. B-ultrasound is a two-dimensional image, and the B-ultrasound probe is rotated by 90 degrees in each catheter plane to obtain a three-dimensional image of the same section, thereby ensuring the accurate position of each catheter. When arranging the catheters, the catheters should be arranged in parallel and penetrate the tumor. According to the size of the peripheral dose to be given to the tumor, the interstitial treatment planning system is used to calculate and determine the position and residence time of each residence point.

[0082] The tumor implantation radiotherapy controls the marginal dose at 10 Gy, and the central dose can reach 250 Gy,

[0083] The treatment time is 32 seconds. Studies have shown that HDRBT has good treatment effect and safety on subcutaneously implanted tumor-bearing nude mice.

[0084] The process of heat ablation condition exploration of animal models is as follows.

[0085] After drug anesthesia, heat ablation treatment is performed. The surgical field is routinely disinfected and draped, the skin at the tumor site is cut, careful operation is paid attention to during the operation to avoid damaging the tumor, the tumor is exposed, the radiofrequency needle is punctured to a depth of about 0.5 cm, the radiofrequency time is 30 s, the power is 10 W, and the tumor lesion is incompletely ablated.

[0086] The two heat ablation treatment methods of microwave ablation and radiofrequency ablation are compared on animal models. According to the incomplete ablation power and time, the body weight and tumor size changes of 7-day tumor-bearing nude mice are observed. The tumor growth curves of subcutaneously implanted tumor-bearing nude mice treated by microwave ablation and radiofrequency ablation are as shown in Figure 2 The results show that the tolerance of nude mice to radiofrequency ablation is not as good as that to microwave ablation, although microwave ablation is considered to have a larger treatment range in clinical practice.

[0087] Example 2:

[0088] The process of evaluating the effectiveness of ultrasound-guided HDRBT combined with heat ablation in treating hepatocellular carcinoma is as follows.

[0089] A subcutaneous mouse model of primary liver cancer was established. Under ultrasound guidance, percutaneous needle insertion was performed, followed by brachytherapy. The thermal ablation group underwent ultrasound-guided needle insertion and thermal ablation therapy. The combined group received thermal ablation followed by HDRBT. The negative control group only underwent tumor biopsy without any other treatment intervention. Figure 3 As shown. CT and contrast-enhanced ultrasound were used to observe tumor development and metastasis in nude mice, and three-dimensional reconstruction was performed to measure tumor volume, plot volume and time growth curves, calculate tumor formation rate, record survival time of tumor-bearing nude mice, and record adverse reactions in each group of nude mice.

[0090] To observe the growth of transplanted tumors in nude mice, growth curves of the transplanted tumors were plotted and tumor weight and tumor inhibition rate were calculated. The long diameter of the transplanted tumor was measured by ultrasound every two days. With short axis And calculate the tumor volume. The formula is Based on the calculated tumor volume, a growth curve of the xenograft tumor in nude mice was plotted. Blood was drawn to measure liver and kidney function before the nude mice were euthanized. The nude mice were euthanized after anesthesia with chloral hydrate, and the tumor tissue was dissected, photographed, and weighed. The average tumor weight and tumor weight inhibition rate were calculated. Tumor weight inhibition rate % = (tumor weight of blank control group - tumor weight of experimental group) / tumor weight of blank control group × 100%.

[0091] Nude mice were subjected to contrast-enhanced ultrasound examination via tail vein injection followed by administration of contrast agent by an experienced sonographer. After intraperitoneal anesthesia, a routine ultrasound scan was performed first to select the optimal scanning location and maximum cross-section of the tumor. The scan was then switched to contrast mode, and contrast agent was injected via tail vein bolus injection at a rate of 0.1 ml / kg body weight, followed by flushing with 2 ml of normal saline. Dynamic images were continuously acquired for 5 minutes and stored on the machine's hard drive.

[0092] The treatment efficacy assessment between each group is as follows.

[0093] The tumors were divided into five groups: Control, Incomplete ablation (iMWA), Complete ablation (cMWA), HDRBT (intercalation radiotherapy), and HDRBT+iMWA (combination group). Tumor growth curves (observed for 28 days post-treatment): From day 9 post-treatment, the HDRBT+iMWA group was significantly lower than the Control and HDRBT groups. From day 21 post-treatment, the HDRBT+iMWA group was significantly lower than the iMWA group, and the difference was statistically significant. Figure 4 As shown in D. There were no differences in tumor weight growth curves among the groups, as shown in Figure D. Figure 4 As shown in Figure E. The bar chart of tumor weight changes after dissection shows a statistically significant difference between the HDRBT+iMWA group and the iMWA group.

[0094] Contrast-enhanced ultrasound can be used to observe the inactivation of subcutaneous tumors, such as... Figure 5The results are shown in FIG. 1.

[0095] The blood samples of the nude mice in each group showed that there was no statistical difference in blood indicators of liver and kidney function between the HDRBT+iMWA group and the iMWA group, such as Figure 6 and Figure 7 The results are shown in FIG. 1.

[0096] According to the weight change trend chart of the nude mice in each group, it was found that the safety of HDRBT combined with thermal ablation was higher. At the same time, according to the HE staining chart of each group, it was found that the combined treatment and simple treatment were safe and effective, such as Figure 8 The results are shown in FIG. 1.

[0097] Example 3:

[0098] This example explains the related mechanism of HDRBT combined with thermal ablation.

[0099] The process of screening differential proteins by mass spectrometry is as follows.

[0100] Sample processing according to SDS In-solution digestion steps: add 40 μl of 50 mM ABC to the EP tube, then add 10 μl of sample; add 5 μl of 100 mM DTT to the EP tube, and place at room temperature for 45 min; add 5 μl of 100 mM IAA to the EP tube, avoid light, and place at room temperature for 45 min; add 20 μl of 100 mM DTT to the EP tube, and place at room temperature for 45 min; add 20 μl of 50 mM ABC to make the total volume 100 μl; add 2 μl of 0.1 μg / μL trypsin, then add 4 μl of 10% SDC, mix well, and place at 37°C overnight; slowly add 5 μL of 10% TFA and mix slowly, and after 2 min, white precipitate is produced; centrifuge at 15000g for 30 min, and carefully take the supernatant to a new tube; desalt with a C18 column, vacuum dry, and add 30 μl of 0.1% FA to redissolve; after redissolving, centrifuge at 15000g for 15 min, carefully take about 25 μL of the supernatant, and then sample, and screen differential proteins by mass spectrometry.

[0101] Dewaxing, citric acid high-temperature repair, primary antibody and secondary antibody incubation, and diaminobenzidine color development were performed according to the kit instruction. Immunohistochemical semi-quantitative analysis was recorded as 0-3 points according to the color depth of cells and the number of positive cells. If the cell nucleus or cell membrane was colored light brown, it was 1 point, brown was 2 points, dark brown was 3 points, and no color was 0 point. The proportion of positive cells to all cells in the field of view was <10% for 1 point, 10%-50% for 2 points, and >50% for 3 points. According to the integral number of the above two indexes, it was divided into 4 levels, 0 point was negative (-), 2-3 points was weakly positive (+), 4 points was positive (++), and 5-6 points was strongly positive (+++).

[0102] The samples were added to 6% gel on 10% SDS polyacrylamide gel, 5 μl of each sample was electrophoresed for 90 min. The membrane was transferred to a polyvinylidene fluoride membrane (150 min for polyvinylidene fluoride membrane with relative molecular mass greater than 130 000; 90 min for polyvinylidene fluoride membrane with relative molecular mass between 40 000 and 130 000; 45-60 min for polyvinylidene fluoride membrane with relative molecular mass below 40 000), and blocked with 5% skim milk for 1 h. The membrane with relative molecular mass between 40 000 and 130 000 was cut into two parts, i.e., a membrane with relative molecular mass between 75 000 and 130 000 and a membrane with relative molecular mass between 40 000 and 75 000. The protein antibody was diluted at 1:1000, and the anti-GAPDH antibody was diluted at 1:1000. The membrane was incubated at 4°C overnight, washed with phosphate-buffered saline for 3 times, each for 10 min, and then the secondary antibody was added to the blocking solution and incubated at room temperature for 2.5 h. The membrane was washed with phosphate-buffered saline for 3 times, each for 10 min. The color was developed by chemiluminescence reagent.

[0103] After the experiment, the tumor was stripped and observed for general morphology and weighed. The tumor tissue was divided into two parts, one part was fixed and embedded for pathological examination, and the remaining tumor tissue was stored at -80°C. Immunohistochemistry and Western blot were used to detect EMT-related proteins such as G6PD, E-cadherin, N-cadherin, Vimentin, IL-6, STAT3, and Snail.

[0104] Mass spectrometry was used to screen differential proteins in tumor samples from each group. G6PD was found to be one of the differential proteins. RT-PCR and immunohistochemistry statistical results showed that the incomplete ablation group showed a high expression trend, which may be a potential target.

[0105] Immunohistochemistry results between samples from each group showed that G6PD was lowly expressed in the HDRBT+iMWA group, which was significantly lower than that in the iMWA group, as shown in FIG. 4. The proliferation and apoptosis indicators were consistent with the treatment effect. Figure 9

[0106] Bioinformatics tools were used to analyze the differential gene G6PD according to the tumor grade. Compared with normal samples, G6PD was increased in tumor samples, and higher expression was also found in metastatic samples. The survival curves of different G6PD expression populations are shown in FIGS. 5 and 6. Figure 10 Figure 11 The correlation between G6PD level and OS in the previously generated microarray data set of 364 liver cancer patients was analyzed. The results showed that high expression of G6PD was associated with poor clinical prognosis of liver cancer patients (P<0.01). The increase in G6PD expression was also associated with tumor stage, and the higher the stage, the higher the G6PD expression.

[0107] ​​In summary, G6PD plays an important role in promoting the progression of liver cancer.

[0108] Example 4:

[0109] A dynamic weighted fusion algorithm is proposed for the complementarity of CT and ultrasound multi-modal images. The algorithm dynamically adjusts the fusion weights of CT and ultrasound images by calculating the edge definition and tissue contrast of the two images, solves the limitations of single modal images, and realizes the precise three-dimensional reconstruction of liver cancer lesions and surrounding blood vessels and bile ducts.

[0110] The core logic of the algorithm is that images with high edge definition and contrast get higher weights to ensure that the fused image considers both anatomical accuracy and real-time performance.

[0111] Enhanced CT images and real-time ultrasound images of liver cancer patients were collected, covering the tumor and surrounding 2 cm range, and the CT images were subjected to Gaussian filter denoising and Sobel operator edge enhancement, and the ultrasound images were subjected to debanding and histogram equalization to improve contrast.

[0112] The gradient amplitude mean value is used to calculate the edge definition, and the formula is:

[0113]

[0114] In the formula, is the edge definition; is the image width; is the image height; is the gradient of the pixel .

[0115] The average gray difference between the target area and the background area is calculated, and the formula is:

[0116]

[0117] In the formula, is the tissue contrast; is the number of tumor area pixels; is the pixel set of the target area; is the gray value of the tumor pixel; is the average gray value of the background area.

[0118] The edge definition and contrast of CT and ultrasound are combined to calculate the dynamic fusion weight :

[0119]

[0120] In the formula, is the edge definition of the CT image; is the edge contrast of the CT image; edge definition of the ultrasound image; edge contrast of the ultrasound image; for preventing the denominator from being 0.

[0121] The fused pixel value is:

[0122]

[0123] In the formula, is the pixel value of the fused image at coordinates ; is a dynamic fusion weight, and the value range is 0 to 1; is the pixel value of the CT image; is the pixel value of the ultrasound image.

[0124] The fused two-dimensional image sequence is imported into three-dimensional visualization software, and a three-dimensional model of a liver cancer lesion, a hepatic artery, a portal vein and a bile duct is reconstructed by using a moving cube algorithm, interactive operations such as rotation, scaling and sectioning are supported, and a file that can be imported into a treatment planning system is output.

[0125] Through 20 cases of HepG2 nude mouse orthotopic liver cancer models, the three-dimensional reconstruction error of the algorithm is , which is significantly lower than that of single-mode CT reconstruction and ultrasound reconstruction. The HDRBT catheter implantation error based on the model is reduced to , and the thermal ablation boundary coverage is increased to 95%. Clinical trials of 10 patients with liver cancer show that the complete ablation rate of the combined treatment group is 92%, and the local recurrence rate within one year is reduced to 8%.

[0126] The results show that the algorithm dynamically weights the high-resolution anatomical information of CT and the real-time guidance information of ultrasound, significantly improves the accuracy and robustness of three-dimensional reconstruction of liver cancer, and the fused three-dimensional model is superior to single-mode reconstruction technology in terms of anatomical structure restoration and real-time positioning consistency. The improvement of reconstruction accuracy directly optimizes the catheter implantation path and the positioning accuracy of the thermal ablation electrode, significantly improves the safety boundary coverage of the combined treatment, and reduces the need for intraoperative adjustment caused by image registration deviation. Ultimately, the algorithm provides a reliable spatial reference for the collaborative planning of HDRBT and thermal ablation, and the complete ablation rate and local control rate of the combined treatment in the clinical trials are systematically improved.

[0127] Example 5:

[0128] Thermal ablation can cause thermal damage to tumor tissue, increase the radiosensitivity of residual tumor cells; at the same time, the thermal coagulation layer can form a thermal protection effect, reducing the penetration of radiation to the surrounding normal tissue. This model determines the thermal ablation range and residual tumor volume through a three-dimensional visualization model, combines the radiosensitivity changes after thermal damage, optimizes the edge dose and central dose of HDRBT, and realizes the synergistic effect of thermal ablation to reduce tumor burden and HDRBT to remove residual lesions.

[0129] On the basis of the three-dimensional visualization model of the foregoing embodiment, based on the enhanced CT image after thermal ablation, the thermal ablation area is segmented by threshold segmentation method, the residual tumor area is segmented by enhanced region identification, and the total tumor volume is calculated.

[0130] Through experiments on 10 nude mice, the clonogenic survival of residual tumor cells after thermal ablation was measured, and the radiosensitization ratio of thermal damaged cells to normal cells was 1.5.

[0131] For the radiosensitization effect of residual tumor area, the optimization formula is:

[0132]

[0133] In the formula, is the edge dose; is the residual tumor volume ratio.

[0134] Based on the dose distribution characteristics of the radiation source, the optimization formula is:

[0135]

[0136] In the formula, is the central dose.

[0137] The dose distribution of the radiation source follows the inverse square law, and the central dose is 25 times the edge dose, which is a specific parameter verified by Monte Carlo simulation.

[0138] The dose distribution of HDRBT is simulated by a treatment planning system to ensure that:

[0139] The dose of the residual tumor area is covered, and 95% of the residual tumor volume receives ≥ edge dose;

[0140] The dose limit of the surrounding normal tissue is that the dose at 2 cm outside the radiation source is ≤ 2 Gy, and the maximum dose of the bile duct is ≤ 5 Gy.

[0141] Through experiments on 20 nude mice, the residual tumor volume of the combined treatment group was , which was significantly lower than that of the simple HDRBT group ; The tumor weight inhibition rate reached 78%, which was significantly higher than the 55% of the simple HDRBT group. The clinical trial of 10 patients showed that the liver and kidney function damage of the combined treatment group was 15%, and the incidence of bile duct injury decreased to 0. In addition, by analyzing the data of 364 liver cancer patients through bioinformatics tools, the 1-year survival rate of patients with low expression of G6PD treated by this model reached 85%, verifying the efficacy prediction value of the model.

[0142] The results show that the model quantifies the change of radiosensitivity of residual tumor after thermal ablation, dynamically optimizes the edge dose and central dose distribution of HDRBT. This dose strategy significantly enhances the clearance efficiency of residual lesions, and the residual tumor volume of the combined treatment group is significantly lower than that of the traditional fixed dose group, and the tumor inhibition rate shows substantial improvement. At the same time, the safety dose constraint based on the three-dimensional visualization model effectively reduces the radiation toxicity of normal tissues, showing a synergistic decrease in the degree of liver function damage and the incidence of bile duct complications. In addition, the expression level of differential protein G6PD verifies the biological rationality of dose optimization, further confirming the clinical value of the model in predicting patient survival benefits, providing a theoretical closed loop for individualized combined treatment.

Claims

1. A multimodal imageomics HDRBT combined with thermal ablation treatment for liver cancer efficacy evaluation system, characterized in that, The system comprises: a multi-modal image data acquisition module for collecting multi-modal image data, providing real-time guidance information and high-resolution anatomical structure information, and fusing the two through a dynamic weighted fusion algorithm; a three-dimensional visualization model construction module for reconstructing the three-dimensional structure of liver tumors, surrounding blood vessels and bile ducts based on the multi-modal image data through an image segmentation algorithm, and fusing tumor edge, tissue density and blood vessel distribution imageomics features to generate an interactive three-dimensional visualization model; a joint treatment plan generation module for integrating the HDRBT dosimetric parameters and the thermal ablation process parameters according to the three-dimensional visualization model, optimizing the catheter implantation path, thermal ablation electrode position and treatment sequence, and generating an individualized joint treatment plan; the joint treatment plan generation module comprises an HDRBT synergistic dose optimization model for residual tumor after thermal ablation, which dynamically adjusts the HDRBT edge dose based on the residual tumor volume ratio, and the dynamic adjustment is to dynamically adjust the edge dose and central dose distribution through a dose optimization formula according to the residual tumor volume ratio and the preset radiosensitization ratio; an efficacy evaluation module for quantitatively analyzing tumor volume changes, ablation boundary integrity and residual tumor activity through accurate registration of multi-modal images before and after treatment; a safety evaluation module for integrating liver and kidney function indicators, pathological HE staining results and adverse reaction records, and evaluating the safety of the joint treatment.

2. The system of claim 1, wherein: the multi-modal image data acquisition module fuses real-time ultrasound imaging and high-resolution CT and magnetic resonance imaging through magnetic positioning technology to achieve dynamic image alignment.

3. The system of claim 1, wherein: the dynamic weighted fusion algorithm comprises: calculating the edge sharpness and tissue contrast of CT images and ultrasound images; dynamically adjusting the fusion weights of CT images and ultrasound images according to the edge sharpness and tissue contrast, so that images with high edge sharpness and good contrast obtain higher weights; performing pixel-level fusion of CT images and ultrasound images based on the adjusted fusion weights to generate a fused image sequence; reconstructing the fused image sequence into a three-dimensional model of liver cancer lesions, hepatic arteries, portal veins and bile ducts through a three-dimensional reconstruction algorithm.

4. The system of claim 1, wherein: the three-dimensional visualization model construction module uses a deep convolutional neural network to segment liver tumors and blood vessels, and optimizes the segmentation results through a variational energy function.

5. The system of claim 1, wherein: the joint treatment plan generation module uses a topology optimization algorithm to optimize the catheter implantation path based on the three-dimensional visualization model.

6. The system of claim 1, wherein: in the dose optimization formula, the adjustment of the edge dose is positively correlated with the residual tumor volume ratio, and the central dose is a preset multiple of the edge dose.

7. The system of claim 1, wherein: the efficacy evaluation module analyzes residual tumor activity through the dynamic enhancement curve characteristics of ultrasound contrast, and judges the completeness of tumor ablation.

8. The system of claim 1, wherein: The safety evaluation module uses analytic hierarchy process to quantize liver and kidney function indexes, pathological results and adverse reactions into safety scores to evaluate safety grades.

Images