System for constructing nanoparticle tumor accumulation prediction model based on CT (Computed Tomography) and ultrasonic image
By constructing a prediction model of nanoparticle tumor accumulation based on CT and ultrasonic imaging omics technology, the differences and prediction problems of nanodrug delivery efficiency in tumors were solved, and the effects of non-invasive prediction and cost saving were achieved.
Patent Information
- Application Number
- CN202411191041.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-28
- Publication Date
- 2024-12-06
AI Technical Summary
There are obvious tumor differences in the tumor delivery efficiency of existing nanodrugs. Traditional pharmacokinetic models are difficult to accurately predict the accumulation of nanoparticles, resulting in high R&D costs and high risk of transformation failure, and lack of non-invasive and prospective prediction solutions.
The nanoparticle tumor accumulation prediction model is constructed based on CT and ultrasonic imaging omics technology, and the non-invasive prediction of the accumulation amount of nanoparticle through nanoparticle synthesis and detection, imaging omic feature extraction, feature analysis and prediction model construction is achieved.
A non-invasive prediction of the accumulation of nanoparticle tumors is achieved, and unsuitable animals can be eliminated before the injection of nanodrugs, which can improve the development speed of nanodrugs, save R&D costs, and improve the conversion success rate.
Smart Images

Figure HDA0005015106400000011 
Figure HDA0005015106400000012 
Figure HDA0005015106400000021
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of nanoparticle delivery, and in particular relates to a system for constructing a nanoparticle tumor accumulation prediction model based on CT and ultrasound images. Background Art
[0002] Nanomedicine is a type of drug loaded with nanoparticles that can achieve targeted drug delivery to tumors by taking advantage of the property of nanoparticles to selectively accumulate in tumors. However, compared with traditional small molecule preparations, nanomedicines have high R&D costs and a higher risk of clinical transformation failure. The main problem is that there are significant differences between tumors in the delivery efficiency of nanoparticles to tumors. Since tumor heterogeneity significantly affects nanoparticle accumulation, traditional pharmacokinetic models cannot accurately predict the accumulation of nanoparticles in tumors, making it difficult to stratify subjects in order to reduce R&D costs and improve the success rate of nanomedicine transformation. The nanodrug accumulation prediction model currently under development mainly relies on pathological samples and cannot non-invasively and prospectively predict the delivery efficiency of nanodrugs.
[0003] Medical imaging can non-invasively display the heterogeneity of tumors. In particular, imaging omics technology can extract quantitative features of tumor images at high throughput, and use artificial intelligence models to screen out tumor imaging features that are most closely related to the accumulation level of nanomedicines, which is expected to improve the accuracy of medical imaging in predicting the accumulation level of nanomedicines in tumors.
[0004] However, there is no research on the prediction of nanoparticle animal tumor accumulation level by imaging omics so far. Therefore, it is an urgent research topic for those skilled in the art to provide a method and tool for predicting nanoparticle animal tumor accumulation based on CT and ultrasound imaging omics technology. This tool is used for animal tumor stratification in the process of nanodrug development, thereby accelerating the speed of nanodrug development and saving development costs. Summary of the invention
[0005] The purpose of the present invention is to solve some of the problems existing in the above-mentioned prior art and to provide a system for constructing a nanoparticle tumor accumulation prediction model based on CT and ultrasound images, so as to solve the problem that there is no prospective and non-invasive prediction scheme for the existing nanoparticle animal tumor delivery.
[0006] In order to achieve the above object, the present invention adopts the following technical solution: a system for constructing a nanoparticle tumor accumulation prediction model based on CT and ultrasound images, comprising:
[0007] Nanoparticle synthesis and detection device, used to synthesize nanoparticles and obtain particle size data;
[0008] A model building module, used to build more than one tumor model and obtain tumor type data;
[0009] CT, for CT imaging of tumors;
[0010] An ultrasound imaging device, used for ultrasound imaging of tumors;
[0011] An imaging genomics feature extraction module is used to segment each tumor area from tumor images obtained from CT and ultrasound imaging devices and extract imaging genomics features;
[0012] The feature analysis module is used to analyze the extracted radiomics features, remove highly collinear features, and screen out features that are closely related to nanoparticle accumulation in tumors;
[0013] A delivery and accumulation classification module is used to deliver nanoparticles so that the nanoparticles accumulate at the tumor site, collect the tumor and detect the accumulation of nanoparticles in the tumor, and classify the tumor according to the median accumulation amount;
[0014] A prediction model module was constructed to establish a composite imaging prediction model based on nanoparticle size, tumor type, CT and ultrasound imaging features, and tumor nanoparticle accumulation data, and to evaluate the performance of the composite prediction model.
[0015] Furthermore, the synthesized nanoparticles are gold nanoparticles, silica nanoparticles, nanoliposomes or nanomicelles, the surface of the nanoparticles is modified with polyethylene glycol, the hydrated particle size of the synthesized nanoparticles is 10 to 300 nm, and the morphology is spherical, rod-shaped or multi-branched.
[0016] Furthermore, the animal type of the tumor model is rat or mouse, the tumor type is pancreatic cancer, liver cancer, breast cancer, lung cancer, bladder cancer, colon cancer or melanoma, and the tumor model site is a subcutaneous tumor or an in situ tumor.
[0017] Furthermore, each tumor region is segmented from the tumor images obtained by CT and ultrasound imaging devices, and the imaging genomics features are extracted by the following specific steps:
[0018] (1) The tumor area of the tumor image was delineated using ITK-SNAP software. The tumor area was delineated layer by layer in the CT image, and the region of interest at the layer with the largest tumor diameter was delineated in the ultrasound image;
[0019] (2) Use the Pyradiomics package in Python to extract tumor CT and ultrasound image features.
[0020] Furthermore, radiomics features include first-order statistics, shape features (2D and 3D), texture features, as well as exponential, logarithmic, square, gradient, square root, Laplace of Gaussian, and wavelet transform features.
[0021] Furthermore, the features closely related to the accumulation of nanoparticles in tumors were screened out, specifically by the following steps:
[0022] (1) Some tumor features were randomly selected from all tumor features for the second delineation. The interval between the two delineations of the region of interest was one month, and the imaging features with an ICC > 0.80 in the second delineation were included in the subsequent feature screening;
[0023] (2) Through the Pearson correlation coefficient analysis method, the feature set with correlation coefficients between image features exceeding 0.95 was simplified into a single representative feature;
[0024] (3) Through LASSO regression analysis, feature selection and coefficient shrinkage were performed on CT and ultrasound imaging features to identify features that are closely related to the prediction of nanoparticle tumor accumulation.
[0025] Furthermore, the amount of nanoparticle accumulation in the tumor was detected 24 h after the delivery of the nanoparticles.
[0026] Furthermore, the establishment of the radiomics composite prediction model adopts the following steps:
[0027] (1) Integrate tumor type, nanoparticle size, SWE average and dynamic enhancement curve quantitative parameters, CT-ultrasound imaging features, and tumor nanoparticle accumulation data;
[0028] (2) The model construction method is logistic regression model;
[0029] (3) Establishment of prediction models: Model 1 is a composite model based on radiomics, and Model 2 is a baseline model used for performance comparison.
[0030] Furthermore, the prediction model building module is also used to establish a baseline model for comparative performance.
[0031] Furthermore, the performance of the radiomics composite prediction model was evaluated using the following steps:
[0032] (1) Use ROC to evaluate the predictive performance of each model and compare the area under the curve (AUC) of each model;
[0033] (2) Use the confusion matrix to further quantify the classification performance of the model and evaluate the accuracy of the model in tumor classification;
[0034] (3) Using the DCA analysis method, by calculating the net benefit ratio, the clinical practicality or economic value of the model at different thresholds can be intuitively demonstrated, and the additional benefits that each model can bring in practical applications can be compared, so as to select the optimal prediction model.
[0035] The advantages of the present invention are: based on CT and ultrasound imaging data, the present invention realizes non-invasive prediction of nanoparticle tumor accumulation; the method can exclude animals that are not suitable for nanomedicine before injecting nanomedicine. It is expected to be used for animal tumor stratification and screening of suitable test animals in the process of nanomedicine research and development, thereby accelerating the speed of nanomedicine research and development and saving research and development costs. BRIEF DESCRIPTION OF THE DRAWINGS
[0036] Figure 1 A flow chart of a method and tool for predicting nanoparticle tumor accumulation based on CT and ultrasound imaging genomics technology provided in an embodiment of the present invention.
[0037] Figure 2 This is a statistical graph of the accumulation of gold nanoparticles of different sizes in different tumors provided by the embodiment of the present invention.
[0038] Figure 3 ROC curves and calibration curves of different prediction models provided by the embodiments of the present invention.
[0039] Figure 4 Confusion matrices of different prediction models provided by embodiments of the present invention.
[0040] Figure 5 DCA curves of different prediction models provided by the embodiments of the present invention.
[0041] Figure 6 The prediction results of typical cases on different models provided in the embodiments of the present invention. DETAILED DESCRIPTION
[0042] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. The following examples are only used to more clearly illustrate the technical solutions of the present invention, and cannot be used to limit the scope of protection of the present invention.
[0043] The present invention is not limited to the following embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely represents the selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.
[0044] The present invention is described in detail below according to specific embodiments and drawings:
[0045] Example:
[0046] In order to achieve the purpose of the present invention, this embodiment provides a method and tool for predicting nanoparticle tumor accumulation based on CT and ultrasound imaging technology, such as Figure 1 As shown, the implementation steps include the following seven links: synthesis link, modeling link, scanning link, grouping link, extraction and screening link, prediction link and evaluation link.
[0047] The synthesis step is used to obtain nanoparticles of different particle sizes;
[0048] The modeling step is used to obtain a variety of tumor models;
[0049] The scanning step is used to obtain CT and ultrasound images of the tumor;
[0050] The grouping step is used to group the nanoparticles according to their accumulation amount;
[0051] The extraction and screening steps are used for imaging omics feature extraction and screening;
[0052] The prediction link is used to construct a variety of prediction models;
[0053] The evaluation step is used to evaluate the prediction model.
[0054] The nanoparticles synthesized in the synthesis step are gold nanoparticles (GNPs). By controlling the reaction conditions, 15nm, 40nm and 70nm GNPs are synthesized respectively. In order to increase the blood circulation time of GNPs and maintain their surface potential suppression, the surface of GNPs is modified with polyethanol to make the potential close to neutrality. In addition, by detecting the hydrated particle size and transmission electron microscopy, it is ensured that the GNPs of the required size are synthesized.
[0055] The various tumor models constructed in the modeling phase used C57BL / 6 mice, and tumors were implanted subcutaneously on the back of the mice. The tumor types constructed included melanoma (B16 cell line), breast cancer (E0771 cell line), liver cancer (Hepa1-6 cell line), lung adenocarcinoma (LLC cell line), bladder cancer (MB49 cell line), colon cancer (MC38 cell line), and pancreatic cancer (PAN02 cell line).
[0056] In the scanning process of obtaining tumor CT and ultrasound images, the scanning parameters that need to be adjusted for CT scanning include layer thickness (1mm), layer spacing (0mm), tube voltage (120kVp) and tube current time product (100mAs). In ultrasound imaging, an abdominal ultrasound probe is used with a frequency of 4-13MHz; imaging modalities include B-mode US, SWE and CEUS imaging, and CEUS imaging is recorded continuously for 90 seconds. From the CEUS image, a time-intensity curve is fitted, and the following parameters are further obtained based on the curve: curve type, perfusion slope, peak intensity, peak time and area under the curve.
[0057] The grouping step uses the GNPs and tumor models synthesized in the synthesis step and the modeling step, and is completed in the scanning step. GNPs of different particle sizes are injected into mice with different tumor types by tail vein injection. After 24 hours, the tumors are peeled off, and the gold mass in each tumor is obtained by inductively coupled plasma optical emission spectroscopy. The GNP accumulation amount of each tumor is calculated according to the formula: intratumoral accumulation amount (%ID / g) = tumor gold mass / (intravenously injected gold mass×tumor mass)×100%. According to the median of the accumulation amount, the tumors are divided into a high accumulation group and a low accumulation group ( Figure 2 ).
[0058] The extraction and screening phase uses the CT and ultrasound images obtained in the scanning phase to manually outline the tumor region of interest (ROI) using the open source software ITK-SNAP. After the outline is completed, the original tumor image and mask file are stored. The open source Pyradiomics package is used to extract radiomic features from each tumor ROI, including first-order statistics, shape features (including 2D and 3D), texture features, and exponential, logarithmic, square, gradient, square root, Gaussian Laplace and wavelet transform features. After the extraction is completed, each radiomic feature is standardized by the standard score (Z-Score) method. The Z-Score calculation is expressed as: Z = X-μσ, where X is the value to be calculated, μ is the mean value of the eigenvalue, and σ is the standard deviation of the eigenvalue. In order to evaluate the internal consistency of the observer, some tumors were randomly selected from all tumors for two outlines, and the two ROI outlines were separated by one month. The radiomic features with ICC>0.80 were included in the subsequent feature screening. By using the Pearson correlation coefficient analysis method, the feature set with correlation coefficients between imaging features exceeding 0.95 was simplified into a single representative feature to reduce multicollinearity. Using LASSO regression analysis, a few CT and ultrasound imaging features that are most critical for nanoparticle tumor accumulation were screened out, thereby improving the interpretability and predictive performance of the model.
[0059] The prediction step constructs multiple prediction models. First, all mouse tumor data are randomly divided into training set and test set (7:3); the model is constructed using the logistic regression method; and the prediction model is established according to the source of features: Model 1 is based on the radiomics composite model (tumor type, nanoparticle size, SWE average value and dynamic enhancement curve quantitative parameters, CT-US radiomics features), and Model 2 is the baseline model (tumor category, nanoparticle size).
[0060] The evaluation phase evaluates various prediction models established by different methods, such as Figure 3 As shown in , ROC is used to evaluate the predictive performance of each model, and the models are compared based on the area under the curve; Figure 4 As shown in Figure 2, the confusion matrix is used to further quantify the classification performance of the model and evaluate the accuracy of the model in tumor classification; Figure 5 As shown in the figure, the DCA curve is used for evaluation. By calculating the net benefit ratio, the clinical practicality or economic value of the model at different thresholds can be intuitively demonstrated, and the additional benefits that each model can bring in actual application can be compared, so as to select the most optimized prediction model.
[0061] Based on the prediction model established above, in order to compare the effectiveness of different models for tumor stratification, two cases (A and B) were used as demonstrations ( Figure 6 ). The radiomics composite model made correct predictions, while the baseline model made incorrect predictions. The radiomics composite model accurately stratified animal tumors by effectively combining cancer type with radiomics features such as US Radscore, CT Radscore, SWE average, and peak intensity of the time-intensity curve. In contrast, the baseline model relied only on GNP size and cancer type, resulting in incorrect predictions. This highlights the importance of combining radiomics features to improve model performance.
[0062] In summary, the method and tool for predicting nanoparticle animal tumor delivery based on CT and ultrasound imaging omics technology provided in this embodiment have the following technical effects:
[0063] This embodiment provides a new scheme for predicting the tumor delivery efficiency of gold nanoparticles in mouse tumors based on radiomics technology, including synthesis, modeling, scanning, grouping, extraction and screening, prediction and evaluation. Through these links, the radiomics features of CT and ultrasound of mouse tumors can be extracted, and then the mouse tumors can be classified according to the model to distinguish between mouse tumors with high and low accumulation of gold nanoparticles, thereby achieving the purpose of non-invasively predicting the tumor delivery efficiency of nanoparticles, and being used for animal model stratification in the process of nano-drug research and development, which can speed up the development of nano-drugs and save research and development costs.
[0064] As used in this article, the term "radiomics" refers to the extraction of a large number of quantitative features (such as shape features, texture features, etc.) of regions of interest from medical images (such as CT images, ultrasound images), and the application of statistical and machine learning methods to analyze them to find patterns related to outcomes (such as tumor growth rate, treatment response, survival rate, etc.).
[0065] As used herein, the term "nanoparticle" refers to tiny particles with a size of 10-300 nm. They can be made of various materials, including metals, semiconductors, polymers, and biomolecules. The following is a further description of some of the nanoparticles covered by the present invention:
[0066] Nanoparticle 1: As used herein, the term "gold nanoparticle" refers to an aggregate composed of gold atoms. The size is between 10-300 nm. Based on the size effect and surface plasmon resonance phenomenon, gold nanoparticles have unique optical, electrical and catalytic properties.
[0067] Nanoparticles 2: As used herein, the term "silicon dioxide nanoparticles" refers to nanoparticles made of silicon dioxide (SiO 2 ) are tiny particles. The size is between 10-300nm. Silica nanoparticles have a porous structure, stable chemical properties, good biocompatibility, and an easily modified surface.
[0068] Nanoparticle 3: As used herein, the term "nanoliposome" refers to a closed spherical structure composed of a phospholipid bilayer. The size ranges from 10 to 300 nanometers. Nanoliposomes mimic the structure of biological membranes, with an aqueous phase space inside and a hydrophobic phospholipid bilayer outside. They have good biocompatibility, targeting and controlled release.
[0069] Nanoparticle 4: As used herein, the term "nano micelle" refers to an aggregate formed spontaneously in water by amphiphilic molecules (i.e., molecules with a hydrophilic end and a hydrophobic end). The size is between 10 and 300 nanometers. Its structure is similar to that of a micro soap bubble, with its hydrophobic end gathering inward to form a core, while the hydrophilic end contacts the water outward to form a stable interface. Nano micelles have the characteristics of self-assembly, uniform size, and drug-carrying capacity.
[0070] As used herein, the term "accumulation" in the field of nanomedicine refers to the concentrated distribution of nanoparticles in a specific biological environment or target site. In tumor diagnosis and treatment, high accumulation of nanoparticles means that more of them reach the tumor area, which is used to enhance the diagnosis and treatment effect while reducing side effects on normal tissues. This process not only depends on the physical and chemical properties of the nanoparticles themselves, but is also closely related to the physiological and pathological characteristics of the tumor.
[0071] As used herein, the term "prediction" refers to a guess about the amount of nanoparticle tumor accumulation, and for the purpose of the present invention, refers to a guess about the amount of nanoparticle accumulation (eg, high accumulation, low accumulation) at the tumor site.
[0072] As used herein, the term "hydrated particle size" refers to the particle size measured after the particles are adsorbed in an aqueous solution to form a hydration layer on their surface. In the present invention, when nanoparticles are dispersed in water, water molecules will form one or more hydration shells around the surface of the particles, which increases the actual size of the particles. The measurement of hydrated particle size reflects the true dispersion state of nanoparticles and the dynamic size in aqueous solution, and is closely related to the stability, dispersibility and behavior of nanoparticles in vivo.
[0073] As used herein, the term "polyethylene glycol" refers to a polyol composed of repeating ethylene oxide units (-CH 2 -CH 2 -O-) linear polymer (Polyethylene Glycol, referred to as PEG). The molecular weight of PEG ranges from a few hundred to several million Daltons, depending on the number of repeating units. PEG is a colorless, odorless, non-toxic water-soluble polymer with excellent chemical stability and biocompatibility. In the present invention, PEG is modified onto the surface of nanoparticles to increase their stability and circulation time, reduce the rate of clearance, and improve the bioavailability of nanoparticles.
[0074] As used herein, the term "stratification" refers to dividing things into different levels or categories according to certain standards or attributes. For the purpose of the present invention, it means dividing tumors into high accumulation group and low accumulation group according to the amount of nanoparticle accumulation in the tumor.
[0075] As used herein, the term "CT" refers to Computed Tomography, which is a medical imaging method that uses X-rays and computer processing technology to generate images of tissue structures based on differences in tissue density.
[0076] As used herein, the term "US" refers to ultrasound imaging, which is a medical imaging method that uses the differences in propagation speed and reflection characteristics of high-frequency sound waves in different tissues of the human body to generate tissue structure images.
[0077] As used herein, the term "B-mode US" refers to B-mode grayscale imaging, which is the most common type of ultrasound imaging. It is an imaging method that generates the brightness (i.e., grayscale value) of each point in the image based on the intensity of the signal reflected back when the ultrasound passes through human tissue.
[0078] As used herein, the term “SWE” refers to Shear Wave Elastography (SWE), which is an imaging method for quantifying the stiffness of tissue based on the propagation velocity of shear waves.
[0079] As used herein, the term "CEUS" refers to contrast-enhanced ultrasound, which is an imaging method that uses microbubble contrast agents to enhance the quality of ultrasound imaging, mainly used to improve the visibility of vascular structures and reflect the blood supply of tissues.
[0080] It should be understood that the various technical means mentioned above can be implemented separately in the form of hardware or software, or the integration of the two. Therefore, the method and apparatus of the present invention, as well as several constituent elements or aspects thereof, can be built into a physical medium, such as a program code (i.e., a series of instructions) carried by a floppy disk, CD-ROM, hard disk, or any other machine-readable storage medium. When this program is loaded onto a device such as a computer and executed, the device is converted into a tool for implementing the present invention.
[0081] The contents shown and described above reveal the basic principles, core features and advantages of the present invention. It should be known that for those skilled in the art, the scope of the present invention is not limited to the above exemplary embodiments. These examples and explanatory documents are intended to illustrate the core principles of the present invention, and without violating the core spirit and overall framework of the present invention, the present invention has a variety of variations and optimization space, all of which fall within the scope of the protection sought by the present invention, and the specific scope of protection is defined by the attached claims and their equivalent expressions.
Claims
1. A system for constructing a nanoparticle tumor accumulation prediction model based on CT and ultrasound images, characterized in that include: Nanoparticle synthesis and detection device, used to synthesize nanoparticles and obtain particle size data; A model building module, used to build more than one tumor model and obtain tumor type data; CT, for CT imaging of tumors; An ultrasound imaging device, used for ultrasound imaging of tumors; An imaging genomics feature extraction module is used to segment each tumor area from tumor images obtained from CT and ultrasound imaging devices and extract imaging genomics features; The feature analysis module is used to analyze the extracted radiomics features, remove highly collinear features, and screen out features that are closely related to nanoparticle accumulation in tumors; A delivery and accumulation classification module is used to deliver nanoparticles so that the nanoparticles accumulate at the tumor site, collect the tumor and detect the accumulation of nanoparticles in the tumor, and classify the tumor according to the median accumulation amount; A prediction model module was constructed to establish a composite imaging prediction model based on nanoparticle size, tumor type, CT and ultrasound imaging features, and tumor nanoparticle accumulation data, and to evaluate the performance of the composite prediction model.
2. The system for constructing a nanoparticle tumor accumulation prediction model based on CT and ultrasound images according to claim 1, characterized in that: The synthesized nanoparticles are gold nanoparticles, silicon dioxide nanoparticles, nanoliposomes or nanomicelles. The surface of the nanoparticles is modified with polyethylene glycol. The hydrated particle size of the synthesized nanoparticles is 10-300nm and the shape is spherical, rod-shaped or multi-branched.
3. The system for constructing a nanoparticle tumor accumulation prediction model based on CT and ultrasound images according to claim 1, characterized in that: The animal type of the tumor model is rats or mice, the tumor type is pancreatic cancer, liver cancer, breast cancer, lung cancer, bladder cancer, colon cancer or melanoma, and the tumor model site is a subcutaneous tumor or an in situ tumor.
4. The system for constructing a nanoparticle tumor accumulation prediction model based on CT and ultrasound images according to claim 1, characterized in that: The steps of segmenting each tumor region from the tumor image obtained by CT and ultrasound imaging devices and extracting imaging genomics features are as follows: (1) The tumor area of the tumor image was delineated using ITK-SNAP software. The tumor area was delineated layer by layer in the CT image, and the region of interest at the layer with the largest tumor diameter was delineated in the ultrasound image; (2) Use the Pyradiomics package in Python to extract tumor CT and ultrasound image features.
5. The system for constructing a nanoparticle tumor accumulation prediction model based on CT and ultrasound images according to claim 1, characterized in that: The radiomics features include first-order statistics, shape features (2D and 3D), texture features, and exponential, logarithmic, square, gradient, square root, Laplace of Gaussian, and wavelet transform features.
6. The system for constructing a nanoparticle tumor accumulation prediction model based on CT and ultrasound images according to claim 1, characterized in that: The screening of features closely related to nanoparticle accumulation in tumors specifically adopts the following steps: (1) Some tumor features were randomly selected from all tumor features for the second delineation. The interval between the two delineations of the region of interest was one month, and the imaging features with an ICC > 0.80 in the second delineation were included in the subsequent feature screening; (2) Through the Pearson correlation coefficient analysis method, the feature set with correlation coefficients between image features exceeding 0.95 was simplified into a single representative feature; (3) Through LASSO regression analysis, feature selection and coefficient shrinkage were performed on CT and ultrasound imaging features to identify features that are closely related to the prediction of nanoparticle tumor accumulation.
7. The system for constructing a nanoparticle tumor accumulation prediction model based on CT and ultrasound images according to claim 1, characterized in that: The detection of the nanoparticle accumulation in the tumor was performed 24 hours after the delivery of the nanoparticles.
8. The system for constructing a nanoparticle tumor accumulation prediction model based on CT and ultrasound images according to claim 1, characterized in that: The establishment of the radiomics composite prediction model adopts the following steps: (1) Integrate tumor type, nanoparticle size, SWE average and dynamic enhancement curve quantitative parameters, CT-ultrasound imaging features, and tumor nanoparticle accumulation data; (2) The model construction method is logistic regression model; (3) Establishment of prediction models: Model 1 is a composite model based on radiomics, and Model 2 is a baseline model used for performance comparison.
9. The system for constructing a nanoparticle tumor accumulation prediction model based on CT and ultrasound images according to claim 1, characterized in that: The prediction model building module is also used to establish a baseline model for comparative performance.
10. The system for constructing a nanoparticle tumor accumulation prediction model based on CT and ultrasound images according to claim 1, characterized in that: The performance evaluation of the radiomics composite prediction model was performed using the following steps: (1) Use ROC to evaluate the predictive performance of each model and compare the area under the curve (AUC) of each model; (2) Use the confusion matrix to further quantify the classification performance of the model and evaluate the accuracy of the model in tumor classification; (3) Using the DCA analysis method, by calculating the net benefit ratio, the clinical practicality or economic value of the model at different thresholds can be intuitively demonstrated, and the additional benefits that each model can bring in practical applications can be compared, so as to select the optimal prediction model.
Citation Information
Cited By
System for constructing prediction model for tumor accumulation of nano-particles based on CT and ultrasound imaging
WO2026045121A1
