Pathological complete remission prediction method and device based on difference radiomics

Abstract

The invention discloses a pathological complete remission prediction method and device based on difference radiomics, and relates to the field of medical images.The method comprises the steps that a preoperative medical image and preoperative clinical data of a patient to be predicted are obtained; inputting the preoperative medical image and the preoperative clinical data into the trained pathology complete remission prediction model, and outputting a pathology complete remission result of the patient to be predicted by the pathology complete remission prediction model; the pathology complete remission prediction model is constructed by a joint prediction model and prediction factors screened from sample clinical data, and the joint prediction model is constructed by utilizing pre-treatment radiomics characteristics, post-treatment radiomics characteristics and difference-radiomics characteristics. Whether a non-small cell lung cancer patient receiving neoadjuvant immunochemotherapy reaches complete pathological remission or not is represented in a non-invasive mode, so that medical staff are assisted in formulating a personalized treatment scheme, and the clinical result of the patient is improved.

Description

Technical Field

[0001] The present invention relates to the field of medical imaging, and particularly to a method and device for predicting pathological complete remission based on differential radiomics. Background Art

[0002] Lung cancer is a common malignant tumor globally. Among them, non-small cell lung cancer (NSCLC) is the most common type of lung cancer. The treatment regimens for NSCLC include surgery, radiotherapy, and systemic treatment, etc. Its treatment has entered the era of precision medicine, emphasizing the formulation of individualized treatment plans for patients based on staging, histological type, and molecular characteristics.

[0003] Neoadjuvant immunochemotherapy (NIT) in neoadjuvant treatment is the combined use of immune checkpoint inhibitors (ICIs) and chemotherapy before surgery, and has become a promising treatment regimen for improving the pathological remission rate of NSCLC patients.

[0004] Accurately predicting the treatment effect of patients after NIT before surgery has important clinical value, and can also assist medical staff in formulating individualized treatment plans for patients to ensure the most appropriate and effective therapy for patients. This prediction result can not only improve the cure success rate of patients, but also reduce unnecessary ineffective treatments for patients. In addition, early identification of patients who do not respond well to NIT can also avoid the implementation of unnecessary NIT treatment regimens.

[0005] Although NIT has shown great promise in NSCLC, predicting the treatment effect of each patient after NIT before surgery is still a clinical challenge. Pathological complete response (pCR) is significantly associated with good prognosis of neoadjuvant immunochemotherapy for various cancers (including non-small cell lung cancer), and can currently be used as a short-term target endpoint for NIT clinical trials, and is defined as no residual tumor cells after evaluation of the resected tumor tissue and regional lymph nodes. Preoperative evaluation often cannot accurately judge, and currently still relies on postoperative pathological confirmation, because the traditional imaging evaluation method based on tumor size changes may not accurately predict the above treatment effect.

[0006] Predicting pCR status before surgery helps select the best treatment plan for patients. For example, for patients who fail to achieve pCR, invasive surgery can be considered to achieve the treatment goal; for patients who can achieve pCR, conservative treatment can be chosen to avoid the harm caused by unnecessary invasive treatment. Similarly, the current determination of a patient's pCR also relies on postoperative pathological confirmation. It is of great clinical significance if it can be accurately identified whether a patient can achieve pCR after NIT before the patient undergoes NIT.

[0007] Therefore, how to provide a non-invasive, preoperative prediction method for pCR of NSCLC to assist medical staff in determining surgical indications and treatment decisions is an important issue that needs to be urgently addressed in the industry. Summary of the invention

[0008] In view of this, the embodiments of the present invention provide a method and device for predicting pathological complete remission based on differential imaging genomics, so as to solve the problem that a non-invasive and preoperative method is urgently needed for predicting pCR of NSCLC.

[0009] According to a first aspect, an embodiment of the present invention provides a method for predicting pathological complete remission based on differential radiomics, the method comprising: Obtain preoperative medical images and preoperative clinical data of the patient to be predicted; Inputting preoperative medical images and preoperative clinical data into the trained pathological complete remission prediction model, the pathological complete remission prediction model outputs the pathological complete remission result of the patient to be predicted; The pathological complete remission prediction model is constructed by a combined prediction model and prediction factors screened from sample clinical data, and the combined prediction model is constructed using pre-treatment radiomics features, post-treatment radiomics features, and difference-radiomics features; The radiomic features before treatment and the radiomic features after treatment are radiomic features extracted from sample medical images based on the region of interest and the lesion to be analyzed; The difference-radiomic features are obtained based on the radiomic features before and after treatment. The difference-radiomic features include: The first difference-radiomic feature used to characterize the absolute change value of the imaging radiomic features before and after neoadjuvant immunochemotherapy, the second difference-radiomic feature used to characterize the percentage change rate of the imaging radiomic features before and after neoadjuvant immunochemotherapy, and the third difference-radiomic feature used to characterize the average percentage change rate of the imaging radiomic features before and after neoadjuvant immunochemotherapy per chemotherapy cycle.

[0010] In combination with the first aspect, in the first implementation manner of the first aspect, the pathological complete remission prediction model is trained through the following steps: Obtain the sample medical images of the sample patients; the sample medical images include contrast-enhanced computed tomography before neoadjuvant immunochemotherapy and contrast-enhanced computed tomography images after neoadjuvant immunochemotherapy; Retain the images that meet the preset criteria in the sample medical images to obtain the medical images to be analyzed; Obtain the sample clinical data of the sample patients and the preset imaging markers, perform differential analysis and logistic regression analysis on the sample clinical data according to the preset imaging markers, screen the prediction factors, and determine the lesions to be analyzed according to the prediction factors; the significance level value of the screened prediction factors is lower than the preset level value; Based on the region of interest, extract the lesion regions containing the lesions to be detected that represent the venous phase and the arterial phase from the images to be analyzed respectively, and use the lesion regions representing the venous phase to perform matching and calibration of the venous phase and arterial phase contours to obtain the sample target lesion regions of the images to be analyzed; Extract the radiomics features from the sample target lesion regions to obtain the pre-treatment radiomics features and the post-treatment radiomics features; Based on the pre-treatment and post-treatment radiomics features, construct the difference-radiomics features; Screen out the model training features from the pre-treatment radiomics features, the post-treatment radiomics features and the difference-radiomics features, and train the combined prediction model according to the model training features; Establish a mapping relationship between the trained combined prediction model and the treatment effect after the sample patients receive neoadjuvant immunochemotherapy according to the prediction factors to obtain the pathological complete remission prediction model.

[0011] In combination with the first implementation manner of the first aspect, in the second implementation manner of the first aspect, the obtaining of the sample clinical data of the sample patients and the preset imaging markers, performing differential analysis on the sample clinical data according to the preset imaging markers, extracting the prediction factors, and determining the lesions to be analyzed according to the prediction factors specifically include: Obtain the sample clinical data of the sample patients and screen and determine the preset imaging markers; Perform differential analysis and logistic regression analysis on the sample clinical data according to the preset imaging markers, screen the sample clinical features with significance level values lower than the preset level value from the sample clinical data, and use the sample clinical features as the prediction factors; Determine the lesions to be analyzed from the preset imaging markers according to the prediction factors.

[0012] Combined with the first implementation of the first aspect, in the third implementation of the first aspect, screening out model training features from the pre-treatment radiomics features, post-treatment radiomics features, and difference-radiomics features, and training a joint prediction model based on the model training features specifically includes: Perform standardization processing on the pre-treatment radiomics features, post-treatment radiomics features, and difference-radiomics features to obtain standardized features; Use the Synthetic Minority Over-sampling Technique (SMOTE) to perform sample balancing processing on the standardized features to obtain balanced features, and delete the features with a variance of 0 and a Spearman rank correlation coefficient exceeding the preset coefficient in the balanced features; Use the maximum relevance - minimum redundancy algorithm to screen out preliminary training features from the balanced features; Use the Least Absolute Shrinkage and Selection Operator (LASSO) algorithm to screen out model training features from the preliminary training features; Based on the model training features and using the logistic regression method, train to obtain a joint prediction model.

[0013] Combined with the first implementation of the first aspect, in the fourth implementation of the first aspect, based on the region of interest, respectively extract the lesion regions containing the lesion to be detected that characterize the venous phase and arterial phase from the images to be analyzed, and use the lesion region characterizing the venous phase to perform matching calibration of the venous phase and arterial phase contours to obtain the sample target lesion region of the images to be analyzed, specifically including: Based on the first preset window, determine the corresponding plain chest window image from each image to be analyzed, and, based on the second preset window, determine the corresponding enhanced mediastinal window image from each image to be analyzed; Based on the region of interest, respectively determine the first target lesion region and the second target lesion region containing the lesion to be detected from the plain chest window image and the enhanced mediastinal window image; the first target lesion region and the second target lesion region respectively contain the first target lesion and the surrounding region within the first preset range centered on the first target lesion, the second target lesion and the surrounding region within the first preset range centered on the second target lesion; the first target lesion is the largest lesion in the plain chest window image, and the second target lesion is the largest lesion in the enhanced mediastinal window image; Eliminate the non-target regions contained in the first target lesion region; Determine the position information of the first target lesion and the second target lesion corresponding to each image to be analyzed respectively, and based on the position information and taking the first target lesion region as the reference, perform matching calibration of the lesion region contours of each image to be analyzed to obtain the sample target lesion region of each image to be analyzed.

[0014] In combination with the fourth implementation manner of the first aspect, in the fifth implementation manner of the first aspect, based on the region of interest, respectively extracting lesion regions containing lesions to be detected that characterize the venous phase and the arterial phase from the image to be analyzed, and using the lesion region characterizing the venous phase to perform matching calibration of the venous phase and the arterial phase contours to obtain a sample target lesion region of the image to be analyzed, further comprising the following steps: Determine the intragroup correlation coefficient of the sample target lesion area, and eliminate the sample target lesion area with a correlation lower than the preset correlation based on the intragroup correlation coefficient.

[0015] In combination with the first implementation of the first aspect, in a sixth implementation of the first aspect, the preset standard is: Patients received only neoadjuvant immunochemotherapy for non-small cell lung cancer; the time interval between the baseline computed tomography scan and the first treatment did not exceed the first preset interval; the time interval between the preoperative computed tomography scan and the surgery did not exceed the second preset interval; there were no conditions in the sample medical images that hindered tumor segmentation and feature extraction.

[0016] According to the second aspect, an embodiment of the present invention further provides a pathological complete remission prediction device based on differential radiomics, the device comprising: A data acquisition module, used to obtain preoperative medical images and preoperative clinical data of the patient to be predicted; A model prediction module is used to input preoperative medical images and preoperative clinical data into a trained pathological complete remission prediction model, and the pathological complete remission prediction model outputs the pathological complete remission result of the patient to be predicted; The pathological complete remission prediction model is constructed by a combined prediction model and prediction factors screened from sample clinical data, and the combined prediction model is constructed using pre-treatment radiomics features, post-treatment radiomics features, and difference-radiomics features; The radiomic features before treatment and the radiomic features after treatment are radiomic features extracted from sample medical images based on the region of interest and the lesion to be analyzed; The difference-radiomic features are obtained based on the radiomic features before and after treatment. The difference-radiomic features include: The first difference-radiomic feature used to characterize the absolute change value of the imaging radiomic features before and after neoadjuvant immunochemotherapy, the second difference-radiomic feature used to characterize the percentage change rate of the imaging radiomic features before and after neoadjuvant immunochemotherapy, and the third difference-radiomic feature used to characterize the average percentage change rate of the imaging radiomic features before and after neoadjuvant immunochemotherapy per chemotherapy cycle.

[0017] According to a third aspect, an embodiment of the present invention further provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, the steps of any one of the above-mentioned pathological complete remission prediction methods based on differential radiomics are implemented.

[0018] According to a fourth aspect, an embodiment of the present invention further provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of any one of the above-mentioned pathological complete remission prediction methods based on differential radiomics are implemented.

[0019] The pathological complete remission prediction method and device based on differential radiomics of the present invention segment and extract radiomics features for the sample image data before and after NIT treatment for the regions of interest and the lesions to be analyzed obtained based on the prediction factors, so as to obtain the radiomics features before and after treatment. Then, three difference-radiomics features are constructed according to the radiomics features before and after treatment. The maximum relevance-minimum redundancy algorithm (mRMR) and the least absolute shrinkage and selection operator (LASSO) are used to screen out meaningful features from these features. Then, the screened features and independent prediction factors are used to construct the final pathological complete remission prediction model. The prediction factors are obtained by combining sample clinical data and conventional imaging markers. The pathological complete remission prediction model constructed in this way combines two types of features, namely difference-radiomics and clinical features, and the combination of these two types of features provides complementary feature supplementation. Compared with constructing a model using only single-modal information, it can provide better prediction performance. The combination of these two types of features provides complementary feature supplementation, significantly improving the prediction accuracy. The pathological complete remission prediction model non-invasively characterizes whether NSCLC patients receiving NIT achieve pCR, so as to assist medical staff in formulating personalized treatment plans and ultimately improving the clinical outcomes of non-small cell lung cancer patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] The features and advantages of the present invention will be more clearly understood by referring to the accompanying drawings. The drawings are schematic and should not be construed as limiting the present invention in any way. In the drawings: Figure 1 A schematic flowchart of the pathological complete remission prediction method based on differential radiomics provided by the present invention is shown; Figure 2 A curve graph of the mean square error during the construction process of the pathological complete remission prediction model in the pathological complete remission prediction method based on differential radiomics provided by the present invention is shown; Figure 3 A curve graph of the regression coefficient during the construction process of the pathological complete remission prediction model in the pathological complete remission prediction method based on differential radiomics provided by the present invention is shown; Figure 4 Shows the bar chart of feature coefficients in the process of constructing the pathological complete remission prediction model in the pathological complete remission prediction method based on differential imaging omics provided by the present invention; Figure 5 Shows the ROC curve of the training set in the process of constructing the pathological complete remission prediction model in the pathological complete remission prediction method based on differential imaging omics provided by the present invention; Figure 6 Shows the ROC curve of the test set in the process of constructing the pathological complete remission prediction model in the pathological complete remission prediction method based on differential imaging omics provided by the present invention; Figure 7 Shows the ROC curve of the validation set in the process of constructing the pathological complete remission prediction model in the pathological complete remission prediction method based on differential imaging omics provided by the present invention; Figure 8 Shows the DCA curve of the training set in the process of constructing the pathological complete remission prediction model in the pathological complete remission prediction method based on differential imaging omics provided by the present invention; Figure 9 Shows the DCA curve of the test set in the process of constructing the pathological complete remission prediction model in the pathological complete remission prediction method based on differential imaging omics provided by the present invention; Figure 10 Shows the DCA curve of the validation set in the process of constructing the pathological complete remission prediction model in the pathological complete remission prediction method based on differential imaging omics provided by the present invention; Figure 11 Shows the structural schematic diagram of the pathological complete remission prediction device based on differential imaging omics provided by the present invention; Figure 12 Shows the hardware structural schematic diagram of the electronic device provided by the present invention. Detailed implementation manners

[0021] To make the objectives, 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 with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0022] Lung cancer is a common malignant tumor globally. Among them, NSCLC is the most common type of lung cancer. The treatment regimens for NSCLC include surgery, radiotherapy, and systemic treatment, etc. Its treatment has entered the era of precision medicine, emphasizing the formulation of individualized treatment plans for patients based on staging, histological type, and molecular characteristics.

[0023] NIT in neoadjuvant therapy is the combined use of ICIs and chemotherapy before surgery, which has become a promising treatment option to improve the pathological remission rate of NSCLC patients.

[0024] Accurately predicting the treatment effect of patients undergoing NIT before surgery has important clinical value and can also assist medical staff in formulating personalized treatment plans for patients to ensure the most appropriate and effective therapy for patients. This prediction result can not only improve the cure success rate of patients but also reduce unnecessary ineffective treatments for patients. In addition, early identification of patients who do not respond well to NIT can also avoid the implementation of unnecessary NIT treatment regimens.

[0025] Although NIT has shown great promise in NSCLC, predicting the treatment effect of each patient undergoing NIT before surgery remains a clinical challenge. pCR is significantly associated with good prognoses of neoadjuvant immunochemotherapy for various cancers, including non-small cell lung cancer, and can currently be used as a short-term objective endpoint in NIT clinical trials and is defined as no residual tumor cells after evaluation of resected tumor tissue and regional lymph nodes. Preoperative evaluation often cannot accurately judge and currently still relies on postoperative pathological confirmation because traditional imaging evaluation methods based on tumor size changes may not accurately predict the above treatment effect.

[0026] Preoperative prediction of pCR status helps select the best treatment plan for patients. For example, for patients who fail to achieve pCR, invasive surgery can be considered to achieve the treatment goal; while for patients who can achieve pCR, conservative treatment can be selected to avoid the harm caused by unnecessary invasive treatment. Similarly, the current determination of pCR in patients also relies on postoperative pathological confirmation. It is of great clinical significance to accurately identify whether a patient can achieve pCR after NIT before the patient undergoes NIT.

[0027] In summary, how to provide a non-invasive, preoperative prediction method for pCR in NSCLC to assist doctors and patients in determining surgical indications and treatment decisions is an important issue that the industry urgently needs to solve.

[0028] Currently, the main methods for preoperative prediction of pCR in NSCLC patients after NIT include the following: Medical imaging: It can be used as a potential biomarker by providing rich quantitative data; Radiomics: By extracting high-dimensional features from medical images such as computed tomography (CT), magnetic resonance imaging (MRI), and positron emission tomography (PET) scans, and based on these high-dimensional features, evaluating tumor invasiveness, characterizing tumor microenvironment, and predicting treatment response, traditional radiomics analysis usually relies on imaging at a single time point, which may lead to ignoring the dynamic changes of tumor characteristics after treatment; Delta-radiomics: By analyzing the temporal changes of radiomic features at multiple time points, it provides a non-invasive method to capture tumor changes caused by NIT and predict clinical outcomes.

[0029] However, the use of any of the above methods alone can no longer meet the industry's increasing demand for pCR prediction accuracy.

[0030] In order to solve the above problems, a pathological complete remission prediction method based on differential radiomics is provided in this embodiment, aiming to provide a model for pCR prediction that combines differential radiomics and clinical characteristics, evaluates the dynamic changes of characteristics before and after treatment by affecting omics characteristics, and then combines clinical data and conventional imaging marker analysis to obtain independent predictive factors, and finally establishes a pathological complete remission prediction model. The prediction results of this prediction model characterize whether NSCLC patients receiving NIT have achieved pCR in a non-invasive manner, so as to assist medical staff in formulating personalized treatment plans and improve patients' clinical outcomes. The pathological complete remission prediction method based on differential radiomics in the embodiment of the present invention can be used in electronic devices, including but not limited to computers, mobile terminals, etc. Figure 1 is a flow chart of a method for predicting pathological complete remission based on differential radiomics according to an embodiment of the present invention, such as Figure 1 As shown, the method may include the following steps: S10, obtaining preoperative medical images and preoperative clinical data of the patient to be predicted. In this embodiment, the preoperative medical images are specifically contrast-enhanced (CE) CT images of the patient before NIT, that is, preoperative contrast-enhanced CT images.

[0031] Preoperative medical images may be stored in advance in the electronic device, or may be acquired by the electronic device from the outside, for example, the electronic device acquires image data from an external acquisition device, or the electronic device acquires the image data from related hospitals, medical institutions, etc.

[0032] There is no specific restriction on the specific acquisition form of preoperative medical images here, as long as it is ensured that the electronic device can acquire preoperative medical images.

[0033] Preoperative clinical data include the type of immune checkpoint inhibitor (ICI), patient age, number of treatment cycles, gender, smoking history, pathological classification (squamous cell carcinoma, adenocarcinoma, and other non-small cell lung cancers), clinical stage, serological markers (such as cytokeratin 19 fragment, neuron-specific enolase, cancer antigen 125, cancer antigen 199, carcinoembryonic antigen, squamous cell carcinoma antigen, etc.), and immune response evaluation criteria in solid tumors (iRECIST), etc. Among them, the above-mentioned preoperative clinical data are obtained from the medical records, pathological records, laboratory test results, imaging examination results, etc. of these sample patients.

[0034] It can be understood that the patient to be predicted must be pathologically diagnosed as NSCLC.

[0035] S20. Input the preoperative medical images and preoperative clinical data into the trained pathological complete response prediction model, and the pathological complete response result of the patient to be predicted is output by the pathological complete response prediction model.

[0036] Among them, the pathological complete response prediction model is constructed by a combined prediction model and prediction factors screened from sample clinical data. The combined prediction model is constructed using pre-treatment radiomics features, post-treatment radiomics features, and difference-radiomics features. Specifically, the pre-treatment radiomics features and post-treatment radiomics features are radiomics features extracted from sample medical images based on the region of interest and the lesion to be analyzed. The difference-radiomics features include the first difference-radiomics feature, the second-radiomics feature, and the third-radiomics feature. Among them, the first difference-radiomics feature (Delta 1) is used to characterize the absolute change value of radiomics features before and after neoadjuvant immunochemotherapy, and its calculation method is the pre-treatment radiomics feature minus the post-treatment radiomics feature; the second difference-radiomics feature (Delta 2) is used to characterize the percentage change rate of radiomics features before and after neoadjuvant immunochemotherapy, and its calculation method is the difference obtained by subtracting the post-treatment radiomics feature from the pre-treatment radiomics feature and then dividing by the pre-treatment radiomics feature; the third difference-radiomics feature (Delta 3) is used for the average percentage change rate of radiomics features per chemotherapy cycle before and after neoadjuvant immunochemotherapy, and its calculation method is the difference obtained by subtracting the post-treatment radiomics feature from the pre-treatment radiomics feature divided by the product of the pre-treatment radiomics feature, that is: (pre - post) / (pre × number of cycles of neoadjuvant immunochemotherapy) Wherein, pre represents the radiomics features before treatment; post represents the radiomics features after treatment.

[0037] In this embodiment, the pathological complete remission prediction model is trained through the following steps: S30. Obtain the sample medical images of the sample patients. In this embodiment, the sample medical images include the sample contrast-enhanced (CE) CT images before NIT, i.e., the preoperative sample contrast-enhanced CT images, and the sample contrast-enhanced CT images after NIT, i.e., the postoperative sample contrast-enhanced CT images.

[0038] It can be understood that these sample patients must be pathologically diagnosed with NSCLC, and the corresponding pathological responses are reflected in the postoperative pathological reports. Since the output result of the pathological complete remission prediction model is the pathological complete remission result, and the pathological complete remission result can be used to predict whether the patient can achieve pCR after receiving NIT, the sample patients can be divided into a pCR sample group and a non-pCR (non-pathological complete remission) sample group when selecting the sample patients.

[0039] To ensure the reliability of the sample medical images, the sample patients need to have received a preset number of cycles (such as 2 - 4 cycles) of NIT.

[0040] Similarly, the sample medical images can be stored in the electronic device in advance, or obtained by the electronic device from the outside. For example, the image data obtained by the electronic device from an external acquisition device, or the data obtained by the electronic device from associated hospitals, medical institutions, etc.

[0041] No specific restrictions are imposed on the specific acquisition form of the sample medical images here, as long as it is ensured that the electronic device can obtain the sample medical images.

[0042] S40. Retain the images in the sample medical images that meet the preset criteria to obtain the medical images to be analyzed.

[0043] In this embodiment, considering that the accuracy of the training data during model training will directly affect the prediction accuracy of the model, when the sample patients have the following situations, the corresponding obtained sample medical images will not be adopted for subsequent model training: Have received other anti-tumor treatments; the time interval between the baseline CT scan and the first NIT treatment exceeds the first preset interval (such as 4 weeks); the time interval between the preoperative CT scan and the surgery exceeds the second preset interval (such as 2 weeks); any situation that hinders tumor segmentation and feature extraction (such as the presence of obvious artifacts, the lesion has reached radiological complete remission after NIT).

[0044] Therefore, the preset criteria are as follows: Only received neoadjuvant immunochemotherapy for NSCLC; the time interval between the baseline CT scan and the first treatment does not exceed the first preset interval; the time interval between the preoperative CT scan and the surgery does not exceed the second preset interval; there are no situations in the sample medical images that hinder tumor segmentation and feature extraction.

[0045] S50. Obtain the sample clinical data of the sample patient and the preset imaging markers, perform differential analysis and logistic regression analysis on the sample clinical data according to the preset imaging markers, screen the predictive factors, and determine the lesions to be analyzed according to the predictive factors. Among them, based on whether the pathological result reaches pCR, perform differential analysis and logistic regression analysis on the sample clinical data with the preset imaging markers, and the significance level value of the predictive factors screened is lower than the preset level value.

[0046] More specifically, step S50 includes: S51. Obtain the sample clinical data of the sample patient and screen and determine the preset imaging markers.

[0047] Similarly, the sample clinical data also includes the type of immune checkpoint inhibitor (ICI) of the sample patient, patient age, number of treatment cycles, gender, smoking history, pathological classification (squamous cell carcinoma, adenocarcinoma, and other non-small cell lung cancers), clinical stage, serological markers (such as cytokeratin 19 fragment, neuron-specific enolase, cancer antigen 125, cancer antigen 199, carcinoembryonic antigen, squamous cell carcinoma antigen, etc.), and data such as iRECIST for solid tumor immunotherapy.

[0048] Conventional imaging markers include the location, size, number, long diameter, short diameter, distribution, edge, lobulation sign, spiculation sign, vacuole sign, bronchial obstruction change, pleural indentation sign, calcification, necrosis, and pleural effusion of the primary lung cancer lesion. In this embodiment, a radiologist with more than ten years of experience in diagnosing chest diseases determines the preset imaging markers with statistical differences and valuable for model prediction based on these conventional imaging markers.

[0049] For example, set the preset level value (p-value) to 0.1, and use the variables with statistical differences (p < 0.1) as the preset imaging markers.

[0050] S52. Perform differential analysis and logistic regression analysis on the sample clinical data according to the preset imaging markers, screen the sample clinical features with significance level values lower than the preset level value from the sample clinical data, and use the sample clinical features as the predictive factors.

[0051] Similarly, differential analysis is performed on the sample clinical data according to the preset imaging markers, and the variables with statistical differences (p<0.1) are used as predictive factors.

[0052] S53. Determine the lesion to be analyzed from the preset imaging markers according to the predictive factors.

[0053] Finally, based on the independent predictive factors, determine the lesions to be analyzed required for subsequent radiomics feature extraction from the previously determined preset imaging markers. Through such settings, the subsequent radiomics feature extraction is obtained based on the independent predictive factors, further improving the prediction effect of the subsequent pathological complete remission prediction model.

[0054] S60. Based on the region of interest (Volume of Interest, VOI), extract the lesion regions containing the lesion to be detected that characterize the venous phase and arterial phase from the images to be analyzed respectively, and use the lesion region characterizing the venous phase to perform matching and calibration of the venous phase and arterial phase contours to obtain the sample target lesion region of the image to be analyzed.

[0055] Specifically, a non-contrast lung window image and a contrast-enhanced mediastinal window image are determined from each image to be analyzed, and based on the region of interest (Volume of Interest, VOI), a first target lesion region is segmented from the non-contrast lung window image and a second target lesion region is segmented from the contrast-enhanced mediastinal window image. The first target lesion region segmented from the non-contrast lung window image and the second target lesion region segmented from the contrast-enhanced mediastinal window image can respectively characterize the lesion and its surrounding region in the venous phase and the lesion and its surrounding region in the arterial phase.

[0056] In this embodiment, the region of interest includes regions such as the axial plane, coronal plane, and sagittal plane. At the same time, the first and second target lesion regions respectively include a first target lesion and a surrounding region (such as blood vessels) within a first preset range centered on the first target lesion, a second target lesion, and a surrounding region (such as enhanced blood vessels) within a second preset range centered on the second target lesion. The first target lesion is the largest lesion in the non-contrast lung window image, and the second target lesion is the largest lesion in the contrast-enhanced mediastinal window image, that is, if there are multiple lesions in the image, the largest one is selected as the target lesion.

[0057] S70. Extract radiomics features from the sample target lesion area to obtain pre-treatment radiomics features and post-treatment radiomics features. Since the images to be analyzed contain pre-treatment and post-treatment imaging data, the extracted radiomics features include pre-treatment radiomics features and post-treatment radiomics features. Pool all radiomics features to obtain a pre-treatment radiomics feature dataset and a post-treatment radiomics feature dataset.

[0058] In this embodiment, each dataset contains 1037 radiomics features, including: 107 original features, 186 Laplacian of Gaussian (LoG) features, and 744 wavelet features. The original features include: 14 shape features, 18 first-order features, 75 texture features, 24 Gray Level Co-occurrence Matrix (GLCM) features, 14 Gray Level Dependence Matrix (GLDM) features, 16 Gray Level RunLength Matrix (GLRLM) features, 16 Gray Level Size Zone Matrix (GLSZM) features, and 5 Neighboring Gray Tone Difference Matrix (NGTDM) features.

[0059] The LoG features and wavelet features are obtained based on the first-order features and texture features.

[0060] Preferably, the radiomics features of the sample target lesion area are extracted by PyRadiomics in Python version 3.6.

[0061] S80. Based on the pre-treatment and post-treatment radiomics features, construct difference-radiomics features.

[0062] The first difference-radiomics feature (Delta 1) is calculated as the pre-treatment radiomics feature minus the post-treatment radiomics feature; the second difference-radiomics feature (Delta 2) is calculated as the difference obtained by subtracting the post-treatment radiomics feature from the pre-treatment radiomics feature and then dividing by the pre-treatment radiomics feature; the third difference-radiomics feature (Delta 3) is calculated as the difference obtained by subtracting the post-treatment radiomics feature from the pre-treatment radiomics feature divided by the product of the pre-treatment radiomics feature, i.e., (pre - post) / (pre × the number of cycles of neoadjuvant immunochemotherapy).

[0063] S90. Select model training features from the pre-treatment radiomics features, post-treatment radiomics features, and difference-radiomics features, and train a combined prediction model based on the model training features.

[0064] S1000. Establish a mapping relationship between the trained combined prediction model and the prediction factors and the treatment effect of the sample patients after receiving NIT to obtain a pathological complete remission prediction model.

[0065] The pathological complete remission prediction model shows good clinical decision-making value in most cases. Compared with the strategies of fully predicting pCR or not predicting pCR at all, it can provide more valuable data for most patients.

[0066] More specifically, a total of 3 models were trained in the training stage, namely the clinical model: trained based on the prediction factors selected from the sample clinical data alone; combined model 1: trained based on all model training features; combined model 2: trained based on all model training features and the prediction factors selected from the sample clinical data. Finally, it was found that after adding the prediction factors, the performance of the overall model was better. The pathological complete remission prediction model is finally combined model 2, that is, trained based on all model training features and the prediction factors selected from the sample clinical data.

[0067] The method for predicting pathological complete remission based on differential imaging omics of the present invention segments and extracts imaging omics features for the sample image data before and after NIT treatment for the lesions to be analyzed based on the region of interest and the predictor, thereby obtaining the imaging omics features before and after treatment. Then, three differential-imaging omics features are constructed based on the imaging omics features before and after treatment. The maximum relevance-minimum redundancy algorithm (mRMR) and the least absolute shrinkage and selection operator (LASSO) are used to screen out meaningful features from these features. Then, the screened features and independent predictors are used to construct the final pathological complete remission prediction model. The predictor is obtained by analyzing the sample clinical data and conventional imaging markers. The pathological complete remission prediction model constructed in this way combines two features, namely differential-imaging omics and clinical features, and the combination of these two features provides complementary feature supplementation. Compared with constructing a model using only single-modal information, it can provide better prediction performance. The combination of these two features provides complementary feature supplementation and significantly improves the prediction accuracy. The pathological complete remission prediction model non-invasively characterizes whether NSCLC patients receiving NIT achieve pCR to assist medical staff in formulating personalized treatment plans and ultimately improve the clinical outcomes of NSCLC patients.

[0068] In this embodiment, step S60 specifically includes: S61. Determine the corresponding plain lung window image from each medical image to be analyzed based on the first preset window, and determine the corresponding enhanced mediastinal window image from each medical image to be analyzed based on the second preset window.

[0069] It can be understood that the first preset window is the lung window, and the second preset window is the mediastinal window, and each of these two windows has its corresponding window parameters.

[0070] S62. Based on the region of interest, determine the first target lesion region and the second target lesion region containing the lesion to be detected from the plain lung window image and the enhanced mediastinal window image respectively.

[0071] S63. Eliminate the non-target regions contained in the first target lesion region. The regions that affect subsequent feature extraction are the thicker blood vessels and trachea regions, that is, the main blood vessels and trachea regions. In this embodiment, the non-target regions are the blood vessels and trachea with a diameter exceeding the preset diameter and the peripheral regions within the third preset range.

[0072] It can be understood that both the first preset range and the second preset range are greater than the third preset range, and the first preset range and the second preset range do not need to be the same.

[0073] Preferably, a non-target region included in the first target lesion region is extracted by a trained object detection model, which can be a prediction model that has been actually applied at present and is used to extract the detection frame and object category information of the object in the image.

[0074] S64. Determine the position information of the first target lesion and the second target lesion corresponding to each image to be analyzed respectively. Based on the position information and taking the first target lesion region as a reference, perform contour matching and calibration on the lesion region contour of each image to be analyzed, and remove noise information such as blood vessels, so as to obtain the sample target lesion region of each image to be analyzed.

[0075] After removing the non-target region included in the first target lesion region, the main blood vessels and trachea can be avoided during the subsequent contour matching and calibration.

[0076] Affected by the breathing amplitude, the image information in the arterial phase may not be accurate enough. Specifically, transfer the second target lesion region representing the lesion in the arterial phase to the first target lesion region representing the lesion in the venous phase according to the contour, and perform contour matching based on the first target lesion region to obtain the contour determined in the venous phase. In this way, the surrounding region that can be displayed in the second preset window can be removed based on the first target lesion region in the first preset window, so as to obtain the final sample target lesion region of each image to be analyzed combined with the region of interest.

[0077] Preferably, a radiologist with more than ten years of experience in diagnosing chest diseases can also fine-tune the contours of the first and second target lesion regions by referring to the historical images in the calibrated historical enhanced CT sequence, so as to manually correct the VOI region. Similarly, the main blood vessels and trachea will be avoided during this process.

[0078] For example, export or save the image to be analyzed in DICOM format, and then determine the plain lung window image and the enhanced mediastinal window image from each image to be analyzed based on the first preset window and the second preset window.

[0079] Subsequently, the plain lung window image and the enhanced mediastinal window image can be imported into the 3D Slicer software, and a radiologist with more than ten years of experience in diagnosing chest diseases can determine the target lesion and perform segmentation of the region of interest.

[0080] In order to avoid the quality of the extracted radiomics features not meeting the standard due to differences in scanning parameters, reconstruction parameters, etc., the method further includes the following steps before radiomics feature extraction: The medical images to be analyzed are preprocessed, and the preprocessed medical images to be analyzed are standardized image data. It can be understood that the required plain scan lung window image and enhanced mediastinum window image are determined from each medical image to be analyzed from the preprocessed medical images to be analyzed.

[0081] In this embodiment, the preprocessing methods include but are not limited to: filtering processing, voxel processing, grayscale processing according to preset parameters, discretization processing, etc.

[0082] To ensure the consistency and accuracy of the lesion region segmented based on the region of interest, in this embodiment, step S50 further includes the following steps: S65. Determine the intraclass correlation coefficient (ICC) of the sample target lesion area, and remove the sample target lesion area with a correlation lower than the preset correlation according to the intraclass correlation coefficient. For example, the preset correlation is configured as an ICC value>0.75. When the ICC value of the sample target lesion area>0.75, it indicates that the extracted features have strong consistency.

[0083] In this embodiment, step S90 specifically includes: S91. Standardize the radiomics features before treatment, the radiomics features after treatment, and the difference-radiological features to obtain standardized features, so that the features are comparable.

[0084] It can be understood that the standardized features include standardized pre-treatment radiomics features, standardized post-treatment radiomics features and standardized difference-radiomics features.

[0085] S92. Use the synthetic minority oversampling technique (SMOTE) to perform sample balancing on the standardized features to obtain balanced features, so that the features of the pCR group and the non-pCR group can achieve a 1:1 balance, and delete the features with a variance of 0 and a Spearman rank correlation coefficient (Spearman correlation coefficient) exceeding the preset coefficient in the balanced features, so as to eliminate the collinearity problem in the features as much as possible.

[0086] It can be understood that the balanced features after sample balancing and feature deletion processing include balanced pre-treatment radiomics features, balanced post-treatment radiomics features, and balanced difference-radiomics features.

[0087] S93. Use the max-Relevance and Min-Redundancy (mRMR) algorithm to screen out the preliminary training features with high importance from the balanced features. The predictive ability of these preliminary training features is the strongest among all the balanced features.

[0088] It can be understood that the preliminary training features include the screened pre-treatment radiomics features, the screened post-treatment radiomics features, and the screened difference - radiomics features.

[0089] S94. Use the Least Absolute Shrinkage and Selection Operator (LASSO) algorithm to screen out the model training features with high importance from the preliminary training features again, that is, a total of two feature selections are performed through the mRMR algorithm and the LASSO algorithm.

[0090] It can be understood that the preliminary training features include the screened pre-treatment radiomics features, the screened post-treatment radiomics features, and the screened difference - radiomics features.

[0091] More specifically, the LASSO algorithm compresses the regression coefficients through L1 regularization, eliminates redundant features, and retains stable features.

[0092] Among them, after the pre-treatment radiomics features are processed through steps S91 to S94, a total of 1 model training feature is obtained; after the post-treatment radiomics features are processed through steps S91 to S94, a total of 2 model training features are obtained. After the difference - radiomics features are processed through steps S91 to S94, the first difference - radiomics features obtain a total of 17 model training features, the second difference - radiomics features obtain a total of 10 model training features, and the third difference - radiomics features obtain a total of 9 model training features.

[0093] S95. According to the model training features and the sample clinical data and using the logistic regression method, establish the mapping relationship with the treatment effect of the sample patients after receiving neoadjuvant immunochemotherapy, construct the pathological complete remission prediction model, and the pathological complete remission result output by the pathological complete remission prediction model is the pathological remission status (pCR or non - pCR) of the patient to be predicted.

[0094] For the features obtained from radiomics in the model training features (pre-treatment radiomics features, post-treatment radiomics features, and difference - radiomics features), the radiomics score, i.e., Rad-score, is used to determine the prediction probability of each feature. Then, by combining the model training features with the sample clinical data, 5-fold cross-validation is used to fine-tune the model parameters and determine the optimal regularization intensity value to ensure the optimal model performance and good generalization ability of the model.

[0095] Please refer to Figures 2 to 10 , and the receiver operating characteristic curve (ROC), the area under the ROC curve (AUC), and the decision curve analysis curve (DCA) are used to evaluate the model performance and clinical practicability. Among them, the optimal cut-off value can be determined by the Youden index of the ROC curve, and then key indicators such as accuracy, sensitivity, specificity, positive predictive value, and F1 value are calculated.

[0096] More specifically, the joint prediction model is trained based on all model training features and sample clinical data in the training stage.

[0097] The following describes the pathological complete remission prediction device based on difference radiomics provided by the embodiments of the present invention. The pathological complete remission prediction device based on difference radiomics described below can be correspondingly referred to the pathological complete remission prediction method based on difference radiomics described above.

[0098] To solve the above problems, in this embodiment, a pathological complete remission prediction device based on difference radiomics is provided, aiming to provide a prediction model that combines difference - radiomics and clinical features. The prediction result of this prediction model non-invasively characterizes whether NSCLC patients receiving NIT achieve pCR to assist medical staff in formulating personalized treatment plans and improving the clinical outcomes of patients. Figure 11 is a schematic structural diagram of the pathological complete remission prediction method based on difference radiomics according to the embodiments of the present invention. As Figure 11 shown, the device may include: A data acquisition module 10, configured to acquire the preoperative medical images and preoperative clinical data of the patient to be predicted. In this embodiment, the preoperative medical images are specifically the contrast-enhanced CT images before the patient receives NIT, i.e., preoperative contrast-enhanced CT images.

[0099] The preoperative medical images can be pre-stored in an electronic device or obtained by the electronic device from the outside world. For example, the image data obtained by the electronic device from an external acquisition device, or the electronic device obtains the images from associated hospitals, medical institutions, etc.

[0100] No specific restrictions are imposed on the specific acquisition form of the preoperative medical images here, as long as it is ensured that the electronic device can obtain the preoperative medical images.

[0101] The preoperative clinical data includes the type of immune checkpoint inhibitor (ICI), patient age, number of treatment cycles, gender, smoking history, pathological classification (squamous cell carcinoma, adenocarcinoma, and other non-small cell lung cancers), clinical stage, serological markers (such as cytokeratin 19 fragment, neuron-specific enolase, cancer antigen 125, cancer antigen 199, carcinoembryonic antigen, squamous cell carcinoma antigen, etc.), and iRECIST data. Among them, the above-mentioned preoperative clinical data is obtained from the medical records, pathological records, laboratory test results, imaging examination results, etc. of these sample patients.

[0102] It can be understood that the patient to be predicted must be pathologically diagnosed as NSCLC.

[0103] The model prediction module 20 is used to input the preoperative medical images and the preoperative clinical data into the trained pathological complete remission prediction model, and the pathological complete remission result of the patient to be predicted is output by the pathological complete remission prediction model.

[0104] Among them, the pathological complete remission prediction model is constructed from a combined prediction model and prediction factors screened from sample clinical data. The combined prediction model is constructed using pre-treatment radiomics features, post-treatment radiomics features, and difference-radiomics features. Specifically, the pre-treatment radiomics features and post-treatment radiomics features are radiomics features extracted from sample medical images based on regions of interest and lesions to be analyzed. The difference-radiomics features include the first difference-radiomics feature, the second difference-radiomics feature, and the third difference-radiomics feature. Among them, the first difference-radiomics feature (Delta 1) is used to characterize the absolute change value of radiomics features before and after neoadjuvant immunochemotherapy, and its calculation method is pre-treatment radiomics features minus post-treatment radiomics features; the second difference-radiomics feature (Delta 2) is used to characterize the percentage change rate of radiomics features before and after neoadjuvant immunochemotherapy, and its calculation method is the difference obtained by subtracting post-treatment radiomics features from pre-treatment radiomics features divided by pre-treatment radiomics features; the third difference-radiomics feature (Delta 3) is used for the average percentage change rate of radiomics features per chemotherapy cycle before and after neoadjuvant immunochemotherapy, and its calculation method is the difference obtained by subtracting post-treatment radiomics features from pre-treatment radiomics features divided by the product of pre-treatment radiomics features.

[0105] The pathological complete remission prediction device based on difference radiomics of the present invention segments and extracts radiomics features from sample image data before and after NIT based on regions of interest and lesions to be analyzed obtained based on prediction factors, thereby obtaining pre-treatment and post-treatment radiomics features. Then, three difference-radiomics features are constructed according to the pre-treatment and post-treatment radiomics features. The maximum relevance-minimum redundancy algorithm (mRMR) and the least absolute shrinkage and selection operator (LASSO) are used to screen out meaningful features from these features. Then, the screened features and independent prediction factors are used to construct the final pathological complete remission prediction model. The prediction factors are obtained by combining sample clinical data and conventional imaging markers. The pathological complete remission prediction model constructed in this way combines two types of features, namely difference-radiomics and clinical features, and the combination of these two types of features provides complementary feature supplementation. Compared with constructing a model using only single-modal information, it can provide better prediction performance. The combination of these two types of features provides complementary feature supplementation, significantly improving the accuracy of prediction. The pathological complete remission prediction model non-invasively characterizes whether NSCLC patients receiving NIT achieve pCR to assist medical staff in formulating personalized treatment plans and ultimately improving the clinical outcomes of non-small cell lung cancer patients.

[0106] Figure 12 Illustrates a schematic diagram of the physical structure of an electronic device, such as Figure 12As shown, the electronic device may include: a processor 1210, a communications interface 1220, a memory 1230, and a communication bus 1240. Among them, the processor 1210, the communications interface 1220, and the memory 1230 complete communication with each other through the communication bus 1240. The processor 1210 may call logical commands in the memory 1230 to execute a method for predicting pathological complete remission based on differential radiomics, and the method includes: Obtain the preoperative medical images and preoperative clinical data of the patient to be predicted; Input the preoperative medical images and preoperative clinical data into the trained pathological complete remission prediction model, and the pathological complete remission prediction model outputs the pathological complete remission result of the patient to be predicted; The pathological complete remission prediction model is constructed by a combined prediction model and prediction factors selected from sample clinical data. The combined prediction model is constructed using pre-treatment radiomics features, post-treatment radiomics features, and differential-radiomics features; The pre-treatment radiomics features and the post-treatment radiomics features are radiomics features extracted from sample medical images based on the region of interest and the lesion to be analyzed; The differential-radiomics features are obtained from the pre-treatment radiomics features and the post-treatment radiomics features. The differential-radiomics features include: A first differential-radiomics feature for characterizing the absolute change value of radiomics features before and after neoadjuvant immunochemotherapy, a second differential-radiomics feature for characterizing the percentage change rate of radiomics features before and after neoadjuvant immunochemotherapy, and a third differential-radiomics feature for the average percentage change rate of radiomics features in each chemotherapy cycle before and after neoadjuvant immunochemotherapy.

[0107] In addition, when the logical instructions in the above-mentioned memory 1230 are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs.

[0108] On the other hand, the present invention also provides a computer program product. The computer program product includes a computer program stored on a non-transitory computer-readable storage medium. The computer program includes program instructions. When the program instructions are executed by a computer, the computer can execute the pathological complete remission prediction method based on differential imaging omics provided by the above-mentioned various methods. The method includes: Obtain the preoperative medical images and preoperative clinical data of the patient to be predicted; Input the preoperative medical images and preoperative clinical data into the trained pathological complete remission prediction model, and the pathological complete remission prediction model outputs the pathological complete remission result of the patient to be predicted; The pathological complete remission prediction model is constructed by a combined prediction model and prediction factors selected from sample clinical data. The combined prediction model is constructed using pre-treatment imaging omics features, post-treatment imaging omics features, and differential-imaging omics features; The pre-treatment imaging omics features and post-treatment imaging omics features are imaging omics features extracted from sample medical images based on the region of interest and the lesion to be analyzed; The differential-imaging omics features are obtained from the pre-treatment imaging omics features and post-treatment imaging omics features. The differential-imaging omics features include: A first differential-imaging omics feature used to represent the absolute change value of the imaging omics features before and after neoadjuvant immunochemotherapy, a second differential-imaging omics feature used to represent the percentage change rate of the imaging omics features before and after neoadjuvant immunochemotherapy, and a third differential-imaging omics feature used to represent the average percentage change rate of the imaging omics features for each chemotherapy cycle before and after neoadjuvant immunochemotherapy.

[0109] In another aspect, the present invention also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements the method for predicting pathological complete remission based on differential imaging omics provided above. The method includes: Obtain the preoperative medical images and preoperative clinical data of the patient to be predicted; Input the preoperative medical images and preoperative clinical data into the trained pathological complete remission prediction model, and the pathological complete remission result of the patient to be predicted is output by the pathological complete remission prediction model; The pathological complete remission prediction model is constructed by a combined prediction model and prediction factors screened from sample clinical data. The combined prediction model is constructed by using pretreatment imaging omics features, post-treatment imaging omics features, and difference-imaging omics features; The pretreatment imaging omics features and post-treatment imaging omics features are imaging omics features extracted from sample medical images based on the region of interest and the lesion to be analyzed; The difference-imaging omics features are obtained from the pretreatment imaging omics features and post-treatment imaging omics features. The difference-imaging omics features include: The first difference-imaging omics feature for characterizing the absolute change value of imaging omics features before and after neoadjuvant immunochemotherapy, the second difference-imaging omics feature for characterizing the percentage change rate of imaging omics features before and after neoadjuvant immunochemotherapy, and the third difference-imaging omics feature for the average percentage change rate of imaging omics features in each chemotherapy cycle before and after neoadjuvant immunochemotherapy.

[0110] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for predicting pathological complete remission based on differential imaging omics, characterized in that The method includes: Obtaining the preoperative medical images and preoperative clinical data of the patient to be predicted; Inputting the preoperative medical images and preoperative clinical data into the trained pathological complete remission prediction model, and outputting the pathological complete remission result of the patient to be predicted by the pathological complete remission prediction model; The pathological complete remission prediction model is constructed by a combined prediction model and prediction factors screened from sample clinical data, and the combined prediction model is constructed by using pretreatment radiomics features, post-treatment radiomics features, and difference-radiomics features; The pretreatment radiomics features and post-treatment radiomics features are radiomics features extracted from sample medical images based on regions of interest and lesions to be analyzed; The difference-radiomics features are obtained from the pretreatment radiomics features and post-treatment radiomics features, and the difference-radiomics features include: The first difference-radiomics feature for characterizing the absolute change value of the imaging radiomics features before and after neoadjuvant immunochemotherapy, the second difference-radiomics feature for characterizing the percentage change rate of the imaging radiomics features before and after neoadjuvant immunochemotherapy, and the third difference-radiomics feature for the average percentage change rate of the imaging radiomics features for each chemotherapy cycle before and after neoadjuvant immunochemotherapy.

2. The method for predicting pathological complete remission based on differential imaging omics according to claim 1, wherein The pathological complete remission prediction model is trained through the following steps: Obtaining the sample medical images of the sample patients; the sample medical images include contrast-enhanced computed tomography images before neoadjuvant immunochemotherapy and contrast-enhanced computed tomography images after neoadjuvant immunochemotherapy; Retaining the images that meet the preset criteria in the sample medical images to obtain the medical images to be analyzed; Obtaining the sample clinical data and preset imaging markers of the sample patients, performing differential analysis and logistic regression analysis on the sample clinical data according to the preset imaging markers, screening the prediction factors and determining the lesions to be analyzed according to the prediction factors; the significance level value of the screened prediction factors is lower than the preset level value; Based on the region of interest, respectively extract the lesion regions containing the lesions to be detected that characterize the venous phase and arterial phase from the images to be analyzed, and use the lesion regions characterizing the venous phase to perform matching calibration of the venous phase and arterial phase contours to obtain the sample target lesion regions of the images to be analyzed; Extracting radiomics features from the sample target lesion regions to obtain pretreatment radiomics features and post-treatment radiomics features; Constructing difference-radiomics features based on the pretreatment and post-treatment radiomics features; Screening out the model training features from the pretreatment radiomics features, post-treatment radiomics features, and difference-radiomics features, and training the combined prediction model according to the model training features; Establishing a mapping relationship between the trained combined prediction model and the prediction factors and the treatment effect after the sample patients receive neoadjuvant immunochemotherapy to obtain the pathological complete remission prediction model.

3. The method for predicting pathological complete remission based on differential imaging omics according to claim 2, wherein The obtaining of the sample clinical data and preset imaging markers of the sample patients, performing differential analysis on the sample clinical data according to the preset imaging markers, extracting the prediction factors and determining the lesions to be analyzed according to the prediction factors specifically includes: Obtain the sample clinical data of the sample patients, and screen and determine the preset imaging markers; Conduct differential analysis and logistic regression analysis on the sample clinical data according to the preset imaging markers, screen the sample clinical features with the significance level value lower than the preset level value from the sample clinical data, and use the sample clinical features as predictors; Determine the lesions to be analyzed from the preset imaging markers according to the predictors.

4. The method for predicting pathological complete remission based on differential imaging omics according to claim 2, wherein The steps of screening out the model training features from the pre-treatment radiomics features, post-treatment radiomics features and difference-radiomics features, and training the joint prediction model according to the model training features specifically include: Perform standardization processing on the pre-treatment radiomics features, post-treatment radiomics features and difference-radiomics features to obtain standardized features; Use the synthetic minority over-sampling technique to perform sample balancing processing on the standardized features to obtain balanced features, and delete the features with a variance of 0 and a Spearman rank correlation coefficient exceeding the preset coefficient in the balanced features; Use the maximum relevance-minimum redundancy algorithm to screen out the preliminary training features from the balanced features; Use the least absolute shrinkage and selection operator algorithm to screen out the model training features from the preliminary training features; Train a joint prediction model according to the model training features by using the logistic regression method.

5. The method for predicting pathological complete remission based on differential imaging omics according to claim 2, characterized in that, Based on the region of interest, extract the lesion regions containing the lesions to be detected representing the venous phase and the arterial phase from the images to be analyzed respectively, and use the lesion region representing the venous phase to perform matching and calibration of the venous phase and arterial phase contours to obtain the sample target lesion regions of the images to be analyzed, specifically including: Determine the corresponding plain lung window images based on the first preset window from each image to be analyzed, and determine the corresponding enhanced mediastinal window images based on the second preset window from each image to be analyzed; Based on the region of interest, determine the first target lesion region and the second target lesion region containing the lesions to be detected from the plain lung window images and the enhanced mediastinal window images respectively; the first target lesion region and the second target lesion region respectively include the first target lesion and the peripheral region within the first preset range centered on the first target lesion, the second target lesion and the peripheral region within the second preset range centered on the second target lesion; the first target lesion is the largest lesion in the plain lung window image, and the second target lesion is the largest lesion in the enhanced mediastinal window image; Eliminate the non-target regions contained in the first target lesion region; Determine the position information of the first target lesion and the second target lesion corresponding to each image to be analyzed respectively, and perform matching and calibration of the lesion region contours of each image to be analyzed based on the position information with the first target lesion region as the reference to obtain the sample target lesion regions of each image to be analyzed.

6. The method for predicting pathological complete remission based on differential imaging omics according to claim 5, wherein Based on the region of interest, extract the lesion regions containing the lesions to be detected representing the venous phase and the arterial phase from the images to be analyzed respectively, and use the lesion region representing the venous phase to perform matching and calibration of the venous phase and arterial phase contours to obtain the sample target lesion regions of the images to be analyzed, and further include the following steps: Determine the intraclass correlation coefficient of the sample target lesion area, and remove the sample target lesion areas with a correlation lower than the preset correlation according to the intraclass correlation coefficient.

7. The method for predicting pathological complete remission based on differential imaging omics according to claim 2, wherein, The preset criteria are as follows: Only received neoadjuvant immunochemotherapy for non-small cell lung cancer; the time interval between the baseline computed tomography scan and the first treatment does not exceed the first preset interval; the time interval between the preoperative computed tomography scan and the surgery does not exceed the second preset interval; there are no situations in the sample medical images that hinder tumor segmentation and feature extraction.

8. A pathological complete remission prediction device based on differential imaging omics, characterized in that, The device includes: A data acquisition module for acquiring the preoperative medical images and preoperative clinical data of the patient to be predicted; A model prediction module for inputting the preoperative medical images and preoperative clinical data into a trained pathological complete response prediction model, and outputting the pathological complete response result of the patient to be predicted by the pathological complete response prediction model; The pathological complete response prediction model is constructed by a combined prediction model and a prediction factor selected from the sample clinical data. The combined prediction model is constructed using pre-treatment radiomics features, post-treatment radiomics features, and difference - radiomics features; The pre-treatment radiomics features and post-treatment radiomics features are radiomics features extracted from the sample medical images based on the region of interest and the lesion to be analyzed; The difference - radiomics features are obtained based on the pre-treatment radiomics features and post-treatment radiomics features. The difference - radiomics features include: The first difference - radiomics feature used to represent the absolute change value of the imaging radiomics features before and after neoadjuvant immunochemotherapy, the second difference - radiomics feature used to represent the percentage change rate of the imaging radiomics features before and after neoadjuvant immunochemotherapy, and the third difference - radiomics feature used to represent the average percentage change rate of the imaging radiomics features per chemotherapy cycle before and after neoadjuvant immunochemotherapy.

9. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the pathological complete response prediction method based on difference radiomics according to any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the pathological complete response prediction method based on difference radiomics according to any one of claims 1 to 7.

Images