A type based on [ 68 Ga]Ga-DOTA-FAPI-04 Method and System for Predicting the Risk of Major Adverse Cardiovascular Events After Myocardial Infarction Using PET / MR Images

Abstract

This application discloses a method based on [ 68 A method and system for predicting the risk of major adverse cardiovascular events after myocardial infarction using Ga-DOTA-FAPI-04 PET / MR images. The method includes: acquiring cardiac images from a patient with myocardial infarction... 68 Ga-DOTA-FAPI-04 PET / MR images; image features calculated based on the cardiac PET / MR images [ 68 Ga]Ga-DOTA-FAPI-04 uptake volume (UV); the image features are input into a pre-built risk prediction model to predict the risk of major adverse cardiovascular events in patients with myocardial infarction. The system includes: an acquisition module, a feature extraction module, and a prediction module. The method and system of this application have extremely high accuracy and system risk stratification capabilities, and can significantly improve the predictive performance of existing models.

Description

Technical Field

[0001] This application relates to a method based on [ 68 The Ga]Ga-DOTA-FAPI-04 method and system for predicting the risk of major adverse cardiovascular events after myocardial infarction using PET / MR images belongs to the field of medical image data processing and auxiliary diagnostic technology. Background Technology

[0002] Acute myocardial infarction (MI) is the leading cause of fatal cardiovascular events worldwide. With the widespread adoption of percutaneous transluminal coronary intervention (PCI), the acute-phase survival rate of MI patients has significantly improved. However, some patients still experience adverse left ventricular remodeling during long-term follow-up after revascularization, ultimately leading to major adverse cardiovascular events (MACE), including cardiovascular death, recurrent MI, stroke, hospitalization for heart failure, and unplanned revascularization. Therefore, accurately predicting the long-term risk of MACE after myocardial infarction is crucial for developing targeted secondary prevention strategies in clinical practice.

[0003] Currently, the gold standard for clinical imaging assessment of myocardial infarction risk is cardiac magnetic resonance imaging (CMR), which primarily relies on late gadolinium enhancement (LGE) technology to quantify the volume of irreversible myocardial necrosis scars. However, the occurrence of MACE is not solely determined by the static volume of necrosis, but is driven by dynamic biological responses within the myocardial tissue, particularly the sustained activation of fibroblasts and excessive extracellular matrix deposition. Traditional parameters such as LGE can only reflect end-stage morphological necrosis and cannot quantify the active fibrotic process in volume.

[0004] To overcome the limitations of traditional morphological imaging, molecular imaging has gradually entered the forefront of cardiovascular research. Fibroblast activating protein (FAP) is a type II transmembrane serine protease. In healthy myocardial tissue of normal adults, FAP is almost not expressed or expressed at extremely low levels; however, during the active extracellular matrix remodeling and tissue repair process after myocardial infarction, FAP exhibits a high degree of overexpression on the surface of activated fibroblasts. In recent years, inhibitors of fibroblast activating protein based on gallium-68 radioisotope labeling ([… 68The Ga-DOTA-FAPI-04 PET imaging technology, with its extremely high target affinity and target-to-background ratio (TBR), has been initially applied to the visualization study of active fibrosis in cardiovascular diseases such as myocarditis, myocardial amyloidosis, radiotherapy and chemotherapy-related myocardial injury, hypertrophic cardiomyopathy, and myocardial infarction.

[0005] Although those skilled in the art can extract parameters from PET images using specific thresholding algorithms, such as the maximum target-to-base ratio (TBR) of a lesion, the total lesion-to-base ratio (TBR) is still high. max ), target-to-base ratio (TBR) mean ), Maximum Standard Intake Value (SUV) max ) and uptake volume (UV), but existing studies are mostly limited to demonstrating the correlation between UV and early myocardial injury or left ventricular remodeling, failing to [ 68 The Ga-DOTA-FAPI-04 PET parameters have been effectively translated into a predictive tool for the clinical hard endpoint MACE, which has led to a standstill in the translation of molecular imaging data into clinical decision-making in this field. Summary of the Invention

[0006] This application aims to overcome the shortcomings of existing technologies and provide a method based on [ 68 A method and system for predicting the risk of major adverse cardiovascular events after myocardial infarction using Ga-DOTA-FAPI-04 PET / MR images, with [ 68 A method and system for predicting long-term MACE risk after myocardial infarction using Ga-DOTA-FAPI-04 uptake volume (UV) as the core imaging feature parameter. The core of this application lies not in the UV parameter extraction algorithm itself, but in the first-ever establishment of [Ga-DOTA-FAPI-04 uptake volume (UV) as the core imaging feature parameter.] 68 Application of Ga-DOTA-FAPI-04 UV parameters as independent predictors of MACE and high-risk early warning cutoff values, and verification [ 68 The incremental predictive value provided by Ga]Ga-DOTA-FAPI-04 when combined with routine clinical laboratory indicators.

[0007] To achieve the above objectives, this application adopts the following technical solution: Firstly, this application provides a method based on [ 68 The Ga-DOTA-FAPI-04 PET / MR image prediction method for the risk of major adverse cardiovascular events after myocardial infarction includes the following steps: Step 1: Obtain the heart from a patient with myocardial infarction. 68 Ga]Ga-DOTA-FAPI-04 PET / MR image; Step 2: Based on the heart [ 68 Image features were calculated from Ga-DOTA-FAPI-04 PET / MR images; the image features are [ 68 Ga]Ga-DOTA-FAPI-04 Uptake Volume (UV), i.e., the total volume of FAPI uptake voxels within the myocardial volume of interest (VOI); the specific steps of the calculation include: using the left ventricular blood pool as a reference tissue, calculating the average standardized uptake value (SUV) within the reference tissue. mean The standard deviation and the volume of voxels within the reference tissue above a set threshold are used as the uptake volume; the set threshold is: the volume of voxels within the reference tissue above the set standard deviation and the volume of voxels above a set standard deviation. mean Add 2 standard deviations; the SUV mean The sum and standard deviation are data obtained based on the same clinical sample, which may be a previous clinical sample; Step 3: Input the image features into a pre-built risk prediction model to predict the risk of major adverse cardiovascular events in patients with myocardial infarction.

[0008] Furthermore, in step 3, the method for constructing the risk prediction model includes the following steps: Obtain the heart of a patient with myocardial infarction. 68 Ga]Ga-DOTA-FAPI-04 PET / MR image dataset, which contains labels of major adverse cardiovascular events; Extract image features from the dataset; The image features are input into a statistical model for fitting to obtain a risk prediction model for predicting the risk of major adverse cardiovascular events after myocardial infarction.

[0009] The warning cutoff value is based on the historical myocardial infarction patients [ 68 The Ga-DOTA-FAPI-04 UV dataset was labeled with MACE data. A receiver operating characteristic (ROC) curve was constructed, and the threshold was dynamically determined by calculating the maximum Yoden index; the warning cutoff value was 181.28 cm⁻¹. 3 .

[0010] Furthermore, step 1 also includes obtaining at least one of the following clinical test indicators in patients with myocardial infarction: peak high-sensitivity troponin I (Peak hs-cTnI), peak N-terminal pro-B-type natriuretic peptide (Peak NT-proBNP), left ventricular ejection fraction (LVEF), and late gadolinium enhancement volume (LGE volume). Accordingly, step 3 is replaced by step 3': based on the image features and clinical test indicators, the results of the pre-constructed risk prediction model are used to predict the risk of major adverse cardiovascular events in patients with myocardial infarction.

[0011] Further, step 3' specifically involves: inputting the imaging features and the clinical laboratory indicators as covariates into a risk prediction model for joint evaluation, and outputting a relative risk assessment indicator for major adverse cardiovascular events in myocardial infarction patients based on the model output; compared to the baseline model that relies solely on the clinical laboratory indicators, the risk prediction model of the joint evaluation has a positive Integrated Discriminant Improvement Index (IDI) and Net Reclassification Improvement Index (NRI).

[0012] Furthermore, in step 3', the method for constructing the risk prediction model includes the following steps: Obtain the heart of a patient with myocardial infarction. 68 Ga]Ga-DOTA-FAPI-04 PET / MR image dataset and clinical laboratory indicators, the dataset containing labels of major adverse cardiovascular events; Extract image features from the dataset; The imaging features and clinical test indicators are input into a statistical model for fitting, resulting in a risk prediction model for predicting the risk of major adverse cardiovascular events after myocardial infarction.

[0013] Furthermore, the statistical model includes: a Cox proportional hazards regression model or a logistic regression model.

[0014] Secondly, this application provides a system for predicting the risk of major adverse cardiovascular events after myocardial infarction, comprising: The acquisition module is used to acquire the heart data of patients with myocardial infarction. 68 Ga]Ga-DOTA-FAPI-04 PET / MR image; Feature extraction module, used for the heart [ 68 Image features were obtained by PET / MR imaging calculation using Ga-DOTA-FAPI-04, and the image features are [ 68 Ga]Ga-DOTA-FAPI-04 UV, which is the total volume of FAPI uptake voxels within the VOI; The prediction module is used to input the image features into a pre-built risk prediction model to predict the risk of major adverse cardiovascular events after myocardial infarction.

[0015] Furthermore, the acquisition module also includes acquiring at least one of the following clinical laboratory indicators for patients with myocardial infarction: peak high-sensitivity troponin I (Peak hs-cTnI), peak N-terminal pro-B-type natriuretic peptide (Peak NT-proBNP), left ventricular ejection fraction (LVEF), and late gadolinium enhancement volume (LGE volume); correspondingly, the prediction module is replaced by: inputting the imaging features and the clinical laboratory indicators into a pre-constructed risk prediction model, and outputting the risk stratification results or relative risk assessment indicators of major adverse cardiovascular events in patients with myocardial infarction.

[0016] Thirdly, this application provides a device for predicting the risk of major adverse cardiovascular events after myocardial infarction, comprising: a memory and a processor; the memory is used to store program instructions; the processor is used to invoke the program instructions, and when the program instructions are executed, to implement the steps in the above method.

[0017] Fourthly, this application provides a computer-readable storage medium storing computer program instructions that, when executed by a processor, implement the steps in the above-described method.

[0018] Compared with the prior art, this application has the following beneficial effects: This application is from [ 68 Key imaging parameters for predicting major adverse cardiovascular events (MACE) after myocardial infarction were identified in Ga-DOTA-FAPI-04 PET / MR images. 68 Ga]Ga-DOTA-FAPI-04 UV, and obtained a warning cutoff value of 181.28 cm for UV used in MACE risk stratification. 3 UV has extremely high accuracy and systematic risk stratification ability in predicting MACE. After adding UV to the baseline model of routine clinical laboratory indicators, the overall predictive performance shows a leap improvement, demonstrating good clinical application prospects. Attached Figure Description

[0019] Figure 1 for[ 68 Comparison of Ga-DOTA-FAPI-04 PET / MR parameters in myocardial infarction patients with and without major adverse cardiovascular events (MACE); Figure 2 for[ 68 Receiver operating characteristic (ROC) curves of Ga]Ga-DOTA-FAPI-04 PET / MR parameters for predicting major adverse cardiovascular events (MACE); Figure 3 To ensure representativeness in the STEMI population with and without major adverse cardiovascular events (MACE) [68 Ga]Ga-DOTA-FAPI-04 PET / MR image. Detailed Implementation

[0020] To make the technical solution of this application clearer and easier to understand, preferred embodiments are described in detail below with reference to the accompanying drawings.

[0021] Unless otherwise specified, the experimental or testing methods described in the following examples are conventional methods; the reagents and materials described are obtained from conventional commercial sources unless otherwise specified.

[0022] Example 1. Study subjects and inclusion criteria This study was approved by the Ethics Committee of Ruijin Hospital, affiliated with Shanghai Jiao Tong University School of Medicine (2020-152), and informed consent was obtained from the participants. Participants included those diagnosed with STEMI at Ruijin Hospital, affiliated with Shanghai Jiao Tong University School of Medicine, between July 2022 and June 2024, and who completed [the study] 4-6 weeks after diagnosis. 68 Patients with Ga]Ga-DOTA-FAPI-04 PET / MR. A total of 164 patients were ultimately included in this study.

[0023] The selection criteria are as follows: Patients eligible to enroll in this study must meet all of the following criteria: (1) Written informed consent must be obtained before any assessment is conducted; (2) Male or female patients aged 18 or older; (3) Based on the generally accepted definition of myocardial infarction, the diagnosis is spontaneous AMI; Spontaneous acute myocardial infarction (AMI) is defined as clinical evidence of myocardial necrosis consistent with myocardial ischemia resulting from a major coronary event. In these cases, spontaneous AMI must be diagnosed according to the following criteria: An elevated and / or decreased cardiac enzyme (cardiac troponin, cTn, or CK-MB) was detected, with at least one value exceeding the 99th percentile of the upper limit of normal (URL) or the local laboratory's diagnostic threshold for MI, and at least one of the following pieces of evidence of myocardial ischemia was present: (1) Ischemic discomfort or other ischemic symptoms; (2) ECG features of STEMI, including new or presumed to be new significant ST-T changes; (3) The presence of newly appearing pathological Q waves or left bundle branch block on ECG (If the patient’s spontaneous MI is secondary to another medical condition, such as anemia, hypotension or arrhythmia, or if the spontaneous MI is thought to be caused by coronary artery spasm and the coronary arteries are normal, the patient is not eligible). Patients who presented with clinical manifestations associated with Takotsubo cardiomyopathy did not meet the inclusion criteria (the time of this AMI visit refers to the time when the patient went to the emergency room / emergency treatment center, was admitted to the cardiovascular intensive care unit or cardiology ward, etc., for the treatment of this AMI).

[0024] Exclusion criteria Patients meeting any of the following criteria are not eligible for enrollment in this study: (1) Pre-existing cardiovascular disease, including coronary artery disease (CAD), previous myocardial infarction, coronary revascularization, heart failure, or severe valvular disease; (2) Systemic complications, including active / chronic infections; renal failure, autoimmune diseases, malignancies or hematologic disorders; (3) There exists a known [ 68 Ga-DOTA-FAPI-04 PET / MR contraindications include severe hypersensitivity to radioactive tracers, claustrophobia, use of pacemakers, metallic joints or mechanical valves, and renal insufficiency with an estimated glomerular filtration rate (eGFR) ≤40 mL / min / 1.73m 2 ; (4) Cardiogenic shock, severe heart failure, or any serious clinical condition that makes it impossible for the patient to tolerate prolonged supine positioning for a [68Ga]Ga-DOTA-FAPI-04 PET / MR scan.

[0025] 2.[ 68 Ga-DOTA-FAPI-04 PET / MR Image Acquisition Method All patients underwent treatment on the Biograph mMR system (Siemens AG, Erlangen, Germany). 68 Ga-DOTA-FAPI-04 PET / MR imaging. DOTA-FAPI-04 was obtained from CSBio Ltd. (Shanghai, China) and radiolabeled according to established protocols. Intravenous injection of 2.2 ± 0.2 MBq / kg 68 Sixty minutes later, a 30-minute cardiac PET scan was performed. PET images were reconstructed using the point spread function algorithm (matrix: 172×172; 4 iterations; 21 subsets; filter full width at half maximum: 2 mm).

[0026] LGE and PET images were acquired simultaneously. LGE images were acquired 10 minutes after contrast agent administration (0.1 mmol / kg; Magnevis, Bayer) using a phase-sensitive inversion recovery gradient recall echo sequence (repetition time: 930.40 ms; echo time: 1.55 ms; slice thickness: 8.0 mm; matrix: 256 × 164; field of view: 350 × 263 mm; voxel size: 1.4 × 1.4 × 8.0 mm). To assess myocardial edema, T2-weighted short tau inversion recovery (STIR) imaging was performed using a black blood sequence (slice thickness: 8.0 mm) prior to contrast agent administration.

[0027] 3.[ 68 Ga-DOTA-FAPI-04 PET / MR Image Analysis Method PET and LGE images were transferred to a Syngo.via workstation (Siemens AG, Erlangen, Germany) for registration, integration, and quantification. Two nuclear medicine physicians with no prior knowledge of the patient's clinical and biological data analyzed and reviewed the images. A spherical volume of interest (VOI) was manually placed in the left ventricle to obtain images of the infarcted myocardium. 68 Ga]Ga-DOTA-FAPI-04's maximum normalized uptake (SUV) max Using reference tissue (left ventricular blood pool) with values ​​above the mean standardized uptake (SUV) mean The threshold of 1 + 2 standard deviations is automatically calculated. 68 The uptake volume (UV) of Ga-DOTA-FAPI-04 is defined as the degree to which it activates cardiomyocytes. The maximum and mean target-to-background ratio (TBR) are used. max and TBR mean Quantifying activation intensity by measuring the SUV of the infarcted myocardium max and SUV mean Calculated by dividing by the corresponding value of the reference organization.

[0028] The endocardial and epicardial boundaries, excluding papillary muscles, were manually traced using Segment software (v3.1, http: / / segment.heiberg.se). Infarct area was represented by LGE volume and its percentage of left ventricular myocardial mass (LVMM), automatically calculated using an expectation-maximization, weighted intensity, and prior information algorithm. The degree of microvascular occlusion (MVO), defined as the percentage of low-signal area on the LGE image, was also automatically calculated. Left ventricular end-diastolic volume (LVEDV), end-systolic volume (LVESV), ejection fraction (LVEF), and left ventricular mass (LVM) were calculated from short-axis cine images using cvi42 software (v5.14, CircleCardiovascular Imaging, Calgary, Canada). Divide LVM, LVEDV, and LVESV by body surface area to obtain the LVM index (LVMI), LVEDV index (LVEDVI), and LVESV index (LVESVI), respectively. Body surface area is calculated as [(height (cm) × weight (kg) / 3600)]. 1 / 2 .

[0029] 4. Study follow-up and endpoint Prospective follow-up was conducted via telephone and face-to-face interviews. Major adverse cardiovascular event (MACE) was the primary endpoint, defined as a composite of all-cause mortality, non-fatal myocardial infarction, stroke, hospitalization for heart failure, or unplanned revascularization. All endpoints were assessed according to the definition from the Cardiovascular Trial Standardization Data Collection Initiative. [The remaining text appears to be incomplete and requires further context.] 68 The Endpoint Determination Committee reviewed the source documents to confirm the event and assess hospitalization needs for the Ga]Ga-DOTA-FAPI-04 PET / MR data uninformed endpoint. Follow-up periods were calculated based on the date of death or the last recorded medical history.

[0030] 5. Statistical methods Statistical analysis was performed using R version 3.5.1 (R Studio, Boston, MA). All tests were two-tailed, and a p-value < 0.05 was considered statistically significant.

[0031] The data distribution was assessed using the Kolmogorov-Smirnov test. Continuous variables were expressed as the mean standard deviation (SD) or median (interquartile range [IQR]). Student's t-test and Mann-Whitney U test were applied to assess the normality of the data. Categorical variables were expressed as percentages. Categorical variables were compared using the chi-square test or Fisher's exact test. Missing values ​​were handled using multiple imputation.

[0032] For outcome analysis, follow-up continued until the primary endpoint was reached or the last recorded visit, whichever occurred first. The proportional hazards assumption for all Cox regression models was assessed using scaled Schoenfeld residuals, and global tests showed no violation of this assumption (all P > 0.05). Univariate and multivariate Cox proportional hazards models were used to assess […]. 68 The association between the PET / MR parameter and MACE risk was studied using Ga-DOTA-FAPI-04, with results expressed as hazard ratios (HR) and 95% confidence intervals (CI). Event-free survival was assessed using Kaplan-Meier analysis, and intergroup comparisons were performed using the Log-rank test.

[0033] Receiver operating characteristic (ROC) analysis was used to assess [ 68 The accuracy of the Ga-DOTA-FAPI-04 PET / MR parameter in predicting MACE was assessed. The Youden index was used to derive the optimal critical values ​​from the ROC curve, and the sensitivity and specificity of these critical values ​​were calculated. The C-statistic calculated using the Cox proportional hazards regression model, along with the Net Reclassification Index (NRI) and Integrated Discriminant Improvement Index (IDI) based on the Logistic model, were used to evaluate […]. 68 Ga]Ga-DOTA-FAPI-04 UV's ability to predict incremental changes in routine clinical laboratory indicators.

[0034] 6. Research Results (1) Result 1: Analysis of baseline clinical characteristics of subjects (as shown in Table 1) Table 1 Baseline clinical characteristics of enrolled patients

[0035] The values ​​are expressed as medians (lower quartile Q1, upper quartile Q3).

[0036] (2) Result 2: The subject's cardiac magnetic resonance imaging and [ 68 Ga-DOTA-FAPI-04 PET / MR characteristics (as shown in Table 2) Table 2 Cardiac MRI and [[] of enrolled patients 68 Ga]Ga-DOTA-FAPI-04 PET / MR results

[0037] Abbreviation Explanation: LVM: Left ventricular mass; LVEDV: Left ventricular end-diastolic volume; LVESV: Left ventricular end-systolic volume; LVEF: Left ventricular ejection fraction; LVMI: Left ventricular mass index; LVEDVI: Left ventricular end-diastolic volume index; LVESVI: Left ventricular end-systolic volume index; PET: Positron emission tomography; LGE: Delayed gadolinium enhancement; LVMM: Left ventricular myocardial mass; MVO: Microvascular occlusion; UV: Gallium-68 labeled FAPI-04 uptake volume; TBR: Target-to-background ratio.

[0038] (3) Result 3: [ 68 Comparison of PET / MR parameters in individuals with and without major adverse cardiovascular events (MACE) using Ga-DOTA-FAPI-04. The median follow-up time was 1.33 (1.00, 2.12) years. A total of 19 patients (11.6%) experienced major adverse cardiovascular events (MACE), including all-cause mortality (n=3, 1.8%), heart failure (n=7, 4.3%), recurrent non-fatal infarction (n=3, 1.8%), stroke (n=2, 1.2%), and repeat unplanned revascularization (n=4, 2.4%). Patients experiencing adverse cardiovascular events […] 68 Ga-DOTA-FAPI-04 UV levels were significantly higher in patients who did not experience such events (192.2 [102.3, 254.8] vs. 102.2 [69.4, 141.6] cm⁻¹). 3 , P <0.001), while TBR did not show any discriminative value (TBR mean :2.7 [2.4,3.0] vs. 2.9 [2.4, 3.3], P = 0.5241; TBR max :4.9 [4.4, 6.2] vs. 5.0 [3.9, 6.2], P =0.5967 (e.g.) Figure 1 ).

[0039] (4) Result 4: [ 68 The predictive value of Ga-DOTA-FAPI-04 PET / MR for the clinical prognosis of patients with myocardial infarction To verify the [methods] used in this application 68 The Ga]Ga-DOTA-FAPI-04 PET / MR demonstrated its technological advantages in predicting major adverse cardiovascular events (MACE) through rigorous univariate and multivariate Cox proportional hazards regression analyses, as shown in Table 3.

[0040] Analysis results show that traditional morphological and tissue feature imaging parameters (such as LGE volume, LGE volume percentage, MVO, and AAR) are not statistically significant in predicting MACE, confirming that static structural damage in the infarct area alone cannot effectively predict the long-term incidence of adverse events. In stark contrast, the PET three-dimensional quantitative parameters proposed in this application […]. 68 The Ga-DOTA-FAPI-04 uptake volume (UV) showed highly significant predictive value in the univariate Cox regression model (HR 1.104 per 10 standard deviations, 95% CI: 1.047-1.164). P <0.001). Furthermore, existing conventional PET image analysis relies on the Local Signal Intensity Index (TBR). mean With TBR max There were no statistically significant differences in the predictions (P values ​​were 0.672 and 0.174, respectively), further illustrating that one-dimensional local intensity signals are easily disturbed and do not meet the conditions to be used as reliable prognostic predictors.

[0041] More importantly, to eliminate the confounding influence of traditional necrotic scar volume, a strict forced correction was applied to the LGE volume in the multivariate Cox regression model. The results showed that, after excluding the interference of static scar area, the UV parameter maintained a highly robust and independent statistically significant predictive power for MACE (HR 1.097, 95% CI: 1.030-1.168 for every 10 standard deviations increase). P = 0.004), UV provides independent prognostic information outside of LGE volume.

[0042] Table 3 [ 68 The predictive value of Ga-DOTA-FAPI-04 PET / MR for the clinical prognosis of patients with myocardial infarction

[0043] Multivariate Cox regression analysis corrected for delayed gadolinium enhancement volume.

[0044] (5) Based on [ 68 MACE identification performance and risk stratification of Ga-DOTA-FAPI-04 UV Combined with receiver operating characteristic (ROC) curves (e.g.) Figure 2 This application further demonstrates the extracted [ 68 The Ga]Ga-DOTA-FAPI-04 uptake volume (UV) demonstrates extremely high accuracy and systematic risk stratification capability in predicting the composite clinical endpoint (MACE).

[0045] Analysis of ROC curves for predictive performance of MACE (e.g.) Figure 2 As can be seen from the data and statistics, the UV parameters claimed in this application exhibit high diagnostic accuracy, with an area under the curve (AUC) as high as 0.85. In contrast, the AUC of the traditional late gadolinium enhancement (LGE) volume parameter in magnetic resonance imaging is only 0.71, while the TBR (transient signal intensity parameter) shows a much higher AUC. mean and TBR max The AUCs of UV radiation were as low as 0.60 and 0.54, respectively, lacking reliable clinical diagnostic capability. Based on the risk prediction model, the system defined the optimal warning cutoff value for MACE from UV radiation as 181.28 cm⁻¹. 3 The prediction sensitivity was 75.0%, and the prediction specificity was 88.8%.

[0046] Similarly, in a patient who subsequently experienced a major adverse cardiovascular event (MACE) (Patient 3, a 65-year-old male with left circumflex artery occlusion and late gadolinium enhancement (LGE) volume of 17.2 mL, as... Figure 3 A larger [A] was observed in (as shown in Figure A). 68 Ga-DOTA-FAPI-04 has an uptake volume (UV) of 199.98 cm⁻¹. 3 In contrast, another patient who did not develop MACE (Patient No. 49, a 73-year-old female with right coronary artery occlusion and an LGE volume of 19.7 mL) Figure 3 The UV value (as shown in B) is 99.72 cm⁻¹. 3 .

[0047] (6) Construction of a multivariate joint risk prediction model The risk prediction model in this application is a multivariate joint risk prediction model constructed statistically based on real clinical follow-up data of 164 patients with myocardial infarction.

[0048] A. Feature training data extraction We obtained a database of clinical laboratory indicators and imaging data from 164 patients with myocardial infarction. For each individual, we extracted their imaging characteristic variables. 68 Ga]Ga-DOTA-FAPI-04 UV (Unit: cm) 3 Simultaneously, extract its clinical laboratory indicator variables. C As covariate inputs to the model. Depending on the different application scenarios of the system, the clinical laboratory indicator variables... CIncludes one of the following: peak high-sensitivity troponin I (Peak hs-cTnI), peak N-terminal pro-B-type natriuretic peptide (Peak NT-proBNP), left ventricular ejection fraction (LVEF), or late gadolinium enhancement volume (LGE volume); whether a MACE event occurred during the follow-up period was extracted as a monitoring label (occurrence = 1, no occurrence = 0).

[0049] B. Bivariate Algorithm Fitting and Formula Fixation The above-mentioned imaging characteristic variable UV was compared with clinical laboratory indicator variables. C The data is fitted to a Cox proportional hazards regression model to generate a fixed value for assessing the patient's time-dependent relative risk. h ( t Prediction formula:

[0050] Based on the fitting of historical survival follow-up data, the fixed weight coefficients are shown in Table 4, where... h 0( t ) is the baseline hazard function.

[0051] Table 4. Fixed Weight Coefficients for Bivariate Joint Risk Prediction Model

[0052] C. Model Performance Validation As shown in Table 5, the efficacy of the bivariate joint risk prediction model was further evaluated. Statistical analysis showed that, in the baseline models constructed using high-sensitivity cardiac troponin I peak (Peak hs-cTnI), peak N-terminal pro-B-type natriuretic peptide (Peak NT-proBNP), echocardiographic left ventricular ejection fraction (LVEF), and late gadolinium enhancement volume (LGE volume), the inclusion of the imaging feature UV parameter significantly improved the C-statistic of the joint risk prediction model compared to the single-variable baseline model. Simultaneously, this joint risk prediction model exhibited significant improvements in overall discrimination and net reclassification (Table 5). For example, the C-statistic for Peak hs-cTnI jumped from 0.710 to 0.831 for Peak hs-cTnI + UV, with an overall discrimination improvement index (IDI) of 0.273 (P=0.001) and a net reclassification improvement index (NRI) of 0.667 (P=0.039). Bivariate risk prediction models can significantly eliminate the under-diagnosis of hidden high-risk patients by single clinical laboratory indicators, and greatly optimize the sensitivity and specificity of risk stratification after myocardial infarction.

[0053] Table 5 [ 68 The Incremental Predictive Value of Ga-DOTA-FAPI-04 Uptake Volume for Major Adverse Cardiovascular Events in Patients with Myocardial Infarction

[0054] Abbreviation Explanation: NA: Not applicable; Ref: Reference group.

[0055] (7) A system for predicting the risk of major adverse cardiovascular events after myocardial infarction A system for predicting the risk of major adverse cardiovascular events after myocardial infarction includes an acquisition module, a feature extraction module, and a prediction module. When deployed in clinical practice, it executes the following fully automated process when dealing with myocardial infarction patients: A. Real-time data acquisition: The system's acquisition module receives the new patient's heart data via an interface. 68 Ga]Ga-DOTA-FAPI-04 PET / MR images (or [ 68 Ga]Ga-DOTA-FAPI-04 PET / MR images and one of the latest clinical laboratory indicators: Peak hs-cTnI, Peak NT-proBNP, LVEF, or LGE volume.

[0056] B. Automatic Feature Extraction: The feature extraction module automatically extracts the lesion targets for this patient based on a set threshold (left ventricular blood pool SUVmean + 2 standard deviations). 68 Ga]Ga-DOTA-FAPI-04 UV.

[0057] C. Risk Index Calculation and Output: The extracted UV values ​​(or UV values ​​combined with clinical laboratory indicators) will be used to calculate and output the risk index. C The variable value is input into the prediction module, which calls a pre-built risk prediction model to automatically calculate the relative risk index of the patient developing long-term MACE.

[0058] In summary,[ 68 The Ga-DOTA-FAPI-04 PET / MR-derived three-dimensional uptake volume (UV) is a completely independent risk factor for long-term major adverse cardiovascular events (MACE) in myocardial infarction patients, completely independent of the traditional morphological necrosis index LGE volume, and clearly defines 181.28 cm⁻¹. 3 The optimal high-risk warning cutoff value.

[0059] Furthermore, compared to the local signal intensity index (TBR) and traditional cardiac magnetic resonance (CMR) parameters relied upon by conventional PET image analysis, UV exhibits extremely superior systematic risk stratification ability and very high discrimination accuracy (AUC up to 0.85) in predicting clinical hard endpoints.

[0060] More importantly, [ 68The Ga]Ga-DOTA-FAPI-04 UV, when combined with existing routine cardiovascular clinical laboratory indicators, yields highly significant incremental predictive value. The overall predictive efficacy, Integrated Discriminant Index (IDI), and Net Reclassification Index (NRI) of the combined risk prediction model are all significantly improved.

[0061] As a cutting-edge molecular imaging targeting technology, [ 68 Previous studies on Ga-DOTA-FAPI-04 PET / MR have largely focused on visualization of early myocardial injury or the correlation with left ventricular remodeling. Currently, few studies, both domestically and internationally, have effectively translated its core quantitative parameters into predictive tools for clinical hard endpoints (MACE). This application establishes for the first time... 68 The application of Ga-DOTA-FAPI-04 UV as an independent predictor of MACE was studied, and an automated prediction system integrating high-dimensional molecular imaging features and routine clinical laboratory indicators was innovatively constructed. This system also currently has the largest sample size related to STEMI. 68 The Ga]Ga-DOTA-FAPI-04 PET / MR study not only broke the stagnation in the clinical translation of molecular imaging data into myocardial infarction prognosis, but also accurately corrected the missed and misdiagnosed cases of occult high-risk patients by traditional clinical testing, providing a commercially and clinically promising solution for the development of targeted secondary prevention strategies after myocardial infarction.

[0062] The above description is merely a preferred embodiment of this application and is not intended to limit this application in any form or substance. It should be noted that those skilled in the art can make several improvements and additions without departing from this application, and these improvements and additions should also be considered within the scope of protection of this application.

Claims

1. A method based on [ 68 A method for predicting the risk of major adverse cardiovascular events after myocardial infarction using Ga-DOTA-FAPI-04 PET / MR images, characterized in that... Includes the following steps: Step 1: Obtain the heart of a patient with myocardial infarction. 68 Ga]Ga-DOTA-FAPI-04 PET / MR image; Step 2: Based on the heart [ 68 Image features were calculated from Ga-DOTA-FAPI-04 PET / MR images; the image features are [ 68 Ga]Ga-DOTA-FAPI-04 uptake volume, which is the total volume of FAPI uptake voxels within the myocardial volume of interest; The specific steps of the calculation include: using the left ventricular blood pool as a reference tissue, calculating the average standardized uptake value and standard deviation within the reference tissue, and extracting the voxel volume within the myocardial volume of interest that is higher than a set threshold as the uptake volume; the set threshold is: the average standardized uptake value within the reference tissue plus 2 standard deviations. Step 3: Input the image features into a pre-built risk prediction model to predict the risk of major adverse cardiovascular events in patients with myocardial infarction; In step 3, the method for constructing the risk prediction model includes the following steps: Obtain the heart of a patient with myocardial infarction. 68 Ga]Ga-DOTA-FAPI-04 PET / MR image dataset, which contains labels of major adverse cardiovascular events; Extract image features from the dataset; The image features are input into a statistical model for fitting to obtain a risk prediction model for predicting the risk of major adverse cardiovascular events after myocardial infarction.

2. The method according to claim 1, characterized in that, Step 1 also includes obtaining at least one of the following clinical test indicators for patients with myocardial infarction: peak high-sensitivity troponin, peak N-terminal pro-B-type natriuretic peptide, left ventricular ejection fraction, and late gadolinium enhancement volume; Accordingly, step 3 is replaced by step 3': inputting the imaging features and clinical test indicators into a pre-constructed risk prediction model to predict the risk of major adverse cardiovascular events in patients with myocardial infarction; In step 3', the method for constructing the risk prediction model includes the following steps: Obtain the heart of a patient with myocardial infarction. 68 Ga]Ga-DOTA-FAPI-04 PET / MR image dataset and clinical laboratory indicators, the dataset containing labels of major adverse cardiovascular events; Extract image features from the dataset; The imaging features and clinical test indicators are input into a statistical model for fitting, resulting in a risk prediction model for predicting the risk of major adverse cardiovascular events after myocardial infarction.

3. The method according to claim 1 or 2, characterized in that, The statistical models include: Cox proportional hazards regression model or logistic regression model.

4. A system for predicting the risk of major adverse cardiovascular events after myocardial infarction, characterized in that, include: The acquisition module is used to acquire the heart data of patients with myocardial infarction. 68 Ga]Ga-DOTA-FAPI-04 PET / MR image; Feature extraction module, used for based on the heart [ 68 Image features were calculated from Ga-DOTA-FAPI-04 PET / MR images, and the image features are [ 68 Ga]Ga-DOTA-FAPI-04 uptake volume, which is the total volume of FAPI uptake voxels within the myocardial volume of interest; The prediction module is used to input the image features into a pre-built risk prediction model to predict the risk of major adverse cardiovascular events after myocardial infarction.

5. The system according to claim 4, characterized in that, The acquisition module also includes acquiring at least one of the following clinical laboratory indicators for patients with myocardial infarction: peak high-sensitivity troponin, peak N-terminal pro-B-type natriuretic peptide, left ventricular ejection fraction, and late gadolinium enhancement volume; Accordingly, the prediction module is replaced by: inputting the image features and the clinical test indicators into a pre-constructed risk prediction model, and outputting the risk stratification results or relative risk assessment indicators of major adverse cardiovascular events in patients with myocardial infarction.

6. A device for predicting the risk of major adverse cardiovascular events after myocardial infarction, characterized in that, include: Memory and processor; The memory is used to store program instructions; the processor is used to invoke the program instructions, and when the program instructions are executed, to implement the steps of the method according to any one of claims 1-3.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer program instructions that, when executed by a processor, implement the steps of the method according to any one of claims 1-3.

Images