Machine learning aided non-invasive diagnosis of coronary heart disease based on tongue features fusion.

BACKGROUND: Coronary heart disease (CHD) is the first cause of death globally. Hypertension is considered to be the most important independent risk factor for CHD. Early and accurate diagnosis of CHD in patients with hypertension can plays a significant role in reducing the risk and harm of hypertension combined with CHD. OBJECTIVE: To propose a non-invasive method for early diagnosis of coronary heart disease according to tongue image features with the help of machine learning techniques. METHODS: We collected standard tongue images and extract features by Diagnosis Analysis System (TDAS) and ResNet-50. On the basis of these tongue features, a common machine learning method is used to customize the non-invasive CHD diagnosis algorithm based on tongue image. RESULTS: Based on feature fusion, our algorithm has good performance. The results showed that the XGBoost model with fused features had the best performance with accuracy of 0.869, the AUC of 0.957, the AUPR of 0.961, the precision of 0.926, the recall of 0.806, and the F1-score of 0.862. CONCLUSION: We provide a feasible, convenient, and non-invasive method for the diagnosis and large-scale screening of CHD. Tongue image information is a possible effective marker for the diagnosis of CHD.

Feasibility of tongue image detection for coronary artery disease: based on deep learning.

Aim: Clarify the potential diagnostic value of tongue images for coronary artery disease (CAD), develop a CAD diagnostic model that enhances performance by incorporating tongue image inputs, and provide more reliable evidence for the clinical diagnosis of CAD, offering new biological characterization evidence. Methods: We recruited 684 patients from four hospitals in China for a cross-sectional study, collecting their baseline information and standardized tongue images to train and validate our CAD diagnostic algorithm. We used DeepLabV3 + for segmentation of the tongue body and employed Resnet-18, pretrained on ImageNet, to extract features from the tongue images. We applied DT (Decision Trees), RF (Random Forest), LR (Logistic Regression), SVM (Support Vector Machine), and XGBoost models, developing CAD diagnostic models with inputs of risk factors alone and then with the additional inclusion of tongue image features. We compared the diagnostic performance of different algorithms using accuracy, precision, recall, F1-score, AUPR, and AUC. Results: We classified patients with CAD using tongue images and found that this classification criterion was effective (ACC = 0.670, AUC = 0.690, Recall = 0.666). After comparing algorithms such as Decision Tree (DT), Random Forest (RF), Logistic Regression (LR), Support Vector Machine (SVM), and XGBoost, we ultimately chose XGBoost to develop the CAD diagnosis algorithm. The performance of the CAD diagnosis algorithm developed solely based on risk factors was ACC = 0.730, Precision = 0.811, AUC = 0.763. When tongue features were integrated, the performance of the CAD diagnosis algorithm improved to ACC = 0.760, Precision = 0.773, AUC = 0.786, Recall = 0.850, indicating an enhancement in performance. Conclusion: The use of tongue images in the diagnosis of CAD is feasible, and the inclusion of these features can enhance the performance of existing CAD diagnosis algorithms. We have customized this novel CAD diagnosis algorithm, which offers the advantages of being noninvasive, simple, and cost-effective. It is suitable for large-scale screening of CAD among hypertensive populations. Tongue image features may emerge as potential biomarkers and new risk indicators for CAD.

Study on the related factors of TCM constitution and hemodynamics in patients with coronary heart disease.

Background: The concepts of "individualization" and "preventive treatment" should be incorporated into the precise diagnosis and treatment of coronary heart disease (CHD). Both hemodynamics and Chinese medicine constitution studies align with these two concepts. Methods: This study utilized data from 81 patients with CHD, including 12 patients with balanced constitution (BC), 20 patients with blood stasis constitution (BSC), 17 patients with phlegm-dampness constitution (PDC), 15 patients with qi-deficiency constitution (QDC), and 17 patients with other constitutions. Clinical data provided information on the patients' blood property, heart function, degree of coronary stenosis, coronary hemodynamics, and so on. These parameters were compared between patients with balanced constitution vs. biased constitutions as well as between those with blood stasis constitution, phlegm-dampness constitution, and qi-deficiency constitution. Results: Compared to biased constitution (BC), patients with balanced constitution exhibited lower total cholesterol (TC) levels and low-density lipoprotein (LDL) levels. Additionally, they had lighter stenosis degrees in the Left anterior descending branch (LAD) and Left circumflex branch (LCX) branches. The hemodynamic condition of the LAD and LCX was better for those with balanced constitution; however there was no difference in heart function. Among the groups categorized by blood stasis, phlegm dampness or qi deficiency constituions, patients classified under phlegm dampness had higher levels of LDL compared to those classified under blood stasis or qi deficiency, while patients classified under qi deficiency had higher levels of blood glucose compared to those classified under blood stasis or phlegm dampness. Hemodynamic environments also differed among the LAD and LCX for each group but there were no significant differences observed in heart function or degree of coronary stenosis among these three groups. Conclusion: The balanced constitution demonstrates superior blood property, degree of coronary artery stenosis, and coronary hemodynamics compared to the biased constitution. Furthermore, among the three constitutions with CHD, variations in blood property and certain hemodynamic parameters are observed. These findings emphasize the significant clinical value of incorporating physical factors into the diagnosis and treatment of patients with CHD.

Effectiveness and safety of moxibustion for vascular dementia: A systematic review and meta-analysis.

BACKGROUND: Vascular dementia (VD) is the only type of dementia that can be prevented and treated. Compared to conventional treatment methods, moxibustion therapy is more effective for VD. This study evaluated the effectiveness and safety of moxibustion in the treatment of VD through a meta-analysis, to provide a complete overview to the advantages of traditional Chinese medicine and provide guidance for clinical application. METHODS: Clinical trials on the therapeutic effects of moxibustion or moxibustion combined with acupuncture on VD were retrieved from the VIP information database, Wanfang, CNKI, PubMed, EMBase, and other resources. The included studies were conducted from January 2000 to October 2020. Among the retrieved studies, the content met the standards upon being collated and extracted, and RevMan5.3 was used for meta-analysis. RESULTS: Thirteen randomized controlled trials (RCTs) were included with 997 patients. The RevMan bias risk assessment revealed that the quality of the studies was generally low. The meta-analysis showed that compared to conventional treatments, moxibution therapy in terms of effective rate, posttreatment Hasegawa Dementia Scale, Mini-Mental State Examination (MMSE), Activity of Daily Living Scale (ADL), Somatostatin (SS), Arginine Vasopressin (AVP), and Syndrome Differentiation Scale of VD were more favorable, and the difference in efficacy was statistically significant. Furthermore, no adverse events were observed in either group. Sensitivity analysis showed strong homogeneity and stable results, whereas funnel plot analysis revealed no significant publication bias. CONCLUSIONS: Moxibustion is effective and safe in the treatment of VD, but more high-quality evidence from further studies is required to support this.

A Novel Method to Determine the Cause of Left Internal Mammary Artery Instant Non-Patency Based on Transit Time Flow Measurement.

Objective: After coronary artery bypass grafting (CABG) surgery, the main causes of poor instant patency of left internal mammary arteries (LIMAs) are competitive flow and anastomotic stenosis, but how to determine the cause of LIMA non-patency without interfering with the native coronary artery is still a difficult problem to be solved urgently. Methods: In this study, a 0D-3D coupled multiscaled CABG model of anastomotic stenosis and competitive flow was constructed. After calculation, the flow waveform of the LIMA was extracted, and the waveform shape, common clinical parameters (average flow, PI, and DF), and graft flow FFT ratio results (F0/H1 and F0/H2) were analyzed. Results: For LIMA, these three common clinical parameters did not differ significantly between the anastomotic stenosis group and competitive flow group. However, the waveform shape and FFT ratio (especially F0/H2) of the competitive flow group were significantly different from those of the anastomotic stenosis group. When the cause was competitive flow, there was systolic backflow, and F0/H2 was too high (>14.89). When the cause was anastomotic stenosis, the waveform maintained a bimodal state and F0/H2 was in a normal state (about 1.17). Conclusion: When poor instant patency of the LIMA is found after CABG, the causes can be determined by graft flow waveform shape and F0/H2.

The Hemodynamic Mechanism of FFR-Guided Coronary Artery Bypass Grafting.

Clinically, fractional flow reserve (FFR)-guided coronary artery bypass grafting (CABG) is more effective than CABG guided by coronary angiography alone. However, no scholars have explained the mechanism from the perspective of hemodynamics. Two patients were clinically selected; their angiography showed 70% coronary stenosis, and the FFRs were 0.7 (patient 1) and 0.95 (patient 2). The FFR non-invasive computational model of the two patients was constructed by a 0-3D coupled multiscaled model, in order to verify that the model can accurately calculate the FFR results. Virtual bypass surgery was performed on these two stenoses, and a CABG multiscaled model was constructed. The flow rate of the graft and the stenosis coronary artery, as well as the wall shear stress (WSS) and the oscillatory shear index (OSI) in the graft were calculated. The non-invasive calculation results of FFR are 0.67 and 0.91, which are close to the clinical results, which proves that our model is accurate. According to the CABG model, the flow ratios of the stenosis coronary artery to the graft of patient 1 and patient 2 were 0.12 and 0.42, respectively. The time-average wall shear stress (TAWSS) results of patient 1 and patient 2 grafts were 2.09 and 2.16 Pa, respectively, and WSS showed uniform distribution on the grafts. The OSI results of patients 1 and 2 grafts were 0.0375 and 0.1264, respectively, and a significantly high OSI region appeared at the anastomosis of patient 2. The FFR value of the stenosis should be considered when performing bypass surgery. When the stenosis of high FFR values is grafted, a high OSI region is created at the graft, especially at the anastomosis. In the long term, this can cause anastomotic blockage and graft failure.