The peak period of cardiovascular disease (CVD) is around the time of awakening in the morning, which may be related to the surge of sympathetic activity at the end of nocturnal sleep. This paper chose 140 participants as study object, 70 of which had occurred CVD events while the rest hadn’t during a two-year follow-up period. A two-layer model was proposed to investigate whether hypnopompic heart rate variability (HRV) was informative to distinguish these two types of participants. In the proposed model, the extreme gradient boosting algorithm (XGBoost) was used to construct a classifier in the first layer. By evaluating the feature importance of the classifier, those features with larger importance were fed into the second layer to construct the final classifier. Three machine learning algorithms, i.e., XGBoost, random forest and support vector machine were employed and compared in the second layer to find out which one can achieve the highest performance. The results showed that, with the analysis of hypnopompic HRV, the XGBoost+XGBoost model achieved the best performance with an accuracy of 84.3%. Compared with conventional time-domain and frequency-domain features, those features derived from nonlinear dynamic analysis were more important to the model. Especially, modified permutation entropy at scale 1 and sample entropy at scale 3 were relatively important. This study might have significance for the prevention and diagnosis of CVD, as well as for the design of CVD-risk assessment system.
Objective To propose an innovative self-supervised learning method for vascular segmentation in computed tomography angiography (CTA) images by integrating feature reconstruction with masked autoencoding. Methods A 3D masked autoencoder-based framework was developed, where in 3D histogram of oriented gradients (HOG) was utilized for multi-scale vascular feature extraction. During pre-training, random masking was applied to local patches of CTA images, and the model was trained to jointly reconstruct original voxels and HOG features of masked regions. The pre-trained model was further fine-tuned on two annotated datasets for clinical-level vessel segmentation. Results Evaluated on two independent datasets (30 labeled CTA images each), our method achieved superior segmentation accuracy to the supervised neural network U-Net (nnU-Net) baseline, with Dice similarity coefficients of 91.2% vs. 89.7% (aorta) and 84.8% vs. 83.2% (coronary arteries). Conclusion The proposed self-supervised model significantly reduces manual annotation costs without compromising segmentation precision, showing substantial potential for enhancing clinical workflows in vascular disease management.
ObjectiveTo investigate the value of proton magnetic resonance spectroscopy (1H-MRS), gradient dual-echo, and triple-echo sequences in the quantitative evaluation of treatment effect of fatty liver at 3.0T MR.MethodsThirty patients with fatty liver diagnosed by CT or ultrasound who admitted in Sichuan Academy of Medical Sciences & Sichuan Provincial People’s Hospital between August 2017 and May 2018, were enrolled and undergone gradient dual-echo, triple-echo, and 1H-MRS examination before and 3 months after treatment. The fat index (FI) and relative lipid content (RLC) were measured. Fatty liver index (FLI) was calculated from blood biochemical indicators, waist circumference, and BMI at the same time. With the reference standard of FLI, the results before and after treatment measured from MRI were analyzed.ResultsThere were significantly differences of FLI, FIdual, FItriple, and RLC before and after treatment (t=5.281, P<0.001; Z=–3.651, P<0.001; Z=–3.630, P<0.001; Z=–4.762, P<0.001), all indexes decreased after treatment. FIdual and FItriple were positively correlated with FLI before (rs=0.413, P=0.023; rs=0.396, P=0.030) and after treatment (rs=0.395, P=0.031; rs=0.519, P=0.003), the highest correlation factor was FItriple to FLI after treatment. There were no significant correlation between RLC and FLI before and after treatment (P>0.05).ConclusionsIt is feasible to quantitatively evaluate the treatment effect of fatty liver by using 1H-MRS, gradient dual-echo, and triple-echo sequences. Gradient triple-echo sequences has better accuracy, which is technically easy to implement and more suitable for clinical development.
Aiming at the problems of obscure clinical auscultation features of pulmonary hypertension associated with congenital heart disease and the complexity of existing machine-aided diagnostic algorithms, an algorithm based on the statistical characteristics of the high-frequency components of the second heart sound signal is proposed. Firstly, an endpoint detection adaptive segmentation method is employed to extract the second heart sounds. Subsequently, the high-frequency component of the heart sound is decomposed using the discrete wavelet transform. Statistical features including the Hurst exponent, Lempel-Ziv information and sample entropy are extracted from this component. Finally, the extracted features are utilized to train an extreme gradient boosting algorithm (XGBoost) classifier, which achieves an accuracy of 80.45% in triple classification. Notably, this method eliminates the need for a noise reduction algorithm, allows for swift feature extraction, and achieves effective multi-classification using only three features. It is promising for early screening of pulmonary hypertension associated with congenital heart disease.
ObjectivesTo investigate simple assess method of the degree of low transvalvular gradient aortic stenosis patients with impaired left ventricular function and to investigate aortic valve replacement indications, short and mid-term outcome of this kind of patients. MethodsWe retrospectively analyzed the clinical data of 21 low-gradient patients with impaired left ventricular function in our hospital from January 2011 through May 2014. There were 15 males and 6 females aged 41-66 (54.6± 10.7) years with mean aortic transvalvular gradient less than 40 mm Hg and left ventricular ejection fraction (LVEF) less than 50%. ResultsIn response to dobutamine echocardiography stress test, 20 patients underwent aortic valve replacement. The result of intraoperative pathology showed 11 patients were with bicuspid aortic valve malformation, 4 patients with degenerative changes, 4 patients with rheumatic disease. During the same period, 3 patients underwent atrial fibrillation ablation, 1 patient with ascending aorta replacement, 2 patients with coronary artery bypass grafting, 1 patient with mitral valvuloplasty. One patient died of multiple organ failure on the fourth day after operation. The remaining patients recovered. The patients were followed up for 3 to 37 months after operation. Heart function of majority improved to gradeⅠorⅡin 3 months after surgery. The result of echocardiogram showed prosthetic valve function was good and LVEF increased (preoperative 35.7%± 8.2% vs. postoperative 49.4%± 7.2%). One patient suffered sudden death of unknown cause in the 11th months after operation. ConclusionsFor patients whose dobutamine echocardiography stress test displayed with true severe aortic stenosis and left ventricular contractile reserve capacity, after aortic valve replacement and relief of the obstruction, the left ventricular afterload decreases significantly, the left ventricular function also improves, LVEF and the quality of life improve significantly after operation.
Objective To observe the midterm haemodynamic manifestation of the home made C-L pugestrut tilting disc mechanical valve in aortic valve replacement, and to evaluate its function. Methods Twenty patients underwent aortic valve replacement over 5 years were collected and divided into two groups, the C-L pugestrut group （n=10）：aortic valve was replaced by home-made C-L pugestrut tilting disc mechanical valve（21mm）; Medtronic-Hall group （n=10）：aortic valve was replaced by Medtronic-Hall mechanical valve （21mm）. The peak transprosthetic gradients （△P）, mean transprosthetic gradients （△Pm）and effective orifice area（EOA） at rest were compared between two groups. Results At rest, △P of the C-L pugestrut group and Medtronic-Hall group were 11.63±3.23mmHg vs. 9. 78±3. 35mmHg; △Pm of the C-L pugestrut group and Medtronic-Hall group were 6. 25±2. 32 mmHg vs. 5.85±2.32mmHg： EOA of the C-L pugestrut group and Medtronic-Hall group were 1.07±0.17 cm2 vs. 1.25±0.27 cm2. There was no statistically significance in △P, △Pm and EOA between two groups（P〉0.05）. Conclusions The midterm haemodynamic results of the home-made C-L pugestrut tilting disc mechanical valve show that it has comparable haemodynamic results to those of Medtronic-Hall mechanical valve ,and it has well-done function. The home-made C-L pugestrut valve is one of the reliable mechanical heart valves.
ObjectiveTo explore the variation of the structure of the intestinal flora between healthy people and patients with obstructive jaundice perioperatively. MethodsFrom February 2013 to August 2014, 20 patients with obstructive jaundice and 10 healthy persons (normal control group) in our hospitol were selected as the research object. The first stool specimens of the research object after admission were obtained and the total fecal bacteria DNA were extracted. After polymerase chain reaction amplification, the changes in the structure of bacterial flora were dynamic observed by using denaturing gradient gel electrophoresis (DGGE), and the gel bands were analyzed by using Quantity One software. The similarity and diversity of flora structure, and principal component analysis (PCA) were analyzed. ResultsSignificant differences of colonic microflora were found between patients with obstructive jaundice and healthy people; advantage intestinal flora in obstructive jaundice patients was significant lower than the normal control group. With the extension of time and degree of obstruction aggravated, a descending trend was found in number, abundance, and diversity of the intestinal microflora (P < 0.05). ConclusionThere is significant differences in the structure of colon bacteria in patients with obstructive jaundice and healthy persons.
The dose data produced by treatment plan system (TPS) in intensity-modulated radiation therapy (IMRT) has many gradient edge points. Considering this feature we proposed a new interpolation algorithm called treatment plan dose interpolation algorithm based on gradient feature in intensity-modulated radiation therapy (TDAGI), which improves the Canny algorithm to detect the gradient edge points and non-edge points by using the gradient information in the dose data plane. For each gradient edge point, the corresponding gradient profile was traced and the profile's sharpness was calculated, and for each non-edge point, the dispersion was calculated. With the sharpness or dispersion, the kernel coefficients of bi-cubic interpolation can be obtained and can be used as the central point to complete the bi-cubic interpolation calculation. Compared with bi-cubic interpolation and bilinear interpolation, the TDAGI algorithm is more accurate. Furthermore, the TDAGI algorithm has the advantage of gradient keeping. Therefore, TDAGI can be used as an alternative method in the dose interpolation of TPS in IMRT.
In developed nations, aortic stenosis (AS) is the most common valvular heart disease presentation, and its prevalence is increasing due to aging populations. Accurate diagnosis of the disease process and determination of its severity are essential in clinical decision-making. Although current guidelines recommend measuring transvalvular gradients, maximal velocity, and aortic valve area in determining the disease severity, inconsistent grading of disease severity remains a common problem in clinical practice. Recent studies suggest that patients with paradoxical low-flow and/or low-gradient, severe AS are at a more advanced stage of the disease process and have a poorer prognosis. This mode of presentation may lead to an underevaluation of symptoms and inappropriate delay of AVR. Therefore, this challenging clinical situation should be carefully assessed in particular in symptomatic patients and clinical decisions should be tailored individually.
Heart sound is one of the common medical signals for diagnosing cardiovascular diseases. This paper studies the binary classification between normal or abnormal heart sounds, and proposes a heart sound classification algorithm based on the joint decision of extreme gradient boosting (XGBoost) and deep neural network, achieving a further improvement in feature extraction and model accuracy. First, the preprocessed heart sound recordings are segmented into four status, and five categories of features are extracted from the signals based on segmentation. The first four categories of features are sieved through recursive feature elimination, which is used as the input of the XGBoost classifier. The last category is the Mel-frequency cepstral coefficient (MFCC), which is used as the input of long short-term memory network (LSTM). Considering the imbalance of the data set, these two classifiers are both improved with weights. Finally, the heterogeneous integrated decision method is adopted to obtain the prediction. The algorithm was applied to the open heart sound database of the PhysioNet Computing in Cardiology(CINC) Challenge in 2016 on the PhysioNet website, to test the sensitivity, specificity, modified accuracy and F score. The results were 93%, 89.4%, 91.2% and 91.3% respectively. Compared with the results of machine learning, convolutional neural networks (CNN) and other methods used by other researchers, the accuracy and sensibility have been obviously improved, which proves that the method in this paper could effectively improve the accuracy of heart sound signal classification, and has great potential in the clinical auxiliary diagnosis application of some cardiovascular diseases.