Objective: Personalization of hemodynamic modeling plays a crucial role in functional prediction of the cardiovascular system (CVS). While reduced-order models of one-dimensional (1D) blood vessel models with zero-dimensional (0D) blood vessel and heart models have been widely recognized to be an effective tool for reasonably estimating the hemodynamic functions of the whole CVS, practical personalized models are still lacking. In this paper, we present a novel 0-1D coupled, personalized hemodynamic model of the CVS that can predict both pressure waveforms and flow velocities in arteries. Methods: We proposed a methodology by combining the multiscale CVS model with the Levenberg-Marquardt optimization algorithm for effectively solving an inverse problem based on measured blood pressure waveforms. Hemodynamic characteristics including brachial arterial pressure waveforms, artery diameters, stroke volumes, and flow velocities were measured noninvasively for 62 volunteers aged from 20 to 70 years for developing and validating the model. Results: The estimated arterial stiffness shows a physiologically realistic distribution. The model-fitted individual pressure waves have an averaged mean square error (MSE) of 7.1 mmHg $^2$ ; simulated blood flow velocity waveforms in carotid artery match ultrasound measurements well, achieving an average correlation coefficient of 0.911. Conclusion: The model is efficient, versatile, and capable of obtaining well-fitting individualized pressure waveforms while reasonably predicting flow waveforms. Significance: The proposed methodology of personalized hemodynamic modeling may therefore facilitate individualized patient-specific assessment of both physiological and pathological functions of the CVS.

High blood pressure (BP) is among significant risk factor for stroke and other vascular occurrences, it experiences nonstop fluctuations over time as a result of a complex interface among cardiovascular control mechanisms. Large blood pressure variability (BPV) has been proved to be promising in providing potential regulatory mechanisms of the cardiovascular system. Although the previous studies also showed that BPV is associated with increased carotid intima-media thickness (IMT) and plaque, whether the correlation between variability in blood pressure and left common carotid artery-intima-media thickness (LCCA-IMT) is stronger than right common carotid artery-intima-media thickness (RCCA-IMT) remains uncertain in hypertension. We conduct a study (78 hypertensive subjects, aged 28–79) to evaluate the relationship between BPV and carotid intima-media thickness in Shenzhen. The blood pressure was collected using the 24 h ambulatory blood pressure monitoring, and its variability was evaluated using standard deviation (SD), coefficient of variation (CV), and average real variability (ARV) during 24 h, daytime and nighttime. All the IMT measurements are collected by ultrasound. As the results showed, 24 h systolic blood pressure variability (SBPV) evaluated by SD and ARV were significantly related to LCCA-IMT (r1 = 0.261, P = 0.021; r1 = 0.262, P = 0.021, resp.). For the daytime diastolic blood pressure variability (DBPV), ARV indices were significantly related to LCCA-IMT (r1 = 0.239, P = 0.035), which differed form BPV evaluated by SD and CV. For the night time, there is no significant correlation between the BPV and IMT. Moreover, for all the subjects, there is no significant correlation between the BPV and RCCA-IMT/number of plaques, whereas, the SD, CV, and ARV of daytime SBP showed a positive correlation with LCCA-IMT (r1 = 0.312, P = 0.005; r1 = 0.255, P = 0.024; r1 = 0.284, P = 0.012, resp.). Moreover, the ARV of daytime SBPV, 24 h SBPV and nighttime DBPV showed a positive correlation with the number of plaques of LCCA (r1 = 0.356, P = 0.008; r1 = 0.297, P = 0.027; r1 = 0.278, P = 0.040, resp.). In addition, the number of plaques in LCCA had higher correlation with pulse pressure and diastolic blood pressure than that in RCCA. And multiple regression analysis indicated LCCA-IMT might not only be influenced by age or smoking but also by the SD index of daytime SBPV (p = 0.035). The results show that SBPV during daytime and 24 h had significant correlation with IMT, for the hypertensive subjects from the southern area of China. Moreover, we also found the daytime SBPV to be the best predictor for the progression of IMT in multivariate regression analysis. In addition, the present study suggests that the correlation between BPV and left common carotid artery—intima-media thickness/number of plaques is stronger than right common carotid artery-intima-media thickness/number of plaques.

Large blood pressure variability (BPV) will not only harm the target organ but also increase the possibility of the cardiovascular events. Since the damage of vascular system always leads to the alteration of the carotid wall, the structure and function of the carotid artery have been extensively examined in previous studies. In this work we conduct a study (60 subjects, aged 33–79) to evaluate the relationship between BPV and carotid intima-media thickness (IMT) in Shenzhen, which is one large city in the southern area of China. In our study, the blood pressure (BP) was collected using the 24 h ambulatory BP monitoring, and the BPV was evaluated using standard deviation (SD), coefficient of variation (CV), and average real variability (ARV) during 24 h, daytime and nighttime. All the IMT measurements are collected by ultrasound. The results show that both the daytime, and 24 h systolic BPV evaluated by three indices are positively associated with IMT. Among them, daytime systolic BPV evaluated with ARV is the best variable to represent the increasing of carotid IMT. In addition, after adjusting by age, sex, smoking, hypertension, and mean BP and PP values, 24 h diastolic BPV evaluated with SD also presents the favorable performance.

Multi-view clustering has attracted significant attention and application. Nonnegative matrix factorization is one popular feature of learning technology in pattern recognition. In recent years, many semi-supervised nonnegative matrix factorization algorithms were proposed by considering label information, which has achieved outstanding performance for multi-view clustering. However, most of these existing methods have either failed to consider discriminative information effectively or included too much hyper-parameters. Addressing these issues, a semi-supervised multi-view nonnegative matrix factorization with a novel fusion regularization (FRSMNMF) is developed in this article. In this work, we uniformly constrain alignment of multiple views and discriminative information among clusters with designed fusion regularization. Meanwhile, to align the multiple views effectively, two kinds of compensating matrices are used to normalize the feature scales of different views. Additionally, we preserve the geometry structure information of labeled and unlabeled samples by introducing the graph regularization simultaneously. Due to the proposed methods, two effective optimization strategies based on multiplicative update rules are designed. Experiments implemented on six real-world datasets have demonstrated the effectiveness of our FRSMNMF comparing with several state-of-the-art unsupervised and semi-supervised approaches.

Model-based reconstruction algorithms have shown potentials over conventional strain-based methods in quasi-static elastographic image by using realistic finite element (FE) or bio-mechanical model constraints. However, it is still difficult to properly handle the discrepancies between the model constraint and ultrasound data, and the measurement noise. In this paper, we explore the usage of Kalman filtering algorithm for the estimation of strain imaging in quasi-static ultrasound elastography. The proposed strategy formulates the displacement distribution through biomechanical models, and the ultrasound-derived measurements through observation equations. Through this filtering strategy, the discrepancies are quantitatively modelled as one Gaussian white noise, and the measurement noise of ultrasound data is modelled as another independent Gaussian white noise. The optimal estimation of kinematic functions, i.e. the full displacement and velocity field, are computed through this Kalman filter. Then the strain images can be easily calculated from the estimated displacement field. The accuracy and robustness of our proposed framework is first evaluated in synthetic data in controlled conditions, and the performance of this framework is then evaluated in the real data collected from elastography phantoms and patients with favourable results. The potential of our algorithm is to provide the distribution of mechanically meaningful strain under a proper biomechanical model constraint. We address the model-data discrepancy and measurement noise by introducing process noise and measurement noise in our framework, and then the mechanically meaningful strain is estimated through the Kalman filter in the minimum mean square error (MMSE) sense.