A permanent pacemaker (PPM) implantation is common after transcatheter aortic valve implantation ... more A permanent pacemaker (PPM) implantation is common after transcatheter aortic valve implantation (TAVI). We sought to evaluate requirements of pacing and incidence of pacemaker dependency during the first year after TAVI. From August 2007 until May 2011, 65 patients underwent TAVI with self-expandable prosthesis. Five patients paced at baseline and two procedure-related deaths were excluded. Evaluation of ventricular pacing percentage (VP%) and look for spontaneous rhythm were performed at 3, 6, and 12 months. PPM implantation was required in 19/58 patients (33%). Mean VP% decreased between assessments (59% at 3 months, 48% at 6 months, 50% at 12 months), but overall VP% at 1 year was high (57% ± 43%) and most patients were paced ≥10% of time. A favorable annulus-to-aorta angle was associated with lower pacing requirements (60% of patients paced ≤10% of time vs 10% of patients paced >10% of time presented an angle ≤30⁰, P = 0.039). Pacemaker dependency was established in 27% of patients and could be predicted by the presence of porcelain aorta (odds ratio = 30, confidence interval 95% 1-638, P = 0.029). New postprocedural left bundle branch block (LBBB) had a negative impact on 1-year survival (58% vs 82% in non-LBBB group, P = 0.111). PPM implantation had no impact on 1-year survival. One third of patients required PPM after TAVI and full recovery of advanced conduction abnormalities seems unlikely. Unfavorable aortic root orientation may hinder the deployment of the valve and contribute to the continued impairment of the conduction system. Porcelain aorta was a strong predictor of pacemaker dependency.
Journal of Cardiovascular Computed Tomography, 2011
Preliminary studies have shown the potential of myocardial computed tomography perfusion (CTP) an... more Preliminary studies have shown the potential of myocardial computed tomography perfusion (CTP) analysis for ischemia detection in both animals and humans. To provide validation data on stress-rest CTP protocols as additive tools to improve the accuracy of multidetector computed tomography (MDCT) for coronary artery disease (CAD) in symptomatic patients. Ninety symptomatic patients with suspected CAD (62 ± 8 years, 66% males) underwent both MDCT and invasive coronary angiography (XA). The MDCT protocol included a prospective calcium score acquisition, a helical acquisition with retrospective gating during infusion of adenosine (140 μg/kg/min) and a prospective scan for computed tomography angiography (CTA) at rest (total effective radiation dose: 5.1 ± 0.8 mSv). Significant and higher-grade CADs were defined by the presence of ≥50% or ≥70% stenosis in at least one coronary artery, as evaluated by quantitative coronary angiography (QCA) using XA images. On a patient-based model, CTA sensitivity, specificity, and positive (PPV) and negative predictive values (NPV) to detect ≥50% or ≥70% stenosis were 98%, 71%, 80%, and 97% (global accuracy 86%) and 100%, 60%, 64%, and 100% (accuracy 77%), respectively. An integrative approach of CTA and CTP results had the best performance for detection of CAD with sensitivity of 83%, specificity of 98%, PPV of 98%, and NPV of 84% (accuracy 84%) for detection of 50% stenosis and 97%, 90%, 88%, and 98% (accuracy 93%), respectively, for the 70% threshold. The integration of results had the best overall performance in all scenarios but was particularly advantageous in the prediction of higher-grade CAD, with an area under the curve of 0.93, compared with 0.80 for isolated CTA and 0.82 for CTP and in patients with severe calcifications (sensitivity 92%, specificity 87%, overall accuracy of 90%). The integration of functional and morphological data using CTA and CTP improved MDCT accuracy for detection of clinically relevant CAD at both thresholds of 50% and 70% in this intermediate to high pretest probability population.
Introduction: Epicardial adipose tissue (EAT) may play an active role in the development of coron... more Introduction: Epicardial adipose tissue (EAT) may play an active role in the development of coronary artery disease (CAD). The aim of this work was to study the relations between EAT, abdominal visceral fat (AVF), and coronary atherosclerotic burden as assessed by multislice computed tomography (MSCT). Population and methods: Two hundred fifteen patients without known CAD referred to 64-SCT during a 6-months period were included. All patients underwent a standardized protocol including quantification of AVF, EAT, coronary artery calcification (CAC), and coronary angiography by MSCT. Results: Two hundred fifteen patients, with mean age of 58 ± 11 years, in which 61% were males, with mean body mass index (BMI) of 28 ± 4 kg/m 2 were included. EAT volume was directly associated with male sex, age, BMI, abdominal circumference, AVF, number of coronary segments with atherosclerotic plaques (p b 0.01 for all), number of segments with significant stenoses, and presence of metabolic syndrome components (p b 0.05). CAC increased by 14.7% per additional 10 ml of EAT volume. Adjusting for age, gender, and AVF changed this increase to 7.5%. After adjusting for all considered confounders, there was still an independent association, with a CAC increase of 3.7% per additional 10 ml of EAT. A significant interaction was found between EAT volume and gender and between EAT volume and obesity: an increase of EAT was associated with an increase of additional 8% of CAC in men, and additional increase of 5% in non-obese individuals (p b 0.001 for both). Conclusion: EAT volume positively relates to coronary atherosclerotic burden, as assessed by CAC; this correlation was shown to be independent of AVF.
This study sought to compare computed tomography delayed enhancement (CTDE) against cardiac magne... more This study sought to compare computed tomography delayed enhancement (CTDE) against cardiac magnetic resonance (CMR) late gadolinium enhancement (LGE) for detection of ischemic scar and to test the additive value of CTDE as part of a comprehensive multidetector computed tomography (MDCT) stress-rest protocol including computed tomography perfusion (CTP) and computed tomography angiography (CTA) for the diagnosis of significant coronary artery disease (CAD). B AC K G R O U N D CTDE has been recently described as a promising tool for noninvasive detection of myocardial scar, similarly to CMR-LGE techniques. Despite its theoretical potential as an adjunctive tool to improve MDCT accuracy for detection of CAD, its clinical performance has not been validated. M E T H O D S One hundred five symptomatic patients with suspected CAD (age 62.0 AE 8.0 years, 67% men) underwent MDCT, CMR, and x-ray invasive coronary angiography.
Results: Univariate analysis showed left atrial volume, the incidence of spontaneous echo contras... more Results: Univariate analysis showed left atrial volume, the incidence of spontaneous echo contrast, male and hypertension were significantly lower and the incidence of paroxysmal AF and chicken wing type were significantly higher in high LAAV group. Multivariate analysis revealed that chicken wing type was an independent predictor for high LAAV (table). Multivariate analysis for predicting high LAAV group Odds Ratio 95% CI P Value Male 0.42 0.14-1.23 0.115 Paroxysmal atrial fibrillation 1.32 0.48-3.63 0.593 Hyper tension 0.51 0.21-1.26 0.147 Left atrial volume (cm 3) 0.99 0.98-1.00 0.156 Spontaneous echo contrast 0.42 0.13-1.34 0.142 Chicken wing type 4.19 1.59-11.09 0.004 Conclusions: The morphology of LAA determined by CCT, chicken wing or nonchicken wing type, can become a new
In this paper we report on a fuzzy-based ranking system for selecting sentences for extractive su... more In this paper we report on a fuzzy-based ranking system for selecting sentences for extractive summarization. The fuzzy knowledge base was automatically generated through a genetic algorithm. A corpus of newswire texts and their corresponding manual summaries was used to generate fuzzy classification rules. ROUGE informativeness measure was adopted as the fitness function of such an algorithm.
Despite rapid advances in transcatheter aortic valve prostheses, anatomical constraints remain th... more Despite rapid advances in transcatheter aortic valve prostheses, anatomical constraints remain that can limit access to this treatment for patients with severe aortic stenosis. The objective of this study was to determine the proportion of patients anatomically suitable for this technique using the different devices and approaches available. We retrospectively analyzed 145 consecutive patients referred to our center for transcatheter aortic valve implantation. Aortic annulus diameter was measured by transesophageal echocardiography and minimum iliofemoral diameter was determined by multidetector computed tomography. We determined the proportion of patients anatomically suitable for current devices (26-mm, 29-mm and 31-mm Medtronic CoreValve for transfemoral, transaxillary or transaortic approaches, and 23-mm, 26-mm and 29-mm Edwards Sapien XT for transfemoral or transapical approaches). The Medtronic CoreValve was suitable for 89% of patients via transfemoral access and 93.8% via transaxillary or transaortic approaches, while the Edwards Sapien XT was suitable for 82.1% of patients via transfemoral and 97.2% via transapical approaches. Only 1.4% of patients were anatomically unsuitable for all devices and approaches. In this population, most patients were anatomically suitable for transcatheter aortic valve implantation if assessed on the basis of multiple devices and multiple access approaches.
Journal of the American College of Cardiology, 2012
A 42-year-old woman with a medical history of dyslipidemia presented with sudden onset of subster... more A 42-year-old woman with a medical history of dyslipidemia presented with sudden onset of substernal chest discomfort 5 days after an episode of acute pharyngitis. An electrocardiogram showed minor concave upward ST-segment elevation in leads I, aVL, and V 5 -V 6 (A). Laboratory test results revealed elevation of troponin-I and C-reactive protein.
Revista portuguesa de cardiologia : orgão oficial da Sociedade Portuguesa de Cardiologia = Portuguese journal of cardiology : an official journal of the Portuguese Society of Cardiology, 2011
Transcatheter aortic valve implantation is an emerging treatment option for severe symptomatic ao... more Transcatheter aortic valve implantation is an emerging treatment option for severe symptomatic aortic stenosis in patients considered unsuitable for surgical valve replacement. The authors review the use of multislice computed tomography in the selection of candidates for transcatheter aortic valve replacement, procedural support and post-interventional follow-up. A single-center experience of the role of this imaging technique is also described. Multislice computed tomography is an essential imaging tool in the selection and exclusion of candidates for transcatheter aortic valve implantation, providing evaluation of coronary anatomy and the relationship of the coronary ostia with the aortic valve structure, and accurate analysis of the valve annulus and aortic root, left ventricular outflow tract, aorta and peripheral vascular access routes. Multislice computed tomography is also central to the choice of appropriate prosthesis size. In addition, it guides arterial puncture by image...
Background-Multislice computed tomography (MSCT) has shown high negative predictive value in ruli... more Background-Multislice computed tomography (MSCT) has shown high negative predictive value in ruling out obstructive coronary artery disease. Preliminary studies in patients with valvular heart disease (VHD) have demonstrated the potential of MSCT angiography (CTA) in such patients, precluding need for invasive angiography (XA). However, larger prospectively designed studies, including patients with atrial fibrillation and incorporating dose reduction algorithms, are needed. Methods and Results-To evaluate the clinical utility of 64-slice CT in the preoperative assessment in patients with VHD, we prospectively studied 452 consecutive patients undergoing routine cardiac catheterization for eligibility. Two hundred thirty-seven patients underwent both MSCT and XA. Segment-based, vessel-based, and patient-based agreement between CTA and XA was estimated assuming that "nonevaluable" segments were positive for significant coronary stenosis. In a patient-based analysis, sensitivity, specificity, positive predictive value, and negative predictive values of CTA were 95%, 89%, 66%, and 99%, respectively; in vessel-based analysis, 90%, 92%, 48%, and 99%, respectively; and in segment-based analysis, 89%, 97%, 38%, and 100%, respectively. No significant differences were found between patients with or without atrial fibrillation. A CAC value of 390 was the best cutoff for the identification of patients with positive or inconclusive CTA (which would not be exempted from XA in the clinical setting). Conclusions-In the preoperative assessment of patients with predominant VHD, the diagnostic accuracy of 64-slice CTA for ruling out the presence of significant coronary artery disease is very good even when including patients with irregular heart rhythm. Using this approach, CAC quantification before CTA can be successfully used to identify patients who should be referred directly to XA, sparing unnecessary exposure to radiation. (Circ Cardiovasc Imaging. 2009;2:306-313.)
O B J E C T I V E S This study sought to compare computed tomography delayed enhancement (CTDE) a... more O B J E C T I V E S This study sought to compare computed tomography delayed enhancement (CTDE) against cardiac magnetic resonance (CMR) late gadolinium enhancement (LGE) for detection of ischemic scar and to test the additive value of CTDE as part of a comprehensive multidetector computed tomography (MDCT) stress-rest protocol including computed tomography perfusion (CTP) and computed tomography angiography (CTA) for the diagnosis of significant coronary artery disease (CAD).
Epicardial adipose tissue (EAT) may play an active role in the development of coronary artery dis... more Epicardial adipose tissue (EAT) may play an active role in the development of coronary artery disease (CAD). The aim of this work was to study the relations between EAT, abdominal visceral fat (AVF), and coronary atherosclerotic burden as assessed by multislice computed tomography (MSCT). Two hundred fifteen patients without known CAD referred to 64-SCT during a 6-months period were included. All patients underwent a standardized protocol including quantification of AVF, EAT, coronary artery calcification (CAC), and coronary angiography by MSCT. Two hundred fifteen patients, with mean age of 58 ± 11 years, in which 61% were males, with mean body mass index (BMI) of 28 ± 4 kg/m(2) were included. EAT volume was directly associated with male sex, age, BMI, abdominal circumference, AVF, number of coronary segments with atherosclerotic plaques (p<0.01 for all), number of segments with significant stenoses, and presence of metabolic syndrome components (p<0.05). CAC increased by 14.7% per additional 10 ml of EAT volume. Adjusting for age, gender, and AVF changed this increase to 7.5%. After adjusting for all considered confounders, there was still an independent association, with a CAC increase of 3.7% per additional 10 ml of EAT. A significant interaction was found between EAT volume and gender and between EAT volume and obesity: an increase of EAT was associated with an increase of additional 8% of CAC in men, and additional increase of 5% in non-obese individuals (p<0.001 for both). EAT volume positively relates to coronary atherosclerotic burden, as assessed by CAC; this correlation was shown to be independent of AVF.
In recent years there has been increasing interest in computational modeling approaches to deal w... more In recent years there has been increasing interest in computational modeling approaches to deal with real-world data streams. Methods and algorithms have been proposed to uncover meaningful knowledge from very large (often unbounded) data sets in principle with no apparent value. This thesis introduces a framework for evolving granular modeling of uncertain data streams. Evolving granular systems comprise an array of online modeling approaches inspired by the way in which humans deal with complexity. These systems explore the information flow in dynamic environments and derive from it models that can be linguistically understood. Particularly, information granulation is a natural technique to dispense unnecessary details and emphasize transparency, interpretability and scalability of information systems. Uncertain (granular) data arise from imprecise perception or description of the value of a variable. Of particular concern to this work are numerical, interval, and fuzzy types of granular data; and interval, fuzzy, and neurofuzzy modeling frameworks. Learning in evolving granular systems is based on incremental algorithms that build model structure from scratch and adapt model parameters whenever necessary. This learning paradigm is meaningful once it avoids redesigning and retraining models all along if the system changes. Application examples in classification, function approximation, time-series prediction and control using real and synthetic data illustrate the usefulness of the granular approaches and framework proposed. The approaches and framework introduced constitute a natural extension of evolving intelligent systems over numeric data streams to evolving granular systems over granular data streams. -Student level: PhD -Date of conclusion: July 19, 2012 -Examining board members: 1)
Weather modeling and prediction has been quite a challenge over the years. Predictions based on c... more Weather modeling and prediction has been quite a challenge over the years. Predictions based on climatic models whose dynamical behavior is nonlinear, nonstationary, and based on high order difference equations is a tough task and usually requires a demanding and non-intuitive tuning expertise. This paper suggests an ensemble of evolving fuzzy models for multivariate time series prediction. The proposed ensemble approach is able to model the weather dynamics from data streams concerning variables such as wet bulb temperature, atmospheric pressure, maximum temperature, and relative humidity of the air. The purpose is to predict rainfalls 5 days ahead while providing a linguistic description of the reasoning used to give the predictions. Empirical results show that the ensemble-based fuzzy evolving modeling approach outperforms other evolving approaches in terms of accurate predictions.
This paper suggests an adaptive fuzzy neural network framework to classify data streams using a p... more This paper suggests an adaptive fuzzy neural network framework to classify data streams using a partially supervised learning algorithm. The framework consists of an evolving granular neural network which processes nonstationary data streams using a one-pass incremental algorithm. The granular neural network evolves fuzzy hyperboxes and uses a new nullnorm-based neuron model. Contrary to the usual parametric approaches, learning performs simultaneous structural and parametric adaptation whenever environment changes are reflected in input data. The approach does not require prior statistical knowledge about data and classes. Computational experiments show that the granular neural network is robust, processes unlabeled examples, and handles different types of concept drift efficiently. Resumo— Este artigo sugere uma abordagem para modelagem neuro-fuzzy adaptativa visando classificação de fluxo de dados usando um algoritmo de aprendizagem parcialmente supervisionado. A abordagem consis...
Massive amounts of streaming data from complex systems motivate rethinking some aspects of the ma... more Massive amounts of streaming data from complex systems motivate rethinking some aspects of the machine learning theory. Data stream mining is concerned with extracting structured knowledge from spatio-temporally correlated data. A profusion of systems and algorithms devoted to this end has been constructed under the conceptual framework of granular computing. This paper outlines a fuzzy set based granular evolving modeling -FBeM -approach for learning from imprecise data. Granulation arises because modeling uncertain data dispenses attention to details. The evolving aspect is fundamental to account for endless flows of nonsta-tionary data and structural adaptation of models. Experiments with Mackey-Glass benchmark data recommend that the FBeM approach outperforms alternative approaches. Resumo— Modelagem on-line de grandes volumes de dados de sistemas complexos motiva a revisão de di-versos aspectos da teoria de aprendizagem de máquina. A mineração de fluxo de dados lida com a extra...
This work outlines a new approach for online learning from imprecise data, namely, fuzzy set base... more This work outlines a new approach for online learning from imprecise data, namely, fuzzy set based evolving modeling (FBeM) approach. FBeM is an adaptive modeling framework that uses fuzzy granular objects to enclose uncer- tainty in the data. The FBeM algorithm is data flow driven and supports learning on an instance-per-instance recursive basis by developing and refining fuzzy mod- els on-demand. Structurally, FBeM models combine Mamdani and functional fuzzy systems to output granular and singular approximations of nonstationary functions. In general, approximand functions can be time series, decision boundaries between classes, and control and regression functions. Linguistic description of the behavior of the system over time is provided by information granules and associated rules. An application example on a reactive control problem, underlining the complemen- tarity of Mamdani and functional parts of the model, illustrates the usefulness of the approach. More specifically, th...
System modeling in dynamic environments needs processing of streams 1 of sensor data and incremen... more System modeling in dynamic environments needs processing of streams 1 of sensor data and incremental learning algorithms. This paper suggests an incre-2 mental granular fuzzy rule-based modeling approach using streams of fuzzy inter-3 val data. Incremental granular modeling is an adaptive modeling framework that uses 4 fuzzy granular data that originate from unreliable sensors, imprecise perceptions, or 5 description of imprecise values of a variable in the form fuzzy intervals. The incre-6 mental learning algorithm builds the antecedent of functional fuzzy rules and the 7 rule base of the fuzzy model. A recursive least squares algorithm revises the para-8 meters of a state-space representation of the fuzzy rule consequents. Imprecision in 9 data is accounted for using specificity measures. An illustrative example concerning 10 the Rossler attractor is given. 11 1 Introduction 12 Data produced by real world systems result from nonlinear, uncertain, and time-13
2007 IEEE International Electric Machines & Drives Conference, 2007
This paper describes a real-time model-based fault Estimation Module (aPE), which includes the al... more This paper describes a real-time model-based fault Estimation Module (aPE), which includes the algorithms: detection and diagnosis software. The Electric Machines Diag-Recursive Least-Squares (RLS) and Extended Kalman Filter nosis System (EMDS) covers field winding shorted-turns fault in (EKF); Dynamic Models Module (oaDM) for faults simulation; alternators and stator windings shorted-turns fault in induction motors. The EMDS has a modular architecture. The modules Computational Intelligence Module (ad), which includes: include: acquisition and data treatment; well-known parameters MLP and SOM Neural Networks and the FCM technique for estimation algorithms, such as Recursive Least Squares (RLS) patterns recognition. These modules working together generate and Extended Kalman Filter (EKF); dynamic models for faults the health state of various electric machines working in parallel simulation; faults detection and identification tools, such as through routing. The system modules are trigged as soon as M.L.P. and S.O.M. neural networks and Fuzzy C-Means (FCM)
This work outlines a new approach for online learning from imprecise data, namely, fuzzy set base... more This work outlines a new approach for online learning from imprecise data, namely, fuzzy set based evolving modeling (FBeM) approach. FBeM is an adaptive modeling framework that uses fuzzy granular objects to enclose uncertainty in the data. The FBeM algorithm is data flow driven and supports learning on an instance-per-instance recursive basis by developing and refining fuzzy models on-demand. Structurally, FBeM models combine Mamdani and functional fuzzy systems to output granular and singular approximations of nonstationary functions. In general, approximand functions can be time series, decision boundaries between classes, and control and regression functions. Linguistic description of the behavior of the system over time is provided by information granules and associated rules. An application example on a reactive control problem, underlining the complementarity of Mamdani and functional parts of the model, illustrates the usefulness of the approach. More specifically, the problem concerns sensor-based robot navigation and localization. In addition to precise singular output values, granular output values provide effective robust obstacle avoidance navigation.
A permanent pacemaker (PPM) implantation is common after transcatheter aortic valve implantation ... more A permanent pacemaker (PPM) implantation is common after transcatheter aortic valve implantation (TAVI). We sought to evaluate requirements of pacing and incidence of pacemaker dependency during the first year after TAVI. From August 2007 until May 2011, 65 patients underwent TAVI with self-expandable prosthesis. Five patients paced at baseline and two procedure-related deaths were excluded. Evaluation of ventricular pacing percentage (VP%) and look for spontaneous rhythm were performed at 3, 6, and 12 months. PPM implantation was required in 19/58 patients (33%). Mean VP% decreased between assessments (59% at 3 months, 48% at 6 months, 50% at 12 months), but overall VP% at 1 year was high (57% ± 43%) and most patients were paced ≥10% of time. A favorable annulus-to-aorta angle was associated with lower pacing requirements (60% of patients paced ≤10% of time vs 10% of patients paced >10% of time presented an angle ≤30⁰, P = 0.039). Pacemaker dependency was established in 27% of patients and could be predicted by the presence of porcelain aorta (odds ratio = 30, confidence interval 95% 1-638, P = 0.029). New postprocedural left bundle branch block (LBBB) had a negative impact on 1-year survival (58% vs 82% in non-LBBB group, P = 0.111). PPM implantation had no impact on 1-year survival. One third of patients required PPM after TAVI and full recovery of advanced conduction abnormalities seems unlikely. Unfavorable aortic root orientation may hinder the deployment of the valve and contribute to the continued impairment of the conduction system. Porcelain aorta was a strong predictor of pacemaker dependency.
Journal of Cardiovascular Computed Tomography, 2011
Preliminary studies have shown the potential of myocardial computed tomography perfusion (CTP) an... more Preliminary studies have shown the potential of myocardial computed tomography perfusion (CTP) analysis for ischemia detection in both animals and humans. To provide validation data on stress-rest CTP protocols as additive tools to improve the accuracy of multidetector computed tomography (MDCT) for coronary artery disease (CAD) in symptomatic patients. Ninety symptomatic patients with suspected CAD (62 ± 8 years, 66% males) underwent both MDCT and invasive coronary angiography (XA). The MDCT protocol included a prospective calcium score acquisition, a helical acquisition with retrospective gating during infusion of adenosine (140 μg/kg/min) and a prospective scan for computed tomography angiography (CTA) at rest (total effective radiation dose: 5.1 ± 0.8 mSv). Significant and higher-grade CADs were defined by the presence of ≥50% or ≥70% stenosis in at least one coronary artery, as evaluated by quantitative coronary angiography (QCA) using XA images. On a patient-based model, CTA sensitivity, specificity, and positive (PPV) and negative predictive values (NPV) to detect ≥50% or ≥70% stenosis were 98%, 71%, 80%, and 97% (global accuracy 86%) and 100%, 60%, 64%, and 100% (accuracy 77%), respectively. An integrative approach of CTA and CTP results had the best performance for detection of CAD with sensitivity of 83%, specificity of 98%, PPV of 98%, and NPV of 84% (accuracy 84%) for detection of 50% stenosis and 97%, 90%, 88%, and 98% (accuracy 93%), respectively, for the 70% threshold. The integration of results had the best overall performance in all scenarios but was particularly advantageous in the prediction of higher-grade CAD, with an area under the curve of 0.93, compared with 0.80 for isolated CTA and 0.82 for CTP and in patients with severe calcifications (sensitivity 92%, specificity 87%, overall accuracy of 90%). The integration of functional and morphological data using CTA and CTP improved MDCT accuracy for detection of clinically relevant CAD at both thresholds of 50% and 70% in this intermediate to high pretest probability population.
Introduction: Epicardial adipose tissue (EAT) may play an active role in the development of coron... more Introduction: Epicardial adipose tissue (EAT) may play an active role in the development of coronary artery disease (CAD). The aim of this work was to study the relations between EAT, abdominal visceral fat (AVF), and coronary atherosclerotic burden as assessed by multislice computed tomography (MSCT). Population and methods: Two hundred fifteen patients without known CAD referred to 64-SCT during a 6-months period were included. All patients underwent a standardized protocol including quantification of AVF, EAT, coronary artery calcification (CAC), and coronary angiography by MSCT. Results: Two hundred fifteen patients, with mean age of 58 ± 11 years, in which 61% were males, with mean body mass index (BMI) of 28 ± 4 kg/m 2 were included. EAT volume was directly associated with male sex, age, BMI, abdominal circumference, AVF, number of coronary segments with atherosclerotic plaques (p b 0.01 for all), number of segments with significant stenoses, and presence of metabolic syndrome components (p b 0.05). CAC increased by 14.7% per additional 10 ml of EAT volume. Adjusting for age, gender, and AVF changed this increase to 7.5%. After adjusting for all considered confounders, there was still an independent association, with a CAC increase of 3.7% per additional 10 ml of EAT. A significant interaction was found between EAT volume and gender and between EAT volume and obesity: an increase of EAT was associated with an increase of additional 8% of CAC in men, and additional increase of 5% in non-obese individuals (p b 0.001 for both). Conclusion: EAT volume positively relates to coronary atherosclerotic burden, as assessed by CAC; this correlation was shown to be independent of AVF.
This study sought to compare computed tomography delayed enhancement (CTDE) against cardiac magne... more This study sought to compare computed tomography delayed enhancement (CTDE) against cardiac magnetic resonance (CMR) late gadolinium enhancement (LGE) for detection of ischemic scar and to test the additive value of CTDE as part of a comprehensive multidetector computed tomography (MDCT) stress-rest protocol including computed tomography perfusion (CTP) and computed tomography angiography (CTA) for the diagnosis of significant coronary artery disease (CAD). B AC K G R O U N D CTDE has been recently described as a promising tool for noninvasive detection of myocardial scar, similarly to CMR-LGE techniques. Despite its theoretical potential as an adjunctive tool to improve MDCT accuracy for detection of CAD, its clinical performance has not been validated. M E T H O D S One hundred five symptomatic patients with suspected CAD (age 62.0 AE 8.0 years, 67% men) underwent MDCT, CMR, and x-ray invasive coronary angiography.
Results: Univariate analysis showed left atrial volume, the incidence of spontaneous echo contras... more Results: Univariate analysis showed left atrial volume, the incidence of spontaneous echo contrast, male and hypertension were significantly lower and the incidence of paroxysmal AF and chicken wing type were significantly higher in high LAAV group. Multivariate analysis revealed that chicken wing type was an independent predictor for high LAAV (table). Multivariate analysis for predicting high LAAV group Odds Ratio 95% CI P Value Male 0.42 0.14-1.23 0.115 Paroxysmal atrial fibrillation 1.32 0.48-3.63 0.593 Hyper tension 0.51 0.21-1.26 0.147 Left atrial volume (cm 3) 0.99 0.98-1.00 0.156 Spontaneous echo contrast 0.42 0.13-1.34 0.142 Chicken wing type 4.19 1.59-11.09 0.004 Conclusions: The morphology of LAA determined by CCT, chicken wing or nonchicken wing type, can become a new
In this paper we report on a fuzzy-based ranking system for selecting sentences for extractive su... more In this paper we report on a fuzzy-based ranking system for selecting sentences for extractive summarization. The fuzzy knowledge base was automatically generated through a genetic algorithm. A corpus of newswire texts and their corresponding manual summaries was used to generate fuzzy classification rules. ROUGE informativeness measure was adopted as the fitness function of such an algorithm.
Despite rapid advances in transcatheter aortic valve prostheses, anatomical constraints remain th... more Despite rapid advances in transcatheter aortic valve prostheses, anatomical constraints remain that can limit access to this treatment for patients with severe aortic stenosis. The objective of this study was to determine the proportion of patients anatomically suitable for this technique using the different devices and approaches available. We retrospectively analyzed 145 consecutive patients referred to our center for transcatheter aortic valve implantation. Aortic annulus diameter was measured by transesophageal echocardiography and minimum iliofemoral diameter was determined by multidetector computed tomography. We determined the proportion of patients anatomically suitable for current devices (26-mm, 29-mm and 31-mm Medtronic CoreValve for transfemoral, transaxillary or transaortic approaches, and 23-mm, 26-mm and 29-mm Edwards Sapien XT for transfemoral or transapical approaches). The Medtronic CoreValve was suitable for 89% of patients via transfemoral access and 93.8% via transaxillary or transaortic approaches, while the Edwards Sapien XT was suitable for 82.1% of patients via transfemoral and 97.2% via transapical approaches. Only 1.4% of patients were anatomically unsuitable for all devices and approaches. In this population, most patients were anatomically suitable for transcatheter aortic valve implantation if assessed on the basis of multiple devices and multiple access approaches.
Journal of the American College of Cardiology, 2012
A 42-year-old woman with a medical history of dyslipidemia presented with sudden onset of subster... more A 42-year-old woman with a medical history of dyslipidemia presented with sudden onset of substernal chest discomfort 5 days after an episode of acute pharyngitis. An electrocardiogram showed minor concave upward ST-segment elevation in leads I, aVL, and V 5 -V 6 (A). Laboratory test results revealed elevation of troponin-I and C-reactive protein.
Revista portuguesa de cardiologia : orgão oficial da Sociedade Portuguesa de Cardiologia = Portuguese journal of cardiology : an official journal of the Portuguese Society of Cardiology, 2011
Transcatheter aortic valve implantation is an emerging treatment option for severe symptomatic ao... more Transcatheter aortic valve implantation is an emerging treatment option for severe symptomatic aortic stenosis in patients considered unsuitable for surgical valve replacement. The authors review the use of multislice computed tomography in the selection of candidates for transcatheter aortic valve replacement, procedural support and post-interventional follow-up. A single-center experience of the role of this imaging technique is also described. Multislice computed tomography is an essential imaging tool in the selection and exclusion of candidates for transcatheter aortic valve implantation, providing evaluation of coronary anatomy and the relationship of the coronary ostia with the aortic valve structure, and accurate analysis of the valve annulus and aortic root, left ventricular outflow tract, aorta and peripheral vascular access routes. Multislice computed tomography is also central to the choice of appropriate prosthesis size. In addition, it guides arterial puncture by image...
Background-Multislice computed tomography (MSCT) has shown high negative predictive value in ruli... more Background-Multislice computed tomography (MSCT) has shown high negative predictive value in ruling out obstructive coronary artery disease. Preliminary studies in patients with valvular heart disease (VHD) have demonstrated the potential of MSCT angiography (CTA) in such patients, precluding need for invasive angiography (XA). However, larger prospectively designed studies, including patients with atrial fibrillation and incorporating dose reduction algorithms, are needed. Methods and Results-To evaluate the clinical utility of 64-slice CT in the preoperative assessment in patients with VHD, we prospectively studied 452 consecutive patients undergoing routine cardiac catheterization for eligibility. Two hundred thirty-seven patients underwent both MSCT and XA. Segment-based, vessel-based, and patient-based agreement between CTA and XA was estimated assuming that "nonevaluable" segments were positive for significant coronary stenosis. In a patient-based analysis, sensitivity, specificity, positive predictive value, and negative predictive values of CTA were 95%, 89%, 66%, and 99%, respectively; in vessel-based analysis, 90%, 92%, 48%, and 99%, respectively; and in segment-based analysis, 89%, 97%, 38%, and 100%, respectively. No significant differences were found between patients with or without atrial fibrillation. A CAC value of 390 was the best cutoff for the identification of patients with positive or inconclusive CTA (which would not be exempted from XA in the clinical setting). Conclusions-In the preoperative assessment of patients with predominant VHD, the diagnostic accuracy of 64-slice CTA for ruling out the presence of significant coronary artery disease is very good even when including patients with irregular heart rhythm. Using this approach, CAC quantification before CTA can be successfully used to identify patients who should be referred directly to XA, sparing unnecessary exposure to radiation. (Circ Cardiovasc Imaging. 2009;2:306-313.)
O B J E C T I V E S This study sought to compare computed tomography delayed enhancement (CTDE) a... more O B J E C T I V E S This study sought to compare computed tomography delayed enhancement (CTDE) against cardiac magnetic resonance (CMR) late gadolinium enhancement (LGE) for detection of ischemic scar and to test the additive value of CTDE as part of a comprehensive multidetector computed tomography (MDCT) stress-rest protocol including computed tomography perfusion (CTP) and computed tomography angiography (CTA) for the diagnosis of significant coronary artery disease (CAD).
Epicardial adipose tissue (EAT) may play an active role in the development of coronary artery dis... more Epicardial adipose tissue (EAT) may play an active role in the development of coronary artery disease (CAD). The aim of this work was to study the relations between EAT, abdominal visceral fat (AVF), and coronary atherosclerotic burden as assessed by multislice computed tomography (MSCT). Two hundred fifteen patients without known CAD referred to 64-SCT during a 6-months period were included. All patients underwent a standardized protocol including quantification of AVF, EAT, coronary artery calcification (CAC), and coronary angiography by MSCT. Two hundred fifteen patients, with mean age of 58 ± 11 years, in which 61% were males, with mean body mass index (BMI) of 28 ± 4 kg/m(2) were included. EAT volume was directly associated with male sex, age, BMI, abdominal circumference, AVF, number of coronary segments with atherosclerotic plaques (p<0.01 for all), number of segments with significant stenoses, and presence of metabolic syndrome components (p<0.05). CAC increased by 14.7% per additional 10 ml of EAT volume. Adjusting for age, gender, and AVF changed this increase to 7.5%. After adjusting for all considered confounders, there was still an independent association, with a CAC increase of 3.7% per additional 10 ml of EAT. A significant interaction was found between EAT volume and gender and between EAT volume and obesity: an increase of EAT was associated with an increase of additional 8% of CAC in men, and additional increase of 5% in non-obese individuals (p<0.001 for both). EAT volume positively relates to coronary atherosclerotic burden, as assessed by CAC; this correlation was shown to be independent of AVF.
In recent years there has been increasing interest in computational modeling approaches to deal w... more In recent years there has been increasing interest in computational modeling approaches to deal with real-world data streams. Methods and algorithms have been proposed to uncover meaningful knowledge from very large (often unbounded) data sets in principle with no apparent value. This thesis introduces a framework for evolving granular modeling of uncertain data streams. Evolving granular systems comprise an array of online modeling approaches inspired by the way in which humans deal with complexity. These systems explore the information flow in dynamic environments and derive from it models that can be linguistically understood. Particularly, information granulation is a natural technique to dispense unnecessary details and emphasize transparency, interpretability and scalability of information systems. Uncertain (granular) data arise from imprecise perception or description of the value of a variable. Of particular concern to this work are numerical, interval, and fuzzy types of granular data; and interval, fuzzy, and neurofuzzy modeling frameworks. Learning in evolving granular systems is based on incremental algorithms that build model structure from scratch and adapt model parameters whenever necessary. This learning paradigm is meaningful once it avoids redesigning and retraining models all along if the system changes. Application examples in classification, function approximation, time-series prediction and control using real and synthetic data illustrate the usefulness of the granular approaches and framework proposed. The approaches and framework introduced constitute a natural extension of evolving intelligent systems over numeric data streams to evolving granular systems over granular data streams. -Student level: PhD -Date of conclusion: July 19, 2012 -Examining board members: 1)
Weather modeling and prediction has been quite a challenge over the years. Predictions based on c... more Weather modeling and prediction has been quite a challenge over the years. Predictions based on climatic models whose dynamical behavior is nonlinear, nonstationary, and based on high order difference equations is a tough task and usually requires a demanding and non-intuitive tuning expertise. This paper suggests an ensemble of evolving fuzzy models for multivariate time series prediction. The proposed ensemble approach is able to model the weather dynamics from data streams concerning variables such as wet bulb temperature, atmospheric pressure, maximum temperature, and relative humidity of the air. The purpose is to predict rainfalls 5 days ahead while providing a linguistic description of the reasoning used to give the predictions. Empirical results show that the ensemble-based fuzzy evolving modeling approach outperforms other evolving approaches in terms of accurate predictions.
This paper suggests an adaptive fuzzy neural network framework to classify data streams using a p... more This paper suggests an adaptive fuzzy neural network framework to classify data streams using a partially supervised learning algorithm. The framework consists of an evolving granular neural network which processes nonstationary data streams using a one-pass incremental algorithm. The granular neural network evolves fuzzy hyperboxes and uses a new nullnorm-based neuron model. Contrary to the usual parametric approaches, learning performs simultaneous structural and parametric adaptation whenever environment changes are reflected in input data. The approach does not require prior statistical knowledge about data and classes. Computational experiments show that the granular neural network is robust, processes unlabeled examples, and handles different types of concept drift efficiently. Resumo— Este artigo sugere uma abordagem para modelagem neuro-fuzzy adaptativa visando classificação de fluxo de dados usando um algoritmo de aprendizagem parcialmente supervisionado. A abordagem consis...
Massive amounts of streaming data from complex systems motivate rethinking some aspects of the ma... more Massive amounts of streaming data from complex systems motivate rethinking some aspects of the machine learning theory. Data stream mining is concerned with extracting structured knowledge from spatio-temporally correlated data. A profusion of systems and algorithms devoted to this end has been constructed under the conceptual framework of granular computing. This paper outlines a fuzzy set based granular evolving modeling -FBeM -approach for learning from imprecise data. Granulation arises because modeling uncertain data dispenses attention to details. The evolving aspect is fundamental to account for endless flows of nonsta-tionary data and structural adaptation of models. Experiments with Mackey-Glass benchmark data recommend that the FBeM approach outperforms alternative approaches. Resumo— Modelagem on-line de grandes volumes de dados de sistemas complexos motiva a revisão de di-versos aspectos da teoria de aprendizagem de máquina. A mineração de fluxo de dados lida com a extra...
This work outlines a new approach for online learning from imprecise data, namely, fuzzy set base... more This work outlines a new approach for online learning from imprecise data, namely, fuzzy set based evolving modeling (FBeM) approach. FBeM is an adaptive modeling framework that uses fuzzy granular objects to enclose uncer- tainty in the data. The FBeM algorithm is data flow driven and supports learning on an instance-per-instance recursive basis by developing and refining fuzzy mod- els on-demand. Structurally, FBeM models combine Mamdani and functional fuzzy systems to output granular and singular approximations of nonstationary functions. In general, approximand functions can be time series, decision boundaries between classes, and control and regression functions. Linguistic description of the behavior of the system over time is provided by information granules and associated rules. An application example on a reactive control problem, underlining the complemen- tarity of Mamdani and functional parts of the model, illustrates the usefulness of the approach. More specifically, th...
System modeling in dynamic environments needs processing of streams 1 of sensor data and incremen... more System modeling in dynamic environments needs processing of streams 1 of sensor data and incremental learning algorithms. This paper suggests an incre-2 mental granular fuzzy rule-based modeling approach using streams of fuzzy inter-3 val data. Incremental granular modeling is an adaptive modeling framework that uses 4 fuzzy granular data that originate from unreliable sensors, imprecise perceptions, or 5 description of imprecise values of a variable in the form fuzzy intervals. The incre-6 mental learning algorithm builds the antecedent of functional fuzzy rules and the 7 rule base of the fuzzy model. A recursive least squares algorithm revises the para-8 meters of a state-space representation of the fuzzy rule consequents. Imprecision in 9 data is accounted for using specificity measures. An illustrative example concerning 10 the Rossler attractor is given. 11 1 Introduction 12 Data produced by real world systems result from nonlinear, uncertain, and time-13
2007 IEEE International Electric Machines & Drives Conference, 2007
This paper describes a real-time model-based fault Estimation Module (aPE), which includes the al... more This paper describes a real-time model-based fault Estimation Module (aPE), which includes the algorithms: detection and diagnosis software. The Electric Machines Diag-Recursive Least-Squares (RLS) and Extended Kalman Filter nosis System (EMDS) covers field winding shorted-turns fault in (EKF); Dynamic Models Module (oaDM) for faults simulation; alternators and stator windings shorted-turns fault in induction motors. The EMDS has a modular architecture. The modules Computational Intelligence Module (ad), which includes: include: acquisition and data treatment; well-known parameters MLP and SOM Neural Networks and the FCM technique for estimation algorithms, such as Recursive Least Squares (RLS) patterns recognition. These modules working together generate and Extended Kalman Filter (EKF); dynamic models for faults the health state of various electric machines working in parallel simulation; faults detection and identification tools, such as through routing. The system modules are trigged as soon as M.L.P. and S.O.M. neural networks and Fuzzy C-Means (FCM)
This work outlines a new approach for online learning from imprecise data, namely, fuzzy set base... more This work outlines a new approach for online learning from imprecise data, namely, fuzzy set based evolving modeling (FBeM) approach. FBeM is an adaptive modeling framework that uses fuzzy granular objects to enclose uncertainty in the data. The FBeM algorithm is data flow driven and supports learning on an instance-per-instance recursive basis by developing and refining fuzzy models on-demand. Structurally, FBeM models combine Mamdani and functional fuzzy systems to output granular and singular approximations of nonstationary functions. In general, approximand functions can be time series, decision boundaries between classes, and control and regression functions. Linguistic description of the behavior of the system over time is provided by information granules and associated rules. An application example on a reactive control problem, underlining the complementarity of Mamdani and functional parts of the model, illustrates the usefulness of the approach. More specifically, the problem concerns sensor-based robot navigation and localization. In addition to precise singular output values, granular output values provide effective robust obstacle avoidance navigation.
Uploads
Papers by Daniel Leite