31 research outputs found
Reti neurali artificiali: identificazione di pazienti ad alto rischio cardiovascolare
No full textL’identificazione dei pazienti ad alto rischio di malattie cardiovascolari (MC) è un importante obiettivo della medicina dei paesi occidentali. Negli ultimi anni numerosi sforzi sono stati effettuati al fine di sviluppare strumenti a basso costo per il riconoscimento precoce di questi soggetti. Importanti studi epidemiologici (PROCAM, FRAMINGHAM) hanno prodotto algoritmi capaci di individuare i soggetti a rischio sulla base dei fattori di rischio. L’ispessimento medio intimale (IMT) delle carotidi extracraniche è un indice precoce di aterosclerosi generalizzata, anch’esso potenzialmente utilizzabile per l’individuazione precoce di soggetti predisposti alla patologia aterosclerotica. Le reti neurali artificiali (RNA) sono algoritmi informatici ispirati ai processi altamente interattivi del cervello umano. Come il cervello, le RNA sono in grado di decifrare i sottili meccanismi che mettono in relazione le diverse variabili in modelli sperimentali complessi e di assolvere a compiti di classificazione. Il presente studio è stato disegnato al fine di valutare la capacità delle RNA di distinguere, sulla base dei fattori di rischio convenzionali, dell’IMT carotideo o di entrambi, tra pazienti a basso o alto rischio per patologie cardiovascolari. Lo studio è stato condotto in 578 soggetti dislipidemici. Fra questi, 114 erano sintomatici per malattie cardiovascolari (infarto miocardico, angina), o cerebrovascolari (ischemia cerebrale transitoria, ictus) o per ateropatie periferiche e in quanto tali sono stati definiti ad alto rischio. I restanti 464 soggetti erano asintomatici e sono stati definiti a basso rischio. L’IMT carotideo, visualizzato mediante ultrasonografia B-mode, è stato misurato in tempo reale utilizzando il calibro elettronico della macchina stessa. Per l’analisi sono stati effettuati numerosi esperimenti utilizzando diverse reti neurali ideate dal Centro Ricerche Semeion. Nel migliore dei casi, utilizzando i fattori di rischio convenzionali come variabili di entrata nel sistema di classificazione è stata ottenuta una accuratezza di classificazione fra soggetti a basso o alto rischio (media ponderata) del 87%. Utilizzando come variabili di entrata le variabili ultrasonografiche, si otteneva una accuratezza del 77%. Aggiungendo a questo set di variabili quelle ottenibili a costo zero (età, sesso, peso, altezza e indice di massa corporea) l’accuratezza di predizione aumentava fino all’86%. L’utilizzo, nel sistema di classificazione, di tutte le variabili ecografiche e di tutti i fattori di rischio come variabili di entrata non migliorava l’accuratezza delle RNA nel compito di classificazione (accuratezza di predizione pari a circa 83%). Infine, permettendo al sistema di selezionare automaticamente le variabili più rilevanti (I.S. system-Semeion ), 31 variabili entravano nel modello e fra queste ben 6 erano variabili ultrasonografiche. Utilizzando questo set di variabili, l’accuratezza delle RNA nella classificazione dei soggetti a basso o ad alto rischio aumentava drammaticamente raggiungendo un’accuratezza globale di predizione del 92% ed un 100% di classificazione corretta dei soggetti ad alto rischio. In conclusione, le RNA sono una tecnologia promettente per lo sviluppo di strumenti diagnostici utilizzabili nella routine clinica per la classificazione di pazienti a basso e ad alto rischio di patologie vascolari.Early recognition of patients at high risk of vascular diseases (VDs) is an important goal in medicine of western countries. Efforts in developing inexpensive screening devices that can assist in the differentiation between low and high risk subjects have been numerous. Large epidemiological studies (PROCAM, FRAMINGHAM) provide algorithms, which assess the individual global risk to develop new vascular events on the basis of vascular risk factors (VRFs). The intima media thickness (IMT) of carotid arteries (CAs) evaluated by ultrasound methodologies is an early manifestation of atherosclerosis, potentially predictive of symptomatic VD. IMT is a reliable index of the presence of atherosclerosis in other vascular districts and can predict new vascular events. Artificial neural networks (ANNs) are computer algorithms inspired by highly interactive processing of human brain. Like the brain, ANNs can recognize patterns and manage data, and when exposed to complex data sets, they have the capability to learn the underlying mechanics relating different variables and to recognise complex patterns and classification tasks. The purpose of the study was to evaluate the performance of ANNs in the recognition of patients at low or high risk of VDs on the basis of conventional VRFs, IMT or both. Patients were arbitrarily assigned to the high risk group when suffering from overt cardio- (myocardial infarction or angina), cerebro- (transient ischemic attack or stroke) or peripheral-VDs. The near and far wall of left and right carotid arteries were measured from 578 patients (464 at low and 114 at high risk for VDs) using B-Mode ultrasound. CA-IMT images were processed in real time by using the electronic caliper of the machine. With optimal settings, a prediction accuracy (weighted mean) of about 87% were obtained when conventional VRFs were used as input variables in the ANN classification system. When only ultrasonic variables were used, a prediction accuracy of about 77% was observed. The addition, to this last set, of variables obtained without any additional cost (gender, age, weight, height and body mass index) led accuracy of prediction to 86%. Pooling data of all ultrasonic variables and all VRFs did not significantly improve the performance of ANNs in the classification task (prediction accuracy = 83%). Finally, when ANNs were allowed to choose automatically the relevant input data (I.S. system), 31 variables were selected and, among these, 6 were ultrasonic variables. By using this set of variables as input data the performance of ANNs in the classification task increased, reaching a prediction accuracy about 92%, with 100% of correct classification of high risk patients. In conclusion, ANN technology is promising in the development of highly specific diagnostic tools to be used for patients’ classification into low or high risk classes
Artificial neural networks in the recognition of the presence of thyroid disease in patients with atrophic body gastritis
No full textAIM: To investigate the role of artificial neural networks in predicting the presence of thyroid disease in atrophic body gastritis patients. METHODS: A dataset of 29 input variables of 253 atrophic body gastritis patients was applied to artificial neural networks (ANNs) using a data optimisation procedure (standard ANNs, T&T-IS protocol, TWIST protocol). The target variable was the presence of thyroid disease. RESULTS: Standard ANNs obtained a mean accuracy of 64.4% with a sensitivity of 69% and a specificity of 59.8% in recognizing atrophic body gastritis patients with thyroid disease. The optimization procedures (T&T-IS and TWIST protocol) improved the performance of the recognition task yielding a mean accuracy, sensitivity and specificity of 74.7% and 75.8%, 78.8% and 81.8%, and 70.5% and 69.9%, respectively. The increase of sensitivity of the TWIST protocol was statistically significant compared to T&T-IS. CONCLUSION: This study suggests that artificial neural networks may be taken into consideration as a potential clinical decision-support tool for identifying ABG patients at risk for harbouring an unknown thyroid disease and thus requiring diagnostic work-up of their thyroid status. (c) 2008 WJG. All rights reserved
B-mode ultrasound measurements of carotid intima media thickness and the assessment of global cardiovascular risk
No full textCarotid intima-media thickness (C-IMT) has been shown to be related to vascular risk factors (VRFs), prevalent cardiovascular disease (CVD), and atherosclerosis in coronary and peripheral arteries. Despite these relationships only a few studies have evaluated the potentiality of C-IMT to identify patient at high risk of CVD. In these, C-IMT included in a risk function for the assessment of global risk does not increases its predictivity. This can be due to a real lack of prediction capacity of C-IMT but also to the use of statistical tools unable to disentangle the non linear relationships connecting IMT to the global risk. Artificial neural networks (ANNs) are highly sophisticated computer algorithms able to recognise even the more hidden non linear relationships relating different variables, and to absolve complex classification tasks. In the present study the potentiality of C-IMT, alone or added to established VRFs, to identify patients at high risk of vascular disease (VD) was investigated by using ANNs and the classical statistical approach based on discriminant analyses. Patients were arbitrarily assigned to the high risk group when suffering from overt cardio- (myocardial infarction or angina), cerebro- (transient ischemic attack or stroke) or peripheral-VD. Arterial near and far wall of left and right carotid arteries were measured from 578 patients (464 at low and 114 at high risk for VD) by using B-Mode ultrasound. The results show that ANNs can be trained to identify low and high risk subjects with a greater accuracy than discriminant analyses. In addition, with optimal settings, a prediction accuracy of about 87% was achieved using conventional VRFs as input variables in the ANN classification task. When only ultrasonic variables were used, a prediction accuracy of about 77% was observed. The addition to this set of variables obtained without any additional cost (gender, age, weight, height and body mass index) led the accuracy of prediction to 86%. Pooling data of all ultrasonic variables and all VRFs did not significantly improve the performance of ANNs in the classification task (prediction accuracy = 83%). Finally, when ANNs were allowed to choose automatically the relevant input data (I.S. system-Semeion), 31 variables were selected and, among these, 6 were ultrasonic variables. By using this set of variables as input data the performance of ANNs in the classification task increased, reaching a prediction accuracy close to 92%, with 100% of correct classification of high risk patients. In conclusion, with the ANN technology C-IMT may increase the discriminant capacity of vascular risk factors in the classification of patients into low or high risk classes
Prediction of substitutive pharmacological treatment of hospitalized drug addicts using neural networks
No full textUse of artificial networks in clinical trials: a pilot study to predict responsiveness to donepezil in Alzheimer's disease
No full textObjectives: To evaluate the accuracy of artificial neural networks compared with discriminant analysis in classifying positive and negative response to the cholinesterase inhibitor donepezil in a group of Alzheimer's disease (AD) patients.
Design: Convenience sample.
Setting: Patients with mild to moderate AD consecutively admitted to a geriatric day hospital and treated with donepezil 5 mg/day.
Participants: Sixty-one older patients of both sexes with AD.
Measurements: Accuracy in detecting subjects sensitive (responders) or not (nonresponders) to 3-month therapy with ANNs. The criterion standard for evaluation of efficacy was the scores of Alzheimer's Disease Assessment Scale-Cognitive portion and Clinician's Interview Based Impression of Change-plus scales.
Results: ANNs were more effective in discriminating between responders and nonresponders than other advanced statistical methods, particularly linear discriminant analysis. The total accuracy in predicting the outcome was 92.59%.
Conclusions: ANNs appear to be a useful tool in detecting patient responsiveness to pharmacological treatment in AD
Artificial Neural Network assessment of substitutive pharmacological treatments in hospitalized intravenous drug users
No full textArtificial neural networks (ANNs) provide better solutions than linear discriminant analysis (LDA) to problems of classification and estimation involving a large number of non-homogeneous (categorical and metric) variables. In this study, we compared the ability of traditional LDA and a feed-forward back-propagation (FF-BP) ANN with self-momentum to predict pharmacological treatments received by intravenous drug users (IDUs) hospitalised for coexisting medical illness. When medical staff considered detoxification appropriate they usually suggested methadone (MET) and (or) benzodiazepines (BDZ). Given four different treatment options (MET, BDZ, MET+BDZ, no treatment) as dependent variables and 38 independent variables, the FF-BP ANN provided the best prediction of the consultant's decision (overall accuracy: 62.7%). It achieved the highest level of predictive accuracy for the BDZ option (90.5%), the lowest for no treatment (29.6), often misclassifying no treatment as BDZ. The LDA yielded a lower mean accuracy (50.3%). When the untreated group was excluded, ANN improved its absolute recognition rate by only 1.2% and the BDZ group remained the best predicted. In contrast, LDA improved its absolute recognition rate from 50.3 to 58.9%, maximum 65.7% for the BDZ group. In conclusion, the FF-BP ANN was more accurate than the statistical model (discriminant analysis) in predicting the pharmacological treatment of IDUs
Possible contribution of artificial neural networks and linear discriminant analysis in recognition of patients with suspected atrophic body gastritis.
No full textIntima media thickness and vascular risk factors for the recognition of patients at high risk of atherosclerosis
No full textWe have previously shown in a large cross-sectional study that intima media thickness (IMT) of carotid arteries, as measured in the normal clinical practice, correlated with most of the vascular risk factors (VRFs), and discriminated well between patients with and without previous history of cardiovascular events. This approach, however, does not provide information on the cardiovascular risk of individual patients, but rather it allows the identification of groups of patients. Artificial neural networks (ANNs) are computer algorithms inspired by highly interactive processing of human brain. ANNs are able to recognize patterns and, when exposed to complex data sets, they have the capability to learn the underlying mechanisms relating different variables and to recognize classification tasks. In present study, we have assessed the performance of ANNs in the recognition of patients at low (n= 464) or high risk (n=114) of vascular disease on the basis of conventional VRFs, IMT or both. Patients were arbitrarily assigned to the high risk group when suffering from overt cardiovascular disease. With optimal settings, a prediction accuracy of about 87% were obtained when conventional VRFs were used as input variables in the ANN classification system, whereas a 77% was obtained with only ultrasonic variables. The addition to ultrasonic variables of gender, age, weight, height and body mass index led this figure to 86%. When ANNs were allowed to choose automatically the relevant input data (I.S. system) 31 variables were selected comprising 6 ultrasonic variables. With this input, the performance of ANNs in the classification task increased reaching a prediction accuracy about 92%, with 100% of correct classification of high risk patients. In conclusion, ANN technology is promising in the development of highly specific diagnostic tools to be used for patients’ classification into low or high risk classes
ARTIFICIAL NEURAL NETWORKS IN THE RECOGNITION OF PATIENTS AT HIGH RISK OF CARDIOVASCULAR DISEASE
No full textWe have previously shown, in a large cross-sectional study, that intima media thickness (IMT) of carotid arteries, as measured in the normal clinical practice, correlated with most of the vascular risk factors (VRFs), and discriminated well between patients with and without previous history of cardiovascular events. This approach, however, does not provide information on the cardiovascular risk of individual patients, but rather it allows the identification of groups of patients. Artificial neural networks (ANNs) are computer algorithms inspired by highly interactive processing of human brain. ANNs are able to recognize patterns and, when exposed to complex data sets, they have the capability to learn the underlying mechanisms relating different variables and to recognize classification tasks. In present study, we have assessed the performance of ANNs in the recognition of patients at low (n= 464) or high risk (n=114) of vascular disease on the basis of conventional VRFs, IMT or both. Patients were arbitrarily assigned to the high risk group when suffering from overt cardiovascular disease. With optimal settings, a prediction accuracy of about 87% were obtained when conventional VRFs were used as input variables in the ANN classification system, whereas a 77% was obtained with only ultrasonic variables. The addition to ultrasonic variables of gender, age, weight, height and body mass index led this figure to 86%. When ANNs were allowed to choose automatically the relevant input data (I.S. system) 31 variables were selected comprising 6 ultrasonic variables. With this input, the performance of ANNs in the classification task increased reaching a prediction accuracy about 92%, with 100% of correct classification of high risk patients. In conclusion, ANN technology is promising in the development of highly specific diagnostic tools to be used for patients’ classification into low or high risk classes
