Paolo Barbini

Automated diagnosis of pigmented skin lesions.

Since advanced melanoma remains practically incurable, early detection is an important step toward a reduction in mortality. High expectations are entertained for a technique known as dermoscopy or epiluminescence light microscopy; however, evaluation of pigmented skin lesions by this method is often extremely complex and subjective. To obviate the problem of qualitative interpretation, methods based on mathematical analysis of pigmented skin lesions, such as digital dermoscopy analysis, have been developed. In the present study, we used a digital dermoscopy analyzer (DBDermo‐Mips system) to evaluate a series of 588 excised, clinically atypical, flat pigmented skin lesions (371 benign, 217 malignant). The analyzer evaluated 48 parameters grouped into 4 categories (geometries, colors, textures and islands of color), which were used to train an artificial neural network. To evaluate the diagnostic performance of the neural network and to check it during the training process, we used the error area over the receiver operating characteristic curve. The discriminating power of the digital dermoscopy analyzer plus artificial neural network was compared with histologic diagnosis. A feature selection procedure indicated that as few as 13 of the variables were sufficient to discriminate the 2 groups of lesions, and this also ensured high generalization power. The artificial neural network designed with these variables enabled a diagnostic accuracy of about 94%. In conclusion, the good diagnostic performance and high speed in reading and analyzing lesions (real time) of our method constitute an important step in the direction of automated diagnosis of pigmented skin lesions.

CADCS simulation of the closed-loop cardiovascular system

A pulsatile simulator of the closed-loop cardiovascular system, designed to solve simulation, identification and control problems in a research and education context, is presented. Its implementation makes use of a command-driven interactive program for simulation of non-linear ordinary differential equations. The flexibility of the simulator is demonstrated by the results presented which refer to a basal steady-state circulatory condition as well as a transient induced by an abrupt change in peripheral resistance.

A new approach for tracking respiratory mechanical parameters in real-time

A new recursive least-squares procedure for on-line tracking of changes in viscoelastic properties of respiratory mechanics is proposed and applied to artificially ventilated patients. Classical least-squares methods based on simple first-order linear models with time-constant parameters generally provide systematic residuals that hardly satisfy standard statistical tests for model validation in terms of residuals. On the other hand, high order and/or nonlinear models introduce parameters whose estimates are of difficult interpretation in a clinical context. The present procedure overcomes these limitations by using the well-known first-order model of respiratory mechanics, wherein variability of resistance and elastance during the breathing cycle is allowed to take into account nonlinear and high-order behavior. Mean and standard deviation of resistance and elastance estimates, relative to a respiratory cycle, are then determined recursively. Feasibility of the method is evaluated by applying it both to experimental and simulated pressure-airflow signals measured in an intensive care unit during mechanical ventilation of patients recovering from heart surgery. Results demonstrate that the proposed procedure provides data description satisfying statistical tests, such as residual whiteness, and reliable estimates of viscoelastic lung parameters even during substantial and fast variations in the respiratory status. In addition, unlike classical methods, the new technique provides the means for on-line evaluation of parameter variability during each respiratory cycle, by the estimate of their standard deviations. This is important in clinical practice, because only the knowledge of reliable parameter values and standard deviations enables significant changes in the respiratory viscoelastic characteristics, and thus in patient status, to be assessed.

Time-Varying Mechanical Properties of the Left Ventricle-A Computer Simulation

A numerical model of left ventricular (LV) pump function, incorporating cardiac muscle mechanics and LV geometry, was used to derive a simple linear model of local LV contractile properties. This simplified model views the ventricle as a pressure generator (related to isovolumic contraction) coupled with two time-varying elements: 1) a viscous term (related to the dissipative properties of the myocardium), and 2) an elastic term (related to the tension-length curve of activated fiber and to LV geometry).

A Versatile Identification Method Applied to Analysis of Respiratory Mechanics

A method for the straightforward estimation of the param eters in lumped continuous linear models of respiratory mechanics is presented in this paper. First, a discrete-time model is obtained from the samples of input-output data and then transformed into a conuous-time model.

DNA-recognition process described by MD simulations of the lactose repressor protein on a specific and a non-specific DNA sequence

The lactose repressor protein may bind DNA in two possible configurations: a specific one, if the DNA sequence corresponds to a binding site, and a non-specific one otherwise. To find its target sequences, the lactose repressor first binds non-specifically to DNA, and subsequently, it rapidly searches for a binding site. Atomic structures of non-specific and specific complexes are available from crystallographic and nuclear magnetic resonance experiments. However, what remains unknown is a detailed description of the steps that transform the non-specific complex into the specific one. Here, how the protein first recognizes its binding site has been studied using molecular dynamics simulations. The picture that emerges is that of a protein that is as mobile when interacting with non-specific DNA sequences as when free in solution. This high degree of mobility allows the protein to rapidly sample different DNA sequences. In contrast, when the protein encounters a binding site, the configuration ensemble collapses, and the protein sliding movements along the DNA sequence become scarce. The binding energies in the specific and non-specific complexes were analysed using the Molecular Mechanics Poisson Boltzmann Surface Area approach. These results represent a first step towards a throughout characterization of the DNA-recognition process.

Real-time tracking of parameters of lung mechanics : emphasis on algorithm tuning

We consider the problem of tracking rapid changes in the viscous and elastic properties of the respiratory system by using mouth flow and transpulmonary pressure data measured during mechanical ventilation. A recursive least-squares algorithm with adjustable compensator is used for online estimation of an R-C model of the breathing mechanics. Specific simulation experiments are presented to provide guidelines to select suitable values for the key variable, which controls the compromise between tracking ability and noise sensitivity. The results obtained confirm the critical role of the optimum tuning in relation to the noise level. Experimental results obtained from data measured on mechanically-ventilated dogs, in which respiratory distress syndrome was intravenously induced by oleic acid, demonstrate that the tuned algorithm is able to track appropriately both the viscous and elastic properties of lung mechanics. Parameter estimates are consistent with those obtained by standard and robust offline algorithms and their time course is in qualitative agreement with known physiopathological behaviour.

A comparative analysis of predictive models of morbidity in intensive care unit after cardiac surgery – Part I: model planning

BackgroundDifferent methods have recently been proposed for predicting morbidity in intensive care units (ICU). The aim of the present study was to critically review a number of approaches for developing models capable of estimating the probability of morbidity in ICU after heart surgery. The study is divided into two parts. In this first part, popular models used to estimate the probability of class membership are grouped into distinct categories according to their underlying mathematical principles. Modelling techniques and intrinsic strengths and weaknesses of each model are analysed and discussed from a theoretical point of view, in consideration of clinical applications.MethodsModels based on Bayes rule, k- nearest neighbour algorithm, logistic regression, scoring systems and artificial neural networks are investigated. Key issues for model design are described. The mathematical treatment of some aspects of model structure is also included for readers interested in developing models, though a full understanding of mathematical relationships is not necessary if the reader is only interested in perceiving the practical meaning of model assumptions, weaknesses and strengths from a user point of view.ResultsScoring systems are very attractive due to their simplicity of use, although this may undermine their predictive capacity. Logistic regression models are trustworthy tools, although they suffer from the principal limitations of most regression procedures. Bayesian models seem to be a good compromise between complexity and predictive performance, but model recalibration is generally necessary. k-nearest neighbour may be a valid non parametric technique, though computational cost and the need for large data storage are major weaknesses of this approach. Artificial neural networks have intrinsic advantages with respect to common statistical models, though the training process may be problematical.ConclusionKnowledge of model assumptions and the theoretical strengths and weaknesses of different approaches are fundamental for designing models for estimating the probability of morbidity after heart surgery. However, a rational choice also requires evaluation and comparison of actual performances of locally-developed competitive models in the clinical scenario to obtain satisfactory agreement between local needs and model response. In the second part of this study the above predictive models will therefore be tested on real data acquired in a specialized ICU.

A dynamic morphometric model of the normal lung for studying expiratory flow limitation in mechanical ventilation

A nonlinear dynamic morphometric model of breathing mechanics during artificial ventilation is described. On the basis of the Weibel symmetrical representation of the tracheobronchial tree, the model accurately accounts for the geometrical and mechanical characteristics of the conductive zone and packs the respiratory zone into a viscoelastic Voigt body. The model also accounts for the main mechanisms limiting expiratory flow (wave speed limitation and viscous flow limitation), in order to reproduce satisfactorily, under dynamic conditions, the expiratory flow limitation phenomenon occurring in normal subjects when the difference between alveolar pressure and tracheal pressure (driving pressure) is high. Several expirations characterized by different levels of driving pressure are simulated and expiratory flow limitation is detected by plotting the isovolume pressure–flow curves. The model is used to study the time course of resistance and total cross-sectional area as well as the ratio of fluid velocity to wave speed (speed index), in conductive airway generations. The results highlight that the coupling between dissipative pressure losses and airway compliance leads to onset of expiratory flow limitation in normal lungs when driving pressure is increased significantly by applying a subatmospheric pressure to the outlet of the ventilator expiratory channel; wave speed limitation becomes predominant at still higher driving pressures.

Morphometric distinction of low- and high-grade dysplasias in gastric biopsies

Substantial agreement has previously been demonstrated between qualitative and morphometric grading of gastric dysplasia. In the present study, a further attempt is made to quantitatively define the dysplastic changes in relation to associated benign or malignant changes of gastric mucosa. In total, 232 cases were studied and were associated with benign peptic ulcer (89 cases), histologically proven gastric cancer (88 cases), and gastritis-associated mild, moderate, and severe dysplasias (55 cases). The analysis showed that one discriminant function consisting of seven morphometric features is sufficient to separate the regenerative and cancer-associated cases. The classification results obtained on the basis of this discriminant function in both training and testing sets are encouragingly similar, indicating that the classification is reliable. This was further confirmed by the results of the application of this rule in the mild, moderate, and severe dysplasia biopsies (the above-mentioned gastritis-associated cases) used in a previous study. The quantitative analysis permits two grades, instead of three, to be distinguished: low-grade and high-grade dysplasia.