Aleksandar Peulic

ARTreat project: Three-dimensional numerical simulation of plaque formation and development in the arteries

Atherosclerosis is a progressive disease characterized by the accumulation of lipids and fibrous elements in arteries. It is characterized by dysfunction of endothelium and vasculitis, and accumulation of lipid, cholesterol, and cell elements inside blood vessel wall. In this study, a continuum-based approach for plaque formation and development in 3-D is presented. The blood flow is simulated by the 3-D Navier-Stokes equations, together with the continuity equation while low-density lipoprotein (LDL) transport in lumen of the vessel is coupled with Kedem-Katchalsky equations. The inflammatory process was solved using three additional reaction-diffusion partial differential equations. Transport of labeled LDL was fitted with our experiment on the rabbit animal model. Matching with histological data for LDL localization was achieved. Also, 3-D model of the straight artery with initial mild constriction of 30% plaque for formation and development is presented.

Thermography based breast cancer detection using texture features and minimum variance quantization.

In this paper, we present a system based on feature extraction techniques and image segmentation techniques for detecting and diagnosing abnormal patterns in breast thermograms. The proposed system consists of three major steps: feature extraction, classification into normal and abnormal pattern and segmentation of abnormal pattern. Computed features based on gray-level co-occurrence matrices are used to evaluate the effectiveness of textural information possessed by mass regions. A total of 20 GLCM features are extracted from thermograms. The ability of feature set in differentiating abnormal from normal tissue is investigated using a Support Vector Machine classifier, Naive Bayes classifier and K-Nearest Neighbor classifier. To evaluate the classification performance, five-fold cross validation method and Receiver operating characteristic analysis was performed. The verification results show that the proposed algorithm gives the best classification results using K-Nearest Neighbor classifier and a accuracy of 92.5%. Image segmentation techniques can play an important role to segment and extract suspected hot regions of interests in the breast infrared images. Three image segmentation techniques: minimum variance quantization, dilation of image and erosion of image are discussed. The hottest regions of thermal breast images are extracted and compared to the original images. According to the results, the proposed method has potential to extract almost exact shape of tumors.

Transient finite element modeling of functional electrical stimulation.

Transcutaneous functional electrical stimulation is commonly used for strengthening muscle. However, transient effects during stimulation are not yet well explored. The effect of an amplitude change of the stimulation can be described by static model, but there is no differency for different pulse duration. The aim of this study is to present the finite element (FE) model of a transient electrical stimulation on the forearm. Discrete FE equations were derived by using a standard Galerkin procedure. Different tissue conductive and dielectric properties are fitted using least square method and trial and error analysis from experimental measurement. This study showed that FE modeling of electrical stimulation can give the spatial-temporal distribution of applied current in the forearm. Three different cases were modeled with the same geometry but with different input of the current pulse, in order to fit the tissue properties by using transient FE analysis. All three cases were compared with experimental measurements of intramuscular voltage on one volunteer.

Electromagnetic field investigation on different cancer cell lines

BackgroundThere is a strong interest in the investigation of extremely low frequency Electromagnetic Fields (EMF) in the clinic. While evidence about anticancer effects exists, the mechanism explaining this effect is still unknown.MethodsWe investigated in vitro, and with computer simulation, the influence of a 50 Hz EMF on three cancer cell lines: breast cancer MDA-MB-231, and colon cancer SW-480 and HCT-116. After 24 h preincubation, cells were exposed to 50 Hz extremely low frequency (ELF) radiofrequency EMF using in vitro exposure systems for 24 and 72 h. A computer reaction-diffusion model with the net rate of cell proliferation and effect of EMF in time was developed. The fitting procedure for estimation of the computer model parameters was implemented.ResultsExperimental results clearly showed disintegration of cells treated with a 50 Hz EMF, compared to untreated control cells. A large percentage of treated cells resulted in increased early apoptosis after 24 h and 72 h, compared to the controls. Computer model have shown good comparison with experimental data.ConclusionUsing EMF at specific frequencies may represent a new approach in controlling the growth of cancer cells, while computer modelling could be used to predict such effects and make optimisation for complex experimental design. Further studies are required before testing this approach in humans.

Comparative analysis of breast cancer detection in mammograms and thermograms

Abstract In this paper, we present a system based on feature extraction techniques for detecting abnormal patterns in digital mammograms and thermograms. A comparative study of texture-analysis methods is performed for three image groups: mammograms from the Mammographic Image Analysis Society mammographic database; digital mammograms from the local database; and thermography images of the breast. Also, we present a procedure for the automatic separation of the breast region from the mammograms. Computed features based on gray-level co-occurrence matrices are used to evaluate the effectiveness of textural information possessed by mass regions. A total of 20 texture features are extracted from the region of interest. The ability of feature set in differentiating abnormal from normal tissue is investigated using a support vector machine classifier, Naive Bayes classifier and K-Nearest Neighbor classifier. To evaluate the classification performance, five-fold cross-validation method and receiver operating characteristic analysis was performed.

Application of data mining algorithms for mammogram classification

One of the leading causes of cancer death among women is breast cancer. In our work we aim at proposing a prototype of a medical expert system (based on data mining techniques) that could significantly aid medical experts to detect breast cancer. This paper presents the CAD (computer aided diagnosis) system for the detection of normal and abnormal pattern in the breast. The proposed system consists of four major steps: the image preprocessing, the feature extraction, the feature selection and the classification process that classifies mammogram into normal (without tumor) and abnormal (with tumor) pattern. After removing noise from mammogram using the Discrete Wavelet Transformation (DWT), first is selected the region of interest (ROI). By identifying the boundary of the breast, it is possible to remove any artifact present outside the breast area, such as patient markings. Then, a total of 20 GLCM features are extracted from the ROI, which were used as inputs for classification algorithms. In order to compare the classification results, we used seven different classifiers. Normal breast images and breast image with masses (total 322 images) used as input images in this study are taken from the mini-MIAS database.

Segmentation for the enhancement of microcalcifications in digital mammograms

Microcalcification clusters appear as groups of small, bright particles with arbitrary shapes on mammographic images. They are the earliest sign of breast carcinomas and their detection is the key for improving breast cancer prognosis. But due to the low contrast of microcalcifications and same properties as noise, it is difficult to detect microcalcification. This work is devoted to developing a system for the detection of microcalcification in digital mammograms. After removing noise from mammogram using the Discrete Wavelet Transformation (DWT), we first selected the region of interest (ROI) in order to demarcate the breast region on a mammogram. Segmenting region of interest represents one of the most important stages of mammogram processing procedure. The proposed segmentation method is based on a filtering using the Sobel filter. This process will identify the significant pixels, that belong to edges of microcalcifications. Microcalcifications were detected by increasing the contrast of the images obtained by applying Sobel operator. In order to confirm the effectiveness of this microcalcification segmentation method, the Support Vector Machine (SVM) and k-Nearest Neighborhood (k-NN) algorithm are employed for the classification task using cross-validation technique.

Parameter optimization of a computer-aided diagnosis system for detection of masses on digitized mammograms

BACKGROUND Reading mammograms is a difficult task and for this reason any development that may improve the performance in breast cancer screening is of great importance. OBJECTIVE We proposed optimized computer aided diagnosis (CAD) system, equipped with reliability estimate module, for mass detection on digitized mammograms. METHODS Proposed CAD system consists of four major steps: preprocessing, segmentation, feature extraction and classification. We propose a simple regression function as a threshold function for extraction of potential masses. By running optimization procedure we estimate parameters of the preprocessing and segmentation steps thus ensuring maximum mass detection sensitivity. In addition to the classification, where we tested seven different classifiers, the CAD system is equipped with reliability estimate module. RESULTS By performing segmentation 91.3% of masses were correctly segmented with 4.14 false positives per image (FPpi). This result is improved in the classification phase where, among the seven tested classifiers, multilayer perceptron neural network achieved the best result including 77.4% sensitivity and 0.49 FPpi. CONCLUSION By using the proposed regression function and parameter optimization we were able to improve segmentation results comparing to the literature. In addition, we showed that CAD system has high potential for being equipped with reliability estimate module.

Arterial stiffness modeling using variations of pulse transit time

In this paper, a finite elements (FE) modeling is used to model effects of the arterial stiffness on the different signal patterns of the pulse transit time (PTT). Four different breathing patterns of the same subject are measured with PTT signal and corresponding finite element model of the straight elastic artery is applied. The computational fluid structure model provides arterial elastic behavior and fitting procedure was applied in order to estimate stiffness of the artery. It was found that same elastic material characteristics were fitted for four different breathing patterns which validate this methodology for possible noninvasive determination of the arterial stiffness.

Wireless Sensor Network Wavelet Signal Processing

Application of information technologies (IT) and wireless networking will liberate people from such confinement and enable continuous real time monitoring of physiological data, which is vital for medical care. To achieve this goal, it is necessary research and development on wearable intelligent sensor devices, sensor miniaturization, signal processing, wireless transmission, and databases for these vital data. Our goal is to implement wavelet transformations in our measured system.