True Negative Fraction Research Articles

Breast Tumor Detection in MR Images Based on Density

Breast cancer is dangerous in women. It is generally found after the symptoms appear. Detecting the breast cancer at an early stage and understanding the treatment are the most important strategies to prevent death from cancer. Generally, for detection of breast cancer, breast Magnetic Resonance Image (MRI) takes place. It is one of the best approaches to detect tumor in women. In this research paper, a combination of selection methods for seed region growing image segmentation is suggested to detect breast tumor. The suggested method has been divided into following parts: First, the pre-processing of breast image is performed. Second, the automatic threshold for binarization process is calculated. Third, the number of seed points and its position in the breast image are determined automatically using density of pixels value. Fourth, a method for calculation of threshold value is proposed for the purpose of region creation in seed region growing. For the evaluation purpose, the proposed method was applied and tested on the RIDER MRI breast dataset from National Biomedical Imaging Archive (NBIA). After the test was performed, it was observed that proposed algorithm gives 90% accuracy, 88% True Negative Fraction, 91% True Positive Fraction, 10% Misclassification Rate, 94% Precision and 86% Relative Overlap which is better than other existing methods. It not only gives better evaluation measure but also provides segmentation method for multiple tumor detection.

Computer-aided segmentation system for breast MRI tumour using modified automatic seeded region growing (BMRI-MASRG).

In this paper, an automatic computer-aided detection system for breast magnetic resonance imaging (MRI) tumour segmentation will be presented. The study is focused on tumour segmentation using the modified automatic seeded region growing algorithm with a variation of the automated initial seed and threshold selection methodologies. Prior to that, some pre-processing methodologies are involved. Breast skin is detected and deleted using the integration of two algorithms, namely the level set active contour and morphological thinning. The system is applied and tested on 40 test images from the RIDER breast MRI dataset, the results are evaluated and presented in comparison to the ground truths of the dataset. The analysis of variance (ANOVA) test shows that there is a statistically significance in the performance compared to the previous segmentation approaches that have been tested on the same dataset where ANOVA p values for the evaluation measures' results are less than 0.05, such as: relative overlap (p = 0.0002), misclassification rate (p = 0.045), true negative fraction (p = 0.0001) and sum of true volume fraction (p = 0.0001).

Application of a neuro-fuzzy model to landslide-susceptibility mapping for shallow landslides in a tropical hilly area

This paper presents landslide-susceptibility mapping using an adaptive neuro-fuzzy inference system (ANFIS) using a geographic information system (GIS) environment. In the first stage, landslide locations from the study area were identified by interpreting aerial photographs and supported by an extensive field survey. In the second stage, landslide-related conditioning factors such as altitude, slope angle, plan curvature, distance to drainage, distance to road, soil texture and stream power index (SPI) were extracted from the topographic and soil maps. Then, landslide-susceptible areas were analyzed by the ANFIS approach and mapped using landslide-conditioning factors. In particular, various membership functions (MFs) were applied for the landslide-susceptibility mapping and their results were compared with the field-verified landslide locations. Additionally, the receiver operating characteristics (ROC) curve for all landslide susceptibility maps were drawn and the areas under curve values were calculated. The ROC curve technique is based on the plotting of model sensitivity — true positive fraction values calculated for different threshold values, versus model specificity — true negative fraction values, on a graph. Landslide test locations that were not used during the ANFIS modeling purpose were used to validate the landslide susceptibility maps. The validation results revealed that the susceptibility maps constructed by the ANFIS predictive models using triangular, trapezoidal, generalized bell and polynomial MFs produced reasonable results (84.39%), which can be used for preliminary land-use planning. Finally, the authors concluded that ANFIS is a very useful and an effective tool in regional landslide susceptibility assessment.

Use of GIS-based fuzzy logic relations and its cross application to produce landslide susceptibility maps in three test areas in Malaysia

Landslides are one of the most frequent and common natural hazards in Malaysia. Preparation of landslide susceptibility maps is one of the first and most important steps in the landslide hazard mitigation. However, due to complex nature of landslides, producing a reliable susceptibility map is not easy. For this reason, a number of different approaches have been used, including direct and indirect heuristic approaches, deterministic, probabilistic, statistical, and data mining approaches. Moreover, these landslides can be systematically assessed and mapped through a traditional mapping framework using geoinformation technologies. Since the early 1990s, several mathematical models have been developed and applied to landslide hazard mapping using geographic information system (GIS). Among various approaches, fuzzy logic relation for mapping landslide susceptibility is one of the techniques that allows to describe the role of each predisposing factor (landslide-conditioning parameters) and their optimal combination. This paper presents a new attempt at landslide susceptibility mapping using fuzzy logic relations and their cross application of membership values to three study areas in Malaysia using a GIS. The possibility of capturing the judgment and the modeling of conditioning factors are the main advantages of using fuzzy logic. These models are capable to capture the conditioning factors directly affecting the landslides and also the inter-relationship among them. In the first stage of the study, a landslide inventory was complied for each of the three study areas using both field surveys and airphoto studies. Using total 12 topographic and lithological variables, landslide susceptibility models were developed using the fuzzy logic approach. Then the landslide inventory and the parameter maps were analyzed together using the fuzzy relations and the landslide susceptibility maps produced. Finally, the prediction performance of the susceptibility maps was checked by considering field-verified landslide locations in the studied areas. Further, the susceptibility maps were validated using the receiver-operating characteristics (ROC) success rate curves. The ROC curve technique is based on plotting model sensitivity—true positive fraction values calculated for different threshold values versus model specificity—true negative fraction values on a graph. The ROC curves were calculated for the landslide susceptibility maps obtained from the application and cross application of fuzzy logic relations. Qualitatively, the produced landslide susceptibility maps showed greater than 82% landslide susceptibility in all nine cases. The results indicated that, when compared with the landslide susceptibility maps, the landslides identified in the study areas were found to be located in the very high and high susceptibility zones. This shows that as far as the performance of the fuzzy logic relation approach is concerned, the results appeared to be quite satisfactory, the zones determined on the map being zones of relative susceptibility.

Classification of burn injuries using near-infrared spectroscopy

Early surgical management of those burn injuries that will not heal spontaneously is critical. The decision to excise and graft is based on a visual assessment that is often inaccurate but yet continues to be the primary means of grading the injury. Superficial and intermediate partial-thickness injuries generally heal with appropriate wound care while deep partial- and full-thickness injuries generally require surgery. This study explores the possibility of using near-infrared spectroscopy to provide an objective and accurate means of distinguishing shallow injuries from deeper burns that require surgery. Twenty burn injuries are studied in five animals, with burns covering <1% of the total body surface area. Carefully controlled superficial, intermediate, and deep partial-thickness injuries as well as full-thickness injuries could be studied with this model. Near-infrared reflectance spectroscopy was used to evaluate these injuries 1 to 3 hours after the insult. A probabilistic model employing partial least-squares logistic regression was used to determine the degree of injury, shallow (superficial or intermediate partial) from deep (deep partial and full thickness), based on the reflectance spectrum of the wound. A leave-animal-out cross-validation strategy was used to test the predictive ability of a 2-latent variable, partial least-squares logistic regression model to distinguish deep burn injuries from shallow injuries. The model displayed reasonable ranking quality as summarized by the area under the receiver operator characteristics curve, AUC = 0.879. Fixing the threshold for the class boundaries at 0.5 probability, the model sensitivity (true positive fraction) to separate deep from shallow burns was 0.90, while model specificity (true negative fraction) was 0.83. Using an acute porcine model of thermal burn injuries, the potential of near-infrared spectroscopy to distinguish between shallow healing burns and deeper burn injuries was demonstrated. While these results should be considered as preliminary and require clinical validation, a probabilistic model capable of differentiating these classes of burns would be a significant aid to the burn specialist.

Statistical Considerations in Combining Biomarkers for Disease Classification

The recent advances in genomics (gene expression arrays and SNPs), proteomics (protein expression using mass spectrometry or antibody arrays) opened the door for biomedical researchers to combine multiple biomarkers measured using noninvasive procedures (e.g., samples from serum, urine, stool) for disease classifications (detection, diagnosis, and prognosis). This is important because for most diseases a single biomarker is not adequate in terms of classification performance, e.g., sensitivity (true positive fraction) and specificity (true negative fraction), for intended clinical use. By combining the biomarkers, we hope to achieve more accurate disease classifications and increased clinical benefits. However, this great promise comes with great challenges. The number of ways of combining multiple biomarkers increases exponentially with the number of biomarkers. As a consequence, searching for the optimal combination of biomarkers is not only computationally demanding but also run the high risk of false positive findings. False positive findings waste precious resources used for further investigations, or potentially cause harm to patients if they are put for clinical use without firm validations. The goal of this paper is to provide biomarker researchers and analysts who evaluate biomarkers for potential clinical use some guidance on how to combine biomarkers. The organization of the paper is as follows: Section 2 discusses the challenges due to the high dimensionality of modern genomic and proteomic data, “the curse of dimensionality”, and the importance of a test set of samples in such investigations. When a test set is not practical, we propose to use Cross-Validation for model selection and a bootstrap procedure for estimating the prediction error of the selected model. The principles of the optimal approach to combine biomarkers when the joint distribution of biomarkers for each of the disease and control groups is known are described in Section 3. This optimality is usually unattainable as we often have little knowledge about these joint distributions. However, the principles provide justifications of choosing one procedure over another in the settings when we do not know the joint distributions of biomarkers. For such settings, three procedures are recommended and contrasted in Section 4: nonparametric method, logistic regression, and boosting. Section 5 discusses a more challenging situation: the identities of biomarkers are unknown, or known with some uncertainty. This is typical for the protein profile data from mass spectrometry, due to both the measurement errors in protein mass (mass/charge) and the nature of such exploratory protein-expression studies mining for disease classification profiles without first identifying proteins/peptides. Concluding remarks are given in Section 6.

Applications of neural network analyses to in vivo 1H magnetic resonance spectroscopy of Parkinson disease patients.

To apply neural network analyses to in vivo magnetic resonance spectra of controls and Parkinson disease (PD) patients for the purpose of classification. Ninety-seven in vivo proton magnetic resonance spectra of the basal ganglia were recorded from 31 patients with (PD) and 14 age-matched healthy volunteers on a 1.5-T imager. The PD patients were grouped as follows: probable PD (N = 15), possible PD (N = 11), and atypical PD (N = 5). Total acquisition times of approximately five minutes were achieved with a TE (echo time) of 135 msec, a TR (repetition time) of 2000 msec, and 128 scan averages. Neural network (back propagation, Kohonen, probabilistic, and radial basis function) and related (generative topographic mapping) data analyses were performed. Conventional data analysis showed no statistically significant differences in metabolite ratios based on measuring signal intensities. The trained networks could distinguish control from PD with considerable accuracy (true positive fraction 0.971, true negative fraction 0.933). When four classes were defined, approximately 88% of the predictions were correct. The multivariate analysis indicated metabolic changes in the basal ganglia in PD. A variety of neural network and related approaches can be successfully applied to both qualitative visualization and classification of in vivo spectra of PD patients.

New Insight into the Comparative Power Quality-Control Rules That Use Control Observations within a Single Analytical Run

The error detection characteristics of quality-control (QC) rules that use control observations within a single analytical run are investigated. Unlike the evaluation of QC rules that span multiple analytical runs, most of the fundamental results regarding the performance of QC rules applied within a single analytical run can be obtained from statistical theory, without the need for simulation studies. The case of two control observations per run is investigated for ease of graphical display, but the conclusions can be extended to more than two control observations per run. Results are summarized in a graphical format that offers many interesting insights into the relations among the various QC rules. The graphs provide heuristic support to the theoretical conclusions that no QC rule is best under all error conditions, but the multirule that combines the mean rule and a within-run standard deviation rule offers an attractive compromise.

Bone and soft-tissue lesions: diagnosis with combined H-1 MR imaging and P-31 MR spectroscopy.

The feasibility of combined magnetic resonance (MR) imaging and surface coil phosphorus-31 MR spectroscopy at 1.5 T was examined in a clinical study of 34 patients before biopsy of bone or soft-tissue lesions of the extremity and trunk. The results confirmed the inability of MR imaging alone to distinguish most benign lesions from malignant ones. Malignant lesions were distinguished from benign lesions on the basis of significantly higher mean peak ratios of phosphomonoester (PME) to beta-nucleoside triphosphate (NTP) and of phosphodiester to NTP, a significantly lower mean peak area ratio of phosphocreatine to NTP, and a higher mean pH. The diagnostic performance of the PME/NTP peak area ratio is characterized by a sensitivity (true-positive fraction) of 1.00 and a specificity (true-negative fraction) of 0.93. This study provided preliminary evidence that P-31 MR spectroscopy may be used to improve diagnostic specificity in bone and soft-tissue lesions.

Structural characteristics of compounds that can be activated to chemically reactive metabolites: use for a prediction of a carcinogenic potential.

Many mutagens and carcinogens act via covalent interaction of metabolic intermediates with DNA in the target cell. This report groups those structural elements which are often found to form the basis for a metabolism to such chemically reactive metabolites. Compounds which are chemically reactive per se and which do not require metabolic activation form group 1. Group 2 comprises of olefins and aromatic hydrocarbons where the oxidation via an epoxide can be responsible for the generation of reactive species. Aromatic amines, hydrazines, and nitrosamines form group 3 requiring an oxidation of a nitrogen atom or of a carbon atom in alpha position to a nitrosated amine. Group 4 compounds are halogenated hydrocarbons which can either give rise to radicals or can form an olefin (group 2) upon dehydrohalogenation. Group 5 compounds depend upon some preceding enzymatic activity either not available in the target cell or acting on positions in the molecule which are not directly involved in the subsequent formation of electrophilic atoms. Examples for each group are taken from the "List of Chemicals and Industrial Processes Associated with Cancer in Humans" as compiled by the International Agency for the Research on Cancer, and it is shown that 91% of the organic carcinogens would have been detected on the basis of structural elements characteristic for group 1-5. As opposed to this very high sensitivity, the specificity (the true negative fraction) of using this approach as a short-term test for carcinogenicity is shown to be bad because detoxification pathways have so far not been taken into account. These competing processes are so complex, however, that either only very extensive knowledge about pharmacokinetics, stability, and reactivity will be required or that in vivo systems have to be used to predict, on a quantitative basis, the damage expected on the DNA. DNA-binding experiments in vivo are presented with benzene and toluene to demonstrate one possible way for an experimental assessment and it is shown that the detoxification reaction at the methyl group available only in toluene gives rise to a reduction by at least a factor of forty for the binding to rat liver DNA. This quantitative approach available with DNA-binding tests in vivo, also allows evaluation as to whether reactive metabolites and their DNA binding are always the most important single activities contributing to the overall carcinogenicity of a chemical.(ABSTRACT TRUNCATED AT 400 WORDS)

Sensitivity, Specificity, and Predictive Accuracy as Measures of Efficacy of Diagnostic Tests

ABSTRACT The proliferation of new tests based on sophisticated technology has given impetus to critical evaluation of the role of clinical decision making. In this presentation selected aspects of ...