Dinesh P. Mital

Texture segmentation using Gabor filters

An unsupervised texture segmentation technique using multi-channel filtering has been proposed. The main advantage of this approach is that it can use simple statistics of gray values in the filtered images as texture features. This simplicity is due to the direct result of decomposition of the original image into several filtered images with limited spectral information. The main issues involved in this approach are: 1) functional characterization of the channels and number of channels, 2) extraction of appropriate texture features from the filtered images, 3) the relationship between the channels, and 4) integration of texture features from different channels to produce a reliable segmentation.

Mathematical modeling of tumor growth and tumor growth inhibition in oncology drug development

Introduction: Approaches aiming to model the time course of tumor growth and tumor growth inhibition following a therapeutic intervention have recently been proposed for supporting decision making in oncology drug development. When considered in a comprehensive model-based approach, tumor growth can be included in the cascade of quantitative and causally related markers that lead to the prediction of survival, the final clinical response. Areas covered: The authors examine articles dealing with the modeling of tumor growth and tumor growth inhibition in both preclinical and clinical settings. In addition, the authors review models describing how pharmacological markers can be used to predict tumor growth and models describing how tumor growth can be linked to survival endpoints. Expert opinion: Approaches and success stories of application of model-based drug development centered on tumor growth modeling are growing. It is also apparent that these approaches can answer practical questions on drug development more effectively than that in the past. For modeling purposes, some improvements are still needed related to study design and data quality. Further efforts are needed to encourage the mind shift from a simple description of data to the prediction of untested conditions that modeling approaches allow.

Identification of heparin samples that contain impurities or contaminants by chemometric pattern recognition analysis of proton NMR spectral data

AbstractChemometric analysis of a set of one-dimensional (1D) 1H nuclear magnetic resonance (NMR) spectral data for heparin sodium active pharmaceutical ingredient (API) samples was employed to distinguish USP-grade heparin samples from those containing oversulfated chondroitin sulfate (OSCS) contaminant and/or unacceptable levels of dermatan sulfate (DS) impurity. Three chemometric pattern recognition approaches were implemented: classification and regression tree (CART), artificial neural network (ANN), and support vector machine (SVM). Heparin sodium samples from various manufacturers were analyzed in 2008 and 2009 by 1D 1H NMR, strong anion-exchange high-performance liquid chromatography, and percent galactosamine in total hexosamine tests. Based on these data, the samples were divided into three groups: Heparin, DSu2009≤u20091.0% and OSCSu2009=u20090%; DS, DSu2009>u20091.0% and OSCSu2009=u20090%; and OSCS, OSCSu2009>u20090% with any content of DS. Three data sets corresponding to different chemical shift regions (1.95–2.20, 3.10–5.70, and 1.95–5.70xa0ppm) were evaluated. While all three chemometric approaches were able to effectively model the data in the 1.95–2.20xa0ppm region, SVM was found to substantially outperform CART and ANN for data in the 3.10–5.70xa0ppm region in terms of classification success rate. A 100% prediction rate was frequently achieved for discrimination between heparin and OSCS samples. The majority of classification errors between heparin and DS involved cases where the DS content was close to the 1.0% DS borderline between the two classes. When these borderline samples were removed, nearly perfect classification results were attained. Satisfactory results were achieved when the resulting models were challenged by test samples containing blends of heparin APIs spiked with non-, partially, or fully oversulfated chondroitin sulfate A, heparan sulfate, or DS at the 1.0%, 5.0%, and 10.0% (w/w) levels. This study demonstrated that the combination of 1D 1H NMR spectroscopy with multivariate chemometric methods is a nonsubjective, statistics-based approach for heparin quality control and purity assessment that, once standardized, minimizes the need for expert analysts.n FigureContour plot from grid search of the optimal values of γ and C for the SVM model

Advances in Biomedical Informatics for the Management of Cancer

Abstract: Increased access to health care, and advances in education and technology have resulted in a larger proportion of the population having longer life expectancy. The strong correlation between age and cancer has resulted in a major healthcare problem for this century, and until recently cancer has defied any long‐lasting cure. However, progress, especially in the field of biomedical informatics, promises a successful prediction and possibly a permanent cure for cancer within the next two decades. Biomedical informatics—with its roots in computer science, biomedical engineering, biostatistics, and mathematics—helps to bring the patient closer to the physician, facilitates access to specialist information and knowledge bases across the world, and makes it possible to identify genetic expression profiles for malignant or cancerous cells. This paper reviews the new research findings in biomedical informatics, working toward the ultimate goal of successfully predicting cancer, solving complex problems in prevention and treatment of cancer, and perhaps completely curing the scourge of cancer.

Electronic dental record information model

Electronic dental records (EDRs) vary widely with regard to the content and structure of the information they contain. At a time when electronic patient records are becoming increasingly essential in delivering high-quality patient care, the Tower of Babel that todays EDRs represent is an increasingly significant drawback. This situation prevents the computational approaches for supporting patient and will also significantly hamper the exchange of patient information within the emerging National Health Information Infrastructure. Incomplete and fragmented EDRs force many practitioners to maintain both paper- and computer-based patient records, potentially contributing to the very limited adoption of completely electronic patient records by general dentists. This paper proposes an electronic dental record information model (EDRIM) design as a reference standard for the content and structure of EDRs.

A novel solution for maze traversal problems using artificial neural networks

In this paper we have addressed the problem of finding a path through a maze of a given size. The traditional ways of finding a path through a maze employ recursive algorithms in which unwanted or non-paths are eliminated in a recursive manner. Neural networks with their parallel and distributed nature of processing seem to provide a natural solution to this problem. We present a biologically inspired solution using a two level hierarchical neural network for the mapping of the maze as also the generation of the path if it exists. For a maze of size S the amount of time it takes would be a function of S (O(S)) and a shortest path (if more than one path exists) could be found in around S cycles where each cycle involves all the neurons doing their processing in a parallel manner. The solution presented in this paper finds all valid paths and a simple technique for finding the shortest path amongst them is also given. The results are very encouraging and more applications of the network setup used in this report are currently being investigated. These include synthetic modeling of biological neural mechanisms, traversal of decision trees, modeling of associative neural networks (as in relating visual and auditory stimuli of a given phenomenon) and surgical micro-robot trajectory planning and execution.

Prediction of hepatitis C using artificial neural network

The main objective of this research project is develop an expert system module, based on a back propagation feed forward artificial neural networks (ANNs), for the diagnosis of hepatitis C and compare its performance with other existing computer based decision support systems. The ANN based system was developed with a commercially available software package (Brain Maker, California scientific Software). Two different types of ANN models, unsupervised and supervised, were developed, compared, and tested. The predictive accuracy and the model training for supervised model was significantly better. The model was able to predict the Hepatitis C in patients very accurately, however performance was not significantly better than the traditional computer model based techniques. Further investigations are needed to understand the impact of this methodology on the outcome analysis. An existing database of hepatitis C infected patient was used. Data of 15 infected and 20 normal individual were collected. Dichotomous variables were coded as present (1) or not present (0). Continuous variable were recorded for patient age, ethnicity, patient number and patient sex. The results have been very interesting, however, some more research work is required to fine-tune the results. The main advantage of the developed system is that it is adaptive and self-adaptive type.

A quantitative analysis of the effectiveness of laparascopy and endoscopy virtual reality simulators

Abstract The increasing use of virtual reality (VR) simulators in surgical training makes it imperative that definitive studies be performed to assess their training effectiveness. Indeed in this paper we report the meta-analysis of the efficacy of virtual reality simulators in (1) the transference of skills from the simulator training environment to the operating room and (2) their ability to discriminate between the experience levels of its users. The task completion time and the error score were the two study outcomes collated and analyzed in this meta-analysis. Sixteen studies were identified from a computer-based literature search (1996–2004). The meta-analysis of the random-effects model (because of the heterogeneity of the data) revealed that training on virtual reality simulators did lessen the time taken to complete a given surgical task as also clearly differentiate between the experienced and the novice trainees. Meta-analytic studies such as the one reported here would be very helpful in the planning and setting up of surgical training programs and for the establishment of reference ‘learning curves’ for a specific simulator and surgical task. If any such programs already exist they can then indicate the improvements to be made in the simulator used such as providing for more variety in their case scenarios based on the state and/or rate of learning of the trainee.

Diagnostic efficacy value in terms of sensitivity and specificity of imaging modalities in detecting the abdominal aortic aneurysm: a systematic review

The purpose of this study was to examine whether duplex ultrasonography (DUS) or MR angiography (MRA) or CT angiography (CTA) is more applicable to use as alternative modality in terms of sensitivity and specificity for detection of abdominal aortic aneurysm (AAA). A search of the medical databases was performed for describing AAA evaluation and detection. Twenty eight studies were found and met the selection criteria. Diameter of aneurysms was categorised by size: ≤ 2.5 cm of the aneurysm diameter. For aneurisms ≤ 2.5 cm, the mean reported sensitivities and specificities were DUS: 81% and 91.1%; CTA: 84.3% and 98.4%; MRA: 95.8% and 95.8%, respectively compared DSA as gold standard. MRA has the highest sensitivity and CTA has the highest specificity reported diagnostic accuracy in detecting the aneurysm ≤ 2.5 cm of AAA diameter and they could be used as a reliable alternative modality to invasive DSA.

Analysis of the incidence of lung cancer based on anatomical sites of lung using NIS data

Analysis concludes that the risk occurrence of lung cancer follows an order from high to low: upper lobe, lower lobe, other-parts of the lung, main bronchus, and middle lobe. We have also examined the association of each category of lung with gender and race too. The overall lung cancer analysis indicate that White and Black Americans have the higher risk of getting lung cancer as compared with other races, but anatomical category of lung cancer indicates White American are more susceptible to each category of lung except other-part of lung as compared with other races. The odds ratio analysis for each anatomical site concludes White male and female [odds ratio of all sub category – male/female 1.03 to 1.08] are equally susceptible to each category of lung. The Black females [odds ratio of lower lobe-male/female 1.08] are highly associated with lower lobe lung cancer as compared with Black males while Hispanic, Asian and Native American females are more associated with middle lobe lung cancer as compared with Hispanic, Asian and native American males [odds ratio for middle lobe – male/female 0.89 to 0.95].