Dilber Uzun Ozsahin

A sub-mm spatial resolution LYSO:Ce detector for small animal PET

Current high-resolution scintillators are fabricated using mechanical pixelation technique. However the fabrication cost of finely pitched scintillator arrays together with their lack of flexibility to accommodate new depth of interaction designs has remained a significant issue with mechanical pixelation. Another pitfall of mechanically pixelated scintillators is their relatively large inter-pixel gaps that adversely affect their sensitivity to the incident gamma-ray. The main objective of our ongoing efforts is to fabricate high-spatial resolution and high sensitivity PET detectors with depth of interaction (DOI) capability and single-side readout in a cost-effective manner using laser-induced optical barriers (LIOB) technique. We have simulated the behavior of simple optical barriers in LYSO:Ce crystal using the DETECT simulation code. We have also created optical barriers with different size and barrier density in LYSO:Ce at various depths up to 20 mm to form pixel-like shapes similar to mechanically pixelated crystals. In order to process 20mm thick crystals we corrected for laser beam defocusing effect and its adverse effect on laser energy density which results in smaller barrier size and reflectivity. The fabrication time for 10×10×1 and 10×10×20 mm3 LYSO crystals (with 1mm pixels) was ~8 and 95 minutes respectively.

One-Year Survival Prediction of Myocardial Infarction

Myocardial infarction is still one of the leading causes of death and morbidity. The early prediction of such disease can prevent or reduce the development of it. Machine learning can be an efficient tool for predicting such diseases. Many people have suffered myocardial infarction in the past. Some of those have survived and others were dead after a period of time. A machine learning system can learn from the past data of those patients to be capable of predicting the one-year survival or death of patients with myocardial infarction. The survival at one year, death at one year, survival period, in addition to some clinical data of patients who have suffered myocardial infarction can be used to train an intelligent system to predict the one-year survival or death of current myocardial infarction patients. This paper introduces the use of two neural networks: Feedforward neural network that uses backpropagation learning algorithm (BPNN) and radial basis function networks (RBFN) that were trained on past data of patients who suffered myocardial infarction to be capable of generalizing the one-year survival or death of new patients. Experimentally, both networks were tested on 64 instances and showed a good generalization capability in predicting the correct diagnosis of the patients. However, the radial basis function network outperformed the backpropagation network in performing this prediction task.

Sliding Window Based Machine Learning System for the Left Ventricle Localization in MR Cardiac Images

The most commonly encountered problem in vision systems includes its capability to suffice for different scenes containing the object of interest to be detected. Generally, the different backgrounds in which the objects of interest are contained significantly dwindle the performance of vision systems. In this work, we design a sliding windows machine learning system for the recognition and detection of left ventricles in MR cardiac images. We leverage on the capability of artificial neural networks to cope with some of the inevitable scene constraints encountered in medical objects detection tasks. We train a backpropagation neural network on samples of left and nonleft ventricles. We reformulate the left ventricles detection task as a machine learning problem and employ an intelligent system (backpropagation neural network) to achieve the detection task. We treat the left ventricle detection problem as binary classification tasks by assigning collected left ventricle samples as one class, and random (nonleft ventricles) objects are the other class. The trained backpropagation neural network is validated to possess a good generalization power by simulating it with a test set. A recognition rate of 100% and 88% is achieved on the training and test set, respectively. The trained backpropagation neural network is used to determine if the sampled region in a target image contains a left ventricle or not. Lastly, we show the effectiveness of the proposed system by comparing the manual detection of left ventricles drawn by medical experts and the automatic detection by the trained network.

Evaluating X-Ray based Medical Imaging Devices with Fuzzy Preference Ranking Organization Method for Enrichment Evaluations

X-rays are ionizing radiation of very high energy, which are used in the medical imaging field to produce images of diagnostic importance. X-ray-based imaging devices are machines that send ionizing radiation to the patient’s body, and obtain an image which can be used to effectively diagnose the patient. These devices serve the same purpose, only that some are the advanced form of the others and are used for specialized radiological exams. These devices have image quality parameters which need to be assessed in order to portray the efficiency, potentiality and negativity of each. The parameters include sensitivity and specificity, radiation dose delivered to the patient, cost of treatment and machine. The parameters are important in that they affect the patient, the hospital management and the radiation worker. Therefore, this paper incorporates these parameters into fuzzy PROMETHEE (Preference Ranking Organization Method for Enrichment Evaluation) multi-criteria decision theory in order to help the decision makers to improve the efficiency of their decision processes, so that they will arrive at the best solution in due course.

Evaluating Cancer Treatment Alternatives using Fuzzy PROMETHEE Method

The aim of this study is to apply the principle of multi-criteria decision making theories on various types of cancer treatment techniques. Cancer is an abnormal cell that divides in an uncontrolled manner, it is a growth (tumor) that starts when alterations in genes make one cell to grow and multiply rapidly. Eventually, these cells may metastasize to other tissues. The primary factors that influence the comprehensive treatment plan of cancer include, but not limited to genetic factors, patient general health condition, explicit characteristic of cancer, and even purpose of the treatment. Other factors which are also essential include treatment duration, cost of treatment, comfortability, side effects and percentage of survival rate. The latter factors play an important role in the course of treatment and are therefore needed in order to evaluate the several treatment procedures. The outcome of the decision-making theories on these treatment procedures will help the concerned parties such as the patients, oncologists, and the hospital management. The most common cancer treatment techniques were evaluated and compared based on certain criteria using Fuzzy PROMETHEE decision-making theory.