1.Prediction of Outcome of Renal Transplant Survival, Application of Artificial Neural Networks.
Abstract: One of the purposes of the evolving of medical informatics is to develop decision support systems that enhance the human ability to diagnose, treat, and assess prognoses of pathologic conditions. This work is to compare the ability of an artificial neural network and a model of logistic regression to predict individual survival status at 2 years after renal transplant. Artificial Neural Networks (ANN s) are new computational tools, which once trained, can recall proper outputs for a specific set of inputs never encountered before. Logistic regression model was built and a neural network was trained on randomly selected 80% of patients (580 patients) to predict individual status at 2 years (status = "1" for "graft loss" and "0" for "graft survival). We classified the risk factors into pre-transplant, transplant (technical), and post-transplant predictors. The performance of the LR and ANN models, revealed a sensitivity (percentage of correctly predicted deaths) of 10.6% and 87.6%, a specifity (percentage of correctly predicted survivors) of 99% and 84%, with an overall accuracy of 85.3% and 85.8% respectively. The results show that neural network has a higher accuracy in predicting the sensitivity at the 2-years survival status. It has also better balance between the correct prediction of losses and survivors. Probably, still some new markers are needed to differentiate those whose survival status was not correctly predicted.
Keywords: Artificial neural networks, Survival Analysis, Regression analysis, Renal transplant.
2. Decision Support System in Biomedical Data using Neural Networks: A Case Study.
Abstract: The field of medical informatics has evolved around structuring, processing, storing and transmitting medical information for a variety of purposes.. The increasing availability of electronic medical databases calls for statistical and artificial intelligent methods that are able to extract important information from patient records and develop accurate predictive models. "Survival analysis" is the phrase used to describe the analysis of data that correspond to the time from a well-defined "time origin" until the occurrence of some particular event or "end-point" The objective of this study is to use artificial neural networks and statistical techniques in predicting 5-years survival of patients with bladder cancer. This is to help the clinician: (1) to identify patients with good prognosis, (2) to identify patients with high risk for whom adjuvant therapy or another therapeutic procedure might be beneficial. 839 patients of bladder cancer, treated by radical cystectomy underwent the study. Variables entered in the study were: age and sex of patient, grade and stage of tumor, lymph node status, histopathology and bilharzias. We test the hypothesis that neural networks may perform better than advanced statistical methods (logistic regression- LR). Both ANN and LR agreed that relevant factors in predicting 5-years survival of patients are grade and stage of the tumor and lymph node status. Neural networks were significantly more sensitive and accurate. The area under the ROC-curve proved that ANN has better capability in discriminating between the two states of the output ("0"= alive, "1"=death).
Keywords: Artificial neural networks, Logistic regression, Survival analysis, Bladder cancer.
3. Bone Mineral Density Assessment of Maxillary and Mandibular Alveolar Bony Septa at Different Ages using Dual Energy Photon Absorptiometry.
Abstract: Immediate dental implants have been placed at the time of extraction with a variety of techniques. The success rate of immediate implant osseointegeration is dependent on many factors as bone mineral density; implant design, site of implant insertion and the age of patient. This study examined the bone mineral density (BMD) of interdental bone septa of both jaws at different five regions weach, in forty healthy dentulous volunteers by using dual X ray energy absorptiometry (DEXA). Statistical analysis was performed to test the significant difference in BMD between four age groups. T test showed a significant difference between the anterior and posterior regions of the maxilla (t=14.2, p<0.001) and also showed a significant difference between the anterior and posterior regions of the mandible (t=1`4.02, p<0.001). The paired T test was used to compare the maxillary and mandibular BMD; a significant difference was noticed also (t=17.66, p<0.001). The analysis of variance (ANOVA) was done to test the difference between age groups; in both maxilla and mandible. A significant difference was noticed between age groups; its value for maxilla was (f=24.2, p<0.001) and for mandible was (f=38.41, p<).001). This dissimilarity in bone mineral density at different maxillary and mandibular sites of alveolar bony septa correlated with age may partly explain the prognosis of immediate implant success.