---

Introduction: Nowadays, cancer diseases are the most important causes of death wor-ldwide and the breast cancer is the most important of them within women. Asses-sing the survival of the patients is one of the most important indices of controlling the cancer. This study aimed to make a compa-rison between the prediction of artificial neural networks (ANN) and Cox regression models for the breast cancer survival. Materials & Methods: The data of the sur-vival study gathered from 344 breast cancer patients between 2005 and 2012 that regist-ered at the Fayyazbakhsh hospital, Tehran, Iran. The status of survival was considered as a dependent variable. Area under rece-iver operative characteristic curve (AUR-OC) and classification accuracy were used for the comparison of artificial neural netw-orks and Cox regression models. Data anal-ysis was performed by R and Minitab sof-tware. Findings: The age of participants expres-sed as Mean ± SD, was 49.9±10.93 years and the median of survival was 44.6 mo-nths. Up to the end study, 45 (13.1) were died. Results showed that AUROC for AN-N and Cox regression were 87.6% and 75.4%, respectively. In addition the clas-sification accuracy of ANN and Cox regres-sion were calculated as 89.42 and 77.68, respectively. Discussion & Conclusion: According to the results, the total classification accuracy of the ANN was better than those of the Cox regression therefore, the ANN model is suggested to predict the survival status of breast cancer disease and also is suggested for diagnostic goals.

---

Introduction: A wide range of factors affect public health and subsequent health status of people is different in different social classes. Nowadays by high acceleration of changes in influencing factors in exacerbation of cancer incidence is predicted to accelerate cancer growth rate and will be more than double the current situation in the next two decades and is considered as a challenge for health systems. The purpose of this study was to investigate the relationship between gastric cancer in men in all provinces of the country and their socio-economic status.

Materials & methods: Drawback of traditional methods for estimating cancer rates is that the coefficients may change with time and for a management system of factors affecting on cancer, in terms of its dynamics, will lose its efficiency. Therefore, applying advanced techniques such as neural networks can be effective in estimating the non-linear and dynamic systems. Socioeconomic data were collected from provincial statistical yearbooks and Data on age-standardized incidence rate (ASIR) of gastric cancer per 100,000 populations stratified by sex were obtained from published reports by Iran Cancer Registry. This study included the combination of sections (all provinces) and the period 2004-2009. MATLAB software was used for data analysis.

Findings: The results showed that socio-economic factors are significantly associated with gastric cancer. Directly relationship of unemployment and family size, and inverse relationship of literacy rate, urbanization ratio and household expenditure with cancer incidence rate are evident in this study.

Discussion & Conclusion: The socioeconomic inequalities in incidence of gastric cancer in Iranian men, requires investigating preventable mechanisms and supporting healthy lifestyles among deprived provinces. In general, this neural network model and non-linear system which obtained from that, can be used to determine the rate incidence of gastric cancer based on the input data in any province or any of the new solar year. This information can be used in planning for prevention and management of cancer incidence rates.

---

Introduction: Nowadays, the artificial neural networks have received much attention in predicting the effects of multiple variables and complex relationships in aparticular variables. In this study, we have focused on the use of artificial neural network versus logistic regression to predict post-traumatic mental disorders.

Materials & methods: In a prospective cohort study, we covered 100 trauma patients admitted to the trauma center of Shahid Beheshti Hospital of Kashan during a six month period. The patients were then randomly divided into two training (n=50) andexperimental(n=50) groups. 14 variablesincluding age, sex, occupation, education level, marital status, socioeconomic status, history ofmental illnessin theimmediate family, history of being hospitalized in neurosurgeryunit, historyof trauma,history ofunderlying disease, history of psychologicaldrug use, history of anesthesia, history of alcohol use, and history of substance abuse were totally investigated. 300artificial neural networksandlogistic regressions were studied in the first group and then the predicted values were compared in the second group using the two models. The ROC curve and classification accuracy tool were applied to estimate the predictive power of mental disorder.

Findings: The results showed that the accurate index for predicting the disorder was90.65% for the neural network model, while it was 75.96% for the logistic regression model.

Discussion & conclusions: The artificial neural network models appeared to bemore powerful in predictingmental disorder versus the logistic regression model.

---

Introduction:  It is of utmost importance to predict cardiovascular diseases correctly. Therefore, it is necessary to utilize those models with a minimum error rate and maximum reliability. This study aimed to combine an artificial neural network with the genetic algorithm to assess patients with myocardial infarction and congestive heart failure.
 
Materials & Methods: This study utilized a multi-layer perceptron artificial neural network and a backpropagation algorithm combined with a genetic algorithm to assess the condition of two patients with cardiovascular diseases. The medical records of 497 patients with cardiovascular diseases at Ayatollah Golpayegani Hospital, Qom, Iran, were collected using a clustering sampling method. The data were analyzed using a Receiver Operating Characteristics Curve. Eventually, the data, including personal and clinical variables of patients (i.e., age, gender, dyspnea, blood pressure variations, and blood test results) were selected using sigmoid-transfer and tangent-sigmoid functions. Following that, the neural network was trained with 19 input neurons and 5 middle-layer neurons.
 
Findings: According to the results, a neural network with 5 middle-layer neurons has more precision, compared to other modes. Therefore, it is possible to predict myocardial infarction in the patients using this neural network with a minimum of 97.7% precision.
 
Discussion & Conclusions: An artificial neural network was combined with a genetic algorithm and proposed as a model to predict myocardial infarction in this study. Moreover, it was attempted to utilize important and cost-effective factors for cardiovascular diseases. As a result, the patients can be aware of their disease at the lowest cost.