Fetal health status prediction based on maternal clinical history using machine learning techniques.

BACKGROUND AND OBJECTIVE: Congenital anomalies are seen at 1-3% of the population, probabilities of which are tried to be found out primarily through double, triple and quad tests during pregnancy. Also, ultrasonographical evaluations of fetuses enhance detecting and defining these abnormalities. About 60-70% of the anomalies can be diagnosed via ultrasonography, while the remaining 30-40% can be diagnosed after childbirth. Medical diagnosis and prediction is a topic that is closely related with e-Health and machine learning. e-Health applications are critically important especially for the patients unable to see a doctor or any health professional. Our objective is to help clinicians and families to better predict fetal congenital anomalies besides the traditional pregnancy tests using machine learning techniques and e-Health applications. METHODS: In this work, we developed a prediction system with assistive e-Health applications which both the pregnant women and practitioners can make use of. A performance comparison (considering Accuracy, F1-Score, AUC measures) was made between 9 binary classification models (Averaged Perceptron, Boosted Decision Tree, Bayes Point Machine, Decision Forest, Decision Jungle, Locally-Deep Support Vector Machine, Logistic Regression, Neural Network, Support Vector Machine) which were trained with the clinical dataset of 96 pregnant women and used to process data to predict fetal anomaly status based on the maternal and clinical data. The dataset was obtained through maternal questionnaire and detailed evaluations of 3 clinicians from RadyoEmar radiodiagnostics center in Istanbul, Turkey. Our e-Health applications are used to get pregnant women's health status and clinical history parameters as inputs, recommend them physical activities to perform during pregnancy, and inform the practitioners and finally the patients about possible risks of fetal anomalies as the output. RESULTS: In this paper, the highest accuracy of prediction was displayed as 89.5% during the development tests with Decision Forest model. In real life testing with 16 users, the performance was 87.5%. This estimate is sufficient to give an idea of fetal health before the patient visits the physician. CONCLUSIONS: The proposed work aims to provide assistive services to pregnant women and clinicians via an online system consisting of a mobile side for the patients, a web application side for their clinicians and a prediction system. In addition, we showed the impact of certain clinical data parameters of pregnant on the fetal health status, statistically correlated the parameters with the existence of fetal anomalies and showed guidelines for future researches.

Detection of laryngeal tumors and tumoral extension by multislice computed tomography-virtual laryngoscopy (MSCT-VL).

Virtual endoscopy is becoming a widely used non-invasive clinical diagnostic tool. The present study was designed to compare the sensitivity and specificity of the conventional endoscopy and virtual laryngoscopy with respect to laryngeal masses. A total of 38 patients (20 males, 18 females, mean age 61 years) with the complaint of hoarseness were included in the study. Laryngeal mucosa, lumen and mass pathology were evaluated initially by direct endoscopy and then by virtual laryngoscopy during multislice CT of the larynx. Histopathologic evaluation of the masses was also made. The main pathology of the patients was found to be laryngeal masses (60% of patients, one mass for each patient), which were polyps (n = 8), papilloma (n = 4) and carcinoma (n = 11) according to histopathologic evaluation. Retrospective evaluation of 6 lesions detected in virtual but not in conventional laryngoscopy resulted with the finding of viscous-dense mucous secretion. On the contrary, three lesions detected by conventional laryngoscopy could not be detected by virtual evaluation. A total of six patients were evaluated and considered as normal both by conventional and virtual laryngoscopic examinations. Sensitivity of the virtual laryngoscopy was 88% (23/26) while its specificity was only 50% (6/12). Positive and negative predictive values were 79% (23/29) and 66% (6/9), respectively. Accuracy of the virtual laryngoscopy was 76% (29/38). Virtual laryngoscopy is not an alternative to conventional laryngoscopy but may assist direct endoscopy without causing additional radiation exposure or discomfort to the patient. The three-dimensional contribution to interpretation of the results and subsequent manipulation of the data can be used for educational and surgical purposes.