Prediction of Successful ETV Outcome in Childhood Hydrocephalus: An Artificial Neural Networks Analysis

Parisa Azimi, Hassan Reza Mohammadi

Abstract


Background: Artificial neural networks (ANNs) can be used as a measure for clinical decision- making process. The aim of this study was to develop an ANN model to predict 6-month success of the endoscopic third ventriculostomy (ETV) and compared it with traditional predictive measures in childhood hydrocephalus.

Materials and Methods: An ANN, ETV Success Score (ETVSS), CURE Children’s Hospital of Uganda (CCHU) ETV Success Score (CCHU ETV) and a logistic regression (LR) models were applied to predict outcomes. The etiology, age, Choroid plexus cauterization (CPC), previous shunt, gender, type of hydrocephalus (TOH), and body weight were considered as input variables for established ANN model. To do so data from childhood hydrocephalus patients who had ETV surgery were trained to predict successful ETV by several input variables. Successful ETV outcome was defined as the absence of ETV failure within 6 months follow-up. Then, sensitivity analysis was performed for the established ANN model to identify the most important variables that predict outcome. The area under a receiver operating characteristic (ROC) curve (AUC), accuracy rate of predicting, and Hosmer-Lemeshow (H-L) statistics were measured in order to test different prediction models.

Results: In all the data for 168 patients (80 male, 88 female, the mean age 1.4 ±2.6 years) were analyzed. They were divided into three groups: training group (n=84), testing group (n=42), and validation group (n=42). Successful ETV outcome was 47% as the absence of ETV failure within 6 months follow-up. The etiology, age, CPC, TOH, and previous shunt were the most important variables that were indicated by the ANN analysis. Compared to the ETVSS, CCHU ETV, and the LR models, the ANN model showed better results: accuracy rate (95.1%); the H-L statistic (41.2 %); and AUC (0.87 %).

Conclusion: The findings show that ANNs can predict 6-month successful ETV with a high level of accuracy in childhood hydrocephalus diseases. Our results will need to be confirmed with further prospective studies

 Keywords: Prediction, Endoscopic Third Ventriculostomy (ETV), Childhood hydrocephalus, Artificial Neural Networks (ANN)


Keywords


Prediction, Endoscopic Third Ventriculostomy (ETV), Childhood hydrocephalus, Artificial Neural Networks (ANN)