An intelligent prenatal screening system for the prediction of Trisomy-21

Abstract

Down syndrome is caused by a trisomy on chromosome 21 of humans. Children with this disease often have varying degrees of physical-mental problems such as mental retardation, communication skills problems, congenital heart disease, and thyroid disease. Early and accurate genetic disorders detection can lead to corrective action and prevent irreversible events. Therefore, accurate and low-risk diagnostic methods to detect these diseases in the fetal period are essential. In this study, a combined artificial neural network (ANN) and genetic algorithm (GA) were used to predict Down syndrome through the first-trimester screening test. In order to examine the proposed model, sample data were collected on 381 pregnant women referred to a private screening reference laboratory for the first-trimester screening test and NT ultrasound between 11 and 13 weeks of gestation. The proposed model in this study was a feedforward neural network for which its structure and input parameters were determined by (GA). The performance of the created model was also evaluated by utilizing the sensitivity, specificity, and accuracy indicators. The average of 10 times experiments showed that the developed model could accurately identify cases of Down syndrome with a specificity of 99.72%, sensitivity of 90.91%, and a mean square error (MSE) of 0.61%. This study showed that the use of GA in optimizing the structure of the neural network technique could increase the accuracy in diagnosing Down syndrome through the information of first-trimester screening tests. Most importantly, this study contributes to clinical knowledge by helping clinicians to early identify the risk of Down syndrome with a non-invasive and low-cost method.