Predicting shunt infection in children with hydrocephalus

• Published in: Intelligence-based medicine 2021, Vol.5, p.100029, Article 100029
• Main Authors: Sabeti, M., Boostani, R., Moradi, E., Habibi, Z., Nejat, F.
• Format: Article
• Language: English
• Subjects: Hydrocephalus; Machine learning; Prediction; Shunt infection
• ISSN: 2666-5212; 2666-5212
• DOI: 10.1016/j.ibmed.2021.100029

Hydrocephalus is defined as the increase in Cerebro Spinal Fluid (CSF) volume, which is usually accompanied by high Intracranial Pressure (ICP). The most common treatment for hydrocephalus is ventriculoperitoneal shunt insertion. Shunt is a tube which drains CSF from the ventricular system to peritoneal cavity. Then, the CSF is absorbed from peritoneum. Infection is considered as one of the most complications of shunt systems, which can cause improper prognosis in patients, especially in children’s neuro development. Hence, identifying shunt infection predictive factors could improve the practice in preventing this event. This study used a dataset containing the features of 68 patients with a history of shunt infection and 80 patients without any history of shunt infection (control group) in Children’s Medical Center hospital of Tehran (Iran). The state-of-art techniques were applied to select the most informative predicting factors (features). The probability (accuracy) of shunt infection was determined with different intelligent and statistical classifiers. The results indicated that history of prematurity and intraventricular hemorrhage, age of the first shunt procedure, number of shunt revisions, brain tumor induced hydrocephalus, birth weight, and coinfection are the best descriptive features. In addition, the best classification results by different techniques varied in the accuracy range of 68%–81% in the dataset. •Hydrocephalus defines as increase in cerebrospinal fluid volume which can accompany with high intracranial pressure.•The most common treatment for hydrocephalus is shunt placement and shunt infection could lead to bad prognosis.•Identification of shunt infection predictive factors could improve our practice in preventing this event.•Different algorithms are compared to select the best one in prediction of shunt infection factors.•We could determine the most important shut infection factors with the best predication rate of 68 to 81%.