Prediction of Port A Cath Complications Using Machine Learning Techniques

Abstract

The use of the Port A Cath in chemotherapy is increasing. It improves the management of cancer patients by giving them a permanent and secure administration route, and although the placement of an implantable port is a common practice, it is sometimes dangerous because of the significance of certain complications that can range from a simple hematoma to septic shock threatening the patient's vital prognosis. It is necessary therefore to have a reliable, accurate, and feasible system to detect these complications in time for appropriate management. Machine learning (ML) techniques have been implemented on diverse medical datasets to facilitate the analysis of large and complex data. Recently, many scientists have used various machine learning techniques to help the medical industry and professionals to diagnose diseases, but rarely to predict complications related to Port A Cath. The main objective of this work is to examine and compare the accuracy of different data mining classifications, applying a set of ML algorithms, for the prediction of Port A Cath complications. The dataset is built by conducting a retrospective study at the Hassan II Oncology Center in Oujda (Morocco). The models are developed using the Python language. The investigated ML techniques are support vector machine (SVM), decision tree (DT), random forest (RF), and logistic regression (LR). The experimental results indicate that LR method performs better than other investigated ML techniques.