A Novel Expert System for Non-Invasive Liver Iron Overload Estimation in Thalassemic Patients

Expert Systems can integrate logic based often on computational intelligence methods and they are used in complex problem solving. In this work an Expert System for classifying liver iron concentration in thalassemic patients is presented. In this work, an ANN is used to validate the output of the L.I.O.MO.T (Liver Iron Overload MOnitoring in Thalassemia) method against the output of the state-of-the-art method based on MRI T2* assessment for liver iron concentration. The model has been validated with a dataset of 200 samples. The experimental Mean Squared Error results and Correlation show interesting performances. The proposed algorithm has been developed as a plugin for OsiriX Dicom Viewer.