DIAGNOSTIC EFFICIENCY OF A STEPWISE NEUROIMAGING ALGORITHM COMPARED WITH STANDARD MRI, MR SPECTROSCOPY AND MR TRACTOGRAPHY IN CHILDREN WITH VIRAL PANENCEPHALITIS

Modern MR techniques considerably expand the diagnosis of viral brain infections in children, but each method used in isolation has limited accuracy, and performing the full multiparametric protocol in every child is time-consuming and not always justified [1]. Prognostic models and machine learning approaches combining imaging and clinical data are increasingly proposed [2, 3]; however, a practical, interpretable algorithm for choosing the scope of MR examination in paediatric panencephalitis is still lacking.