Background: Multiple system atrophy is a neurodegenerative disease with α-synuclein aggregation in glial cytoplasmic inclusions, leading to dysautonomia, parkinsonism, and cerebellar ataxia. Objective: The aim of this study was to validate the accuracy of the International Parkinson and Movement Disorder Society Multiple System Atrophy clinical diagnostic criteria, particularly considering the impact of the newly introduced brain magnetic resonance imaging (MRI) markers. Methods: Diagnostic accuracy of the clinical diagnostic criteria for multiple system atrophy was estimated retrospectively in autopsy-confirmed patients with multiple system atrophy, Parkinson's disease, progressive supranuclear palsy, and corticobasal degeneration. Results: We identified a total of 240 patients. Sensitivity of the clinically probable criteria was moderate at symptom onset but improved with disease duration (year 1: 9%, year 3: 39%, final ante mortem record: 77%), whereas their specificity remained consistently high (99%–100%throughout). Sensitivity of the clinically established criteria was low during the first 3 years (1%–9%), with mild improvement at the final ante mortem record (22%), whereas specificity remained high (99%–100% throughout). When MRI features were excluded from the clinically established criteria, their sensitivity increased considerably (year 1: 3%, year 3: 22%, final ante mortem record: 48%), and their specificity was not compromised (99%–100% throughout). Conclusions: The International Parkinson and Movement Disorder Society multiple system atrophy diagnostic criteria showed consistently high specificity and low to moderate sensitivity throughout the disease course. The MRI markers for the clinically established criteria reduced their sensitivity without improving specificity. Combining clinically probable and clinically established criteria, but disregarding MRI features, yielded the best sensitivity with excellent specificity and may be most appropriate to select patients for therapeutic trials.