Aim: L -3,4-dihydroxy-6–18F-fluorophenylalanine (18F-DOPA PET may be used to distinguish subjects with Parkinsonism from those with symptoms not originating from impaired dopaminergic transmission. However, it is not routinely utilized to discriminate Idiopathic Parkinson's disease (IPD) from Atypical Parkinsonian Disorders (APD). We investigated the potential of FDOPA PET to discriminate between IPD and APD, with a focus on the anterior-to-posterior decline in het striatum, considered to be more specific for IPD. Materials and methods: 18F-DOPA PET data from a total of 58 subjects were retrospectively analyzed. 28 subjects had idiopathic Parkinson's disease (14 male, 14 female; age at scan 61 +- 11,5), 13 atypical Parkinsonian disease (7 male, 6 females; age at scan: 69,6 +- 6,4) and 17 were controls (6 male, 11 female; age at scan 65,3 +-8,6). Regional striatal-to-occipital ratio's (RSOR's) were calculated, as well as multiple in-line VOI's from the caudate nucleus to the posterior part of the putamen. The linearity of anteroposterior decline was determined by a linear regression fit and associated R squared values. ROC curves were calculated to assess the diagnostic performance of these measurements. Data contralateral to the clinically most affected side were used for analysis. Results: ROC curve analysis for differentiation between controls and Parkinsonism patients showed the highest AUC for the caudate nucleus-to-posterior putamen ratio (AUC = 0.930; p < 0.00) and for the R squared value for the linear regression fit (AUC = 0.948; p = 0.006). For discrimating IPD from APD, the highest AUC was found for the caudate nucleus-to-anterior putamen ratio (0.824; p < 0.001) Conclusions: Subregional analysis of the striatum in F-DOPA PET scans may provide additional diagnostic information in patients screened for a presynaptic dopaminergic deficit. A more linear decrease from the head of the caudate nucleus to the posterior putamen was present in patients with IPD, although this feature did not have additional diagnostic value over the RSOR analysis.