Background: An ideal imaging biomarker for a neurodegenerative disorder should be able to
measure abnormalities in the earliest stages of the disease.
Objective: We investigated metabolic network changes in two independent cohorts of drugnaïve
Parkinson's disease (PD) patients who have not been exposed to dopaminergic medication.
Methods: We scanned 85 de novo, drug-naïve PD (dPD) patients and 85 age-matched healthy
control subjects from Italy (n=96) and the United States (US, n=74) with [18F]-
fluorodeoxyglucose (FDG) PET. All patients had clinical follow-ups to verify the diagnosis of
idiopathic PD. Spatial covariance analysis was used to identify and validate de novo PD-related
metabolic patterns (dPDRPs) in the Italian and US cohorts. We compared the dPDRPs to the
original PDRP that was identified in more advanced patients who had been on chronic
dopaminergic treatment.
Results: De novo PD-related metabolic patterns, dPDRPs, were identified in each of the two
independent cohorts of drug-naïve PD patients, and each differentiated PD patients from healthy
control subjects. Expression values for these disease patterns were elevated in dPD patients
relative to healthy controls in the identification as well as in each of the validation subgroups.
The two dPDRPs were topographically very similar to each other and to the original PDRP.
Conclusions: Reproducible PDRP patterns are expressed in de novo, drug-naïve PD patients. In
PD, disease-related metabolic patterns have stereotyped topographies that develop independently
of chronic levodopa treatment.
