Table 3.

Characteristics the studies included in the meta-analysis.

Author, year	Country	Study design	Subjects (n)	Sex (% male)	Mean age (years )	Mean disease duration (years)	Scan time (min)	FDG dose (MBq)	Pattern analysis method	Classification algortihm	Reference standard
Tang, 201016	New York, USA	Cohort, PDvsAPD	167	58.7	60.4	5.98	NR	NR	SSM/PCA	Two-level algorithm based on logistic regression of individual patterns scores that quantify expression on specific covariance patterns	Final clinical diagnosis by movement disorders specialist using published clinical diagnosis criteria
Tripathi, 2016	New Delhi, India	Cohort, PDvsAPD	129	69.8	56.1	2.67	20	185–296	SSM/PCA	Two-level algorithm based on logistic regression of individual patterns scores that quantify expression on specific covariance patterns	Final clinical diagnosis by movement disorders specialist using consensus criteria
Rus, 2020	Ljubljana, Slovenia	Cohort, PDvsAPD	56	55.4	67.1	4.06	NR	250	SSM/PCA	Automated two level-algorithm based on the PDRP, MSARP, PSPRP as developed at Feinstein Institute	Clinical diagnosis by movement disorders specialist at least 1 year after FDG-PET, blinded to previous clinical work-up
Marti-Andres, 2020	Pamplona, Spain	Multicenter cohort, PSPvsPD	105	58.9	66.8	2.75	6–15	200	SSM/PCA	Based on the expression of metabolic pattern PSPRP, cutoff Z-score vs. PD patients	Final clinical diagnosis was used as the gold standard
Stockholm Cohort, 2021	Stockholm, Sweden	Cohort, PDvsAPD	35	45.7	65.9	4.2	10	125–250	SSM/PCA	Automated two level-algorithm based on the PDRP, MSARP, PSPRP as developed at Feinstein Institute	All patients enrolled were assessed and investigated by movement disorders specialists