Differential Diagnosis Tool for Parkinsonian Syndrome Using Multiple Structural Brain Measures

Table 4

Classification results.

			Predicted group membership
			MSA-C	MSA-P	PD	Control	Total
Stepwise method (88.9% of original grouped cases correctly classified)
Original data	Count	MSA-C	15	0	0	3	18
		MSA-P	0	9	3	0	12
		PD	0	2	19	0	21
		Control	0	0	0	21	21
	%	MSA-C	83.3	0	0	16.7	100
		MSA-P	0	75.0	25.0	0	100
		PD	0	9.5	90.5	0	100
		Control	0.0	0	0	100.0	100
Two independent variables (84.7% of original grouped cases correctly classified)
Original data	Count	MSA-C	15	0	0	3	18
		MSA-P	0	10	2	0	12
		PD	0	4	17	0	21
		Control	1	1	0	19	21
	%	MSA-C	83.3	0.0	0	16.7	100
		MSA-P	0	83.3	16.7	0	100
		PD	0	19.0	81.0	0	100
		Control	4.8	4.8	0.0	90.5	100
Three independent variables (88.9% of original grouped cases correctly classified)
Original data	Count	MSA-C	14	0	0	4	18
		MSA-P	0	10	2	0	12
		PD	0	1	20	0	21
		Control	0	1	0	20	21
	%	MSA-C	77.8	0.0	0	22.2	100
		MSA-P	0	83.3	16.7	0	100
		PD	0	4.8	95.2	0	100
		Control	0.0	4.8	0.0	95.2	100

MSA-C: cerebellar form of multiple system atrophy; MSA-P: parkinsonism forms of multiple system atrophy; PD: Parkinson’s disease.