TABLE 2

Use of PET for Differential Diagnosis of AD

Radiopharmaceutical	Diagnostic standard	Subjects	Major findings	Comments	Reference
¹⁸F-FDG	C	69 AD	Sens = 93%, Spec = 58%, Accu = 79% (Spec = 80% when patients with Parkinson’s dementia are excluded). Severity-stratified analysis shows for mild AD (av. MMSE = 26) Sens = 87.5%; for mod./severe AD (av. MMSE = 10) Sens = 96%.	AD av. age = 66, av. duration = 2.5 y. Groups well matched for level of severity (AD av. CDR = 2.1, Non-AD av. CDR = 2.1, by visual analysis).	(47)
		48 Non-AD
¹⁸F-FDG	P	13 AD	Sens = 92%, Spec = 71%, Accu = 85%.	Pooled analysis (44) across 3 studies providing small groups of pathologically confirmed cases.	(46)
		7 Non-AD				(47)
						(48)
¹⁸F-FDG	C	20 AD	PET Accu = 90%, SPECT Accu = 67%. A Stratified analysis to look at early AD shows that for subjects with MMSE > 20, PET Accu = 87%, SPECT Accu = 63%.	Receiver-operating characteristic area-under-curve analysis was performed for both SPECT and PET on same 45 patients to determine accuracy of each method	(43)
^99mTc-HMPAO		12 Non-AD
		13 NI
¹⁸F-FDG	L	66 AD	Group analysis shows significant differences for very early AD (av. MMSE = 25) vs. NI (av. MMSE = 28) Posterior cingulate significantly fell (by 21%–22%, P = 0.0007), as did parietal and temporal areas.	To obtain very early AD cases, minimally impaired patients were scanned and then followed longitudinally to determine whether NINCDS criteria for probable AD developed.	(53)
		23 MI
		22 NI
¹⁸F-FDG	C	9 AD	Group analysis by statistical mapping demonstrates significant PET differences between AD and Parkinson’s dementia (Table 1).	Groups well matched for (fairly mild) level of severity (AD av. CDR = 1.2, Non-AD av. CDR = 1.3).	(73)
		9 Non-AD
¹⁸F-FDG	C	19 AD	Group analysis by statistical mapping demonstrates significant PET differences between AD and dementia with Lewy bodies (Table 1).	Groups well matched for level of severity and duration (AD av. MMSE = 18, 24 mo, Non-AD av. MMSE = 18, 24 mo)	(74)
		19 Non-AD
¹⁸F-FDG	C	21 AD	Group analysis by region-of-interest method demonstrates significant differences between AD and frontotemporal dementia (Table 1).	Groups well matched for (fairly mild) level of severity (AD av. MMSE = 20, Non-AD av. MMSE = 19). All subjects had normal MRI findings	(75)
		21 Non-AD
		21 NI
¹⁸F-FDG	P	16 AD	AD identified in 13/14 (Sens = 93% of AD-only and 1/2 AD+ cases (overall Sens = 88%). Absence of AD confirmed in 4/6 cases (Spec = 67%).	14 patients had AD as the only pathologic diagnosis, 1 had AD + Lewy bodies, 1 had AD + PSP.	(49)
		6 NI
¹⁸F-FDG	P	97 AD	AD identified in 85/89 (Sens = 96% of AD-only and 6/8 AD+ cases (overall Sens = 94%). Absence of AD confirmed in 30/41 cases (Spec = 73%), including 23 cases with other neurodegenerative dementias. Absence of neurodegenerative disease confirmed in 14/18 cases (Spec = 78%).	Relatively early dementia group, with 70% having mild or questionable dementia; 89 patients had AD as the only pathologic diagnosis; 5 had AD + Lewy bodies; 1 each had AD + PSP, + Parkinson’s disease, or + cerebrovascular disease.	(50)
		41 Non-AD
H₂¹⁵O	C	16 AD	Scans with H₂¹⁵O show decreased perfusion for AD in parietal and lateral temporal regions. Scans with C¹⁵O reveal no cerebral blood volume differences between AD and NI groups.	Relatively mild AD group, having av. MMSE = 21.	(76)
C¹⁵O		10 NI		Relatively mild AD group, having av. MMSE = 21.

C = diagnosis based on clinical evaluation near time of scan; P = diagnosis based on histopathology; L = diagnosis based on longitudinal clinical follow-up of at least 2 y; AD = cognitively impaired secondary to AD; Non-AD = cognitively impaired but not secondary to AD; MI = isolated memory impairment; NI = cognitively normal; Sens = sensitivity with respect to correctly identifying presence of AD; Spec = specificity with respect to correctly specifying that AD is absent; Accu = overall accuracy; av. = average; mod. = moderate; MMSE = Mini-Mental State Examination; CDR = Clinical Dementia Rating.