By Michael T. Lin, MD, PhD

Associate Professor of Neurology and Neuroscience, Weill Cornell Medical College

Dr. Lin reports no financial relationships relevant to this field of study.

SYNOPSIS: In a prospective, observational study in multiple centers in Italy, amyloid PET imaging was shown to be negative in 35% of patients who met clinical criteria for a diagnosis of Alzheimer’s disease.

SOURCES: Boccardi M, Altomare D, Ferrari C, et al. Assessment of the incremental diagnostic value of florbetapir F18 imaging in patients with cognitive impairment: The incremental diagnostic value of amyloid PET with [18F]-florbetapir (INDIA-FBP) study. JAMA Neurol 2016;73:1417-1424.

Caselli RJ, Woodruff BK. Clinical impact of amyloid positron emission tomography — is it worth the cost? JAMA Neurol 2016;73:1396-1398.

Amyloid plaques are a defining pathological hallmark of Alzheimer’s disease (AD), the most common cause of cognitive impairment with aging, but they are microscopic and, until recently, were not readily detectable pre-mortem. In 2012, the FDA approved the first PET ligand, florbetapir, for detecting amyloid plaques in the living brain. Since then florbetaben and flutemetamol have also become commercially available. Unfortunately, amyloid PET scans are very expensive (~$5,000), and Medicare does not pay for it, arguing that it makes little practical difference and is not worth the cost (multiplied by the high incidence and prevalence of AD with aging), particularly in the absence of any effective therapy for AD. The recently published Incremental Diagnostic Value of Amyloid PET with [18F]-Florbetapir (INDIA-FBP) study and accompanying editorial addressed this issue of cost-effectiveness, looking at the effect of amyloid PET results on diagnosis, diagnostic confidence, and drug treatment.

The INDIA-FBP study is an open-label, multicenter study conducted in 18 AD evaluation units in northern Italy. Patients were eligible if they were between ages 50 and 85 years, had a pre-scan diagnostic confidence of AD between 15% and 85%, had at worst moderate dementia, had no contraindications to PET scan, and had no prior participation in a trial of an anti-amyloid agent. The authors estimated that these criteria selected 14-27% of new patients referred to their memory clinics. All eligible patients were consecutively enrolled. Subjects underwent routine evaluation by the local dementia expert (neurologist or geriatrician) as well as florbetapir amyloid PET. Prior to and immediately after disclosure of PET results, the local experts were asked to formulate their diagnosis, rate their confidence that the cognitive impairment was due to AD, and decide drug treatment. Pre- and post-disclosure diagnoses, confidence, and treatment plans were compared.

From Aug. 5, 2013, to Dec. 31, 2014, 228 consecutively eligible subjects were enrolled. Prior to disclosure of amyloid PET results, 165 were diagnosed with AD and 63 with a non-AD condition (37 with frontotemporal degeneration [FTD], 26 with other [largely vascular and Lewy body dementia]).

Of the 165 subjects with a pre-scan diagnosis of AD, a substantial proportion (35%) had a discordant PET result (amyloid negative), and in the majority (79%) of these cases, the diagnosis was changed to a non-AD condition. In the 21% of cases in which the diagnosis remained AD despite the negative PET result, the pre-scan confidence in the AD result was higher, suggesting that the presentation was very typical of AD. The authors noted that this false-negative rate (35% x 21% = 7%) matches the approximate sensitivity of florbetapir compared to autopsy (92%).

Of the 63 subjects with a pre-scan diagnosis of a non-AD condition, nearly half (48%) had a discordant PET result (amyloid positive). This proportion was similar in the 37 subjects diagnosed with FTD (48.6% were amyloid positive) and the 26 subjects diagnosed with some other condition (46.2% were amyloid positive). In the prescan-FTD/amyloid-positive subjects, 72% had their diagnosis changed to AD, consistent with the ability of AD to mimic FTD. However, in the pre-scan-non-FTD/amyloid-positive subjects, only 25% had their diagnosis changed to AD, consistent with the known high co-occurrence of AD pathology with vascular dementia and Lewy body dementia at autopsy.

Finally, AD treatment (cholinesterase inhibitors and memantine) was initiated in 65% of patients who were amyloid positive but not previously on AD treatment, and AD treatment was withdrawn in 33% of patients who were amyloid negative and previously receiving treatment.


These results show that even among dementia experts there can be substantial discordance between biomarker results and clinical impression of AD, and that amyloid PET results can make a substantial difference in diagnosis and AD drug treatment plan. Presumably, the discordance would be even larger with general neurologists or internists. A similar study is underway in the United States (Imaging Dementia--Evidence for Amyloid Scanning [IDEAS] study,

Although encouraging, it remains to be seen whether amyloid PET results make a difference in long-term outcomes, such as caregiver burden, healthcare utilization, and overall costs. In the absence of strongly effective therapy, it is arguable whether a difference in treatment with cholinesterase inhibitors and memantine is clinically significant. In their editorial, Caselli and Woodruff mentioned circumstances in which the economic justification might be clearer (e.g., applying for disability in someone still working), but these are highly selected circumstances. However, the field is rapidly evolving, with new biomarkers and new therapies that may actually modify the course of disease.1 Thus, the cost-benefit analysis must be continually reassessed.


  1. Sevigny J, Chiao P, Bussiere T, et al. The antibody aducanumab reduces Aß plaques in Alzheimer's disease. Nature 2016;537:50-56.