Diagnostic impact of [ 18 F]flutemetamol PET in early-onset dementia
The aim of the present study was to assess the diagnostic impact of [18F]flutemetamol PET in patients suspected of early-onset dementia, who were visiting a tertiary memory clinic. We found that, after a standardized clinical work-up, [18F]flutemetamol PET results led to changes in clinical diagnosis, increases in diagnostic confidence, and alteration of the initial patient management plan in a substantial number of patients.
This prospective study predominantly included patients suspected of early-onset AD. In almost a quarter of these patients, PET showed no evidence of amyloid pathology, comparable with proportions found in clinical pathological comparison studies [19–21]. In line with previous studies, PET results were most often in agreement with the initial clinical diagnosis and overall resulted in increased diagnostic confidence. Subsequently, in these patients PET results more often led to the initiation of AD medication, as reported previously [7]. These findings suggest an additive but primarily confirmatory role for amyloid PET as a diagnostic marker in patients suspected of early-onset AD.
The overall impact on diagnosis seems to be somewhat lower compared with prior studies, although these finding were highly variable, ranging from 9 to 73% [5–11]. Lower impact might be explained by selection of a different patient population or the more liberal method of patient selection in the present study (diagnostic certainty??90%), because we found that patients with less diagnostic certainty prior to PET were more likely to have their diagnosis changed after PET.
Of major interest were patients with inconsistent PET results according to their pre-PET diagnosis. In patients with an AD diagnosis and negative amyloid PET, clinicians remained uncertain about the underlying etiology. This probably explains that predominantly for these cases the clinicians requested further investigations after amyloid PET, most often [18F]FDG PET, to seek evidence for an alternative (nonamyloid) cause of the dementia.
In both FTD patients and patients classified as ‘non-neurodegenerative disease’, AD was often part of the differential diagnosis and subsequently positive PET results frequently led to a change in diagnosis to AD and increased confidence in the post-PET diagnosis, and often led to prescription of symptomatic treatment. In contrast, in patients classified as ‘other dementia’ prior to PET, scan results did not increase overall diagnostic confidence and rarely led to a change in pre-PET diagnosis. A possible explanation for this finding could be that in these patients AD was less often considered as a differential diagnosis prior to PET. An alternative explanation lies in the composition of the ‘other dementia’ group, because almost half of the patients were diagnosed with DLB prior to PET. In these patients, neither a positive nor a negative PET scan changed the initial DLB diagnosis, because amyloid pathology is known to occur in half of the DLB patients [22] and to a lesser extent in other non-AD dementias [23]. In only a few AD patients who turned out to have a negative amyloid PET scan was the AD diagnosis maintained. This is an interesting subset of AD patients that may give us more insight in the various underlying neuropathologies of AD phenotypes and warrant further investigation in future studies [24].
Appropriate use criteria (AUC) for clinical use of amyloid PET were published [4]. The preamble states that the dementia expert must expect that determination of amyloid status would both increase the level of diagnostic confidence and alter the plan for patient management. The present study included a large memory clinic patient sample suspected of mild and early-onset dementia, in which uncertainty in diagnostic confidence remained after standardized work-up. These inclusion criteria generally align with the AUC. However, part of our patients showed no increase in diagnostic confidence but did have a changed diagnosis, consequently resulting in an altered plan for patient management. Thus even without increase in diagnostic confidence, patients may benefit from amyloid PET, implying a more liberal application of the AUC.
We used [18F]flutemetamol PET as a surrogate marker for brain amyloid deposition. Previous studies have shown high correlation between [18F]flutemetamol retention and neuropathology findings [25–27]. In the present study an amyloid-positive PET scan often supported or changed a diagnosis into AD, nevertheless amyloid pathology was present in a few patients diagnosed with another dementia and interpreted as mixed or copathology and not the primary cause of the clinical manifestation of dementia. Because of the clinician’s awareness of the patients age prior to PET, the a-priori probability of detecting amyloid pathology related to age was part of the diagnostic decision-making [28].
Future analysis in this ongoing study will involve diagnostic accuracy of [18F]flutemetamol PET after a 2-year clinical follow-up period. Furthermore, health economic consequences for the use of amyloid PET in this setting are of great socioeconomic interest and will be assessed after clinical follow-up. In this respect, we would like to mention recent efforts to evaluate the effect of amyloid status disclosure, which showed a positive effect on caregivers [29].
Conducting a prospective study in a clinical cohort is accompanied by several limitations. First, investigations other than amyloid PET necessary for clinical diagnosis could have been ordered prior to PET (decisions that were not made based on PET results). More specifically, the lumbar puncture procedure is part of our standardized work-up. The results for amyloid 1–42, total tau, and p-tau, however, are not used during our multidisciplinary meeting when clinical diagnosis is made. The knowledge of availability of CSF biomarkers after post-PET diagnosis, however, may have had an effect on clinical decisions. This may have led to an underestimation of the impact on patient management in this study.
On the other hand, clinicians were aware of the fact that patients were included in the present study, which may have clinicians decide to postpone decision-making about patient management plan until PET results were disclosed, resulting in a relative overestimation.
Second, the vast majority of patients were included at the VUmc Alzheimer Center, which is a tertiary referral center with a high proportion of young patients with complex clinical presentations. The results of the present study are therefore probably not an accurate reflection of the effect of the use of amyloid PET in a general, often older aged, memory clinic population. Instead, these results support the notion in the appropriate use criteria for amyloid PET, describing a potential added value of amyloid PET in patients with early-onset dementia with unclear clinical presentation.