Incremental value of amyloid PET in a tertiary memory clinic setting in China

Abstract

INTRODUCTION

The objective of this study is to investigate the incremental value of amyloid positron emission tomography (Aβ‐PET) in a tertiary memory clinic setting in China.

METHODS

A total of 1073 patients were offered Aβ‐PET using ¹⁸F‐florbetapir. The neurologists determined a suspected etiology (Alzheimer's disease [AD] or non‐AD) with a percentage estimate of their confidence and medication prescription both before and after receiving the Aβ‐PET results.

RESULTS

After disclosure of the Aβ‐PET results, etiological diagnoses changed in 19.3% of patients, and diagnostic confidence increased from 69.3% to 85.6%. Amyloid PET results led to a change of treatment plan in 36.5% of patients. Compared to the late‐onset group, the early‐onset group had a more frequent change in diagnoses and a higher increase in diagnostic confidence.

DISCUSSION

Aβ‐PET has significant impacts on the changes of diagnoses and management in Chinese population. Early‐onset cases are more likely to benefit from Aβ‐PET than late‐onset cases.

Highlights

• Amyloid PET contributes to diagnostic changes and its confidence in Chinese patients.

• Amyloid PET leads to a change of treatment plans in Chinese patients.

• Early‐onset cases are more likely to benefit from amyloid PET than late‐onset cases.

1. BACKGROUND

Alzheimer's disease (AD) is characterized by the pathological accumulation of amyloid‐β (Aβ) and tau in the brain. The biochemical and imaging biomarkers allow us to identify AD in its preclinical and clinical phases. ¹ With the advent of disease‐modifying therapies for AD, ² , ³ a high‐confidence etiological diagnosis and early disease recognition have become increasingly imperative. Due to concerns regarding invasive procedures and associated risks, ⁴ Amyloid positron emission tomography imaging (Aβ‐PET) has emerged as a widely preferred option for both clinicians and patients. ⁵ , ⁶ Currently, there are several tracers available for clinical practice, such as ¹¹C‐Pittsburgh compound B (PiB), ¹⁸F‐florbetapir, ¹⁸F‐flutemetamol, and ¹⁸F‐florbetaben. ⁷ In China, Aβ‐PET has been an integral part of research for diagnosing AD. ⁸ , ⁹

While Alzheimer's disease pathology and the associated deposition of β‐amyloid plaques can occur in individuals both with and without cognitive impairment, ¹⁰ the significant impact of Aβ‐PET on diagnosis, diagnostic confidence, and treatment in cognitively impaired subjects has been reported in various studies. ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ Aβ‐PET might provide incremental diagnostic value beyond clinical assessment, ¹⁸F‐fluorodeoxyglucose (FDG) PET, ¹⁶ and cerebrospinal fluid (CSF) biomarkers. ¹⁷ According to a large prospective multicenter study, the Aβ‐PET has emerged as a valuable diagnostic tool in routine clinical practice, with 61.5% of patients exhibiting cognitive impairment experiencing a change in treatment and 35.6% experiencing an alteration of etiologic diagnosis as a result. ¹³ The clinical utility of Aβ‐PET has been systematically studied in many other countries, ¹¹ , ¹² , ¹³ , ¹⁵ , ¹⁸ whereas only a few studies with small sample sizes were performed in Chinese population. ¹⁹ , ²⁰ , ²¹ The diagnosis was changed in more than 30% of patients with dementia after considering ¹⁸F‐FDG with or without PiB results. ¹⁹ , ²⁰ A recent study has shown that ¹⁸F‐flutemetamol PET showed added value in Chinese patients, particularly reflected in the change of diagnosis of individuals with unclear etiology and AD‐suspected patients. ²¹ However, due to the lack of conclusive evidence supporting its clinical impact and the uniqueness of the medical insurance system in China, Aβ‐PET has yet to be reimbursed by the Chinese medical insurance system.

Early‐onset cognitive impairment (EOCI), defined as the onset of cognitive impairment before the age of 65 years, comprises a heterogeneous range of phenotypic presentations, associated with delayed diagnosis and misdiagnosis. ²² Aβ‐PET was recommended for patients with atypically early age of onset based on the appropriate use criteria (AUC) for Aβ‐PET. ²³ A few studies that focused on EOCI have shown that Aβ‐PET contribute to changes in diagnosis, increases in diagnostic confidence, and impacts on treatment management plan for patients with EOCI. ²⁴ , ²⁵ , ²⁶ But the additional value of Aβ‐PET in Chinese EOCI patients was not entirely clear and still needed further investigation.

In the present work, we investigated the contribution of Aβ‐PET, in addition to routine assessment, to changes in diagnosis, diagnostic confidence, and treatment plans in a large, unselected Chinese memory clinic cohort. Our aim is to address the current evidence gap by providing robust evidence on the clinical utility of Aβ‐PET in China. Moreover, we also examined where there were any between‐group differences in early‐onset and late‐onset groups.

2. METHODS

2.1. Participants

Cognitively impaired patients were consecutively recruited at the memory clinic at Huashan Hospital affiliated to Fudan University between March 2019 and March 2023. The patients had a cognitive complaint considered to be possibly due to AD and might benefit from the knowledge of brain amyloid status by the dementia specialist (eMethods in the Supplement). We offered Aβ‐PET to all these patients who were willing and medically feasible to complete it. The study was approved by the Ethics Committee of Huashan Hospital. Written informed consent was obtained from all participants or their guardians.

2.2. Study design

The study design is illustrated in Figure S1. All participants were assessed with a baseline diagnostic workup, including medical history, neuropsychological assessment, clinical laboratory testing, electroencephalogram, and magnetic resonance imaging (MRI) or computed tomography (CT). Pre‐PET diagnoses were established by a consensus at weekly multidisciplinary meetings using conventional clinical criteria without knowledge of Aβ‐PET. During multidisciplinary meetings, the neurologists determined the cognitive stage (subjective cognitive decline [SCD], mild cognitive impairment [MCI], or dementia) and the suspected etiology of AD (i.e., typical or atypical AD, or mixed AD) or non‐AD. Then, non‐AD diagnoses were further grouped into frontotemporal lobe dementia (FTD), vascular cognitive impairment, other neurodegenerative diseases (e.g., Parkinson's disease dementia, dementia with Lewy bodies, multiple system atrophy, progressive supranuclear palsy, corticobasal degeneration), and non‐neurodegenerative disease (e.g., normal pressure hydrocephalus, sleep disorders, anxiety or depression). Neurologists then estimated their level of diagnostic confidence in the suspected etiology on a visual scale that ranged from 0% to 100% and provided medication prescription, including cognition specific (i.e., acetylcholinesterase inhibitors or memantine) and non‐cognition‐specific (i.e., anxiolytics, hypnotics, antidepressants, antipsychotics, and anticonvulsants). More details were described in the eMethods.

RESEARCH IN CONTEXT

1. Systematic review: The incremental value of amyloid positron emission tomography (Aβ‐PET) in patients with cognitive impairment has been reported in previous studies. However, only several small sample studies focused on Chinese population. Our study evaluates the added value of Aβ‐PET in an unselected cohort in China.

2. Interpretation: Aβ‐PET has significant impacts on the changes in diagnosis, diagnostic confidence, and treatment plan among Chinese patients. Early‐onset patients are more likely to derive utility from Aβ‐PET compared to late‐onset patients.

3. Future directions: To better understand the impact of Aβ‐PET, future studies should adopt large‐scale multicenter prospective clinical protocols and investigate the diagnostic benefits of combining Aβ‐PET with other PET modalities. Additional economic data are required to be combined to assess the cost‐effectiveness of Aβ‐PET in clinical routine.

Aβ‐PET was performed shortly after the routine diagnostic workup. An additional follow‐up visit was required after the ¹⁸F‐florbetapir PET scan, and the PET result was also disclosed to the patients or their legally authorized representative. At the post‐PET visit, the neurologists reevaluated the syndrome diagnosis, etiological diagnosis, diagnostic confidence, as well as patient treatment. Outcomes were changes in diagnosis (from AD to non‐AD or vice versa), diagnostic confidence, and treatment plan. Patient treatment in terms of (1) ancillary investigations (e.g., FDG‐PET, dopamine transporter [DaT] scan [¹⁸F‐N‐(3‐fluoropropyl)−2β‐carboxymethoxy‐3β‐(4‐iodophenyl) nortropane PET], tau PET, genetic screening, and CSF analysis) (2) initiation or withdrawal of cognition/noncognition specific medications.

2.3. Aβ‐PET acquisition and interpretation

All participants underwent ¹⁸F‐florbetapir PET as well as 3.0T multimodal brain MRI. The interval between the two imaging examinations did not exceed 1 month. ¹⁸F‐florbetapir PET was scanned using Biograph 128 mCT, Siemens, Germany and uPMR790 TOF, Untied Imaging, China. Intravenous injection of ¹⁸F‐florbetapir 7.4 MBq/kg was administered according to the subject's body mass, and brain PET/CT visualization was performed after 50 min of quiet rest.10‐s low‐dose head CT scan at one bed position.20‐min brain PET scan in 3D mode; PET image reconstruction using filtered inverse projection; MR scan using a 3.0T Siemens Prisma MR scanner, Germany; T1 structural imaging using Magnetization Prepared Rapid Gradient Echo imaging (MPRAGE).

Postacquisition images were visually assessed using FBP reconstruction. Reconstructed ¹⁸F‐florbetapir PET images were visually analyzed by an intermediate (W.‐Y.W.) and a senior PET diagnostician (J.‐J.G.). In the case of disagreement, a third senior diagnostician (F.X.or C.‐T.Z.) made the judgment. Based on the distribution of cortical Aβ deposits, the diagnostic results were categorized as positive and negative (Figure S2). ²⁷ , ²⁸

2.4. Statistical analyses

The distribution of the variables was assessed through histogram inspection and the Shapiro–Wilk normality test. They have been described as mean and standard deviation for continuous data and as numbers and percentages for categorical data. We performed nonparametric analyses using Mann–Whitney U‐test or Kruskal–Wallis test for continuous variables and χ ² test for categorical variables. In case the number of groups in the comparison was larger than 2, post hoc pairwise comparisons (i.e., SCD vs. MCI, MCI vs. dementia, and SCD vs. dementia) were adjusted using Bonferroni corrections to assess the main outcome. For comparison of pre‐ and post‐ PET, we used the McNemar test or paired Wilcoxon signed rank tests. Post hoc pairwise comparisons were adjusted using Bonferroni corrections. All statistical tests were two‐sided with a significance level of 0.05. Statistical computations were performed using SPSS version 26 and R version 4.2.3 (R Foundation for Statistical Computing).

3. RESULTS

3.1. Demographic and clinical characteristics

A total of 1073 patients were included in the final analysis data set. The baseline characteristics of the participants included in the analyses were illustrated in Table 1. Based on the pre‐PET clinical workup, AD was the leading suspected etiology of cognitive impairment in 69.5% of all patients. The diagnostic confidence of the whole sample was 69.3% [10.5]. At baseline, 10.6% of patients were diagnosed with SCD (mean age [SD], 62.9 [10.5] years; 62.3% female), 37.0% with MCI (mean age [SD], 65.5 [10.0] years; 56.2% female), and 52.4% with dementia (mean age [SD], 64.3 [9.6] years; 58.5% female), with no differences in age and sex among groups (Table S1). There was a decrease in education years (12.1 [3.7], 10.7 [3.8], 9.8 [4.1], p < 0.001) from SCD to dementia group. The detailed baseline etiological diagnoses were presented in Table S2.

TABLE 1.

Baseline characteristics of patients

Characteristic	Whole sample (n = 1073)	EOCI (n = 534)	LOCI (n = 539)
Sociodemographic
Age at diagnosis, mean (SD), years	64.6 (9.9)	56.3 (5.5)	72.7 (5.3)
Disease duration, mean (SD), years	2.5 (1.9)	2.4 (1.7)	2.8 (2.1)
Age at onset, mean (SD), years	59.5 (9.6)	53.0 (5.8)	68.9 (5.2)
Sex, no. (%)
Male	450 (41.9)	226 (42.3)	224 (41.6)
Female	623 (58.1)	308 (57.7)	315 (58.4)
Education, mean (SD), years	10.4 (4.0)	10.5 (3.9)	10.3 (4.2)
Mental status, mean (SD)
MMSE	20.0 (6.7)	19.6 (6.9)	20.3 (6.5)
MoCA	13.8 (6.8)	13.7 (7.2)	13.9 (6.4)
HAMA	6.9 (5.8)	7.3 (6.1)	6.5 (5.3)
HAMD	5.2 (5.0)	5.6 (5.3)	4.7 (4.6)
Patient management
Cognitive stage at baseline, No. (%)
SCD	114 (10.6)	72 (13.5)	42 (7.8)
MCI	397 (37.0)	184 (34.5)	213 (39.5)
Dementia	562 (52.4)	278 (52.0)	284 (52.7)
Etiological diagnosis at baseline, No. (%)
AD	746 (69.5)	360 (67.4)	386 (71.6)
non‐AD	327 (30.5)	174 (32.6)	153 (28.4)
Diagnostic confidence at baseline, mean (SD), %
In AD etiological diagnoses	70.6 (9.8)	67.9 (10.4)	73.1 (8.4)
In non‐AD etiological diagnoses	66.2 (11.5)	64.9 (12.6)	67.6 (10.0)
Medication management, No. (%)
Cognition‐specific medications	472 (44.1)	235 (44.0)	237 (44.0)
Non‐cognition‐specific medications	112 (10.4)	75 (14.0)	37 (6.8)
Combined use a	244 (22.7)	106 (19.9)	138 (25.6)
Nonpharmacological interventions	245 (22.8)	118 (22.1)	127 (23.6)

Note: Data are presented as mean (SD) for continuous variables and number (proportions, %) for categorical variables.

Abbreviations: AD, Alzheimer disease; EOCI, early‐onset cognitive impairment; HAMA, Hamilton Anxiety Scale; HAMD, Hamilton Depression Scale; LOCI, late‐onset cognitive impairment; MCI, mild cognitive impairment.; MMSE, Mini‐Mental State Examination; MoCA, Montreal Cognitive Assessment; SCD, subjective cognitive decline.

^a Combined use of medication management means the combination of cognition specific and noncognition specific drugs.

Based on age of onset, 534 (49.8%) participants were clinically diagnosed with EOCI, and 539 (50.2%) participants were diagnosed with LOCI. The age of symptom onset was significantly earlier in patients with EOCI than LOCI (53.0 [5.8] vs. 68.9 [5.2]; p < 0.001). Symptom duration was longer in the LOCI group than in the EOCI group (2.8 [2.1] years vs. 2.4 [1.7] years, p = 0.04). There were no differences in sex, education, Mini‐Mental State Examination (MMSE), Montreal Cognitive Assessment (MoCA), Hamilton Anxiety Scale, and Hamilton Depression Scale scores between early and late‐onset cases (Table 1). Patients with SCD in the EOCI were more than that in the LOCI (13.5% vs. 7.8%; p = 0.007). Diagnostic confidence with a baseline etiological diagnosis of AD in LOCI was higher than EOCI (73.1% [8.4%] vs. 67.9% [10.4%]; p < 0.001).

3.2. Prevalence of amyloid positivity

Figure 1A illustrates the prevalence of Aβ‐PET positivity across cognitive stages. There were amyloid positive in 703 of the whole 1073 cases (65.5%). Aβ‐PET was positive in 586 of 746 patients (78.6%) with suspected AD etiology and in 117 of 327 (35.8%) patients with a pre‐PET diagnosis of non‐AD. The prevalence of amyloid positivity increased with the severity of cognitive stage: 26 of 114 participants (22.8%) with SCD, 236 of 397 (59.4%) with MCI, and 441 of 562 participants (78.5%) with dementia.

FIGURE 1.

Prevalence of amyloid‐PET positivity across cognitive stages in the whole sample (A) and disaggregating by age of onset (B) and (C). EOCI, early‐onset cognitive impairment; LOCI, late‐onset cognitive impairment; MCI, mild cognitive impairment; PET, positron emission tomography; SCD, subjective cognitive decline

Aβ‐PET was positive in 347 of 534 patients (65.0%) with EOCI and in 356 of 539 participants (66.0%) with LOCI. Disaggregating by baseline cognitive stage, we observed a trend consistent with that of the whole sample (Figure 1B and C).

3.3. Change in etiological diagnosis after Aβ‐PET

Figure 2, Figure S3, and Table S3 indicate the etiological diagnostic change in the whole sample and by age‐of‐onset and cognitive‐stage groups. Aβ‐PET results contributed to a change of suspected etiological diagnosis in 207 of 1073 patients (19.3%). Diagnoses changed from AD to non‐AD in 135 participants (18.1% of participants with a pre‐PET diagnosis of AD) and from non‐AD to AD in 72 participants (22.0% of participants with a pre‐PET diagnosis of non‐AD). Diagnosis changed more often because of negative (36.5%, 135/370) than positive PET results (10.2%, 72/703; p < 0.001). Disaggregating by baseline cognitive stage, the diagnosis changed more often in MCI (113/397, 28.5%; p < 0.001) and SCD (24/114, 21.1%; p < 0.001) than in dementia (70/562, 12.5%).

FIGURE 2.

Change in etiological diagnosis after the amyloid PET scan in the whole sample (A) and disaggregating by age of onset (B) and (C). AD, Alzheimer's disease; EOCI, early‐onset cognitive impairment; LOCI, late‐onset cognitive impairment; PET, positron emission tomography

Post‐PET changes in the suspected etiology occurred more frequently in EOCI (25.3%, 135/534) than in LOCI (13.4%, 72/539; p < 0.001). Participants with a baseline diagnosis of AD (EOCI: 81/360, 22.5% vs. LOCI: 54/386, 14.0%; p = 0.003) and non‐AD (EOCI: 54/174, 31.0% vs. LOCI: 18/153, 11.8%; p < 0.001) were both reclassified more frequently in EOCI than in LOCI. In the EOCI group, the diagnosis changed more often in SCD (21/72, 29.2%; p < 0.001) and MCI (71/184, 38.6%; p < 0.001) than in dementia (43/278, 15.5%). Whereas, in the LOCI group, the diagnosis changed more often in MCI (42/213, 19.7%) than dementia (27/284, 9.5%; p = 0.002), but no statistical difference in SCD (3/42,7.1%) between MCI or dementia.

3.4. Change in diagnostic confidence

In the total population, the overall confidence of diagnosis increased from 69.3% [10.5%] to 85.6% [13.1%] (p < 0.001), both as a result of positive (from 69.5% [10.4%] to 87.3% [10.7%]; p < 0.001) and negative (from 68.9% [10.8%] to 82.3% [16.3%]; p < 0.001) scan results. In participants for whom a baseline etiological diagnosis of AD was confirmed, diagnostic confidence increased from 70.7% [9.8%] to 88.5% [10.1%] (+17.8%; p < 0.001; Figure 3A), and the same trend occurred in participants for whom a baseline etiological diagnosis of non‐AD was confirmed, from 68.6% [10.8%] to 85.0% [14.8%] (+16.4%; p < 0.001; Figure 3D). The confidence was increased in the participants with a pre‐PET non‐AD diagnosis having a positive scan result and diagnostic change to AD where confidence increased from 57.6% [9.9%] to 80.2% [11.4%] (+22.6%; p < 0.001; Figure 3C), and in the patients with a pre‐PET AD diagnosis having a negative scan result and diagnostic change to non‐AD where confidence increased from 70.2% [9.6%] to 76.1% [16.9%] (+5.9%; p = 0.001; Figure 3B). Disaggregating by baseline cognitive stage, the highest increases in diagnostic confidence were observed in the MCI group (+17.3%, from 66.5% [10.1%] to 83.8% [14.0%]) than the SCD group (+14.3%, from 72.5% [10.3%] to 82.9% [11.6%]; p < 0.001) and dementia group (+16.0%, from 70.6% [10.4%] to 86.6% [12.7%]; p < 0.001).

FIGURE 3.

Change in diagnostic confidence with the added information from the visual read of the amyloid PET scan in the whole sample. The diagnostic confidence that cognitive impairment due to the AD/Non‐AD all significantly increased with a confirmed diagnosis in (A) AD→AD and (C) Non‐AD→Non‐AD; the consistent trend was also observed in patients whose diagnosis changed in (B) AD→Non‐AD and (D) Non‐AD→AD. AD, Alzheimer's disease; PET, positron emission tomography

Diagnostic confidence increased more in EOCI (+18.4%, from 66.9% [11.3%] to 85.3% [13.4%]) than in LOCI (+14.2%, from 71.6% [9.2%] to 85.8% [12.9%]; p < 0.001) (Table S3 and Figure S4). In the analysis by baseline cognitive stage, similar trends were observed in MCI (EOCI: +18.5%, from 63.8% [10.7%] to 82.3% [14.8%] vs. LOCI: +16.2, from 68.8% [10.0%] to 85.0% [13.0%]; p = 0.002) and dementia (EOCI: +19.3%, from 67.7% [10.9%] to 87.0% [12.3%] vs. LOCI: +12.8%, from 73.4%[9.1%] to 86.2%[13.2%]; p < 0.001), but not in SCD (EOCI: +14.7%, from 72.2% [11.6%] to 86.8% [12.4%] vs. LOCI: +13.6%, from 73.0% [7.7%] to 86.5% [10.3%]; p = 0.883).

3.5. Change in cognition‐specific and non‐cognition‐specific medications

Aβ‐PET results led to a change of treatment plan in 392 of 1073 patients (36.5%, Figure 4), including 62 patients with SCD (5.8%), 179 with MCI (16.7%), and 161 with dementia (15.0%) (Figure S5). After a positive PET result, the use of cognition‐specific drugs increased significantly from 77.7% to 95.6% (p < 0.001) in the whole sample, and in patients with SCD from 19.2% to 69.2% (p = 0.001), MCI from 62.7% to 91.5% (p < 0.001), and dementia from 89.1 to 99.3% (p < 0.001). A similar introduction was observed in the use of a non‐cognition‐specific drug after an Aβ‐negative PET result: +10.6% (p = 0.004) in the whole sample, +4.5% (p = 0.557) in SCD, +11.8% (p = 0.002) in MCI, and +13.2% (p = 0.029) in dementia. However, we did not observe that negative Aβ‐PET results lead to cognition‐specific drug discontinuation. Changes in cognition‐specific and non‐cognition‐specific drugs after Aβ‐PET were similar in EOCI and LOCI (Figure 4 and Table S3).

FIGURE 4.

Change in the use of (A) cognition‐specific and (B) noncognition specific medications. AD, Alzheimer's disease; EOCI, early‐onset cognitive impairment; LOCI, late‐onset cognitive impairment

Ancillary investigations were more often for patients with a negative PET (99/370, 26.8%) scan than those with a positive PET scan (69/703, 9.8%; p < 0.001). Neurologists performed ancillary investigations (Table S4), such as FDG‐PET (6 [3.6%]), DaT scan (7 [4.2%]), tau PET (74[44.0%]), genetic screening (42 [25%]), CSF analysis (22 [13.1%]), or others (17 [10.1%]). Tau PET was the most common ancillary investigation, and the percentage was higher in the group having a negative Aβ‐PET scan result (34/370, 9.2%) than a positive result (40/703, 5.7%, p = 0.03).

4. DISCUSSION

In this large, real‐world study, we found that Aβ‐PET led to the changes in etiological diagnosis, diagnostic confidence, and treatment plan, providing new evidence for the use of Aβ‐PET in Chinese population. In about two thirds of these patients suspected of cognitive impairment, PET showed evidence of amyloid pathology, comparable with proportions found in other studies. ⁹ , ²⁸ PET results frequently aligned with the initial clinical diagnosis, leading to heightened confidence in the diagnostic accuracy. In the current study, 19.3% of participants experienced a change in their etiological diagnoses, and the diagnostic confidence overall increased. Subsequently, positive PET results more often led to the initiation of cognition‐specific medications, while negative results led to the administration of non‐cognition‐specific medications, as reported previously. ¹¹ , ¹² , ²⁹ , ³⁰ , ³¹ A total of 15.7% of participants had received further ancillary investigations, especially tau PET, which was often used to confirm AD etiology in amyloid‐positive cases and to detect non‐AD tauopathies in amyloid‐negative cases. ³²

Compared to previous Chinese studies, the overall impact on diagnostic change in our study appears to be relatively lower, ¹⁹ , ²⁰ , ²¹ although these findings in other countries were highly variable, ranging from 19% to 79%. ³³ This might be due to differences in study design and enrollment population. Shea et al. and Liu et al. analyzed the value of amyloid PET combined with FDG‐PET in cognitive impairment diagnosis. ¹⁹ , ²⁰ The high proportion of patients with uncertain dementia was recruited in Bao et al. study. ²¹ Whereas we opted to use an unselected memory clinic sample, rather than imposing restrictive inclusion and exclusion criteria, which may account for the lower proportional change.

Discontinuation of cognition‐specific medications was associated with negative PET results, ¹¹ , ²⁹ which was not observed in this study. There were more than half of the dementia patients in our study. Prior studies have shown that even after negative scans, dementia patients tend to continuing using cholinesterase inhibitor or memantine for treatment. ²⁹ An alternative explanation maybe the difference in prescription patterns in China, where the prevalence of cognition‐specific drug medication prescription in memory clinics in China is much higher than in other countries. ³⁴

This study shows previous findings, as we also compared the additional value in early‐ and late‐onset patients. The early‐onset cases showed a significantly higher rate of diagnostic changes compared with the late‐onset cases. When we analyzed the data by baseline cognitive stage (i.e., SCD, MCI, and dementia), we observed a trend that aligns with the findings from the entire sample. To the best of our knowledge, only a few previous studies have examined the diagnostic changes resulting from Aβ‐PET patients with early‐ and late‐onset conditions. Apostolova et al. ³⁵ reported less frequent diagnostic changes (EOCI vs. LOCI: 17% vs. 43%) and more frequent therapeutic changes (EOCI vs. LOCI: 79% vs. 59%) in early‐onset cases. Similar to Apostolova et al., an unselected memory clinic cohort study, the Alzheimer's biomarkers in daily practice (ABIDE) project, found that changes in the suspected etiology occurred more frequently in older (29%) than younger (20%) patients. ¹² However, Hellwig et al. ¹⁶ observed the clinical diagnosis changed with comparable frequencies in the 2 age groups (early‐onset 22%, late‐onset 23%). The discrepancy may be explained by different rates of negative amyloid scans, which often led to diagnostic changes in cases with a pre‐PET diagnosis of AD. The rates of negative amyloid scans in this population were 34.5% of the whole sample, 35.0% of EOCI, and 34.0% of LOCI, respectively. However, the rates were higher in de Wilde et al. (52.3%) ¹² and Hellwig et al. (41.7%) ¹⁶ studies, and similar (32.7%) in Apostolova et al. ³⁵ study, with a significantly greater proportion of late‐onset compared with early‐onset cases (42.8% vs. 20.8%).

The diagnostic confidence in EOCI increased more than LOCI, especially in MCI and dementia groups in our data. This result was consistent with previous investigations, ³⁶ as there is less probability of “age‐related” brain amyloidosis in younger patients. ³⁷ Collectively, it suggested that Aβ‐PET for both EOCI and LOCI was meaningful for establishing the underlying etiology and increasing diagnostic confidence. The EOCI, which fulfilled the AUC by age, was likely to have greater benefit through Aβ‐PET. ³⁸

This study has several limitations. First, the patients were included in the tertiary hospitals of Shanghai. The sample may not be fully representative of the local population, therefore limiting the generalizability of the findings to primary care and local memory clinics. Second, since the results of Aβ‐PET were assessed by neuroradiologists, the interpretations of images could be largely dependent on personal training and experience. Third, our study allowed patients to self‐select for Aβ‐PET, which may have introduced a bias. Younger people are often more concerned about their cognitive functions and the risk of dementia, ³⁹ potentially leading to a higher prevalence of EOCI observed in our study. Fourth, while the sample size was the largest reported to date in China, it may still be considered modest in scale for ensuring broad generalizability of the findings.

In conclusion, our study has shown that Aβ‐PET has significant impacts on the changes in diagnoses and management among unselected memory clinic patients in China. Early‐onset patients are perhaps more likely to benefit from Aβ‐PET than late‐onset patients. Future studies employing large‐scale multicenter prospective clinical protocols are needed to further clarify the impact of Aβ‐PET in Chinese population. Additional economic data are required to be combined to assess the cost‐effectiveness of Aβ‐PET in clinical routine.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest. Author disclosures are available in the supporting information.

CONSENT STATEMENT

This study was approved by approved by the Ethics Committee of Huashan Hospital. In accordance with the Declaration of Helsinki, written informed consent was obtained from all participants or their guardians.

Supporting information

Supporting Information

Supporting Information

Chen K‐L, Wang M‐Y, Wu J, et al. Incremental value of amyloid PET in a tertiary memory clinic setting in China. Alzheimer's Dement. 2024;20:2516–2525. 10.1002/alz.13728