Predictive Accuracy of Blood-Derived Biomarkers for Amyloid-β Brain Deposition Along with the Alzheimer’s Disease Continuum: A Systematic Review

Abstract

Background:

An amyloid-β (Aβ) positron emission tomography (Aβ-PET) scan of the human brain could lead to an early diagnosis of Alzheimer’s disease (AD) and estimate disease progression. However, Aβ-PET imaging is expensive, invasive, and rarely applicable to cognitively normal subjects at risk for dementia. The identification of blood biomarkers predictive of Aβ brain deposition could help the identification of subjects at risk for dementia and could be helpful for the prognosis of AD progression.

Objective:

This study aimed to analyze the prognostic accuracy of blood biomarkers in predicting Aβ-PET status along with progression toward AD.

Methods:

In accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, we searched bibliographic databases from 2010 to 2020. The quality of the included studies was assessed by the Quality Assessment of Diagnostic Accuracy Studies 2 (QUADAS-2) tool.

Results:

A total of 8 studies were retrieved. The prognostic accuracy of Aβ-PET status was calculated by obtaining ROCs for the following biomarkers: free, total, and bound Aβ₄₂ and Aβ₄₀; Aβ_42/40 ratio; neurofilaments (NFL); total tau (T-tau); and phosphorylated-tau181 (P-tau181). Higher and lower plasma baseline levels of P-tau181 and the Aβ_42/40 ratio, respectively, showed consistently good prognostication of Aβ-PET brain accumulation. Only P-tau181 was shown to predict AD progression.

Conclusion:

In conclusion, the Aβ_42/40 ratio and plasma P-tau181 were shown to predict Aβ-PET status. Plasma P-tau181 could also be a preclinical biomarker for AD progression.

Keywords

Alzheimer’s disease amyloid-β biomarkers mild cognitive impairment preclinical Alzheimer’s disease positron emission tomography prognostic accuracy sensitivity specificity

INTRODUCTION

Alzheimer’s disease (AD) is an age-dependent neurodegenerative disorder and the most prevalent form of dementia among elderly individuals [1]. AD was initially defined as a clinicopathological entity, diagnosed in life as possible or probable AD, and definitely at autopsy [2]. In AD, the onset of the symptoms is subtle and occurs in an already advanced stage of the disease, making accurate and early diagnosis difficult in vivo. The pathological hallmarks of AD are 1) the deposition of Aβ protein in senile plaques of the brain parenchyma, which appears many years before the appearance of AD clinical symptoms, and 2) phosphorylated tau accumulation in neurofibrillary tangles in cerebral neurons, which appear in the later stages of the disease [3, 4]. In this context, the identification of Aβ pathology biomarkers in vivo represents an important advance for early diagnosis and prognosis. The determination of biomarkers that can predict the clinical progression to AD is crucial to accurately identify AD at-risk subjects, particularly for the amnestic subtype of mild cognitive impairment (aMCI), which is associated with an increased risk of developing AD [5, 6]. In this respect, AD conversion prediction over a shorter time span can be considered clinically relevant. Over time, in combination with clinical examinations, neuroimaging and biochemical biomarkers have assumed a pivotal role in the understanding of the neuropathogenic mechanisms along the continuum of the pathological processes from cognitively normal to abnormal and dementia. Such a multidisciplinary approach can be crucial for early diagnosis, differential diagnosis, and prediction of AD progression. In vivo biomarkers include cerebrospinal fluid (CSF) biomarkers and Aβ tracers by PET. Aβ-PET has been approbated for clinical use [7], introducing a new important step for early AD diagnosis in vivo and for monitoring disease progression. In recent years, abnormal levels of CSF Aβ₄₂, total tau (T-tau), and phosphorylated tau (P-tau) [8, 9] have been introduced as research diagnostic criteria due to their high diagnostic accuracy [4]. However, performing radiological and CSF biomarkers in non-demented patients is often not practical and cost-effective. Such limitations could be overcome by blood-based biomarkers predictive of Aβ-PET status. Several studies have analyzed the potential of blood-derived biomarkers to be predictive of AD; however, individual studies of biomarker validity may vary greatly, and their diagnostic performance is not always assessed. Given the high diagnostic value of Aβ-PET for AD diagnosis, in this systematic review, we aimed to determine the prognostic test accuracy of blood-derived biomarkers in predicting the Aβ-PET status between different diagnostic groups and during the transition toward AD. Our primary outcome was to determine the prognostic test accuracy of blood-derived biomarkers in predicting the Aβ-PET status in subjects diagnosed with cognitive normal (CN), subjective memory complaint (SMC), mild cognitive impairment (MCI), or Alzheimer’s disease (AD) at the time of performing the test. As secondary outcomes, we aimed to investigate the heterogeneity of the biomarkers in the included studies by evaluating the spectrum of people, clinical and neuroradiological diagnostic criteria; the techniques for the measurements of the index tests; the reference standards used; the duration of follow-up; the aspects of study quality.

METHODS

The paper was reported in accordance with the Preferred Reporting Items for Systematic Reviews (PRISMA) guidelines [10, 11]. All data originate from previously published papers in international peer-reviewed journals. The PRISMA checklist is shown in Supplementary Table 1.

Search strategy

Relevant studies were identified through electronic searches in PubMed, Embase, Cochrane Library, and Web of Science for eligible studies published from 2010 to 2020. We have restricted the search strategy to the last ten years due to the more sensitive techniques developed in recent decades for the measurement of blood-derived biomarkers in neurodegenerative diseases. The last search was performed on 29 October 2020. The search strategy was conducted by the combination of the following terms for titles and abstracts: (Alzheimer disease OR AD) AND (mild cognitive impairment OR MCI) AND (biomarker OR biomarkers) AND (blood OR serum OR plasma) AND (positron emission tomography OR pet) AND (amyloid OR β-amyloid OR beta-amyloid) NOT (cancer). Target events were a prediction of MCI (from CN to MCI) and AD (from MCI to AD) conversion. We decided to be stringent with search criteria to only include papers focused on the usefulness of circulating biomarkers and PET in AD. The search was limited to English-language articles. The reference lists of retrieved articles were screened for additional studies.

Study selection

The title and abstract of studies were first evaluated independently by two authors of this paper (AC, LC). The authors independently reviewed the selected published articles. A third reviewer (PM) independently reviewed the search strategy. Any disagreements that arose between the reviewers were resolved through discussion with another author (MS. If the abstracts fulfilled the inclusion criteria, the full text was reviewed.

Inclusion and exclusion criteria

The outcome of the study was to determine which blood biomarkers predict the onset of AD in CN subjects at risk for AD and subjects with MCI in association with Aβ-PET. We included longitudinal and prospective studies that evaluated the prognostic accuracy of one or more biomarkers in predicting Aβ-PET status along with AD disease progression.

The inclusion criteria were as follows:

The prognostic performance of the biomarkers in predicting the Aβ-PET status was calculated by ROC curves almost at 1) baseline, 2) follow-up, 3) the transition from CN/SMC to MCI, if applicable, and 4) the transition from MCI to AD, if applicable.

Participants were diagnosed with CN, SMC, MCI, and AD at the time of enrollment.

The participants underwent at least one Aβ-PET scan imaging.

The subjects included in the studies were diagnosed according to one or more of the following diagnostic criteria: Petersen criteria, National Institute on Aging and Alzheimer’s Association (NIA-AA), National Institute of Neurological and Communicative Disorders and Stroke, the Alzheimer’s Disease and Related Disorders Association (NINCDS-ADRDA) and Diagnostic and Statistical Manual of Mental Disorders (DSM).

Data analysis at follow-up was performed according to the baseline biomarker values.

Studies reporting only measures of association, such as a relative risk, hazard ratio, and odds ratio, were included if calculated using the same data used for ROC curves.

The exclusion criteria were as follows:

Follow-up duration less than 1 year.

Sample size < 10 individuals.

No biological data related to blood-derived biomarkers available (i.e., urine, saliva, liquor).

Not-quantitative methods used to assess blood biomarker concentrations (i.e., explorative proteomics or western blot).

Data extraction

The following data were extracted from each included study:

Bibliographic details of primary paper: authors, year, DOI.

Clinical and demographic details: number of participants; the number of subjects classified as Aβ-PET positive and Aβ-PET negative; diagnostic criteria; age; gender; sources of referral; duration of follow-up from the time of the biomarker’s measurement; APOE ɛ4 carrier status (if applicable); Mini-Mental State Examination (MMSE) score (if applicable); years of education (if applicable).

Aβ-PET imaging features: method of image analysis; measures of radiotracer Aβ retention; Aβ retention detection regions; radiotracer used; thresholds used to define Aβ positive and Aβ negative tests.

Index test features sampling source and procedures; measurement methods; thresholds used (if applicable); values of the index tests at baseline and at follow-up (if applicable); overall performance of the test in predicting the Aβ-PET status in terms of AUCs. We recorded the number of participants lost to follow-up and data necessary for the quality assessment, as defined below.

Quality assessment

The diagnostic accuracy of a study is the ability of a medical test, such as a laboratory test, to discriminate between two conditions of interest [12]. We assessed the methodological quality of the included studies using the Quality Assessment of Diagnostic Accuracy Studies 2 tool (QUADAS-2) [13]. This tool comprises four domains: patient selection, index test, reference standard, and patient flow. Briefly, the test whose accuracy is evaluated is called the index test. The ability of an index test to correctly classify the target condition is compared to the best method used for correctly classifying the target condition; this method is the reference standard. The patient selection tool describes the methods of patient selection, and patient flow refers to subjects lost to follow-up for not receiving the index test or the reference standard. We assessed each domain in terms of risk of bias and considered the first three domains in terms of applicability concerns. Studies were scored as “L” for low risk of bias/low concerns regarding applicability, “H” for high risk/high concerns, and “U” for unclear items for each domain. All domains with at least 1 negative response were scored as H regarding applicability, whereas domains with no negative responses but at least 1 unsure response were scored as U. Domains with no negative and no unsure responses were scored as L. Two blinded independent raters (AC and LC) performed the QUADAS-2 assessment. Any disagreements were resolved by consensus or arbitration. We produced a narrative summary that described each included study as having a high, low, or unclear risk of bias, as well as concerns regarding applicability. The risk of bias and applicability judgments in QUADAS-2 items are shown in Supplementary Table 2.

RESULTS

Search results

The selection process and study flow diagram are depicted in Fig. 1. After the primary search, 997 titles of studies were evaluated. After removing duplicates (n = 44), the titles of the remaining 953 studies were evaluated. At this step, 716 were excluded. The remaining 237 studies fulfilled the general aspect of the review and were selected for title and abstract screening. A total of 194 papers did not fulfill the inclusion criteria, and 43 studies were selected for full-text analysis. Of these, 35 studies were excluded: 18 papers were cross-sectional studies [14 –31]; 10 studies did not obtain ROCs curves [17 , 32–39]; 3 studies did not correlate the index tests levels with Aβ-PET imaging [16 , 41]; 1 study performed the follow-up on 6 patients [42]; 1 study validated the baseline data on a different independent cohort [43]; and two studies were consecutive studies [44, 45]. Finally, 8 studies were included in the final analysis.

Fig. 1

Flow diagram of identification, screening, and inclusion of the studies.

Characteristics of included studies

All the included studies evaluated the performance of one or more blood biomarkers in predicting the Aβ-PET status. The index tests were measured, and their diagnostic performance in predicting the Aβ-PET status was calculated at baseline in 5 studies [27 , 47], starting from month 18 in 1 study [48] and within 18 months in 1 study [40]. Follow-up ranged from 18 months to 8 years. In total, the levels of plasma Aβ₄₂ (n = 6 studies), Aβ₄₀ (n = 6 studies), Aβ_42/40 ratio (n = 4 studies), free, total, and bounded Aβ_42/40 ratio (n = 1 study), Aβ_40/42 ratio (n = 1 study), plasma T-tau (n = 1 studies), plasma P-tau181 (n = 3 studies), and NFL (n = 2 studies) were measured to analyze the accuracy prediction of Aβ-PET status. Blood biomarkers were measured and correlated with Aβ-PET status (positive or negative) or radiomic data (standard uptake value ratio SUVR) in CN (n = 5 studies), SMC (n = 3 studies), MCI (n = 4 studies), and AD (n = 4 studies) participants. Three studies made 1 measurement of the index tests at baseline [27 , 49], 4 studies made 2 [44], 3 [46, 48] and 6 [50] measurements of the index tests at baseline and at follow-up time points, and 1 study performed PET scans within 18 months [40]. In the study by Verberk, paired biomarker/Aβ-PET data were available both at baseline and follow-up, and a correlation was made in vivo for both time points [49]. In the study by Rembach and colleagues, the participants underwent two different measurements of Aβ fragments and two Aβ-PET investigations. Statistical analysis between the biomarker levels and Aβ-PET was performed at both times; therefore, we included it in the analysis without ROC analysis. The number of participants who converted to MCI or AD was reported in 3 studies [27 , 49]. Among a total of 1,493 subjects comprising CN, SMC, and MCI participants, 66 CN converted to AD, 23 SMC to MCI, 14 CN to MCI, and 51 MCI to AD for a total of 154 subjects. Schindler et al. reported the number of Aβ-PET conversions in a subcohort. At follow-up, a neuroradiological Aβ-PET reclassification was performed in only 3 studies [27 , 48]; in all the other studies, the reclassification of the patients at follow-up was based only on the clinical and neuropsychological diagnosis. The included studies varied in radiotracer used and in how the scans were performed and interpreted. Table 1 summarizes the data regarding the Aβ-PET imaging protocols.

Table 1

Aβ-PET characteristics and setting

Study	Radiotracer	Threshold/prespecified (Yes/No/NA)	Measure of radiotracer amyloid retention	Method of image analysis	Retention detecting region
Vergallo et al., 2019 [46]	¹⁸F-florbetapir	Mean cortical SUVR. > 0.7918/Yes	Neocortical global average SUVR to the pons and whole cerebellum mean	CATI	Cingulum posterior right and left, cingulum anterior right and left, frontal superior right and left, parietal inferior right and left, precuneus right and left, temporal mid right and left
Fandos et al., 2017 [48]	¹¹C-Pittsburgh Compound-B, ¹⁸F-flutemetamol	Mean cortical SUVR/BeCKeT:> 1.5/Yes	ROI to-cerebellar cortex and the pons ratio and VOI to cerebellar cortex ratio	Manual	Frontal, superior parietal, lateral temporal, lateral occipital, and anterior and posterior cingulate regions
Rembach et al., 2014 [44]	¹¹C-Pittsburgh Compound-B or ¹⁸F-flutemetamol	NA/NA	NA	NA	Global neocortical region
Janelidze et al., 2020 [27]	¹⁸F-flutemetamol	NA/NA	Neocortical weighted mean SUVR to white matter, cerebellum and brainstem	SPM12	Global neocortical region
Thijssen et al, 2020 [31]	¹¹C-Pittsburgh Compound-B, ¹⁸F-florbetapir	Not reported (visual reads)/NA	Average cortical SUVR to cerebellar gray matter and whole cerebellum	SPM12	Frontal, temporal, parietal, and posterior cingulate cortex for PiB, not specified for florbetapir
Karikari et al., 2020 [50]	¹⁸F-florbetapir	SUVR > 1.11/Yes	SUVR to the full cerebellum	Manual	Brain
Verberk et al., 2018 [49]	¹⁸F-florbetaben, ¹⁸F-florbetapir, ¹⁸F-flutemetamol, ¹¹C-Pittsburgh Compound-B	Not reported (visual reads)/NA	NA	Manual	NA
Schindler et al., 2019 [40]	¹¹C-Pittsburgh Compound-B	SUVR > 1.42/Yes SUVR > 1.22/Yes	Average cortical SUVR to left and right lateral orbitofrontal, medial orbitofrontal, precuneus, rostral middle frontal, superior frontal, superior temporal, and middle temporal cortices	NA	Cerebellar gray

Quality assessment

The overall methodological quality of the studies according to the QUADAS-2 scores is displayed in Fig. 2. The review authors’ judgments regarding each methodological quality item for each included study are presented in Supplementary Table 3.

Fig. 2

Risk of bias and applicability concerns summary: review authors’ judgments about each domain for each included study.

Participants’ characteristics

The selected papers combined results for a total of 3,605 subjects, 85.4%of whom were not diagnosed with dementia at the time of the measurement of the biomarkers. In total, the studies included participants diagnosed at baseline with CN (n = 1,690, 47%), SMC (n = 524, 14.5%MCI (n = 863, 24%), and AD (n = 486, 13). A cohort was composed of 42 mixed MCI/mild to moderate AD patients [31]. Aβ-PET was performed in 2,225 subjects, of which 50%were Aβ-PET positive (CN positive n = 242, SMC positive n = 96, MCI positive n = 308, AD positive n = 462). Female gender represented 57%of the total sample size (1,268), age ranged from 60.6 (s.d.±22) to 78 (s.d.±7.8) years, the APOE ɛ4 carrier frequency was 12%in CN, 27.3%in SMC, 31.75%for MCI and 46.3%for AD. The MMSE score ranged from≥27 for all CN and SMC participants and from 26.8 (s.d.±2.4) to 28.4 (s.d.±1.79) for MCI participants. The MMSE score for AD ranged from 19.34 (s.d.±5.3) to 23.7 (s.d.±1.9). Baseline characteristics are shown in Table 2. Detailed characteristics of the included studies are shown in Supplementary Table 3.

Table 2

Baseline demographic characteristics of the included studies

Author	Disease group	No of cases	Aβ-PET+ (N)	Gender (female, %)	Age (y, mean (s.d.))	APOE ɛ4 carriers (%)	MMSE score mean (s.d.)	Years of education (s.d.)	Diagnostic criteria	Follow-up (y)	Transition (N)	Biomarkers	Method
Vergallo et al., 2019 [46]	SMC	276	73	61	Aβ+: 77.3 (3.2)	Aβ+: 10	≥27	n.a.	Cognitive measures	3	n.a.	Aβ_1–42, Aβ_1–40, Aβ_1–40/Aβ_1–42 ratio, t-Tau, NFL	SIMOA
					Aβ–: 76.6 (3.4)	Aβ–: 10
Fandos et al., 2017 [48]	CN	76	27	51	Aβ+: 74.4 (7.4)	Aβ+: 22	n.a.	Aβ+: 12.2 (2.7)	Neuropsychological test battery and cognitive measures	4.5	n.a.	Free, total, and bounded plasma Aβ₄₂/Aβ₄₀ ratio	ELISA
					Aβ–: 72 (7.3)	Aβ–: 20^*		Aβ–: 12.5 (2.3)
Rembach et al., 2014 [44]	CN	768	n.a.	5	69 (6.8)	2	28.911 (1.1)	n.a.	Neuropsychological test battery	1.5	HC - > MCI: 14	Aβ_1–40, Aβ_1–42, Aβ_n–40, Aβ_n–42	INNO-BIA assays
	MCI	133		26	75 (8.2)	27	26.275 (2.4)		NINCDS-ADRDA		MCI - > AD: 23
	AD	211		36	78 (7.8)	39	19.344 (5.3)				Stable HC: 559
											Stable MCI: 41
Janelidze et al., 2020 [27]	CN	219	93	57	Aβ–: 71 (68-75)	n.a.	Aβ–: 29 (28–30)	Aβ–: 12 (9–14)	Petersen criteria, NIA-AA, DSM-5, NIA–Reagan criteria, neuropsychological battery	Up to 8	CN- > AD: 62	Plasma P-tau181, t-Tau, Aβ₄₀, Aβ₄₂, NFL	MSD platform
	MCI	125	81	38	Aβ+: 72 (69–75)		Aβ+: 29 (28–30)	Aβ+: 12 (10–14)			MCI- > AD: 51		Elecsys immunoassay
					Aβ–: 69 (63–74)		Aβ–: 28 (26–28)	Aβ–: 11 (8–12)					SIMOA
					Aβ+: 73 (69-76)		Aβ+: 27 (26–28)	Aβ+: 10 (8–13)
Thijssen et al., 2020 [31]	CN	69	11	46	60.6 (22)	26	29 (1)	n.a.	NIA-AA, CDR, neuropsychological battery	2	n.a.	Plasma P-tau181, Aβ₄₂, Aβ₄₀, Aβ_42/40 ratio, NFL	SIMOA
	MCI	47	18	44	60.8 (14)	12	26.8 (3)
	AD	56	51	59	65 (9)	35	20.3 (6)
	MCI/AD	42	n.a.	59	73 (8)	n.a.	22 (4)
Karikari et al., 2020 [50]	CN	400	68	43	Aβ+: 76.9 (6.2)	Aβ+: 8	Aβ+: 28.8 (1)	Aβ+: 16.6 (2.2)	NINCDS-ADRDA, Petersen criteria, neuropsychological battery	Up to 8	n.a.	Plasma P-tau181	SIMOA
	MCI	558	209	24	Aβ–: 73.5 (5.6)	Aβ–: 15	Aβ–: 29.1 (1.3)	Aβ–: 16.7 (2.7)
	AD	219	137	56	Aβ+: 73.9 (6.7)	Aβ+: 25	Aβ+: 27.6 (1.9)	Aβ+: 15.9 (2.8)
					Aβ–: 71.4 (8)	Aβ–: 16	Aβ–: 28.4 (1.79)	Aβ–: 16.3 (2.6)
					Aβ+: 77.3 (7)	Aβ+: 49	Aβ+: 22.9 (2.7)	Aβ+: 15.7 (2.7)
					Aβ–: 73.4 (8.2)	Aβ–: 6	Aβ–: 23.7 (1.9)	Aβ–: 15.8 (2.4)
Verberk et al., 2018 [49]	SMC	248	23	41	61 (9)	36	28.31 (0.5)	n.a.	Petersen criteria, NIA-AA, NINCDS-ADRDA	3 (±2)	23 SMC- > MCI	Aβ₄₂, Aβ₄₀, Aβ_42/40 ratio, t-Tau	SIMOA
											4 SMC- > AD
											8 SMC- > non-AD dementia
Schindler et al., 2019 [40]	CN	156	43	61.5	Aβ+: 71.4 (6.8)	Aβ+: 27.63	Aβ+: 29 (1.6)	Aβ+: 15.2 (3.2)	CDR,	Up to 9	66 stable Aβ-PET	Aβ_42/40 ratio	IPMS
					Aβ–: 60.8 (6.7)	Aβ–: 39.35	Aβ–: 29.4 (0.8)	Aβ–: 15.9 (2.2)	MMSE		8 > Aβ–PET conversion

MMSE, Mini-Mental State Examination; SMC, subjective memory complaint; CN, cognitive normal; MCI, mild cognitive impairment; AD, Alzheimer’s disease; Aβ-PET, amyloid-beta positron emission tomography; Aβ, amyloid-beta; Aβ+ or Aβ–, Aβ positive or negative subjects regarding categorical Aβ-PET measurement; NFL, neurofilament; NINCDS-ADRDA, National Institute of Neurological and Communicative Diseases and Stroke/Alzheimer’s Disease and Related Disorders Association; NIA-AA, The National Institute on Aging and Alzheimer’s Association; DSM, Diagnostic and Statistical Manual of Mental Disorders; CDR, Clinical Dementia Rating; ELISA, enzyme-linked immunosorbent assay; MSD, Meso Scale Discovery; SIMOA, single molecule array; IPMS, immuno-precipitation mass spectrometry; n.a., not available. ^*p < 0.05 between PET positive/negative groups. ^**p < 0.001 between PET positive/negative groups

Biomarker accuracy in the prediction of Aβ-PET status of the included studies

Biomarker test accuracy was evaluated according to the target condition.

Table 3 shows the AUCs for the studies included in this systematic review.

Table 3

Prognostic accuracy of Aβ₄₂, Aβ₄₀, the Aβ_42/40 ratio, the Aβ_40/42 ratio, plasma T-tau, plasma P-tau 181, and NFL in prognosticating Aβ-PET status

						AUCs (sensitivity-specificity, %)
Author	Target event	No of cases	Disease group	Groups comparison	Aβ₄₂	Aβ₄₀	Aβ_42/40-Aβ_40/42	T-tau	P-tau181	NFL
Vergallo et al., 2019 [46]	Prediction of Aβ-PET status and disease progression	276	SMC	SMC+ versus SMC–	0.680 (52–80%)	–	0.794 (78–75%)	–	–	–
Fandos et al., 2017 [48]	Prediction of Aβ-PET status and disease progression	76	CN	CN+ versus CN–	–	–	NA (71–78%)^*	–	–	–
Rembach et al., 2014 [44]	Correlation between Aβ fragments levels and Aβ-PET status	184	CN, MCI, AD	Baseline versus follow-up		NA	NA
Janelidze et al., 2020 [27]	Prediction of Aβ-PET status and disease progression	344	CN, MCI	Aβ+ versus Aβ–			0.77 (75–69%)	0.61 (44–77%)	0.81 (79–75%)
Thijssen et al., 2020 [31]	Prediction of Aβ-PET status and disease progression	112	CN, MCI, AD, MCI/AD	Aβ+ versus Aβ–			0.768 (57–92%)		0.944 (94–86%)	0.559 (43–88%)
				MCI+ versus MCI–					0.859 (82–83%)
				CN+ versus CN–
Karikari et al., 2020 [50]	Prediction of Aβ-PET status and disease progression	1000	CN, MCI, AD	AD+ versus CN–					0.85 (NA)	0.707 (NA)
				AD+ versus MCI–					0.84 (NA)	0.63 (NA)
				Aβ+ versus Aβ –					0.85 (NA)	0.605 (NA)
				MCI+ versus MCI–					0.80 (NA)
				CN+ versus CN–					0.70 (NA)
Verberk et al., 2018 [49]	Prediction of Aβ-PET status and disease progression	69	SMC	SMC+ versus SMC–	0.66 (NA)		0.79 (NA)
Schindler et al., 2019 [40]	Prediction of brain amyloidosis	156	CN	Aβ+ versus Aβ–		0.88 (0.82–0.93)

SMC, subjective memory complaint; CN, cognitive normal; MCI, mild cognitive impairment; AD, Alzheimer’s disease; AUC, area under the curve; Aβ-PET, amyloid-beta positron emission tomography; Aβ, amyloid-beta; NFL, neurofilament; T-tau, total-tau; P-tau181, phosporilathed-tau181. ^*TP 42/40.

Plasma Aβ fragments prognostic accuracy of Aβ PET

The prognostic accuracy of Aβ fragments in predicting the Aβ-PET status was calculated in 2086 subjects, including CN, MCI, and AD [27 , 49]. Higher AUCs in distinguishing between Aβ-PET-positive and Aβ-negative samples were obtained for the Aβ_42/40 ratio (AUC 0.88 (082–0.93)) followed by Aβ₄₂ (AUC 0.680) [44 , 49]. In the study by Fandos, ROCs analysis of total Aβ_42/40 ratio (TP_42/40) was performed for all plasma-Aβ-PET paired data, and by including plasma TP_42/40 in the demographic models (age positive APOE ɛ4), the AUCs value improved with respect to the model with only age and APOE ɛ4 (p = 0.0017 at DeLong test) [48]. Older age and APOE ɛ4 carrier status were associated with a lower plasma Aβ_42/40 ratio in AβPET-positive subjects, improving the AUC from 0.88 to 0.994 in CN subjects; however, the significance was lost when the two variables were added to the model [40]. In all the studies, lower Aβ_42/40 ratios and Aβ₄₂ levels distinguished Aβ-PET-positive from Aβ-PET-negative patients, independent of diagnostic groups [27 , 49]. All the studies showed a negative correlation between the Aβ_42/40 ratio and SUVRs or Aβ-PET status at baseline [27 , 49] within 18 months from the Aβ-PET scan [40] and at M18 [44, 48]; however, no differences in the rate of change among groups who remained stable or who progressed from HC to MCI and from MCI to AD were observed [44], and no improvement in differential diagnosis emerged [31]. A lower baseline Aβ_42/40 ratio was associated with Aβ-PET conversion (p < 0.01) [42] and with progression from SMC to MCI or dementia (p = 0.002); however, the significance was lost after correcting for age and sex [49]. In the study by Vergallo et al. [46], a longitudinal prediction of the plasma Aβ_40/42 ratio of brain Aβ-PET status was analyzed in SMC by using a machine learning approach (classification and regression trees and random forest analysis) by comparing the training set (time point-1) and the test set (time points-2 and 3). Data on the diagnostic performance of the Aβ_40/42 ratio for Aβ-PET status were confirmed, although no follow-up on Aβ-PET data was available; thus, the number of SMCs who progressed to positive Aβ-PET was unknown. Only one study reported, longitudinally, a significant correlation of Aβ_1–40 and Aβ_n40 fragments with Aβ-PET (p < 0.001) in subjects who progressed from MCI to AD [44]. Incomplete data were available for the remaining selected study [27, 48]. In all the studies, longitudinal analyses were performed according to the baseline data.

Plasma P-tau181 prognostic accuracy of Aβ-PET status

Plasma P-tau 181 accuracy in predicting the Aβ-PET status was evaluated as the event of interest in three studies [27 , 50]. All three studies assessed the ability of P-tau181 to predict Aβ accumulation by ROCs. In all the studies, increased P-tau181 was significantly associated with increased Aβ-PET; furthermore, its levels increased over the continuum from CN to AD [27]. The accuracy of plasma P-tau181 in predicting the Aβ-PET status was measured at baseline. At follow-up, the participants were clinically re-evaluated, and the prognostic values of the biomarkers in predicting cognitive decline were evaluated on the basis of the baseline biomarker data. Plasma P-tau181 had high accuracy in predicting Aβ-PET accumulation in the comparison between Aβ-PET positive (all) and Aβ-PET negative (all) [27, 31], CN positive and CN negative [50], MCI positive and MCI negative [50], and AD positive and Aβ negative (all) [50]. All 3 studies used a cutoff above which plasma P-tau181 levels were considered abnormal [27] and able to predict Aβ-PET positivity and negativity [31, 50]. The relationship between plasma P-tau181 and Aβ-PET status during the conversion to dementia was described in only 1 study [27], and the other 2 studies predicted disease progression based on baseline plasma P-tau181 levels. Along the AD continuum, plasma P-tau181 predicts the Aβ-PET status with high prognostic accuracy and distinguishes AD-positive from CN-negative and MCI-negative patients [27, 50]. At a cutoff > 14.5 pg/mL, plasma P-tau181 classified Aβ-PET CN-, MCI-, and AD-positive samples independently of the diagnostic groups, while at concentrations > 17.7 pg/mL, AD-positive samples were classified as CN negative and MCI negative. Concentrations > 17.7 pg/mL during the follow-up were associated with an increased risk of AD in MCI- and CN-positive patients compared to CN-negative patients [50]. At the cutoff of 8 pg/mL, plasma P-tau181 discriminates strongly between Aβ-positive and Aβ-negative patients [31]. In the longitudinal analysis, the studies were linear in the conclusion that higher levels of P-tau181 at baseline were associated with cognitive decline. Higher plasma P-tau181 baseline was associated with a higher risk of AD dementia in MCI and CN positive, as compared with CN negative [50], to a faster decline in AD and MCI [31] and in predicting the future progression to cognitive decline [27]. In longitudinal analysis, plasma P-tau181 levels remained moderately stable over time in all the diagnostic groups, suggesting that the accuracy in predicting the Aβ-PET status was stable over the study period [50]. In another study, plasma P-tau181 levels increased in 62 subjects without dementia who developed dementia compared to those who did not develop dementia [27].

Total tau prognostic accuracy of Aβ-PET

Total tau was investigated in 2 studies. In the study by Janelidze, for plasma T-tau, the AUC was 0.61 (sensitivity 44%, specificity 77%) [27]. In the study conducted by Verberk, plasma T-tau did not differ between normal and abnormal Aβ CSF in SMC subjects, and no ROC curves were generated [49].

NFL prognostic accuracy of Aβ-PET status

NFL accuracy in predicting the Aβ-PET status was assessed in two studies. At baseline, plasma NFL distinguishes Aβ-positive from CN-negative and MCI-positive from MCI-negative. Data on sensitivities and specificities were computed by ROC 95%confidence intervals (Youden Index). At the cross-sectional level, baseline NFL levels were associated with plasma P-tau 181 according to the Aβ-PET status but not with pathological progression when CN positive, MCI negative, and MCI positive were compared to CN negative (p > 0.05) [50]. A nonsignificant value of prognostic accuracy was found in the second study [31], where NFL distinguished between positive and negative Aβ-PET with an AUC of 0.559 (p = 0.276).

DISCUSSION

Main findings and their relationship to the literature

In this systematic review, we aimed to summarize the evidence on the prognostic performance of blood-derived biomarkers for predicting brain amyloid-β deposition from the earliest manifestations of AD pathology to the dementia stage. Important contributions proving the usefulness of circulating plasma biomarkers in predicting Aβ deposition have been published during the last ten years. In particular, despite being excluded by our search criteria, it is important to consider the work by Nakamura et al. [26], which focused on plasma Aβ biomarker detection in peripheral blood to predict individual brain Aβ-positive Aβ-negative status. This work strongly demonstrated the potential clinical utility of plasma biomarkers in predicting brain Aβ burden at an individual level with respect to PET imaging. Despite this important contribution, conflicting results for the diagnosis of AD using circulating biomarkers have been published. According to our systematic analysis, we found different results for the different biomarkers. Specifically, in subjects without dementia, higher levels of P-tau181 had the best prognostic performance in predicting the Aβ-PET positive status at baseline and at follow-up, totally and when stratified for different diagnostic groups. The minimum AUC value was 0.705 in distinguishing CN positive from CN negative [50] and reached a value of 0.994 in distinguishing MCI positive from MCI negative [31]. At follow-up, higher P-tau181 levels correlated with Aβ-PET baseline were associated with a higher incidence of dementia [27, 31] and with a higher risk of developing dementia at a defined cutoff value [50], suggesting a prognostic role for AD progression. In all 3 studies, a threshold value above which P-tau 181 discriminated between the diagnostic groups was used. A lower Aβ_42/40 ratio was the second biomarker that significantly discriminated between Aβ-PET-positive and Aβ-PET negative subjects without dementia, although the AUC values were lower than that registered for P-tau 181 (maximum AUC value 0.88). However, a lower Aβ_42/40 ratio discriminates Aβ-PET-positive (all) versus Aβ-PET-negative (all) patients, but no differences in the rate of change among diagnostic groups were observed during the transition from HC to MCI and from MCI to AD [44], differentiating between different types of dementia [31]. Similar data were obtained in the study of Schindler et al., where the plasma Aβ_42/40 ratio declined over 18 months of longitudinal Aβ-PET scan and was significantly lower in the Aβ-PET converter; however, longitudinally, no differences in the rate of change of the plasma Aβ_42/40 ratio were observed between PET-positive and PET-negative patients [40]. Aβ_42/40 ratio levels predict Aβ-PET status after 3 years of follow-up in subjects without dementia who progressed to positive Aβ-PET; however, authors refer only to positive Aβ-PET without distinguishing between CN, MCI, or AD groups [46]. The same bias was presented in the study of Schindler et al., where the results were obtained on PET converters without any clinical information [40]. Data on plasma T-tau, NFL, Aβ₄₀, Aβ₄₂, and Aβ_n42 were inconsistent or not significant. In 1 study, Aβ_n40 levels were predictive of disease progression [44]. The small number of studies that investigated T-tau and NFL did not allow us to make a final judgment. No consistent data were found for T-tau, although only 1 study had complete data [27], while for NFL, there were not sufficient data for a definitive evaluation of the ability to predict the Aβ-PET status. Plasma levels of Aβ₄₀ were not associated with AD progression [51] or Aβ-PET accumulation. There are few studies on longitudinal examinations of plasma biomarkers in relation to in vivo Aβ-PET images. In two recent papers published subsequently to the final search of this study, plasma P-tau181 was significantly associated with Aβ-PET pathology in early and late accumulating Aβ brain regions after 6 years of follow-up in CN and MCI subjects [52] and gradual increases along the AD continuum, other than differentiates between different neurodegenerative diseases [50]. In another study, high plasma P-tau 181 was associated with hypometabolism and cortical atrophy at baseline and over time in CN subjects, and longitudinally increased tau was associated with accelerated atrophy [53]. The Aβ₄₂, Aβ₄₀, and relative Aβ_42/40 ratios have been extensively investigated in blood and proven to be associated with AD pathogenesis; however, Aβ fragments in the blood can be derived from other tissues and thus are not specific, and the data are not always reproducible. The relationship of Aβ₄₂ fragments, especially the Aβ_42/40 ratio, with Aβ-PET imaging has been investigated in some cross-sectional studies, and whatever data show the ability of Aβ_42/40 ratio to predict the amyloid PET brain burden, other variables, such as the APOE ɛ4 genotype, are determinant elements [14 , 22]. Plasma NFL is considered a disease-nonspecific marker of neuronal injury. Higher levels of plasma NFL have been found at baseline in MCI and AD [54], with the greatest increase in patients with AD dementia [55], and predicts Aβ-PET load and cognitive performance in AD patients [56] and correlates with Aβ deposition; however, the rate of change predicts future tau PET but not Aβ-PET. In another study, NFL was associated with neurodegeneration in brain regions exceeding the AD-typical pattern in Aβ- participants [57], indicating low specificity for AD. In CN subjects, NFL, but not its rate of change, is higher in CN+ subjects than in CN- subjects and predicts hippocampal atrophy in preclinical AD subjects [58]. In another study, plasma NFL was not increased in response to Aβ pathology in CN subjects but only in MCI and AD subjects; furthermore, it was not associated with reduced white matter in CN subjects [54]. Such scientific evidence suggests a nonspecific predictive role for plasma NFl along the AD spectrum, which could explain the few and inconsistent data we obtained for this biomarker. Literature data on the role of plasma T-tau are controversial. Elevated plasma T-tau levels are associated with neurodegeneration and lower gray matter density in regions that are both AD specific as well as more generally [38, 59] and with Aβ+ [60], other than with widespread reductions of cortical glucose uptake and memory deficits in normal elderly subjects [61]. CN+ and AD subjects had higher plasma T-tau levels [17, 62] and were associated with an increased risk of MCI among CN participants with elevated brain Aβ but not significantly associated with the risk of dementia in MCI [33]. In longitudinal studies, higher levels of plasma T-tau were associated with significant neurological decline over a median follow-up of 3.0 years but not within 15 months of follow-up in CN [33], suggesting a role dependent on the progression of the disease. Data were confirmed by another study in which a combination of plasma T-tau and Aβ₄₂ levels was highly associated with cerebral amyloid deposition, brain glucose metabolism, and hippocampal volume change not at baseline but at 2 years of follow-up [63]. On the other hand, other studies did not find any differences in T-tau levels between AD and HC [64], no correlation between Aβ-PET SUVR within CN, MCI, AD [33, 62], no association with regional cortical thickness [68] and no correlation with all-cause of AD dementia risk in subjects without dementia [51].

Strengths and limitations

Our work has both strengths and limitations. The strengths of our study include using data from large cohorts. Most of the subjects included in the analysis were subjects without dementia at baseline and biomarker values and were blinded to the diagnosis in 5 of the 8 studies [27 , 50]. Generally, the selected papers are of good quality. QUADAS-2 evaluation evidenced a low risk of bias for all the domains except for study flow at follow-up. This limitation is due to the method of reclassification of the diagnostic groups based only on longitudinal clinical examinations, and the number and disease groups of participants who progressed to AD were reported only in 3 studies [27 , 49]. Therefore, at follow-up, the number of participants who progressed toward AD and the corresponding diagnostic categories are not always clear, introducing an important bias for applicability. Furthermore, in 4 studies, biomarker measurements and Aβ-PET images were performed only at baseline and not longitudinally [27 , 50]. Regarding the limitation of our study, it is the low number of papers included. This is due to our decision to investigate the prognostic power of blood biomarkers specifically in longitudinal studies to establish which biomarker better predicts the onset of AD in two main categories, cognitively normal subjects at risk for AD and subjects at an early stage of the disease. Another limitation of this systematic review could arise when considering the different radiotracers (PiB, flumetamol, and florbetapir) used across studies. However, although they present different white and gray matter retention characteristics, they have a good correlation for imaging analysis [65 –67].

In addition, although sensitive, robust, and reproducible biomarker assays were used, different analytical platforms (SIMOA, ELISA, INNO-BIA) and antibodies used to measure biomarkers, as well as the different biosample handling and procedures, could influence the overall results [68]. For example, in the case of SIMOA and INNOBIA technologies that are ultrasensitive detection methods, a more accurate and sensitive result is expected in comparison to the classical colorimetric ELISA test.

Implication and future perspectives

This study has several applications and implications for the findings. Amyloid imaging is useful for differential diagnosis in early-onset dementia; however, it is invasive and expensive and is not sufficient to make a certain AD diagnosis in an individual. A biomarker able to predict Aβ-PET accumulation could ameliorate the selection of patients for Aβ-PET scans, improving the chance for a correct diagnosis in vivo and reducing time and costs. In this view, a multidisciplinary approach of noninvasive markers with the capability to prognosticate AD may be valuable tools for early diagnosis and for monitoring disease progression from a cognitive normal condition to dementia stages. In conclusion, the results of our study suggest that higher plasma P-tau181 reflects amyloid Aβ-PET pathology and is prognostic of disease onset and progression better than the Aβ_42/40 ratio, which distinguishes between Aβ-PET positive and negative but is not useful for differential diagnosis and for predicting AD progression. Plasma P-tau181 but not the Aβ_42/40 ratio is predictive of Aβ-PET status independent of other factors, such as APOE ɛ4 carrier status, age, and sex of the participants. However, the small number of the selected studies, as well as the small number of subjects who transitioned to well-defined diagnostic groups, suggest the further research should be done and that these conclusions should be interpreted with caution.

Footnotes

ACKNOWLEDGMENTS

This work was supported by “Progetti di Ricerca Corrente” and “Rete Italiana Neurologica” funded by the Italian Ministry of Health.

Authors’ disclosures available online ().

The supplementary material is available in the electronic version of this article: .

References

Reitz

, Brayne

, Mayeux

(2011) Epidemiology of Alzheimer disease. Nat Rev Neurol 7, 137–152.

McKhann

, Drachman

, Folstein

, Katzman

, Price

, Stadlan

(1984) Clinical diagnosis of Alzheimer’s disease: Report of the NINCDS-ADRDA Work Group under the auspices of Department of Health and Human Services Task Force on Alzheimer’s Disease. Neurology 34, 939–944.

Scheltens

, Blennow

, Breteler

MMB

, de Strooper

, Frisoni

, Salloway

, Van der Flier

(2016) Alzheimer’s disease. Lancet 388, 505–517.

Blennow

, Zetterberg

, Fagan

(2012) Fluid biomarkers in Alzheimer disease. Cold Spring Harb Perspect Med 2, a006221.

Rozzini

, Chilovi

, Conti

, Bertoletti

, Delrio

, Trabucchi

, Padovani

(2007) Conversion of amnestic Mild Cognitive Impairment to dementia of Alzheimer type is independent to memory deterioration. Int J Geriatr Psychiatry 22, 1217–1222.

Schmidtke

, Hermeneit

(2008) High rate of conversion to Alzheimer’s disease in a cohort of amnestic MCI patients. Int Psychogeriatr 20, 96–108.

Herholz

, Ebmeier

(2011) Clinical amyloid imaging in Alzheimer’s disease. Lancet Neurol 10, 667–670.

Jack

, Bennett

, Blennow

, Carrillo

, Dunn

, Haeberlein

, Holtzman

, Jagust

, Jessen

, Karlawish

, Liu

, Molinuevo

, Montine

, Phelps

, Rankin

, Rowe

, Scheltens

, Siemers

, Snyder

, Sperling

, Contributors (2018) NIA-AA Research Framework: Toward a biological definition of Alzheimer’s disease. Alzheimers Dement 14, 535–562.

Olsson

, Lautner

, Andreasson

, Öhrfelt

, Portelius

, Bjerke

, Hölttä

, Rosén

, Olsson

, Strobel

, Wu

, Dakin

, Petzold

, Blennow

, Zetterberg

(2016) CSF and blood biomarkers for the diagnosis of Alzheimer’s disease: A systematic review and meta-analysis. Lancet Neurol 15, 673–684.

10.

Liberati

, Altman

, Tetzlaff

, Mulrow

, Gøtzsche

, Ioannidis

JPA

, Clarke

, Devereaux

, Kleijnen

, Moher

(2009) The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: Explanation and elaboration. PLoS Med 6, e1000100.

11.

Moher

, Liberati

, Tetzlaff

, Altman

, PRISMA Group (2009) Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement. PLoS Med 6, e1000097.

12.

Šimundić

A-M

(2009) Measures of diagnostic accuracy: Basic definitions. EJIFCC 19, 203–211.

13.

Whiting

, Rutjes

AWS

, Westwood

, Mallett

, Deeks

, Reitsma

, Leeflang

MMG

, Sterne

JAC

, Bossuyt

PMM

, QUADAS-2 Group (2011) QUADAS-2: A revised tool for the quality assessment of diagnostic accuracy studies. Ann Intern Med 155, 529–536.

14.

Lin

S-Y

, Lin

K-J

, Lin

P-C

, Huang

C-C

, Chang

C-C

, Lee

Y-C

, Hsiao

I-T

, Yen

T-C

, Huang

W-S

, Yang

B-H

, Wang

P-N

(2019) Plasma amyloid assay as a pre-screening tool for amyloid positron emission tomography imaging in early stage Alzheimer’s disease. Alzheimers Res Ther 11, 111.

15.

Risacher

, Fandos

, Romero

, Sherriff

, Pesini

, Saykin

, Apostolova

(2019) Plasma amyloid beta levels are associated with cerebral amyloid and tau deposition. Alzheimers Dement (Amst) 11, 510–519.

16.

Park

J-C

, Han

S-H

, Lee

, Jeong

, Byun

, Bae

, Kim

, Lee

, Yi

, Shin

, Kim

, Hwang

, Lee

S-W

, Mook-Jung

(2019) Prognostic plasma protein panel for Aβ deposition in the brain in Alzheimer’s disease. Prog Neurobiol 183, 101690.

17.

Palmqvist

, Insel

, Stomrud

, Janelidze

, Zetterberg

, Brix

, Eichenlaub

, Dage

, Chai

, Blennow

, Mattsson

, Hansson

(2019) Cerebrospinal fluid and plasma biomarker trajectories with increasing amyloid deposition in Alzheimer’s disease. EMBO Mol Med 11, e11170.

18.

Ashton

, Nevado-Holgado

, Barber

, Lynham

, Gupta

, Chatterjee

, Goozee

, Hone

, Pedrini

, Blennow

, Schöll

, Zetterberg

, Ellis

, Bush

, Rowe

, Villemagne

, Ames

, Masters

, Aarsland

, Powell

, Lovestone

, Martins

, Hye

(2019) A plasma protein classifier for predicting amyloid burden for preclinical Alzheimer’s disease. Sci Adv 5, eaau7220.

19.

Han

S-H

, Park

J-C

, Byun

, Yi

, Lee

, Mook-Jung

, KBASE Research Group (2019) Blood acetylcholinesterase level is a potential biomarker for the early detection of cerebral amyloid deposition in cognitively normal individuals. Neurobiol Aging 73, 21–29.

20.

Nourhashemi

, Hooper

, Cantet

, Féart

, Gennero

, Payoux

, Salabert

, Guyonnet

, De Souto Barreto

, Vellas

, Multidomain Alzheimer Preventive Trial/Data sharing Alzheimer (DSA) study group (2018) Cross-sectional associations of plasma vitamin D with cerebral β-amyloid in older adults at risk of dementia. Alzheimers Res Ther 10, 43.

21.

Hooper

, De Souto Barreto

, Coley

, Caussé

, Payoux

, Salabert

, Cesari

, Andrieu

, Bowman

G-L

, Weiner

, Vellas

(2017) Cross-sectional associations of total plasma homocysteine with cortical β-amyloid independently and as a function of omega 3 polyunsaturated fatty acid status in older adults at risk of dementia. J Nutr Health Aging 21, 1075–1080.

22.

Tateno

, Sakayori

, Kim

, Koeda

, Kumita

, Suzuki

, Okubo

(2017) Effect of apolipoprotein E phenotype on the association of plasma amyloid β and amyloid positron emission tomography imaging in Japan. Alzheimers Dement (Amst) 9, 51–56.

23.

Toledo

, Toledo

, Weiner

, Jack

, Jagust

, Lee

VMY

, Shaw

, Trojanowski

, Alzheimer’s Disease Neuroimaging Initiative (2012) Cardiovascular risk factors, cortisol, and amyloid-β deposition in Alzheimer’s Disease Neuroimaging Initiative. Alzheimers Dement 8, 483–489.

24.

Pérez-Grijalba

, Arbizu

, Romero

, Prieto

, Pesini

, Sarasa

, Guillen

, Monleón

, San-José

, Martínez-Lage

, Munuera

, Hernández

, Buendía

, Sotolongo-Grau

, Alegret

, Ruiz

, Tárraga

, Boada

, Sarasa

, AB255 Study Group (2019) Plasma Aβ42/40 ratio alone or combined with FDG-PET can accurately predict amyloid-PET positivity: A cross-sectional analysis from the AB255 Study. Alzheimers Res Ther 11, 96.

25.

Sun

H-L

, Li

W-W

, Zhu

, Jin

W-S

, Liu

Y-H

, Zeng

, Wang

Y-J

, Bu

X-L

(2018) The correlations of plasma and cerebrospinal fluid amyloid-beta levels with platelet count in patients with Alzheimer’s disease. BioMed Res Int 2018, 7302045.

26.

Nakamura

, Kaneko

, Villemagne

, Kato

, Doecke

, Doré

, Fowler

, Li

Q-X

, Martins

, Rowe

, Tomita

, Matsuzaki

, Ishii

, Arahata

, Iwamoto

, Ito

, Tanaka

, Masters

, Yanagisawa

(2018) High performance plasma amyloid-β biomarkers for Alzheimer’s disease. Nature 554, 249–254.

27.

Janelidze

, Mattsson

, Palmqvist

, Smith

, Beach

, Serrano

, Chai

, Proctor

, Eichenlaub

, Zetterberg

, Blennow

, Reiman

, Stomrud

, Dage

, Hansson

(2020) Plasma P-tau181 in Alzheimer’s disease: Relationship to other biomarkers, differential diagnosis, neuropathology and longitudinal progression to Alzheimer’s dementia. Nat Med 26, 379–386.

28.

Verdile

, Laws

, Henley

, Ames

, Bush

, Ellis

, Faux

, Gupta

, Li

Q-X

, Masters

, Pike

, Rowe

, Szoeke

, Taddei

, Villemagne

, Martins

, AIBL Research Group (2014) Associations between gonadotropins, testosterone and β amyloid in men at risk of Alzheimer’s disease. Mol Psychiatry 19, 69–75.

29.

Gupta

, Laws

, Villemagne

, Ames

, Bush

, Ellis

, Lui

, Masters

, Rowe

, Szoeke

, Taddei

, Martins

, AIBL Research Group (2011) Plasma apolipoprotein E and Alzheimer disease risk: The AIBL study of aging. Neurology 76, 1091–1098.

30.

Craig-Schapiro

, Perrin

, Roe

, Xiong

, Carter

, Cairns

, Mintun

, Peskind

, Li

, Galasko

, Clark

, Quinn

, D’Angelo

, Malone

, Townsend

, Morris

, Fagan

, Holtzman

(2010) YKL-40: A novel prognostic fluid biomarker for preclinical Alzheimer’s disease. Biol Psychiatry 68, 903–912.

31.

Thijssen

, La Joie

, Wolf

, Strom

, Wang

, Iaccarino

, Bourakova

, Cobigo

, Heuer

, Spina

, VandeVrede

, Chai

, Proctor

, Airey

, Shcherbinin

, Duggan Evans

, Sims

, Zetterberg

, Blennow

, Karydas

, Teunissen

, Kramer

, Grinberg

, Seeley

, Rosen

, Boeve

, Miller

, Rabinovici

, Dage

, Rojas

, Boxer

, Advancing Research and Treatment for Frontotemporal Lobar Degeneration (ARTFL) investigators (2020) Diagnostic value of plasma phosphorylated tau181 in Alzheimer’s disease and frontotemporal lobar degeneration. Nat Med 26, 387–397.

32.

Vergallo

, Houot

, Cavedo

, Lemercier

, Vanmechelen

, De Vos

, Habert

M-O

, Potier

M-C

, Dubois

, Lista

, Hampel

, INSIGHT-preAD study group, Alzheimer Precision Medicine Initiative (APMI) (2019) Brain Aβ load association and sexual dimorphism of plasma BACE1 concentrations in cognitively normal individuals at risk for AD. Alzheimers Dement 15, 1274–1285.

33.

Mielke

, Hagen

, Wennberg

AMV

, Airey

, Savica

, Knopman

, Machulda

, Roberts

, Jack

, Petersen

, Dage

(2017) Association of plasma total tau level with cognitive decline and risk of mild cognitive impairment or dementia in the Mayo Clinic Study on Aging. JAMA Neurol 74, 1073–1080.

34.

Burnham

, Faux

, Wilson

, Laws

, Ames

, Bedo

, Bush

, Doecke

, Ellis

, Head

, Jones

, Kiiveri

, Martins

, Rembach

, Rowe

, Salvado

, Macaulay

, Masters

, Villemagne

, Alzheimer’s Disease Neuroimaging Initiative, Australian Imaging, Biomarkers and Lifestyle Study Research Group (2014) A blood-based predictor for neocortical Aβ burden in Alzheimer’s disease: Results from the AIBL study. Mol Psychiatry 19, 519–526.

35.

Lui

, Laws

, Li

Q-X

, Villemagne

, Ames

, Brown

, Bush

, De Ruyck

, Dromey

, Ellis

, Faux

, Foster

, Fowler

, Gupta

, Hudson

, Laughton

, Masters

, Pertile

, Rembach

, Rimajova

, Rodrigues

, Rowe

, Rumble

, Szoeke

, Taddei

, Trounson

, Ward

, Martins

, AIBL Research Group (2010) Plasma amyloid-beta as a biomarker in Alzheimer’s disease: The AIBL study of aging. J Alzheimers Dis 20, 1233–1242.

36.

Koch

, DeKosky

, Fitzpatrick

, Furtado

, Lopez

, Kuller

, Mackey

, Hughes

, Mukamal

, Jensen

(2018) Apolipoproteins and Alzheimer’s pathophysiology. Alzheimers Dement (Amst) 10, 545–553.

37.

Benedet

, Labbe

, Lemay

, Zimmer

, Pascoal

, Leuzy

, Mathotaarachchi

, Mohades

, Shin

, Dionne-Laporte

, Beaudry

, Picard

, Gauthier

, Poirier

, Rouleau

, Rosa-Neto

, Alzheimer’s Disease Neuroimaging Initiative (2015) Epistasis analysis links immune cascades and cerebral amyloidosis. J Neuroinflammation 12, 227.

38.

Mattsson

, Zetterberg

, Janelidze

, Insel

, Andreasson

, Stomrud

, Palmqvist

, Baker

, Tan Hehir

, Jeromin

, Hanlon

, Song

, Shaw

, Trojanowski

, Weiner

, Hansson

, Blennow

, ADNI Investigators (2016) Plasma tau in Alzheimer disease. Neurology 87, 1827–1835.

39.

Thambisetty

, Simmons

, Velayudhan

, Hye

, Campbell

, Zhang

, Wahlund

L-O

, Westman

, Kinsey

, Güntert

, Proitsi

, Powell

, Causevic

, Killick

, Lunnon

, Lynham

, Broadstock

, Choudhry

, Howlett

, Williams

, Sharp

, Mitchelmore

, Tunnard

, Leung

, Foy

, O’Brien

, Breen

, Furney

, Ward

, Kloszewska

, Mecocci

, Soininen

, Tsolaki

, Vellas

, Hodges

, Murphy

DGM

, Parkins

, Richardson

, Resnick

, Ferrucci

, Wong

, Zhou

, Muehlboeck

, Evans

, Francis

, Spenger

, Lovestone

(2010) Association of plasma clusterin concentration with severity, pathology, and progression in Alzheimer disease. Arch Gen Psychiatry 67, 739–748.

40.

Schindler

, Bollinger

, Ovod

, Mawuenyega

, Li

, Gordon

, Holtzman

, Morris

, Benzinger

TLS

, Xiong

, Fagan

, Bateman

(2019) High-precision plasma β-amyloid 42/40 predicts current and future brain amyloidosis. Neurology 93, e1647–e1659.

41.

Lim

, Maruff

, Kaneko

, Doecke

, Fowler

, Villemagne

, Kato

, Rowe

, Arahata

, Iwamoto

, Ito

, Tanaka

, Yanagisawa

, Masters

, Nakamura

(2020) Plasma amyloid-β biomarker associated with cognitive decline in preclinical Alzheimer’s disease. J Alzheimers Dis 77, 1057–1065.

42.

Tatebe

, Kasai

, Ohmichi

, Kishi

, Kakeya

, Waragai

, Kondo

, Allsop

, Tokuda

(2017) Quantification of plasma phosphorylated tau to use as a biomarker for brain Alzheimer pathology: Pilot case-control studies including patients with Alzheimer’s disease and down syndrome. Mol Neurodegener 12, 63.

43.

Westwood

, Baird

, Hye

, Ashton

, Nevado-Holgado

, Anand

, Liu

, Newby

, Bazenet

, Kiddle

, Ward

, Newton

, Desai

, Tan Hehir

, Zanette

, Galimberti

, Parnetti

, Lleó

, Baker

, Narayan

, van der Flier

, Scheltens

, Teunissen

, Visser

, Lovestone

(2018) Plasma protein biomarkers for the prediction of CSF amyloid and tau and [18F]-Flutemetamol PET scan result. Front Aging Neurosci 10, 409.

44.

Rembach

, Faux

, Watt

, Pertile

, Rumble

, Trounson

, Fowler

, Roberts

, Perez

, Li

Q-X

, Laws

, Taddei

, Rainey-Smith

, Robertson

, Vandijck

, Vanderstichele

, Barnham

, Ellis

, Szoeke

, Macaulay

, Rowe

, Villemagne

, Ames

, Martins

, Bush

, Masters

, AIBL research group (2014) Changes in plasma amyloid beta in a longitudinal study of aging and Alzheimer’s disease. Alzheimers Dement 10, 53–61.

45.

Rembach

, Watt

, Wilson

, Villemagne

, Burnham

, Ellis

, Maruff

, Ames

, Rowe

, Macaulay

, Bush

, Martins

, Masters

, Doecke

, AIBL Research Group (2014) Plasma amyloid-β levels are significantly associated with a transition toward Alzheimer’s disease as measured by cognitive decline and change in neocortical amyloid burden. J Alzheimers Dis 40, 95–104.

46.

Vergallo

, Mégret

, Lista

, Cavedo

, Zetterberg

, Blennow

, Vanmechelen

, De Vos

, Habert

M-O

, Potier

M-C

, Dubois

, Neri

, Hampel

, INSIGHT-preAD study group, Alzheimer Precision Medicine Initiative (APMI) (2019) Plasma amyloid β 40/42 ratio predicts cerebral amyloidosis in cognitively normal individuals at risk for Alzheimer’s disease. Alzheimers Dement 15, 764–775.

47.

Karikari

, Pascoal

, Ashton

, Janelidze

, Benedet

, Rodriguez

, Chamoun

, Savard

, Kang

, Therriault

, Schöll

, Massarweh

, Soucy

J-P

, Höglund

, Brinkmalm

, Mattsson

, Palmqvist

, Gauthier

, Stomrud

, Zetterberg

, Hansson

, Rosa-Neto

, Blennow

(2020) Blood phosphorylated tau 181 as a biomarker for Alzheimer’s disease: A diagnostic performance and prediction modelling study using data from four prospective cohorts. Lancet Neurol 19, 422–433.

48.

Fandos

, Pérez-Grijalba

, Pesini

, Olmos

, Bossa

, Villemagne

, Doecke

, Fowler

, Masters

, Sarasa

, AIBL Research Group (2017) Plasma amyloid β 42/40 ratios as biomarkers for amyloid β cerebral deposition in cognitively normal individuals. Alzheimers Dement (Amst) 8, 179–187.

49.

Verberk

IMW

, Slot

, Verfaillie

SCJ

, Heijst

, Prins

, van Berckel

BNM

, Scheltens

, Teunissen

, van der Flier

(2018) Plasma amyloid as prescreener for the earliest Alzheimer pathological changes. Ann Neurol 84, 648–658.

50.

Karikari

, Benedet

, Ashton

, Lantero Rodriguez

, Snellman

, Suárez-Calvet

, Saha-Chaudhuri

, Lussier

, Kvartsberg

, Rial

, Pascoal

, Andreasson

, Schöll

, Weiner

, Rosa-Neto

, Trojanowski

, Shaw

, Blennow

, Zetterberg

, Alzheimer’s Disease Neuroimaging Initiative (2021) Diagnostic performance and prediction of clinical progression of plasma phospho-tau181 in the Alzheimer’s Disease Neuroimaging Initiative. Mol Psychiatry 26, 429–442.

51.

de Wolf

, Ghanbari

, Licher

, McRae-McKee

, Gras

, Weverling

, Wermeling

, Sedaghat

, Ikram

, Waziry

, Koudstaal

, Klap

, Kostense

, Hofman

, Anderson

, Goudsmit

, Ikram

(2020) Plasma tau, neurofilament light chain and amyloid-β levels and risk of dementia; a population-based cohort study. Brain 143, 1220–1232.

52.

Moscoso

, Grothe

, Ashton

, Karikari

, Rodriguez

, Snellman

, Suárez-Calvet

, Zetterberg

, Blennow

, Schöll

, for the Alzheimer’s Disease Neuroimaging Initiative (2021) Time course of phosphorylated-tau181 in blood across the Alzheimer’s disease spectrum. Brain 144, 325–339.

53.

Hansson

, Cullen

, Zetterberg

, the Alzheimer’s Disease Neuroimaging Initiative, Blennow

, Mattsson-Carlgren

(2021) Plasma phosphorylated tau181 and neurodegeneration in Alzheimer’s disease. Ann Clin Transl Neurol 8, 259–265.

54.

Andersson

, Janelidze

, Lampinen

, Nilsson

, Leuzy

, Stomrud

, Blennow

, Zetterberg

, Hansson

(2020) Blood and cerebrospinal fluid neurofilament light differentially detect neurodegeneration in early Alzheimer’s disease. Neurobiol Aging 95, 143–153.

55.

Mattsson

, Cullen

, Andreasson

, Zetterberg

, Blennow

(2019) Association between longitudinal plasma neurofilament light and neurodegeneration in patients with Alzheimer disease. JAMA Neurol 76, 791–799.

56.

Rauchmann

, Schneider-Axmann

, Perneczky

, Alzheimer’s Disease Neuroimaging Initiative (ADNI) (2021) Associations of longitudinal 1047 plasma p-tau181 and NfL with tau-PET, Aβ-PET and cognition. J Neurol Neurosurg Psychiatry, doi: 10.1136/jnnp-2020-325537

57.

Moscoso

, Grothe

, Ashton

, Karikari

, Lantero Rodríguez

, Snellman

, Suárez-Calvet

, Blennow

, Zetterberg

, Schöll

, Alzheimer’s Disease Neuroimaging Initiative (2021) Longitudinal associations of blood phosphorylated Tau181 and neurofilament light chain with neurodegeneration in Alzheimer disease. JAMA Neurol 78, 396–406.

58.

, Chen

K-L

, Ou

Y-N

, Cao

X-P

, Chen

S-D

, Cui

, Dong

, Tan

, Yu

J-T

, Alzheimer’s Disease Neuroimaging Initiative (2019) Neurofilament light chain plasma concentration predicts neurodegeneration and clinical progression in nondemented elderly adults. Aging 11, 6904–6914.

59.

Chiu

M-J

, Chen

Y-F

, Chen

T-F

, Yang

S-Y

, Yang

F-PG

, Tseng

T-W

, Chieh

J-J

, Chen

J-CR

, Tzen

K-Y

, Hua

M-S

, Horng

H-E

(2014) Plasma tau as a window to the brain-negative associations with brain volume and memory function in mild cognitive impairment and early Alzheimer’s disease. Hum Brain Mapp 35, 3132–3142.

60.

Deters

, Risacher

, Kim

, Nho

, West

, Blennow

, Zetterberg

, Shaw

, Trojanowski

, Weiner

, Saykin

, Alzheimer Disease Neuroimaging Initiative (2017) Plasma tau association with brain atrophy in mild cognitive impairment and Alzheimer’s disease. J Alzheimers Dis 58, 1245–1254.

61.

Cantero

, Atienza

, Ramos-Cejudo

, Fossati

, Wisniewski

, Osorio

(2020) Plasma tau predicts cerebral vulnerability in aging. Aging 12, 21004–21022.

62.

Mielke

, Hagen

, Xu

, Chai

, Vemuri

, Lowe

, Airey

, Knopman

, Roberts

, Machulda

, Jack

, Petersen

, Dage

(2018) Plasma phospho-tau181 increases with Alzheimer’s disease clinical severity and is associated with tau- and amyloid-positron emission tomography. Alzheimers Dement 14, 989–997.

63.

Park

J-C

, Han

S-H

, Yi

, Byun

, Lee

, Jang

, Ko

, Jeon

, Lee

Y-S

, Kim

, Lee

, Mook-Jung

(2019) Plasma tau/amyloid-β1-42 ratio predicts brain tau deposition and neurodegeneration in Alzheimer’s disease. Brain 142, 771–786.

64.

Illán-Gala

, Lleo

, Karydas

, Staffaroni

, Zetterberg

, Sivasankaran

, Grinberg

, Spina

, Kramer

, Ramos

, Coppola

, La Joie

, Rabinovici

, Perry

, Gorno-Tempini

, Seeley

, Miller

, Rosen

, Blennow

, Boxer

, Rojas

(2021) Plasma tau and neurofilament light in frontotemporal lobar degeneration and Alzheimer disease. Neurology 96, e671–e683.

65.

Landau

, Thomas

, Thurfjell

, Schmidt

, Margolin

, Mintun

, Pontecorvo

, Baker

, Jagust

; Alzheimer’s Disease Neuroimaging Initiative, (2014) Amyloid PET imaging in Alzheimer’s disease: A comparison of three radiotracers. Eur J Nucl Med Mol Imaging 41, 1398–1407.

66.

Villemagne

, Mulligan

, Pejoska

, Ong

, Jones

, O’Keefe

, Chan

, Young

, Tochon-Danguy

, Masters

, Rowe

(2012) comparison of 11C-PiB and 18F-florbetaben for Aβ imaging in ageing and Alzheimer’s disease. Eur J Nucl Med Mol Imaging 39, 983–989.

67.

Vermunt

, van Paasen

AJL

, Teunissen

, Scheltens

, Visser

, Tijms

, Alzheimer’s Disease Neuroimaging Initiative (2019) Alzheimer disease biomarkers may aid in the prognosis of MCI cases initially reverted to normal. Neurology 92, e2699–e2705.

68.

Coppola

, Cianflone

, Grimaldi

, Incoronato

, Bevilacqua

, Messina

, Baselice

, Soricelli

, Mirabelli

, Salvatore

(2019) Biobanking in health care: Evolution and future directions. J Transl Med 17, 172.