Relationships between PET and blood plasma biomarkers in corticobasal syndrome

Abstract

INTRODUCTION

Corticobasal syndrome (CBS) can result from underlying Alzheimer's disease (AD) pathologies. Little is known about the utility of blood plasma metrics to predict positron emission tomography (PET) biomarker‐confirmed AD in CBS.

METHODS

A cohort of eighteen CBS patients (8 amyloid beta [Aβ]+; 10 Aβ−) and 8 cognitively unimpaired (CU) individuals underwent PET imaging and plasma analysis. Plasma concentrations were compared using a Kruskal–Wallis test. Spearman correlations assessed relationships between plasma concentrations and PET uptake.

RESULTS

CBS Aβ+ group showed a reduced Aβ42/40 ratio, with elevated phosphorylated tau (p‐tau)181, glial fibrillary acidic protein (GFAP), and neurofilament light (NfL) concentrations, while CBS Aβ− group only showed elevated NfL concentration compared to CU. Both p‐tau181 and GFAP were able to differentiate CBS Aβ− from CBS Aβ+ and showed positive associations with Aβ and tau PET uptake.

DISCUSSION

This study supports use of plasma p‐tau181 and GFAP to detect AD in CBS. NfL shows potential as a non‐specific disease biomarker of CBS regardless of underlying pathology.

Highlights

• Plasma phosphorylated tau (p‐tau)181 and glial fibrillary acidic protein (GFAP) concentrations differentiate corticobasal syndrome (CBS) amyloid beta (Aβ)− from CBS Aβ+.

• Plasma neurofilament light concentrations are elevated in CBS Aβ− and Aβ+ compared to controls.

• Plasma p‐tau181 and GFAP concentrations were associated with Aβ and tau positron emission tomography (PET) uptake.

• Aβ42/40 ratio showed a negative correlation with Aβ PET uptake.

1. BACKGROUND

Corticobasal syndrome (CBS) is a progressive neurodegenerative disease within the umbrella of atypical parkinsonian disorders. ¹ CBS patients typically present with motor deficits in the form of dystonia, myoclonus, and asymmetric rigidity, along with higher cognitive impairments, including ideomotor apraxia, cortical sensory loss, and behavioral and cognitive changes. ¹ , ² , ³ Heterogeneity in the underlying pathology of CBS has been increasingly recognized, with corticobasal degeneration pathology (CBD) accounting for less than half of CBS cases and Alzheimer's disease (AD) and progressive supranuclear palsy (PSP) being the most common non‐CBD pathology. In fact, AD accounts for ≈ 30% of CBS cases. ⁴ , ⁵ , ⁶ , ⁷ , ⁸ , ⁹

Positron emission tomography (PET) ligands that can detect amyloid beta (Aβ) plaques in vivo have been used to identify CBS cases that may have underlying AD pathology. Studies using Aβ‐PET have found that 25% to 44% of CBS patients have evidence of Aβ deposition on PET (i.e., Aβ+). ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ Many, although not all, of these patients also show AD levels of tau uptake using PET ligands such as 18F‐flortaucipir. Patients that are both Aβ and tau positive on PET are considered to have biomarker‐confirmed AD. ¹⁹ Likewise, cerebrospinal fluid (CSF) markers for Aβ and phosphorylated tau (p‐tau) ¹⁹ have also been used to detect AD ²⁰ , ²¹ and have been proposed for use in refining a CBS diagnosis. ¹ However, the invasiveness of lumbar punctures for CSF and the high cost and need for a complex infrastructure associated with PET imaging have limited their use to specialized clinics only. The ability to diagnose AD in CBS patients during life is critically important for patient management and treatment, given the availability of anti‐Aβ pharmaceuticals. It will also be important for identifying AD patients for inclusion in future treatment trials.

Recent AD research has focused on blood plasma biomarkers, which provide an inexpensive and clinically accessible alternative to CSF and PET imaging, with the most widely studied biomarkers being Aβ42/40 ratio; ²² , ²³ , ²⁴ p‐tau181; ²⁵ , ²⁶ , ²⁷ and neurodegeneration markers such as total tau, ²⁸ glial fibrillary acidic protein (GFAP), ²⁹ , ³⁰ and neurofilament light (NfL). ³¹ , ³² Several studies have reported a significant reduction in the Aβ42/40 ratio, along with an increase in p‐tau181, total tau, GFAP, and NfL concentrations in AD dementia cohorts compared to controls. ²⁹ , ³² , ³³ , ³⁴ , ³⁵ , ³⁶ , ³⁷ , ³⁸ A combination of p‐tau181 and the Aβ42/40 ratio and a combination of NfL and the Aβ42/40 ratio have shown potential in monitoring disease progression and early diagnosis. ³³ , ³⁹ Furthermore, in AD dementia cohorts, plasma Aβ and p‐tau concentrations, in particular, relate well with their corresponding CSF, ²³ , ⁴⁰ PET ²² , ²³ , ⁴¹ metrics, and post mortem ⁴² , ⁴³ measures, while GFAP and NfL correlate well to their corresponding CSF measures. ³⁰ , ⁴⁴ Therefore, these plasma biomarkers may prove useful in identifying PET biomarker–confirmed AD in CBS. A few studies have reported abnormal levels of GFAP and NfL in plasma or CSF in CBS, ⁴⁵ , ⁴⁶ , ⁴⁷ , ⁴⁸ , ⁴⁹ and one study has shown that a plasma p‐tau measure can identify Aβ positivity within CBS. ⁵⁰ However, it is unknown whether GFAP, NfL, or Aβ42/40 differ according to Aβ status in CBS.

The primary aim of this study was to determine whether blood plasma biomarkers differed between CBS Aβ−, CBS Aβ+, and cognitively unimpaired (CU) individuals. We hypothesized that plasma Aβ, p‐tau181, GFAP, and NfL levels would be significantly different between CBS Aβ− and CBS Aβ+, because both Aβ and p‐tau are primarily associated with amyloid deposition, GFAP is closely associated with amyloid pathology, and NfL is a non‐specific marker associated with amyloid‐related neuronal injury. ²³ , ²⁵ , ³⁰ , ⁵¹ Further, we hypothesized there would be a significant difference in all biomarkers between patients with CBS and CU individuals. The secondary aim of this study was to determine the relationships between the plasma biomarkers and both Aβ and tau deposition measured on PET. We hypothesized that plasma Aβ, p‐tau, and GFAP would be associated with Aβ PET because of their close relation to amyloid pathology, while only p‐tau would relate to tau PET. These findings will advance our knowledge of the clinical utility of plasma biomarkers for diagnosing AD in patients with CBS.

2. METHODS

2.1. Patients

Eighteen patients that fulfilled clinical diagnostic criteria for possible or probable CBS ¹ (8 CBS Aβ+ and 10 CBS Aβ−) were recruited by the Neurodegenerative Research Group (NRG) from the Department of Neurology, Mayo Clinic, Rochester, Minnesota, between April 10, 2019, and December 13, 2022. All patients were enrolled regardless of age and sex into the study and underwent extensive neurological evaluations by one of two behavioral neurologists (K.A.J. or J.G.R.). All patients underwent a blood draw, [¹¹C] Pittsburgh compound B (PiB) PET to assess Aβ, and [¹⁸F] flortaucipir PET to assess tau deposition. Per protocol, all clinical diagnoses were rendered prior to, and hence blinded to, imaging results. A cohort of 8 CU Aβ− individuals were recruited by the NRG from the Department of Neurology, Mayo Clinic, Rochester, Minnesota, between November 21, 2022, and March 27, 2023, and underwent the identical protocols.

2.2. Blood plasma biomarker testing

All patients underwent a blood draw, the blood was centrifuged and aliquoted and then stored in the Biospecimens Accessioning and Processing (BAP) laboratory at −80°C. Frozen plasma samples were sent to the Mayo Clinic, Florida, for analysis. Measurements of total tau, p‐tau181, NfL, GFAP, Aβ1–40 and Aβ1–42 were performed using the Simoa HD‐X analyzer (Quanterix), as previously published. ²⁸ , ⁵² NfL, GFAP, Aβ1–40, and Aβ1–42 levels were analyzed using the Neurology 4‐plex E, while total‐tau and p‐tau181 levels were analyzed using the Tau 2.0 Advantage kit and pTau Advantage V2 kit. On the day of the experiment, the Simoa Calibrators, Simoa controls, and plasma samples were allowed to equilibrate to room temperature. Resorufin β‐D‐galactopyranoside (RGP) vials were placed in a heated shaker set at 30°C and agitated at 79 x g for 30 minutes. Subsequently, the Simoa calibrators and controls were dispensed into 96‐well plates as follows: 230, 182, and 230 µL of each calibrator for the Neurology 4‐Plex E, Tau 2.0, and pTau‐181 kits, respectively. For the controls, 100, 120, and 100 µL were loaded for the Neurology 4‐Plex E, Tau 2.0 v2, and pTau‐181 kits as well. Before loading, plasma samples were centrifuged at 10,000 x g for 5 minutes, after which 225 µL of each sample was loaded onto the plate. The plate was sealed using X‐Pierce XP‐100 to prevent any potential sample evaporation, and then set aside. The bead reagents were vortexed for 30 seconds and then loaded into the Simoa HD‐X analyzer alongside the plasma sample diluent, detector reagent, and Streptavidin‐β‐galactosidase (SBG) reagent. Meanwhile, the RGP vials were positioned in a separate RGP rack. The plate was then loaded into the Simoa system. All analyses were performed in duplicate, and the average value was used in the statistical analyses for each participant. We only included sample concentrations with coefficients of variance (CV) < 10%. The mean ± standard deviation CV% for all included samples was 0.05% ± 0.03%. Note no samples were excluded.

2.3. Clinical testing

The neurological evaluations performed on this cohort included the Montreal Cognitive Assessment (MoCA) for assessing general cognitive function, ⁵³ Movement Disorders Society (MDS)–sponsored revision of the Unified Parkinson's Disease Rating Scale III (UPDRS III) to assess parkinsonism, ⁵⁴ and the Western Aphasia Battery ideomotor apraxia (WAB praxis) subtest to assess for ideomotor apraxia. ⁵⁵

2.4. Image acquisition

All patients underwent PET scanning at Mayo Clinic, Rochester, Minnesota. All PET scans were acquired on a PET/computed tomography (CT) scanner. For Aβ PET, patients were injected with ≈ 628 MBq (range 385–723 MBq) of PiB followed by a 40 minute uptake period. For flortaucipir PET, patients were injected with ≈ 370 MBq (range 333–407 MBq) of [¹⁸F] flortaucipir, followed by an 80 minute uptake period. For both Aβ and flortaucipir PET, acquisition was 20 minutes consisting of four, 5 minute dynamic frames after a low‐dose CT image. Standard corrections were applied and the PET sinograms were reconstructed into a 256 mm field of view. Dynamic PET frames were coregistered and averaged. Detailed acquisition details have been previously published. ⁵⁶

RESEARCH IN CONTEXT

1. Systematic review: The authors reviewed literature from PubMed that assessed the utility of blood plasma biomarkers in corticobasal syndrome (CBS). There are a few reports that have showed abnormal blood plasma profiles for glial fibrillary acidic protein (GFAP), neurofilament light (NfL), and phosphorylated tau (p‐tau) measures in CBS. However, it is unknown whether GFAP, NfL, or amyloid beta (Aβ) 42/40 differ according to Aβ status in CBS. The most relevant articles have been cited appropriately.

2. Interpretation: This study contributes to our knowledge of the clinical utility of plasma biomarkers for diagnosing Alzheimer's disease (AD) in patients with CBS. Our findings demonstrate that plasma p‐tau181 and GFAP biomarkers may be capable of detecting positron emission tomography (PET) biomarker‐confirmed AD in CBS patients. Additionally, plasma NfL may be an appropriate biomarker of disease in CBS regardless of underlying pathology. Therefore, plasma concentrations may potentially be incorporated in clinical trials targeting CBS patients with underlying AD pathology.

3. Future directions: These findings warrant future studies that replicate these results in larger CBS Aβ+ and Aβ− cohorts. These findings also warrant replication of these results in longitudinal cohorts.

2.5. Image processing

The Aβ and flortaucipir PET images were registered to their corresponding subject‐space magnetization prepared rapid gradient echo (MPRAGE) T1‐weighted magnetic resonance imaging using SPM12. Regional PET values were calculated by using ANTs ⁵⁷ to propagate the Mayo Clinic Adult Lifespan Template (MCALT) ADIR122 atlas to native MPRAGE space. Unified segmentation ⁵⁸ in SPM12 was used to determine the tissue probabilities of each MPRAGE scan with MCALT tissue priors and settings. ⁵⁹ Median PiB and flortaucipir‐PET uptake were calculated for each patient across gray and white matter and were divided by the cerebellar crus gray matter median uptake value to generate standard uptake value ratios (SUVRs). For the Aβ‐PET, a global PiB SUVR was calculated as previously described. ⁶⁰ An SUVR cut‐point of 1.48 was used to define Aβ positivity, and the target cortical meta‐region of interest (ROI) used to define this positivity included the prefrontal, orbitofrontal, parietal, temporal, anterior and posterior cingulate, and the precuneus. A cut‐point of 1.25 was used to define flortaucipir‐PET positivity, and the target temporal‐lobe meta‐ROI used to define this positivity included the amygdala, entorhinal cortex, fusiform, parahippocampal, and inferior temporal and middle temporal gyri. ⁶¹ Additionally, flortaucipir‐PET uptake was calculated for the basal ganglia (which was a weighted average of the uptake in caudate, putamen, and pallidum) and the precentral gyrus, as these regions show evidence of greater flortaucipir retention in CBS patients. ¹⁷ Detailed image processing details have been previously published. ⁶⁰ , ⁶¹

2.6. Statistical analysis

Group comparisons for patient demographics and disease characteristics were performed using the Fisher exact test for categorical variables and Mann–Whitney test for continuous variables. Group comparisons for assessing differences in plasma levels across the three groups were performed using the Kruskal–Wallis test corrected for multiple comparisons using a Dunn test. Adjustments for age and sex were not performed as these were comparable across groups. This analysis was replicated by classifying CBS patients based on tau positivity.

Area under the receiver operator characteristic curve (AUROC) analysis was used to assess the utility of the plasma biomarkers to differentiate between the CBS Aβ− and CBS Aβ+ groups. The Youden index was also calculated to define maximum potential effectiveness of the plasma biomarkers. Last, Spearman correlations were calculated to evaluate the relationship between PET uptake and plasma biomarker concentrations across the CBS group. For tau‐PET uptake, these correlations were evaluated against three target regions: namely, the temporal lobe meta‐ROI to target AD‐related tau, and the basal ganglia and precentral gyrus to assess CBS‐related tau. Given the large number of correlations (n = 20) a P value of P < 0.0025 was considered significant. These correlations were also calculated within the CBS Aβ+ group only. Spearman correlations were also calculated to evaluate the relationship between plasma biomarkers and clinical features in the CBS group. Given the large number of correlations (n = 15) a P value of P < 0.0034 was considered significant. The Aβ42/40 ratio was untransformed while total tau, p‐tau 181, GFAP, and NfL plasma measures were log‐transformed to reduce skewness as measurement variation tends to increase with concentration. All analyses were performed using statistical software GraphPad Prism version 9.2.0.

3.

3.1. Patient characteristics

The demographic and clinical features of the cohorts are shown in Table 1. CBS Aβ+ and CBS Aβ− did not differ on sex, education, age of onset, age at exam, and time from onset to scan. On clinical testing, the two groups did not differ on the MDS‐UPDRS III, but CBS Aβ+ performed worse on the MoCA. Compared to CU, the CBS groups did not differ on sex, education, and age at exam, but performed worse on the MoCA, UPDRS III, and WAB praxis.

TABLE 1.

Participant demographics and disease characteristics.

	Disease cohort (N = 18)
	CBS Aβ+ (N = 8)	CBS Aβ− (N = 10)	P value CBS Aβ+ versus CBS Aβ−	Cognitively unimpaired (CU) (N = 8)	P value CBS Aβ+ versus CU	P value CBS Aβ− versus CU
Female, n (%)	4 (50%)	6(60%)	> 0.99	5 (63%)	> 0.99	> 0.99
Education, years	16 (15.5, 16.8)	14 (12.5, 16)	0.120	17 (15.5, 18)	0.721	0.087
Age at onset, years	59.5 (57, 69)	63 (59.8, 65.8)	0.778	–	–	–
Age at exam, years	63 (62.1, 73.3)	67 (62.6, 69.8)	0.965	64.6 (61.4, 67.3)	0.823	0.359
Time from onset to scan, years	3.8 (2.52, 5.17)	2.45 (1.85, 3.35)	0.314	–	–	–
Diagnosis certainty	Probable (14%) Possible (86%)	Probable (33%) Possible (67%)	–	–	–	–
MoCA (30)	20 (16.5, 22)	24 (21.8, 25)	0.019	27 (25.8, 29)	0.0003	0.023
UPDRS III (15)	23 (9, 24.5)	31.5 (15.3, 33)	0.226	1.5 (0, 3.25)	0.046	< 0.0001
WAB praxis (60)	47 (45, 53)	55 (53, 55)	0.133	60 (58.5, 60)	0.0002	< 0.0001
Aβ positivity, n (%)	100%	Absent	–	–	–	–
Tau positivity, n (%)	75%	10%	–	–	–	–
Plasma Aβ42/40 ratio	0.063 (0.058, 0.065)	0.068 (0.065, 0.080)	0.106	0.073 (0.071, 0.082)	0.020	>0.99
Plasma p‐tau181 (pg/mL)	2.855 (2.395, 3.578)	1.594 (1.379, 1.922)	0.021	1.052 (1.021, 1.233)	0.0004	0.537
Plasma total tau (pg/mL)	2.185 (1.689, 2.735)	1.795 (1.579, 2.331)	> 0.99	1.645 (1.317, 2.078)	0.350	> 0.99
Plasma GFAP (pg/mL)	142.33 (120.66, 171.93)	76.64 (66.36, 92.49)	0.027	59.13 (50.13, 67.96)	0.001	0.716
Plasma NfL (pg/mL)	28.03 (24.71, 31.46)	33.31 (21.47, 49.82)	> 0.99	9.43 (7.83, 11.98)	0.005	0.0004

Note: Data shown are n (%) or median (first and third quartiles). For continuous variables, P values are from Mann–Whitney test. For categorical variables, P values are from a Fisher exact test.

Abbreviations: Aβ, amyloid beta; CBS, corticobasal syndrome; GFAP, glial fibrillary acidic protein; MoCA, Montreal Cognitive Assessment; NfL, neurofilament light; p‐tau, phosphorylated tau; UPDRS III, Unified Parkinson's Disease Rating Scale III; WAB, Western Aphasia Battery.

3.2. Plasma biomarker differences across groups

CBS Aβ+ had a lower plasma Aβ42/40 ratio, and higher plasma levels for p‐tau181, GFAP, and NfL than CU, while CBS Aβ− differed from CU only on plasma NfL with greater concentrations. CBS Aβ+ had higher plasma p‐tau181 and GFAP concentrations compared to the CBS Aβ− group, with no differences in the total tau plasma concentration across the three groups (Table 1 and Figure 1). Two of the Aβ+ CBS patients were negative on tau PET (highlighted in Figure 1). These two patients showed a low Aβ42/40 ratio and elevated GFAP similar to the rest of the Aβ+ patients but did not show elevated p‐tau181 concentrations. Furthermore, when CBS patients were grouped based on tau positivity, we saw similar results for all biomarkers, apart from GFAP concentrations being unable to differentiate CBS tau+ and CBS tau– groups.

FIGURE 1.

Difference in blood plasma biomarker levels between CBS Aβ−, CBS Aβ+, and CU. Group comparisons assessing differences in plasma levels across the three groups were performed using a Kruskal–Wallis test corrected for multiple comparisons using a Dunn test. The blue dots within the CBS Aβ+ group represent the two patients that were negative on tau PET. The level of significance is indicated by asterisks (^* P < 0.05, ^** P < 0.01, and ^*** P < 0.001). Aβ, amyloid beta; CBS, corticobasal syndrome; CU, cognitively unimpaired; GFAP, glial fibrillary acidic protein; NfL, neurofilament light; PET, positron emission tomography; p‐tau, phosphorylated tau

Plasma p‐tau181 and GFAP concentrations showed very good differentiation of CBS Aβ+ and CBS Aβ− groups, with an AUROC of 0.95 (cutpoint = 0.3/sensitivity = 100%/specificity = 80%) for p‐tau181 and an AUROC of 0.91 (cutpoint = 2.1/sensitivity = 75%/specificity = 100%) for GFAP concentrations (Table 2). The plasma NfL concentration showed complete differentiation of both CBS Aβ+ (AUROC = 1.00, cutpoint = 1.2/sensitivity = 100%/specificity = 100%) and CBS Aβ− (AUROC = 1.00, cutpoint = 1.2/sensitivity = 100%/specificity = 100%) groups from CU, with plasma p‐tau181 (AUROC = 0.97, cutpoint = 0.3/sensitivity = 100%/specificity = 87.5%) and GFAP (AUROC = 0.97, cutpoint = 2.0/sensitivity = 87.5%/specificity = 100%) concentrations also showing very good differentiation of CBS Aβ+ from CU (Table 2).

TABLE 2.

AUROC scores for plasma biomarkers across all groups.

Group contrast	Aβ42/40 ratio	p‐tau181	Total tau	GFAP	NfL
CBS Aβ+ versus CBS Aβ−	0.81	0.95	0.63	0.91	0.61
CBS Aβ+ versus CU	0.88	0.97	0.70	0.97	1.00
CBS Aβ− versus CU	0.62	0.76	0.65	0.71	1.00

Note: Scores > 0.80 are considered good AUROC scores, while > 0.90 are considered very good AUROC scores (highlighted in gray). The Aβ42/40 ratio was untransformed while total tau, p‐tau181, GFAP, and NfL plasma measures were log‐transformed to reduce skewness as measurement variation tends to increase with concentration.

Abbreviations: Aβ, amyloid beta; AUROC, area under the receiver operator characteristic curve; CBS, corticobasal syndrome; CU, cognitively unimpaired; GFAP, glial fibrillary acidic protein; NfL, neurofilament light; p‐tau, phosphorylated tau.

3.3. Relationship to PET metrics and clinical features

Reduced plasma Aβ42/40 ratio (P = 0.02) and elevated p‐tau181 (P = 0.0002) and GFAP (P = 0.0002) concentrations were associated with increased Aβ PET uptake in the CBS group (Figure 2). However, the findings for the Aβ42/40 ratio did not survive correction for multiple comparisons.

FIGURE 2.

Spearman correlations between plasma biomarkers and PET uptake in CBS. These plots represent the relationship between plasma biomarker concentrations and PET SUVR in CBS. The CBS Aβ− and CBS Aβ+ patients have been highlighted in different colors. All significant values are highlighted in bold, with green color reflecting findings that survive correction and orange color reflecting findings that did not survive correction for multiple comparisons. Aβ, amyloid beta; CBS, corticobasal syndrome; GFAP, glial fibrillary acidic protein; NfL, neurofilament light; PET, positron emission tomography; p‐tau, phosphorylated tau; SUVR, standard uptake value ratio

For tau PET, increased temporal lobe and precentral gyrus tau uptake was associated with elevated p‐tau181 (temporal lobe P < 0.0001; precentral P = 0.0002) and GFAP (temporal lobe P = 0.003; precentral P = 0.03) concentrations in the CBS group, with precentral tau uptake also showing associations with elevated total tau concentrations (P = 0.01; Figure 2). However, the precentral tau findings for GFAP and total tau did not survive correction for multiple comparisons. No significant correlations were seen between the basal ganglia tau uptake and plasma biomarkers (Figure 2). Within the CBS Aβ+ group, elevated plasma p‐tau181 concentrations were associated with increased Aβ PET (P = 0.08) and temporal‐lobe tau (P = 0.03) PET uptake, while reduced Aβ42/40 ratio was associated with reduced basal ganglia tau (P = 0.04) PET uptake. However, none of these findings survive correction.

No significant correlations were seen between blood plasma biomarkers and the MoCA, UPDRS III, and WAB praxis scores in the CBS group.

4. DISCUSSION

In this study, we explored the utility of plasma biomarkers in CBS, particularly in identifying PET biomarker–confirmed AD in CBS. We found that p‐tau181 and GFAP were elevated in the CBS Aβ+ group with these metrics showing very good differentiation of CBS Aβ+ and CBS Aβ− groups, with an AUROC of 0.95 for p‐tau181 and an AUROC of 0.91 for GFAP concentrations. Plasma NfL was elevated in both groups and was the only biomarker differentiating CBS Aβ− from CU. Only plasma p‐tau181 and GFAP concentrations were positively correlated with both Aβ and temporal lobe tau PET uptake, with the Aβ42/40 ratio showing a negative correlation to Aβ PET uptake.

Our findings show that both p‐tau181 and GFAP provide very good biomarkers of Aβ positivity in CBS, with both elevated in CBS Aβ+ patients compared to CU and both able to distinguish CBS Aβ− from CBS Aβ+ with very good differentiation (AUROC > 0.90). This concurs with previous studies showing that both plasma metrics can differentiate Aβ+ from Aβ− negative individuals in populations consisting of CU and impaired individuals (i.e., AD spectrum). ²⁵ , ³⁰ A similar finding has been reported by one CBS study, in which p‐tau181 showed potential to differentiate CBS Aβ+ from CBS Aβ−. ⁶² Most of our CBS Aβ+ patients were also positive on tau PET, and hence these findings suggest that both p‐tau181 and GFAP may be useful in identifying PET biomarker–confirmed AD in CBS. Interestingly, however, two patients in the CBS Aβ+ group showed no evidence of tau deposition on PET (i.e., were tau negative), and these two patients showed high plasma GFAP concentrations but relatively normal p‐tau181 levels. This suggests that plasma GFAP concentrations may be associated with Aβ regardless of the degree of tau deposition, while p‐tau181 concentrations are associated predominantly with tau deposition rather than Aβ deposition in CBS patients. GFAP and p‐tau181 were both strongly associated with Aβ and temporal lobe tau‐PET meta‐ROIs in our cohort, although p‐tau181 showed a slightly stronger relationship to tau‐PET and GFAP showed a slightly stronger relationship to Aβ‐PET. This also explains why GFAP concentrations were unable to differentiate the CBS groups when classified based on tau positivity. Similar associations with p‐tau181 were noted within the CBS Aβ+ group as well, although they did not survive correction. Both p‐tau181 and GFAP levels were closely associated with Aβ deposition in previous AD spectrum studies, ²⁵ , ²⁶ , ²⁷ , ³⁰ , ⁶³ with plasma p‐tau181 also associated with tau PET uptake. ⁶³ , ⁶⁴ , ⁶⁵ Furthermore, the associations between GFAP concentrations and tau PET match well with the literature as tau PET also shows close associations with neurodegeneration markers, ⁶³ although it is often difficult to dissociate the influence of Aβ and tau in these cohorts. It is important to note that despite p‐tau181 being associated with both Aβ and tau deposition, it has a stronger relation to Aβ deposition in AD, ⁶³ while in CBS patients we found a stronger relation between p‐tau181 and tau deposition on PET. We also observed an association between both p‐tau181 and GFAP and tau uptake in the precentral gyrus. This suggests that these plasma biomarkers are associated with increased tau uptake in both AD‐related and cortical CBS‐related regions, likely reflecting the widespread cortical tau uptake observed in CBS patients with AD. ⁶⁶ No associations were observed between plasma biomarkers and tau uptake in the basal ganglia, which may be unsurprising because AD is not associated with greater involvement of basal ganglia in CBS. ¹⁷

The plasma Aβ42/40 ratio was also reduced (i.e., abnormal) in the CBS Aβ+ patients, although the differences from CU were not as pronounced as the differences observed with GFAP and p‐tau181. Additionally, the Aβ42/40 ratio did not differ between the Aβ+ and Aβ− patients. There are no studies investigating this association using plasma but a CSF study in CBS reported reduced CSF Aβ42/40 ratio in the CBS Aβ+ group, with a significant difference between the Aβ+ and Aβ− patients. ⁶⁷ The reason for this discrepancy is unclear but may be due to the small sample size of our CBS groups. Although it does not survive correction, reduced Aβ42/40 ratio in the CBS group was closely associated with high Aβ burden on Aβ PET, which matches well with the literature for impaired individuals (i.e., AD spectrum). ²³

In contrast to the other plasma metrics, NfL was abnormal in both CBS groups, regardless of Aβ status, and was the only plasma metric that was abnormal in the CBS Aβ− patients compared to CU. The NfL concentration was also not associated with either Aβ or tau‐PET uptake. These findings suggest that elevated plasma NfL may be a non‐specific feature in CBS regardless of underlying pathology. Previous studies have reported an increase in both plasma and CSF NfL concentrations ⁶⁸ , ⁶⁹ in CBS compared to controls with no difference compared to AD, ⁷⁰ although no previous study has investigated whether NfL was influenced by AD pathology in CBS.

Plasma total tau concentrations were comparable across the groups and were therefore not associated with CBS or affected by Aβ status. Further, they were associated with neither Aβ nor temporal‐tau tau‐PET uptake. Although an association was noted between precentral tau uptake and total tau, it did not survive correction. There are discrepancies in the literature regarding CSF total tau in CBS, with some studies reporting an increase in total tau concentrations in CBS ⁷⁰ , ⁷¹ compared to controls, and others reporting no difference between CBS and controls. ⁷² , ⁷³ One study also reported a significant difference in CSF total tau concentrations based on Aβ status in CBS. ⁶⁷ This disparity in the findings may have been caused by several factors, including the severity of the cohorts, difference in the type of assay used for assessment, and small sample size. Our findings are, however, in line with previous studies of AD spectrum patients showing p‐tau181 is more strongly associated with Aβ and tau PET measures compared to total tau. ⁴¹ Furthermore, total tau is generally regarded as a marker of neurodegeneration but is less related to neuroimaging and clinical outcomes than NfL. ⁵²

There are discrepancies in the literature regarding the relationship between plasma biomarkers and cognitive outcomes, with studies reporting a significant correlation between cognitive outcomes and plasma concentrations of GFAP, NfL, and p‐tau181 ⁷⁴ , ⁷⁵ and others reporting lower correlation with plasma NfL concentrations ⁷⁵ and no correlation with plasma GFAP and p‐tau181 concentrations in AD. ⁷⁶ In this study, we did not note any significant correlations suggesting that plasma biomarker levels may not track well with the degree of cognitive or motor impairment in CBS. However, the lack of associations could have been due to the small sample size of the CBS cohort, and we may therefore need larger effect sizes for these associations to be detectable.

Strengths of this study include the consistent neuroimaging protocols and that we performed both Aβ and tau PET, which allowed us to examine how underlying AD pathology influences plasma biomarkers. Limitations of this study include the relatively small sample size, although the sample size provided adequate power to detect strong relationships with p‐tau181 and GFAP. Longitudinal data were also not available to assess how these plasma biomarker levels progress over time and if they could predict disease progression in CBS. Other limitations include the Quanterix Aβ42/40 assay being unable to differentiate between CBS Aβ+ and CBS Aβ− groups. However, use of other more sensitive Aβ42/40 plasma assays may be able to separate the two groups. In this study, NfL was considered a good non‐specific marker in CBS compared to controls, but it important to note that it may not be helpful in differentiating CBS from other neurodegenerative diseases which also show elevated NfL. ⁷⁶ Similarly, it is also important to note that GFAP is a marker for astrocytic reactivity that has also been associated with amyloid plaques and neuronal damage ²⁹ but this study does not measure neuropathological measures of astrocytic reactivity, preventing us from investigating this important issue. Lack of neuropathological confirmation in this cohort is a general limitation of this study. However, inclusion of other p‐tau markers such as p‐tau217 and p‐tau231 may offer a different insight as they are well known for their utility in preclinical AD stages. ⁷⁷ Use of spectroscopy‐based methods may further improve the performance of discrimination for these plasma biomarkers.

5. CONCLUSION

In summary, CBS is associated with several underlying neuropathologies, which creates challenges for patient diagnosis and prognosis, and uncertainty over appropriate therapies or treatments that can be used in these patients. This study supports the use of p‐tau181 and GFAP biomarkers to detect PET biomarker–confirmed AD in CBS patients and suggests that plasma NfL may be an appropriate biomarker of disease in CBS regardless of underlying pathology. The use of plasma markers would greatly reduce the costs usually associated with CSF and PET measures while still offering a promising diagnostic marker and has the potential to be incorporated in clinical trials for identifying PET biomarker–confirmed AD in CBS.

CONFLICT OF INTEREST STATEMENT

Dr. Singh and Alla Alnobani have no disclosures to report. Dr. Whitwell, Dr. Machulda, Dr. Schwarz, and Dr. Josephs reports receiving research funding from the NIH. Dr. Graff‐Radford reports receiving research support from the NIH and DSMB for StrokeNET. He is an investigator in a trial sponsored by USC and EISAI. Dr. Kanekiyo reports receiving research funding from the NIH and Cure Alzheimer's Fund. Matthew Senjem reports holding stock in Gilead Sciences, Inc., Inovio Pharmaceuticals, Medtronic, Oncothyreon, Inc., and PAREXEL International. Dr. Jack reports serving on an independent data monitoring board for Roche, has consulted for and served as a speaker for Eisai, and consulted for Biogen, but he receives no personal compensation from any commercial entity. He receives research support from NIH, the GHR foundation, and the Alexander Family Alzheimer's Disease Research Professorship of the Mayo Clinic. Dr. Lowe reports consulting for Bayer Schering Pharma, Piramal Life Sciences, Life Molecular Imaging, Eisai Inc., AVID Radiopharmaceuticals, and Merck Research and receiving research support from GE Healthcare, Siemens Molecular Imaging, AVID Radiopharmaceuticals, and the NIH (NIA, NCI). Dr. Mielke reports consulting for Biogen, Eisai, Lilly, Merck, Roche, and Siemens Healthineers and receives research support from the NIH, Department of Defense, and Alzheimer's Association. Author disclosures are available in the supporting information.

CONSENT STATEMENT

The study was approved by the Mayo Clinic Institutional Review Board. All patients gave written informed consent to participate in this study.

Supporting information

Supporting Information

Singh NA, Alnobani A, Graff‐Radford J, et al. Relationships between PET and blood plasma biomarkers in corticobasal syndrome. Alzheimer's Dement. 2024;20:4765–4774. 10.1002/alz.13914

DATA AVAILABILITY STATEMENT

The data that supports the findings of this study will be available from the corresponding author on request.