Biomarkers stewardship in parkinsonism: integrating alpha-synuclein seed amplification assays and neurofilament light chain into diagnostic pathways and patient communication Authors: Vincent Leclercq Published: 2026 ABSTRACT Parkinsonism encompasses Parkinson’s disease (PD) and atypical syndromes including dementia with Lewy bodies (DLB), multiple system atrophy (MSA), progressive supranuclear palsy (PSP), and corticobasal syndrome (CBS), where early clinical differentiation remains challenging. This narrative review synthesizes evidence for two complementary biomarker families: alpha-synuclein seed amplification assays (α-synuclein SAA; including real-time quaking-induced conversion and protein misfolding cyclic amplification) as pathology-proximal markers of α-synuclein seeding and neurofilament light chain (NfL) as a marker of neuroaxonal injury and neurodegeneration burden. CSF α-synuclein SAA demonstrates high diagnostic accuracy for PD and DLB, but its performance is context-dependent, with reduced sensitivity in normosmic PD, some genetic subgroups (especially LRRK2-associated PD), very early or anatomically restricted Lewy pathology, and assay protocols not optimized for MSA strains. Recent biological classification frameworks, including SynNeurGe and neuronal α-synuclein disease integrated staging system (NSD-ISS), place SAA centrally in research definitions; however, clinical implementation requires a stewardship layer that preserves clinical judgment, recognizes S-negative genetic or atypical presentations, and avoids equating SAA positivity with a single clinical diagnosis. NfL reliably differentiates PD/DLB from atypical parkinsonian disorders, but interpretation requires age-, matrix-, and platform-specific cutoffs and attention to confounders, such as renal function, body mass index, diabetes, peripheral neuropathy, and vascular comorbidity. This review proposes pragmatic algorithms integrating pre-test probability, α-synuclein SAA, NfL, imaging, genetics, and patient communication. It emphasizes on how to explain positive, negative, and prodromal biomarker results without overdiagnosis, while acknowledging mixed pathologies and uncertain prognostic implications. Introduction Parkinsonism comprises a diverse group of neurodegenerative disorders, including Parkinson’s disease (PD), dementia with Lewy bodies (DLB), multiple system atrophy (MSA), progressive supranuclear palsy (PSP), and corticobasal syndrome (CBS) [1–3]. These conditions share overlapping motor phenotypes yet differ significantly in their underlying pathology, prognosis, and treatment strategies [1, 4–6]. In the initial stages of the disease and in atypical presentations, pure clinical diagnosis remains challenging, and diagnostic uncertainty may persist for years [1, 6–8]. This can delay appropriate counseling, access to targeted symptomatic interventions, and enrollment into mechanism-based clinical trials [7, 9]. As approaches targeting disease modification and precision medicine advance rapidly, there is an increasing necessity to transition from syndromic labels to biologically grounded diagnostic pathways that can accurately identify the predominant molecular pathology and the pattern of neurodegeneration [10–12]. Two biomarker families have emerged as particularly promising and complementary for this purpose: alpha-synuclein seed amplification assays (α-synuclein SAA) and neurofilament light chain (NfL) [13–15]. α-synuclein SAA techniques (including RT-QuIC- and PMCA-derived platforms) aim to detect misfolded, self-propagating α-synuclein species, thereby providing pathology-proximal evidence of synucleinopathy [7, 10, 13, 16, 17]. Across multiple cohorts, CSF α-synuclein SAA demonstrates high diagnostic performance for PD versus healthy controls, with sensitivities and specificities commonly in the high-80 s to mid-90 s range [7, 10, 15, 16]. Importantly, SAA positivity also extends to prodromal synucleinopathies (e.g., REM sleep behavior disorder with hyposmia), supporting the concept that molecular pathology may be detectable before classical motor criteria are met [10, 16, 18]. However, assay performance is not uniform across all clinical and biological subgroups: sensitivity appears influenced by phenotype (notably olfactory function), genetic background, and disease stage, with reduced detection in very early pathological stages and in certain biological subtypes [10, 16, 19]. These limitations highlight that α-synuclein SAA is best viewed not as a standalone “yes/no PD test”, but as a molecular anchor whose interpretation must be contextualized within clinical probability and disease biology [10, 13, 14, 20]. In parallel, NfL, measured in CSF or blood, reflects neuroaxonal injury and disease intensity, providing a downstream marker of neurodegeneration that is particularly useful for differential diagnosis [21–23]. In clinical practice, markedly elevated NfL values tend to support atypical parkinsonism (such as MSA or PSP) over PD, and studies consistently show strong discrimination between PD and atypical parkinsonian disorders when NfL is incorporated into diagnostic models [21–23]. When combined with α-synuclein SAA, multi-marker panels leverage orthogonal biological information, molecular pathology (α-synuclein seeding) and neurodegeneration burden (NfL), to improve separation among PD, MSA, PSP and related phenotypes [14, 15, 19, 24]. Beyond CSF-based testing, less-invasive approaches (e.g., skin biopsy SAA, or combined peripheral matrices, such as serum and saliva) are increasingly explored, with the goal of expanding access to biology-based diagnosis while minimizing reliance on lumbar puncture [10, 17, 25–27]. A critical next step is translating these advances into pragmatic diagnostic pathways that maximize yield, minimize invasiveness, and remain implementable across real-world care settings. Sequential strategies illustrate this principle: an initial inexpensive screen that enriches pre-test probability (for example, smell testing with age/sex adjustment) can triage which patients truly benefit from confirmatory CSF α-synuclein SAA, reducing unnecessary lumbar punctures while maintaining high accuracy. In specialist clinics, biomarker results may also support diagnostic revision, reduce uncertainty in atypical cases, and inform downstream testing (e.g., imaging or genetic assessment). Nevertheless, implementation barriers remain substantial, including inter-laboratory variability, protocol heterogeneity, and the need for standardized quality control and proficiency frameworks before widespread adoption. In this article, we synthesize current evidence on α-synuclein SAA and NfL in parkinsonism and focus specifically on how these biomarkers can be integrated into clinical diagnostic workflows. We discuss where each biomarker adds the most value (screening, confirmation of synucleinopathy, differential diagnosis, and early/prodromal detection), summarize performance across key disease subtypes and stages, and propose practical, stepwise algorithms that align biomarker selection with clinical question and pre-test probability. Finally, we address real-world considerations, standardization, accessibility, and interpretation in mixed or co-pathology cases, to support responsible translation into routine care and clinical trials. Biological classification frameworks: SynNeurGe, NSD-ISS, and clinical stewardship The rapid rise of α-synuclein SAA has coincided with two influential biological frameworks for PD and DLB. The SynNeurGe research diagnostic criteria classify PD according to three biological axes: evidence of α-synuclein pathology (S), neurodegeneration or dopaminergic dysfunction (N) and genetic status (G) [12]. In parallel, the neuronal α-synuclein disease integrated staging system (NSD-ISS) defines neuronal α-synuclein disease from biological anchors, beginning with pathological neuronal α-synuclein detected in vivo and integrating dopaminergic dysfunction and clinical functional stages [28]. Both frameworks recognize that molecular pathology can precede classical clinical syndromes and both are intended primarily to support research stratification and therapeutic development rather than to replace bedside diagnosis [12, 28, 29]. The algorithms proposed in this review should therefore be read as clinical stewardship algorithms, not as alternative biological diagnostic criteria. They deliberately diverge from a purely biological definition in four ways. First, α-synuclein SAA positivity is interpreted as evidence of seeding-component α-synuclein, not as an automatic diagnosis of PD. Second, a negative SAA result does not exclude clinically meaningful parkinsonism, particularly in normosmic or LRRK2-associated PD, early or anatomically restricted Lewy pathology and assays not optimized for MSA-type strains. Third, SAA positivity in Alzheimer’s disease, PSP/CBS, idiopathic normal pressure hydrocephalus or after traumatic brain injury may reflect co-pathology, incidental pathology or a modifier of phenotype rather than the dominant disease process. Fourth, patient communication must preserve uncertainty: biomarkers can increase or decrease diagnostic probability, but clinical syndrome, trajectory, imaging, genetics, and patient goals remain essential to interpretation. Diagnostic performance of α-synuclein seed amplification assays CSF-based α-synuclein SAA CSF α-synuclein SAA demonstrated consistently high diagnostic performance across multiple large-scale studies. The Parkinson's Progression Markers Initiative (PPMI) cohort, comprising 1123 participants, reported 87.7% sensitivity (95% CI 84.9–90.5%) and 96.3% specificity (95% CI 93.4–99.2%) for distinguishing PD from healthy controls [16]. These findings were corroborated in a meta-analysis of 4925 PD patients showing a pooled sensitivity of 0.91 (95% CI 0.89–0.92) and a specificity of 0.95 (95% CI 0.94–0.96) [15]. Performance varied substantially across clinical phenotypes. Sporadic PD patients with typical olfactory deficits achieved 98.6% sensitivity (95% CI 96.4–99.4%), while those without olfactory dysfunction showed reduced sensitivity of 78.3% (95% CI 69.8–86.7%) [16]. LRRK2-associated PD demonstrated markedly lower SAA positivity at 67.5% (95% CI 59.2–75.8%), decreasing further to 34.7% (95% CI 21.4–48.0%) in LRRK2 carriers with normal olfaction [16]. This molecular heterogeneity suggests distinct pathophysiological mechanisms or different α-synuclein strains in genetic versus sporadic PD. Sensitivity also varied by pathological distribution of Lewy bodies. Neuropathological validation studies demonstrated 97.8% sensitivity and 98.1% specificity for detecting limbic/neocortical pathology, but only 14.3% sensitivity for amygdala-predominant Lewy body disease [30]. This pattern was confirmed in another autopsy series showing 100% sensitivity in limbic and neocortical stages but 37.5% in Braak stages 1–2 and 73.3% in Braak stage 3 [31]. The correlation between number of positive SAA replicates and disease stage suggests SAA detects a threshold level of α-synuclein seeds that accumulates as pathology progresses beyond restricted brainstem regions. Performance in dementia with Lewy bodies CSF α-syn SAA showed excellent performance in DLB cohorts. In a Spanish memory clinic study of 640 cognitively impaired patients, SAA exhibited 95.7% sensitivity (95% CI 91.6%-99.8%) and 93.2% specificity (95% CI 90.4%–96.0%) for DLB diagnosis, with 93.6% overall accuracy [32]. A multi-laboratory European study reported average sensitivity of 78.8% (range 55–100%) and average specificity of 77.5% (range 60–100%) [33] though substantial inter-laboratory variability was noted with fair agreement (average κ = 0.246). This heterogeneity likely reflects differences in assay protocols, substrate preparations, and analytical parameters across laboratories. Notably, 9.5% of Alzheimer's disease patients tested positive by α-syn SAA, and these individuals more frequently presented with core DLB features, higher APOE ε4 frequency, and elevated plasma p-tau217. Among DLB patients, 50% showed AD copathology, associated with lower cognitive scores and higher NfL levels [32]. These findings demonstrate SAA's ability to detect α-synuclein pathology even when overshadowed by concurrent Alzheimer's pathology, with important implications for therapeutic targeting. Performance in multiple system atrophy α-Synuclein SAA performance in MSA showed marked variability across studies and assay protocols. A novel SAA protocol reported 87% overall sensitivity for MSA (95% CI 80–93%), with higher sensitivity in real-life cohorts (91%, 95% CI 80–97%) compared to research cohorts (84%, 95% CI 71–92%). However, specificity for MSA-type 2 seeds was lower in real-life settings (68%, 95% CI 53–81%) versus research cohorts (87%, 95% CI 74–95%) [34]. Several studies reported very low MSA sensitivity with standard protocols optimized for Lewy body disease. One cohort found only 12% sensitivity for MSA [35], while another classic RT-QuIC protocol showed MSA samples were predominantly negative, with all but two cases showing no α-synuclein seeding activity [36]. This discordance reflects fundamental differences in α-synuclein conformational strains between MSA and Lewy body disorders, with MSA-associated oligodendroglial cytoplasmic inclusions potentially harboring distinct seed properties. Modified protocols specifically optimized for MSA achieved better performance. One study using protein misfolding cyclic amplification reported 95.4% sensitivity for MSA [37], while immunoprecipitation-based RT-QuIC detected serum α-synuclein seeds in 64.1% of MSA cases [38]. Importantly, MSA samples showed distinct kinetic signatures, with later aggregation onset but higher maximum fluorescence compared to PD samples, enabling perfect separation between MSA and Lewy body synucleinopathies [39]. These kinetic differences suggest that protocol optimization targeting MSA-specific α-synuclein strains may substantially improve diagnostic utility. Skin biopsy-based α-synuclein SAA Skin biopsy emerged as a promising minimally invasive alternative to CSF sampling. In Chinese populations, skin RT-QuIC achieved 93.3% sensitivity and 100% specificity [40], while a two-center European study reported 90.2% sensitivity and 91.4% specificity with Cohen's kappa of 0.87 indicating excellent inter-laboratory agreement [41]. A comprehensive skin-specific protocol optimization study involving 332 PD patients and 285 controls demonstrated 92.46% sensitivity and 93.33% specificity within 24 h [41]. Skin SAA showed comparable performance to CSF in DLB patients, achieving 94% sensitivity and 98% specificity. However, agreement between CSF and skin varied by condition, with almost perfect agreement for overall synucleinopathies (κ = 0.86) [42]. One study combining skin RT-QuIC with serum NfL reported 100% sensitivity and 100% specificity for distinguishing MSA from PD [43], demonstrating potential for multi-biomarker diagnostic algorithms. Meta-analysis of skin-based approaches showed pooled sensitivity of 0.89 (95% CI 0.85–0.93) with SAA methods compared to 0.82 (95% CI 0.78–0.85) for immunofluorescence and 0.70 (95% CI 0.66–0.74) for immunohistochemistry [44]. While highly sensitive for PD versus healthy controls, skin SAA showed limited ability to differentiate PD from MSA, with specificity only 0.25 (95% CI 0.20–0.31). This likely reflects α-synuclein deposition in peripheral autonomic nerves across multiple synucleinopathies. Distribution patterns of α-synuclein deposition varied regionally. Seeding activity was most prominent in sympathetic adrenergic fibers innervating arrector pili muscles, though also present in sudomotor cholinergic fibers. Higher α-synuclein seeding activity correlated with longer disease duration, more advanced stage, and presence of REM sleep behavior disorder, cognitive impairment, and constipation. All PD patients showed significantly higher cutaneous α-synuclein than controls even without autonomic dysfunction, supporting early peripheral nervous system involvement [45, 46]. Blood-based α-synuclein SAA Blood-based α-synuclein SAA represents the least-invasive approach but showed variable performance. Serum immunoprecipitation-based RT-QuIC demonstrated high diagnostic accuracy for PD versus controls with AUC 0.96 (95% CI 0.95–0.99) in internal cohorts and 0.86 (95% CI 0.74–0.99) in external validation. For DLB, serum assays achieved 96.4% sensitivity and 92.2% specificity. However, MSA detection remained challenging with only 64.1% sensitivity [38]. Combining serum and saliva α-syn SAA substantially improved performance. Individual bio-specimens showed 80.49% sensitivity and 90.48% specificity for serum and 74.70% sensitivity and 97.92% specificity for saliva. When results from both specimens were integrated, sensitivity increased to 95.83% with 96.15% specificity and 0.98 AUC, approaching CSF SAA performance [47]. This synergistic effect suggests α-synuclein seeds distribute heterogeneously across body fluids, with concurrent sampling capturing broader pathology. Blood-based assays using neuronal exosome immunoprecipitation showed α-synuclein seeding in 98.8% of PD patients with specificity approaching 100%. Seeding activity negatively correlated with disease duration both cross-sectionally and longitudinally over 5–9 years [48], suggesting declining peripheral α-synuclein levels as disease advances, potentially reflecting exhaustion of neurons capable of producing seeds or redistribution to central compartments. Alternative bio-specimens Performance in other bio-specimens varied considerably. Olfactory mucosa SAA showed 47% sensitivity and 100% specificity in DLB, substantially lower than CSF or skin [42]. A meta-analysis confirmed olfactory mucosa sensitivity of only 0.64 (95% CI 0.49–0.76) [49]. Technical challenges with tissue quality and inhibitory substances in nasal samples may explain this reduced performance [50]. Urine testing achieved 22% sensitivity despite 100% specificity [42], representing early proof-of-concept for non-invasive sampling but requiring substantial optimization. Saliva showed 70.30% sensitivity and 92.45% specificity when used alone, improving to 75% sensitivity and 91.67% specificity when combined with miRNA-29a-3p biomarkers [51]. Submandibular gland biopsies demonstrated 81.1% sensitivity and 100% specificity by RT-QuIC though this invasive procedure offers limited advantages over CSF sampling [52]. Rectal biopsies showed 100% sensitivity and 100% specificity in a small cohort of 24 PD patients, with seeding activity correlating with motor symptom severity [53] (Table 1).Table 1Summary of α-synuclein seed amplification assay performance and interpretation by biospecimen and settingBiospecimen/settingMost supported useRepresentative performance signalInterpretation caveatsCSF in PD and DLBConfirmation of Lewy body synucleinopathy in clinically selected patients. Support for prodromal enrichment in RBD/hyposmiaOften high sensitivity and specificity for PD/DLB versus controls. Lower sensitivity in normosmia, LRRK2-associated PD and early or restricted pathologySAA-positive does not distinguish PD from DLB by itself. SAA-negative does not exclude PD in selected subgroups. Interpret with olfaction, genetics, DAT imagining and phenotypeCSF in MSAPotential detection of MSA-type α-synuclein strains when protocols are optimizedHighly protocol-dependent. Standard Lewy body disease protocols may miss MSA, whereas modified protocols and kinetic readouts perform betterReport protocol details and kinetic pattern. Combine with NfL, MRI, autonomic features and MSA-specific clinical red flagsSkin biopsyMinimally invasive detection of peripheral α-synuclein seeding or phosphorylated α-synuclein depositionPromising sensitivity and specificity for synucleinopathy versus controls. Useful when CSF is unavailable or declinedLimited specificity for differentiating PD from MSA. Biopsy site, tissue processing and autonomic involvement influence yieldBlood, serum, extracellular vesicles, salivaLeast-invasive exploratory or future screening approach. Possible combined-matrix strategiesPerformance remains variable. Some serum/saliva combinations approach CSF performance in selected cohortsPre-analytical variability and compartment-specific kinetics remain unresolved. Currently less mature for routine individual diagnosisOlfactory mucosa, urine, rectal or salivary gland tissueProof-of-concept research applications in selected settingsPerformance varies widely by matrix and protocolSampling quality, invasiveness and limited validation restrict routine useNon-synucleinopathy contexts (Alzheimer’s disease, PSP/CBS, normal pressure hydrocephalus, traumatic brain injury)Detection of possible α-synuclein co-pathology or clinically relevant biological heterogeneitySAA positivity has been reported in several non-classical contextsDo not equate positivity with PD. Ask whether α-synuclein is the dominant pathology, a co-pathology or an incidental age-related finding Table 1Summary of α-synuclein seed amplification assay performance and interpretation by biospecimen and settingBiospecimen/settingMost supported useRepresentative performance signalInterpretation caveatsCSF in PD and DLBConfirmation of Lewy body synucleinopathy in clinically selected patients. Support for prodromal enrichment in RBD/hyposmiaOften high sensitivity and specificity for PD/DLB versus controls. Lower sensitivity in normosmia, LRRK2-associated PD and early or restricted pathologySAA-positive does not distinguish PD from DLB by itself. SAA-negative does not exclude PD in selected subgroups. Interpret with olfaction, genetics, DAT imagining and phenotypeCSF in MSAPotential detection of MSA-type α-synuclein strains when protocols are optimizedHighly protocol-dependent. Standard Lewy body disease protocols may miss MSA, whereas modified protocols and kinetic readouts perform betterReport protocol details and kinetic pattern. Combine with NfL, MRI, autonomic features and MSA-specific clinical red flagsSkin biopsyMinimally invasive detection of peripheral α-synuclein seeding or phosphorylated α-synuclein depositionPromising sensitivity and specificity for synucleinopathy versus controls. Useful when CSF is unavailable or declinedLimited specificity for differentiating PD from MSA. Biopsy site, tissue processing and autonomic involvement influence yieldBlood, serum, extracellular vesicles, salivaLeast-invasive exploratory or future screening approach. Possible combined-matrix strategiesPerformance remains variable. Some serum/saliva combinations approach CSF performance in selected cohortsPre-analytical variability and compartment-specific kinetics remain unresolved. Currently less mature for routine individual diagnosisOlfactory mucosa, urine, rectal or salivary gland tissueProof-of-concept research applications in selected settingsPerformance varies widely by matrix and protocolSampling quality, invasiveness and limited validation restrict routine useNon-synucleinopathy contexts (Alzheimer’s disease, PSP/CBS, normal pressure hydrocephalus, traumatic brain injury)Detection of possible α-synuclein co-pathology or clinically relevant biological heterogeneitySAA positivity has been reported in several non-classical contextsDo not equate positivity with PD. Ask whether α-synuclein is the dominant pathology, a co-pathology or an incidental age-related finding Summary of α-synuclein seed amplification assay performance and interpretation by biospecimen and setting Diagnostic performance of neurofilament light chain CSF neurofilament light chain CSF NfL emerged as a powerful biomarker for differentiating atypical parkinsonian syndromes from PD. Meta-analysis revealed CSF NfL discriminated PD from atypical parkinsonism with pooled AUC of 0.941 (95% CI 0.916–0.965), sensitivity of 86% (95% CI 79–90%), and specificity of 88% (95% CI 82–92%) [54]. Individual studies confirmed these findings, with NfL alone achieving AUC 0.93 for separating PD from atypical parkinsonian disorders [55]. Performance varied by specific diagnosis. For MSA versus PD, CSF NfL showed AUC values ranging from 0.91 [15] to 0.99[56]. For PSP versus PD, CSF NfL achieved AUC 0.94 [56, 57]. Unlike α-synuclein SAA, CSF NfL levels remained stable over serial measurements despite progressive clinical deterioration in MSA patients [58]. This stability suggests NfL reflects structural neuronal damage rather than dynamic disease activity, limiting utility as a progression biomarker despite excellent diagnostic discrimination. Plasma/serum neurofilament light chain Blood-based NfL showed strong but slightly reduced performance compared to CSF. Plasma NfL discriminated PD from atypical parkinsonism with AUC 0.91 in the Lund cohort, 0.85 in the London cohort, and 0.81 in early disease cohorts (≤ 3 years duration) [59]. Meta-analysis confirmed circulating NfL differentiated PD from atypical parkinsonism with pooled AUC 0.874 (95% CI 0.802–0.946), sensitivity 91% (95% CI 86–95%), and specificity 76% (95% CI 62–85%) [54]. Correlation between blood and CSF NfL was consistently strong (ρ ≥ 0.73–0.84, p ≤ 0.001) [59], though not perfect. Four independent assays showed comparable precision for CSF NfL (AUC range 0.966–0.974) but lower accuracy for plasma NfL (AUC range 0.917–0.924) [60]. This suggests blood–brain barrier transport and peripheral contributions introduce noise in plasma measurements, though correlation remains sufficient for diagnostic purposes. For MSA specifically, plasma NfL achieved excellent discrimination from PD with AUC 0.877–0.97 [56, 61]. Levels were significantly elevated in MSA (mean 23.8 ± 10.3 ng/L) compared to PD (10.4 ± 4.9 ng/L) and controls (11.5 ± 6.5 ng/L), with 86% sensitivity and 85% specificity, yielding 76% positive predictive value and 92% negative predictive value [61]. Age-adjusted analysis showed NfL increased with age in PD and controls but remained markedly elevated in atypical parkinsonism regardless of age [61]. Plasma NfL showed particular utility in specific PD motor subtypes. In the postural instability gait disorder (PIGD) subtype, higher baseline NfL associated with worse global cognition and motor scores and predicted both motor and cognitive decline over 2 years. NfL was significantly higher in PIGD versus tremor-dominant PD at 2-year follow-up (18.4 ± 14.5 versus 12.6 ± 4.4 pg/mL), suggesting NfL may identify patients at higher risk for aggressive disease progression [62] (Table 2).Table 2Neurofilament light chain: discriminative performance, confounders and implementation caveatsMatrix/assay contextBest-supported clinical useTypical discriminative signalKey caveats for thresholdsCSF NfLDifferentiating PD/DLB from atypical parkinsonism, especially MSA and PSP/CBSHigher accuracy than blood NfL in head-to-head parkinsonism cohorts. Strong discrimination of PD versus atypical parkinsonismLumbar puncture required. Values are not disease-specific and should be interpreted with clinical course and imagingPlasma/serum NfLLess-invasive triage marker for atypical or aggressive neurodegeneration. Useful when CSF is not availableStrong but usually slightly lower accuracy than CSF. Useful negative predictive value when low in the right contextAffected by age, BMI/hemodilution, renal function, peripheral neuropathy, diabetes, vascular disease and recent neurological injurySimoa, ELLA, CLEIA and other platformsClinical implementation and multicenter studiesHigh correlations between platforms but non-identical absolute values. Platform-specific cutoffs are neededDo not transfer thresholds between assays without validation. Prefer platform-specific reference intervals, z scores or percentilesNfL combined with SAAResolving whether α-synuclein seeding occurs in a low-versus high-neurodegeneration contextSAA plus CSF NfL improves PD versus atypical parkinsonism discrimination in several cohortsA high NfL does not identify the proteinopathy by itself. A positive SAA plus high NfL should prompt MSA, mixed pathology or aggressive comorbidity work-up Table 2Neurofilament light chain: discriminative performance, confounders and implementation caveatsMatrix/assay contextBest-supported clinical useTypical discriminative signalKey caveats for thresholdsCSF NfLDifferentiating PD/DLB from atypical parkinsonism, especially MSA and PSP/CBSHigher accuracy than blood NfL in head-to-head parkinsonism cohorts. Strong discrimination of PD versus atypical parkinsonismLumbar puncture required. Values are not disease-specific and should be interpreted with clinical course and imagingPlasma/serum NfLLess-invasive triage marker for atypical or aggressive neurodegeneration. Useful when CSF is not availableStrong but usually slightly lower accuracy than CSF. Useful negative predictive value when low in the right contextAffected by age, BMI/hemodilution, renal function, peripheral neuropathy, diabetes, vascular disease and recent neurological injurySimoa, ELLA, CLEIA and other platformsClinical implementation and multicenter studiesHigh correlations between platforms but non-identical absolute values. Platform-specific cutoffs are neededDo not transfer thresholds between assays without validation. Prefer platform-specific reference intervals, z scores or percentilesNfL combined with SAAResolving whether α-synuclein seeding occurs in a low-versus high-neurodegeneration contextSAA plus CSF NfL improves PD versus atypical parkinsonism discrimination in several cohortsA high NfL does not identify the proteinopathy by itself. A positive SAA plus high NfL should prompt MSA, mixed pathology or aggressive comorbidity work-up Neurofilament light chain: discriminative performance, confounders and implementation caveats Performance in prodromal synucleinopathies α-Synuclein SAA demonstrated strong performance in prodromal populations, supporting early biomarker-based diagnosis. Among 51 PPMI participants with REM sleep behavior disorder (RBD) or hyposmia, 86% tested positive by CSF α-syn SAA [16]. Isolated RBD patients showed particularly high positivity rates, 100% in some cohorts [36] and 90.4% in others [63]. One study reported 64% overall sensitivity across four iRBD cohorts though this varied from. 39% in community-detected early cases to approximately 90% in clinic-based cohorts (64), indicating ascertainment bias affects apparent performance. Among non-manifesting carriers of PD-associated genetic variants, 8% tested SAA-positive (9% of LRRK2 carriers, 7% of GBA carriers) [16]. Longitudinal follow-up of prodromal participants showed 14 of 45 (31%) converted to manifest synucleinopathy, with 64% of converters showing baseline SAA positivity [64]. Faster time-to-threshold and higher area under the curve at baseline predicted phenoconversion [65], with SAA-positive prodromal participants showing higher conversion rates (log-rank p = 0.001) [65]. These data demonstrate α-syn SAA can identify at-risk individuals years before motor symptoms emerge. The moderate sensitivity in prodromal populations (64–86% versus 88–96% in manifest PD) may reflect lower α-synuclein seed concentrations in early disease or incomplete propagation to CSF. Two cases of scans without evidence of dopaminergic deficit (SWEDD) tested SAA-positive and subsequently developed PD with abnormal dopamine transporter imaging [66], supporting SAA's ability to detect molecular pathology preceding imaging abnormalities. Differentiation between parkinsonism subtypes α-Synuclein SAA in differential diagnosis Standard α-syn SAA protocols showed limited ability to differentiate between Lewy body disorders (PD and DLB) but successfully distinguished these from non-synucleinopathies. CSF SAA achieved 89% sensitivity and 96% specificity for PD, with comparable 97.1% sensitivity for DLB and 94.4% for PD [36]. Both conditions showed similar fluorescence patterns, with quantification of α-synuclein seeding activity revealing no significant differences between LB-related clinical syndromes [36]. However, kinetic parameters showed promise for subtype differentiation. MSA samples demonstrated significantly longer time-to-threshold (T50) and lower maximum velocity (Vmax) compared to PD [64]. One study reported MSA and PD/DLB could be distinguished based on intermediate versus high fluorescence amplification patterns [34], while another found MSA-type seeds showed lower and slower kinetics as markers of potential Lewy body copathology in PSP [19]. For progressive supranuclear palsy and corticobasal syndrome, α-syn SAA positivity ranged from 3 to 25%. In PSP, 8.5% of cases tested positive for PD-type seeds, with positive cases showing older age and shorter disease duration [67]. CBS showed 25.7% SAA positivity, while another cohort reported 35.9% positivity in CBS and 28.6% in PSP [68]. This likely represents Lewy body copathology, which autopsy studies confirm occurs in 10–30% of tauopathy cases. Young-onset patients showed associations between SAA positivity and AD biomarker status [68], suggesting complex interactions between multiple proteinopathies. SAA positivity beyond classical synucleinopathies A major implication for any diagnostic algorithm is that α-synuclein SAA positivity can occur outside clinically defined PD, DLB, MSA, or pure autonomic failure [10, 20, 69]. In AD cohorts, SAA-positive patients may have Lewy body co-pathology and a more mixed cognitive phenotype [70–72]. In PSP and CBS, positivity can reflect concomitant Lewy pathology, older age or diagnostic heterogeneity rather than conversion of the primary diagnosis into PD [20, 67]. Recent reviews and clinicopathological work also emphasize SAA positivity in idiopathic normal pressure hydrocephalus and traumatic brain injury, where the clinical relevance of seeding activity may range from a genuine contributor to symptoms to an incidental or age-related co-pathology [10, 20, 69, 73–75]. For this reason, the first step of a biomarker algorithm should be phrased as “Is there evidence of α-synuclein seeding?” rather than “Does the patient have PD?”. The subsequent question is whether that biological signal is concordant with the syndrome and trajectory. Discordance should trigger a structured search for mixed pathology, alternative primary diagnoses or technical limitations rather than automatic diagnostic reclassification. Neurofilament light chain in differential diagnosis NfL excelled at distinguishing atypical parkinsonism from PD, outperforming α-synuclein measurements. CSF NfL showed minimal overlap between MSA and PD, achieving AUC 0.97–0.99 [56, 76] with 97% sensitivity and 100% specificity in some cohorts [58]. The combination of elevated NfL (> 14.66 ng/L) with planimetric MRI measures (pontine-to-midbrain diameter ratio > 2.06) achieved 83.7% accuracy for differentiating PD, MSA, and PSP [77]. Plasma NfL similarly distinguished atypical parkinsonism, though with reduced specificity. Blood NfL was markedly increased in MSA, PSP, and CBS compared to both PD and controls across multiple cohorts [59]. For MSA versus PD specifically, plasma NfL achieved AUC 0.877 with 83.3% sensitivity and 91.5% specificity [78], while for PSP versus PD, AUC reached 0.904 when combined with MRI markers. NfL performed poorly for differentiating between atypical parkinsonism subtypes. MSA versus PSP comparisons showed suboptimal discrimination, with NfL elevated in both conditions [79]. Similarly, NfL could not reliably distinguish between PD and healthy controls in most studies [55, 78] though modest elevations were noted in some PD cohorts [80]. Combined biomarker approaches α-Synuclein SAA plus NfL integration Several studies demonstrated that combining α-synuclein SAA with NfL substantially improved diagnostic accuracy. CSF RT-QuIC plus CSF NfL achieved AUC 0.97 for PD versus atypical parkinsonism, AUC 0.97 for MSA, and AUC 0.99 for PSP/CBS [35]. This outperformed RT-QuIC alone (p = 0.047 versus atypical parkinsonism, p = 0.002 versus MSA, p = 0.007 versus PSP/CBS) and NfL alone (p = 0.011 versus atypical parkinsonism, p = 0.0001 versus PSP/CBS) [35]. For plasma biomarkers, combining NfL with other markers improved discrimination. Plasma panels including NfL, GFAP, and α-synuclein achieved AUC up to 1.000 for distinguishing atypical parkinsonism from controls [81]. The combination of NfL, Aβ42, and total tau showed high accuracy for separating PD from atypical parkinsonism [81], while NfL plus GFAP plus disease duration plus pontine volume achieved AUC 0.874 for PD versus atypical parkinsonism [82]. The mechanistic complementarity is evident: α-syn SAA directly detects pathological protein aggregates characteristic of synucleinopathies, while NfL reflects downstream axonal degeneration more pronounced in MSA and PSP. A proposed diagnostic algorithm uses positive α-syn SAA to confirm synucleinopathy, with NfL levels then distinguishing Lewy body disease (normal/mildly elevated NfL) from MSA/PSP (markedly elevated NfL) [43, 60]. Multi-biomarker panels Comprehensive panels incorporating multiple pathology markers showed incremental diagnostic gains. In CSF, combining α-synuclein, NfL, soluble amyloid precursor protein α (sAPPα), and tau achieved AUC 0.95 for distinguishing PD from atypical parkinsonism [83]. Adding markers of neuroinflammation (MCP-1, YKL-40) further improved accuracy [83]. For plasma biomarkers, optimal panels varied by clinical question. Distinguishing parkinsonian syndromes from healthy controls was best achieved with α-synuclein, Aβ42, Aβ40, Aβ42/40 ratio, and NfL (AUC 0.983) [81]. Separating PD specifically from controls showed optimal performance with the same panel (AUC 0.977) [81]. Differentiating frontotemporal dementia from parkinsonian syndromes used Aβ42 × p-tau181 with AUC 0.932 [84]. Combining biomarkers with imaging enhanced diagnostic precision. Integration of plasma NfL and GFAP with quantitative susceptibility mapping iron measurements and MRI planimetry improved accuracy beyond individual markers [77, 78]. Lower NfL levels (< 20 ng/L) with elevated p-tau181 indicated AD, while NfL < 40 ng/L suggested PD [85], demonstrating how threshold-based algorithms can guide differential diagnosis. Quantitative SAA parameters and disease severity Beyond binary positive/negative results, SAA kinetic parameters showed potential as quantitative biomarkers. A novel Lewy-fold pathology (LFP) score based on positive signals across a dilution series significantly correlated with Hoehn and Yahr stage, MDS-UPDRS III, and Montreal Cognitive Assessment scores [86]. Longitudinal tracking in seven PD cases showed progressive LFP score increases corresponding with clinical deterioration [86]. CSF-seeded α-synuclein oligomer concentrations measured by ELISA correlated with UPDRS motor scores (r = 0.58, p < 0.001) and Hoehn and Yahr scores (r = 0.43, p < 0.01) in discovery cohorts, with similar correlations in validation cohorts [87]. Receiver operating characteristic analysis at 20 h yielded 91.9% sensitivity and 85.3% specificity (AUC 0.969) for CSF-seeded oligomers versus 80.7% sensitivity and 76.5% specificity (AUC 0.860) for thioflavin T fluorescence alone [87]. Number of positive replicates (Nrep) increased longitudinally in asymptomatic LBD (β = 0.15, 95% CI 0.09–0.21) and PD without dementia (β = 0.07, 95% CI 0.02–0.12) [88]. Baseline Nrep predicted subsequent dementia in the whole cohort (HR 1.57, 95% CI 1.19–2.07) and PD subgroup (HR 1.83, 95% CI 1.17–2.85) [88]. Conversely, time-to-threshold (lag phase) decreased over time in asymptomatic LBD (β = − 0.24, 95% CI − 0.42 to − 0.06) [88]. Fast versus slow time-to-threshold at baseline predicted cognitive decline defined as Montreal Cognitive Assessment < 22 (HR 2.51, 95% CI 1.50–4.20, p = 0.001) [19]. This prognostic value persisted in a subgroup analysis of Alzheimer's disease biomarker-negative participants (HR 1.80, 95% CI 1.03–3.18) [19], indicating α-synuclein pathology independently drives cognitive deterioration separate from amyloid/tau effects. For MSA, quantitative RT-QuIC parameters showed opposite relationships to disease severity. All kinetic parameters (Nrep, lag phase, Vmax) correlated with worse clinical progression measured by Unified Multiple System Atrophy Rating Scale [64], contrasting with PD where no such correlations existed. This divergence may reflect differences in seed propagation dynamics between oligodendroglial and neuronal α-synuclein pathology. Synthesis: reconciling heterogeneity in biomarker performance Context-dependent performance of α-synuclein SAA The substantial heterogeneity in α-synuclein SAA sensitivity (ranging from 12 to 100% for MSA, 37.5% to 100% for PD depending on subgroup) resolves when examined through lenses of pathological stage, anatomical distribution, and genetic background. Three mechanisms explain apparent discrepancies:First, pathological burden thresholds: Studies showing highest sensitivity (95–100%) predominantly examined limbic/neocortical Lewy body disease [30, 31, 36], while those reporting lower sensitivity included amygdala-predominant variants (14.3–50% sensitivity) [30, 31] or early Braak stages 1–2 (37.5% sensitivity) [31]. CSF SAA appears to require α-synuclein pathology extending beyond restricted brainstem or amygdala regions to achieve reliable detection, explaining reduced sensitivity in early disease stages and focal pathology patterns.Second, olfactory phenotype stratification: Among PD patients, those with typical anosmia showed near-perfect SAA sensitivity (98.6%) [16], while normosmia associated with 78.3% sensitivity. Hyposmia showed very high sensitivity (95%) for predicting SAA positivity across all groups [89], with participants scoring ≤ 15th percentile on smell testing having 18.3-fold higher odds of positive SAA. This likely reflects greater olfactory bulb and limbic α-synuclein deposition in hyposmic patients, increasing seed shedding into CSF.Third, genetic variant effects: LRRK2-associated PD consistently showed lower SAA positivity (67.5%) [16], particularly when combined with preserved olfaction (34.7%). SAA-negative LRRK2-PD patients were older at symptom onset (63.0 versus 55.4 years) and had higher baseline serum NfL (20.4 versus 13.8 pg/mL) [19], suggesting a distinct pathophysiological subtype potentially involving alternative neurodegenerative mechanisms beyond α-synuclein aggregation. GBA-associated PD showed opposite pattern with 95.9% SAA positivity [22, 90] and faster time-to-threshold [19], consistent with accelerated α-synuclein aggregation in GBA mutation carriers. First, pathological burden thresholds: Studies showing highest sensitivity (95–100%) predominantly examined limbic/neocortical Lewy body disease [30, 31, 36], while those reporting lower sensitivity included amygdala-predominant variants (14.3–50% sensitivity) [30, 31] or early Braak stages 1–2 (37.5% sensitivity) [31]. CSF SAA appears to require α-synuclein pathology extending beyond restricted brainstem or amygdala regions to achieve reliable detection, explaining reduced sensitivity in early disease stages and focal pathology patterns. First, pathological burden thresholds: Studies showing highest sensitivity (95–100%) predominantly examined limbic/neocortical Lewy body disease [30, 31, 36], while those reporting lower sensitivity included amygdala-predominant variants (14.3–50% sensitivity) [30, 31] or early Braak stages 1–2 (37.5% sensitivity) [31]. CSF SAA appears to require α-synuclein pathology extending beyond restricted brainstem or amygdala regions to achieve reliable detection, explaining reduced sensitivity in early disease stages and focal pathology patterns. Second, olfactory phenotype stratification: Among PD patients, those with typical anosmia showed near-perfect SAA sensitivity (98.6%) [16], while normosmia associated with 78.3% sensitivity. Hyposmia showed very high sensitivity (95%) for predicting SAA positivity across all groups [89], with participants scoring ≤ 15th percentile on smell testing having 18.3-fold higher odds of positive SAA. This likely reflects greater olfactory bulb and limbic α-synuclein deposition in hyposmic patients, increasing seed shedding into CSF. Second, olfactory phenotype stratification: Among PD patients, those with typical anosmia showed near-perfect SAA sensitivity (98.6%) [16], while normosmia associated with 78.3% sensitivity. Hyposmia showed very high sensitivity (95%) for predicting SAA positivity across all groups [89], with participants scoring ≤ 15th percentile on smell testing having 18.3-fold higher odds of positive SAA. This likely reflects greater olfactory bulb and limbic α-synuclein deposition in hyposmic patients, increasing seed shedding into CSF. Third, genetic variant effects: LRRK2-associated PD consistently showed lower SAA positivity (67.5%) [16], particularly when combined with preserved olfaction (34.7%). SAA-negative LRRK2-PD patients were older at symptom onset (63.0 versus 55.4 years) and had higher baseline serum NfL (20.4 versus 13.8 pg/mL) [19], suggesting a distinct pathophysiological subtype potentially involving alternative neurodegenerative mechanisms beyond α-synuclein aggregation. GBA-associated PD showed opposite pattern with 95.9% SAA positivity [22, 90] and faster time-to-threshold [19], consistent with accelerated α-synuclein aggregation in GBA mutation carriers. Third, genetic variant effects: LRRK2-associated PD consistently showed lower SAA positivity (67.5%) [16], particularly when combined with preserved olfaction (34.7%). SAA-negative LRRK2-PD patients were older at symptom onset (63.0 versus 55.4 years) and had higher baseline serum NfL (20.4 versus 13.8 pg/mL) [19], suggesting a distinct pathophysiological subtype potentially involving alternative neurodegenerative mechanisms beyond α-synuclein aggregation. GBA-associated PD showed opposite pattern with 95.9% SAA positivity [22, 90] and faster time-to-threshold [19], consistent with accelerated α-synuclein aggregation in GBA mutation carriers. Protocol-dependent performance in MSA The wide variation in MSA detection primarily reflects assay optimization for different α-synuclein strains [10, 14, 20]. From a stewardship perspective, a negative result obtained with a Lewy body disease-optimized protocol should not be used to rule out MSA when the phenotype, autonomic failure, MRI pattern or rapid progression are compelling [14, 20]. Kinetic measures are especially important in MSA because the diagnostic signal may lie in amplification rather than binary positivity alone [10, 20, 34, 39]. Laboratories should specify the intended diagnostic target (Lewy body disease versus MSA), substrate and buffer conditions, replicate criteria and whether strain-sensitive kinetic parameters are validated for clinical interpretation [10, 20, 91]. The clinical implication is that MSA diagnosis requires either specialized SAA protocols targeting MSA-type strains or integration with orthogonal biomarkers, such as NfL and MRI [14, 22, 39]. In practice, elevated NfL in an SAA-positive patient should raise the possibility of MSA, mixed pathology or another aggressive neurodegenerative process rather than being interpreted as uncomplicated PD [14, 22, 39, 92]. Inter-laboratory and inter-assay variability Beyond differences in shaking conditions, plate readers, buffers and analytical cutoffs, recent evidence identifies recombinant α-synuclein substrate preparation as an addressable determinant of SAA variability [10, 33, 91, 93]. Al-Azzawi and colleagues compared four α-synuclein purification methods across two laboratories and found that osmotic shock-purified monomer had the lowest de novo aggregation tendency and the best inter-batch reproducibility, translating into more consistent amplification reactions [93]. This finding is directly relevant to clinical translation because substrate instability can generate false-positive or uninterpretable curves, while batch heterogeneity can reduce reproducibility across laboratories [33, 93]. Clinical and research reports should therefore document substrate source, purification method, batch qualification, storage conditions and acceptance criteria for negative and positive controls [10, 91, 93]. Harmonization efforts should include reference materials, replicate-level reporting, inter-laboratory proficiency testing and transparent reporting of substrate purification, with osmotic shock purification considered a strong candidate for standardization when compatible with the validated assay protocol [10, 91, 93, 94]. Multi-laboratory studies revealed both reproducibility and variability. Three independent laboratories testing identical PPMI CSF samples showed remarkably similar results with sensitivity 86–96% and specificity 93–100% [66], demonstrating reproducibility when expert laboratories use optimized protocols. However, a four-laboratory European study of DLB samples showed fair agreement (κ = 0.246) with sensitivity ranging 55–100% and specificity 60–100% across sites [33]. This variability traces to methodological differences. Laboratory-specific variations in recombinant α-synuclein substrate concentration, plate reader settings, incubation protocols, and analytical cutoffs all influence results [33]. Even biospecimen preparation affects outcomes. One study found different skin biopsy homogenate preparations contributed to assay discordance [95]. Standardization efforts using universal control fluids showed promise for reducing variability while maintaining 94% accuracy [42]. For clinical implementation, this heterogeneity presents challenges. A commercial α-synuclein SAA validated under CLIA/CAP certification showed 93.6% accuracy using clinical diagnosis with DAT-SPECT as reference standard but only 83.9% accuracy versus clinical diagnosis alone [90], indicating some interdependence between imaging and molecular biomarkers. Future clinical use requires protocol harmonization, standardized reagents, proficiency testing programs, and established quality control measures comparable to those developed for prion disease RT-QuIC assays. Clinical correlations and phenotypic associations α-Synuclein SAA positivity showed robust associations with specific clinical features across studies. In DLB cohorts, SAA-positive patients had lower Montreal Cognitive Assessment scores (18.8 ± 5.7 versus 21.2 ± 5.2, p = 0.01), worse parkinsonism on MDS-UPDRS part III (33.8 ± 15.1 versus 25.6 ± 16.4, p = 0.001), and higher rates of REM sleep behavior disorder (86% versus 62%, p < 0.0001) [89]. Hyposmia was particularly predictive, with 90% of SAA-positive patients scoring below 15th percentile on smell testing versus only 32% of SAA-negative patients [89]. Core DLB features, fluctuations, visual hallucinations, RBD, and parkinsonism, showed ≥ 79% accuracy for predicting SAA positivity [96], while intact olfactory function associated with lower SAA risk. These phenotypic correlations suggest α-synuclein pathology drives specific symptom clusters, with SAA identifying a biologically homogeneous subgroup regardless of clinical diagnostic label. Serum α-synuclein seeding activity correlated inversely with Montreal Cognitive Assessment in males (p = 0.04) and positively with Hamilton Depression Rating Scale in females (p = 0.04) [47]. Saliva seeding activity correlated inversely with age at diagnosis in males and < 70 age group (p = 0.04) [47]. These sex- and age-dependent associations suggest biological heterogeneity in α-synuclein propagation that may inform personalized diagnostic approaches. Plasma NfL correlations differed from α-synuclein patterns. Higher NfL associated with worse global cognition and UPDRS motor scores in PD, with particularly strong associations in the PIGD motor subtype [62]. Plasma Aβ42 levels correlated positively with UPDRS part III in MSA (p < 0.05), while the Aβ42/40 ratio associated with cognitive function in atypical parkinsonism [81]. These distinct correlation patterns support mechanistic differences between biomarkers, with NfL reflecting structural damage versus α-synuclein SAA indicating molecular pathology. Integration into clinical diagnostic pathways Two-step diagnostic algorithms A validated two-step workflow combining smell function screening with confirmatory CSF α-syn SAA testing demonstrated practical efficiency. Using University of Pennsylvania Smell Identification Test (UPSIT) scores with age and sex for initial risk stratification, followed by CSF SAA only in high-risk individuals, predicted postmortem Lewy body pathology with 94% overall accuracy [97]. This approach achieved 95% accuracy in clinical parkinsonism, 94% in clinical AD, and 93% in clinically unimpaired participants [97]. Critically, the two-step method reduced CSF testing by 43% overall, 23% in parkinsonism patients, 35% in AD patients, and 80% in clinically unimpaired individuals [97, 98]. Positive predictive values reached 96% in parkinsonism where Lewy body pathology prevalence was highest (62.7%), while negative predictive values approached 98% in AD [98]. This demonstrates how sequential testing can maintain accuracy while substantially reducing invasive procedures and costs. Predictive models using only clinical variables achieved comparable performance. UPSIT percentile, sex, constipation status, and LRRK2/GBA genetic status predicted CSF α-syn SAA status with AUROC 0.920 internally and 0.976 on external validation [99]. Models using UPSIT alone performed similarly [99], indicating olfactory function captures much of the predictive information. These models enable pre-test probability estimation to guide CSF sampling decisions. Integration with existing diagnostic methods Studies examining biomarker integration with imaging revealed complementary information. α-Synuclein SAA identified two SWEDD subjects as positive who later developed PD with abnormal DAT-SPECT [66], suggesting molecular pathology precedes dopaminergic imaging abnormalities. Among PD patients, SAA correctly identified cases later determined clinically inconsistent with PD [66], supporting diagnostic revision based on molecular evidence. The combination of CSF α-syn SAA, CSF NfL, and midbrain/pons MRI planimetry yielded high discriminant ability across PD, MSA, and tauopathies (AUC 0.924–0.983) [35]. For PSP specifically, tau-PET imaging combined with CSF NfL and midbrain/pons ratio achieved AUC 0.94 versus controls and 0.92 versus α-synucleinopathies [57]. These integrated approaches leverage orthogonal information from molecular pathology, neurodegeneration, and structural atrophy. Clinical implementation in tertiary care showed feasibility. One behavioral neurology clinic used commercial SAA testing in 21 patients over one year, with 43% testing positive [100]. Meeting probable DLB criteria predicted SAA positivity accurately, while possible DLB criteria showed poor predictive value [100]. SAA results influenced clinical decisions including initiation of targeted therapies and consideration of amyloid-targeting treatments in SAA-positive/amyloid-positive cases [100]. Diagnostic revision and uncertainty reduction Biomarker results enabled diagnostic reclassification in multiple studies. In one cohort, biomarker-guided diagnostic revision increased α-syn RT-QuIC sensitivity for PD from 75 to 81% while decreasing MSA sensitivity from 12 to 9%, with specificity improving from 89 to 91% against tauopathies [35]. This suggests biomarkers help identify clinically atypical presentations that nonetheless harbor expected molecular pathology. CSF NfL predicted longitudinal diagnostic change with 88% accuracy versus 77% for clinical assessment alone [101], potentially enabling earlier accurate diagnosis in 11% of patients. Among autopsy-confirmed cases, integrating α-syn SAA with clinical diagnosis improved identification of Lewy body pathology, with SAA correctly predicting neuropathology even when clinical diagnosis was uncertain [102]. For”suggestive of PSP” cases representing early disease, worse saccade variables and negative/low-fluorescence-positive α-syn SAA at baseline predicted longitudinal diagnostic reinforcement[103]. In probable/possible PSP, positive CSF α-syn SAA suggested copathology requiring pathological confirmation. These applications demonstrate biomarkers both increase diagnostic confidence and identify mixed pathology cases requiring nuanced interpretation (Fig. 1).Fig. 1Illustrates a biomarker stewardship algorithm for α-synuclein SAA and NfL in parkinsonism. CBS Corticobasal syndrome, DLB dementia with Lewy bodies, iRBD isolated REM sleep behaviour disorder, MSA multiple system atrophy, NfL neurofilament light chain, PAF pure autonomic failure, PD Parkinson’s disease, PSP progressive supranuclear palsy, SAA seed amplification assay Illustrates a biomarker stewardship algorithm for α-synuclein SAA and NfL in parkinsonism. CBS Corticobasal syndrome, DLB dementia with Lewy bodies, iRBD isolated REM sleep behaviour disorder, MSA multiple system atrophy, NfL neurofilament light chain, PAF pure autonomic failure, PD Parkinson’s disease, PSP progressive supranuclear palsy, SAA seed amplification assay Diagnostic accuracy by disease stage Early and prodromal disease detection α-Synuclein SAA showed capacity for early detection but with stage-dependent sensitivity. In newly diagnosed PD patients recruited through population screening (median 38 days from diagnosis), CSF SAA achieved 82.6% sensitivity and 88.2% specificity [104]. Among early PD with disease duration 1.2 ± 1.6 years in PPMI, sensitivity reached 88% [16]. However, sensitivity was only 37.5% in Braak stages 1–2, increasing to 73.3% in stage 3 and 100% in stages 4–6 [31]. This stage-dependent performance has important implications. SAA reliably detects patients with sufficient pathology burden for clinical manifestation (Braak ≥ 3, predominantly limbic involvement) but shows reduced sensitivity in preclinical stages. Among prodromal RBD patients, 86% positivity [16] suggests most have already accumulated detectable α-synuclein seeds despite lacking motor symptoms, consistent with RBD representing an advanced prodromal state. Longitudinal data from prodromal cohorts showed SAA status remained stable over serial testing, with no changes in 48 prodromal participants with repeated assays [65]. Among 26 participants converting from SAA-negative to SAA-positive (LBD-converters), baseline testing showed lower number of positive replicates and longer lag phase [88], indicating lower seed concentrations near detection threshold. These converters likely represent a transitional window where seed levels cross from undetectable to detectable. Performance across disease duration Several studies examined how biomarker performance varies with disease chronicity. Plasma NfL showed age-corrected elevation early in PD (mean duration 1.1 years) [62], maintaining diagnostic utility even in earliest symptomatic phases. However, blood-based α-synuclein SAA seeding activity showed inverse correlation with disease duration, with longer duration associated with lower seeding activity (F60 p < 0.001) [48]. Longitudinal analysis confirmed this gradual decrease over 5–9 years [48]. This temporal decline in peripheral α-synuclein contrasts with stable or increasing CSF levels, suggesting differential kinetics across compartments. Potential explanations include peripheral nerve loss reducing seed production, sequestration of aggregates in tissue deposits, or altered trafficking from CNS to periphery as disease advances. The clinical implication is blood-based SAA may have optimal sensitivity in early-moderate disease but declining utility in advanced stages. For MSA, plasma NfL remained stable over serial measurements despite progressive clinical worsening [58, 105]. Baseline levels predicted disease progression, but longitudinal changes did not track with clinical deterioration [105]. This dissociation between biomarker stability and clinical progression suggests NfL reflects cumulative structural damage rather than ongoing disease activity, limiting its utility as a progression monitor while maintaining diagnostic value. Synthesis: optimal biomarker selection by clinical context PD diagnosis and subtyping For confirming PD diagnosis in patients with typical parkinsonian features and hyposmia, CSF α-syn SAA provides near-definitive molecular confirmation (98.6% sensitivity) [16]. Combined serum and saliva SAA offers comparable accuracy (95.83% sensitivity, 96.15% specificity) [47] with less invasiveness. Skin biopsy represents an intermediate option balancing invasiveness and accessibility (90–94% sensitivity, 91–100% specificity) [40, 41]. In atypical presentations (young onset, no olfactory deficit, LRRK2 carriers), CSF SAA sensitivity declines to 35–78% [16], necessitating complementary approaches. Adding NfL excludes atypical parkinsonism (which shows markedly elevated levels) while genetic testing identifies monogenic forms. For LRRK2 carriers without olfactory loss, the 34.7% SAA positivity [16] suggests many represent an alternative pathophysiological subtype where biomarker-based diagnosis requires different strategies. Prodromal PD identification benefits from SAA in high-risk populations (RBD, hyposmia) where 86% positivity substantially exceeds the 8% positivity in asymptomatic genetic carriers [16]. The 18.3-fold increased odds of SAA positivity with UPSIT ≤ 15th percentile supports smell testing as a highly effective pre-screening tool to enrich populations for confirmatory SAA testing [89]. Differentiating atypical parkinsonism For distinguishing PD from MSA/PSP, NfL substantially outperforms α-syn SAA. CSF NfL alone achieves AUC 0.91–0.99 for this discrimination [15, 56], while plasma NfL reaches AUC 0.87–0.91 [59, 61]. The combination of NfL with MRI planimetry further improves accuracy, with pontine-to-midbrain diameter ratio plus serum NfL achieving 83.7% accuracy [77]. However, NfL cannot reliably distinguish between MSA and PSP, as both show marked elevation (AUC suboptimal) [79]. Here, clinical features, imaging patterns, and potentially 4-repeat tau SAA for PSP become necessary [67]. The proposed algorithm uses: (1) α-syn SAA to confirm synucleinopathy; (2) if positive with normal/mildly elevated NfL, diagnose PD/DLB; (3) if positive with markedly elevated NfL, consider MSA; (4) if negative with elevated NfL, evaluate for PSP/CBD with tau biomarkers [60]. For DLB versus AD, α-syn SAA provides superior discrimination. SAA achieved 95.7% sensitivity and 93.2% specificity for DLB [32], while the tau/α-synuclein ratio in CSF showed AUC 0.707–0.867 for separating DLB from AD [106]. Adding NfL and amyloid/tau markers enables identification of mixed pathology cases, with 9.5% of AD patients showing SAA positivity indicating Lewy body copathology and 48–50% of DLB patients showing AD biomarker abnormalities [32, 107]. Addressing diagnostic uncertainty In patients with unclear clinical presentations, biomarkers reduced diagnostic uncertainty. CSF α-syn SAA achieved 84% diagnostic accuracy in parkinsonism of uncertain etiology (sensitivity 75%, specificity 94%), with positive predictive value 93% indicating high confidence when positive [108]. The combination of NfL, sAPPα, and α-synuclein distinguished uncertain parkinsonism cases with AUC 0.95 [83]. Studies consistently found biomarker-confirmed diagnoses showed better long-term stability than clinical diagnoses alone. Follow-up of SAA-tested patients demonstrated 99% had identical results on repeat testing [89], indicating stable molecular signatures. Conversely, some clinically diagnosed PD patients with negative SAA were later determined clinically inconsistent with PD [66], supporting biomarkers' ability to identify misdiagnosed cases. The frequency of SAA positivity in clinically diagnosed DLB ranged from 72 to 96% [32, 89], with SAA-negative cases showing distinct features including less severe parkinsonism, fewer RBD symptoms, and better olfaction [89]. This 28% SAA-negative rate among clinical DLB diagnoses likely includes both false-positive clinical diagnoses and pathologically confirmed cases with insufficient CSF seed concentrations (particularly amygdala-predominant variants) [30]. Longitudinal clinical follow-up, alternative bio-specimen testing, and consideration of mixed pathologies help resolve these discordant cases. Practical implementation considerations Technical and operational requirements α-Synuclein SAA requires specialized equipment including plate readers with precise temperature control and shaking capabilities, recombinant α-synuclein substrate production, thioflavin T fluorescent dye, and automated analysis software [109]. Turnaround time varies from 24 h for optimized protocols [41] to 48 h for standard methods [36, 42], substantially faster than the 5–13-day protocols initially developed [109]. Sample requirements are modest: 7–15 µL CSF volume achieved near-optimal diagnostic accuracy [94], while skin punch biopsies (3–4 mm) provide sufficient tissue [40, 46]. For blood assays, 2 µL serum suffices using ultrasensitive detection [110]. However, sample stability varies, α-synuclein seeds in skin remain stable at − 80 °C but deteriorate with short-term exposure to non-ultra-low temperatures or excessive grinding [41]. Quality control measures are critical given inter-laboratory variability. Universal control fluids enable uniform preparation of sample and control reactions [94], while proficiency testing across expert laboratories demonstrated reproducibility when standardized protocols are followed [66]. Commercial assays validated under CLIA/CAP certification provide regulatory framework [90], though most current testing occurs in research settings. For NfL, highly sensitive immunoassays (Simoa, ELLA, CLEIA) are required [60]. These platforms show high correlation between methods (r ≥ 0.73–0.84 for blood–CSF) [59] and good reproducibility across platforms [58], with coefficient of variation < 10% for quality control samples [62]. Automated analyzers enable rapid processing [60] though absolute concentrations differ between platforms requiring assay-specific cutoffs [111]. Cost-effectiveness and resource allocation While specific cost analyses were not provided, several studies addressed resource considerations. Blood-based NfL assays are described as cost-effective alternatives to CSF sampling [112], avoiding expenses and complications of lumbar puncture. Plasma testing offers potential for primary care screening to identify candidates for specialty referral [59, 111]. Two-step algorithms provide cost savings by targeting confirmatory CSF testing to high-risk individuals. Reducing CSF procedures by 43% while maintaining 94% accuracy [97] translates to substantial cost reductions and reduced patient burden from invasive procedures. For skin biopsies, minimal invasiveness and outpatient feasibility [46] offer advantages over repeated CSF sampling, though tissue processing and RT-QuIC analysis require specialized facilities. Predictive clinical models enable optimization of resource allocation. Models predicting SAA status from clinical variables (AUROC 0.920–0.976) [99] allow identification of patients most likely to benefit from biomarker testing, avoiding unnecessary procedures in very low or very high probability cases. This precision medicine approach maximizes diagnostic yield while minimizing costs. Current clinical applications Threshold-based NfL algorithms require caution in elderly parkinsonian populations. Blood NfL is consistently influenced by age and is also affected by non-neurodegenerative or non-parkinsonian factors that are common in routine clinics, including lower body mass index or blood volume effects, impaired renal function, chronic kidney disease, diabetes and diabetic sensorimotor polyneuropathy, peripheral nerve injury, cerebrovascular disease, cardiovascular comorbidity, recent stroke or trauma and other inflammatory or degenerative neurological disorders [113–128]. Some of these factors can also coexist with parkinsonism and mimic atypical progression [21, 126]. Accordingly, NfL should not be interpreted with a universal numerical threshold. A clinically responsible report should specify the matrix (CSF, plasma or serum), analytical platform, age-adjusted reference range and whether body mass index, renal function and relevant peripheral or vascular comorbidities were considered. A high NfL result should be communicated as evidence of increased neuroaxonal injury rather than as a diagnosis of MSA, PSP or CBS. Conversely, a low or age-appropriate NfL can reduce the probability of aggressive atypical parkinsonism but cannot exclude it when the clinical phenotype is strongly suggestive. Commercial α-synuclein SAA testing is now available to clinicians, with one validation study reporting 87.3% sensitivity and 97.2% specificity for patients with positive DaT SPECT, decreasing to 78.7% sensitivity and 89.5% specificity referencing clinical diagnoses alone [129]. In real-world tertiary clinic use, SAA testing was performed in patients across the cognitive spectrum from subjective impairment to dementia, with varied clinical syndromes including amnestic, dysexecutive, and parkinsonian presentations [100]. Results influenced clinical management through medication changes (including lecanemab consideration in SAA-positive/amyloid-positive cases) and additional diagnostic testing [100]. However, optimal use guidelines remain absent, with testing decisions currently based on individual clinician judgment. Development of formal clinical use guidelines incorporating appropriate patient selection, result interpretation frameworks, and integration with existing diagnostic criteria represents a critical need [20]. CSF NfL testing has achieved broader clinical implementation. The NULISA multiplex panel including NfL among 124 markers enables simultaneous assessment of diverse neurodegenerative targets [130], while Simoa platforms provide standardized plasma NfL measurement [111]. Age-dependent reference ranges are essential, with data-driven cutoffs showing plasma NfL performs best using concentration thresholds to rule out neurodegeneration across age groups [111]. Training and standardization needs Several studies highlighted need for standardization before widespread clinical adoption. Protocol variations in substrate concentration, plate reader settings, and analytical cutoffs contributed to inter-laboratory variability [33], while differences in biospecimen preparation affected reproducibility [95]. Lack of standardized diagnostic criteria for α-synuclein RT-QuIC results [64] and absence of harmonized protocols across research groups [33] impede clinical translation. Ongoing international efforts aim to establish reference ranges and standardized methodologies [101]. For α-synuclein SAA specifically, technical complexity and non-linear assay behavior necessitate specialized expertise [66]. Training needs extend beyond laboratory technique to clinical interpretation, particularly for understanding discordant results, integrating biomarkers with phenotype, and recognizing limitations in specific populations (LRRK2 carriers, amygdala-predominant pathology) [16, 30]. Quality control programs comparable to those for prion disease diagnostics are essential. Proficiency testing, standardized reagents (including universal control fluids [94], harmonized data analysis pipelines, and inter-laboratory comparison studies will enable consistent performance approaching that achieved in expert research settings. Regulatory frameworks for laboratory-developed tests versus commercial assays require clarity [95], with FDA oversight potentially ensuring quality while maintaining accessibility. Communicating SAA and NfL biomarkers results to patients: a stewardship framework Biomarker stewardship must extend beyond test selection. Before testing, clinicians should explain what the biomarker can and cannot answer: SAA detects seeding-competent α-synuclein, not a complete clinical diagnosis, NfL reflects neuroaxonal injury, not a specific proteinopathy, and neither result automatically determines prognosis or treatment. Pre-test counseling should also address potential emotional consequences, incidental or prodromal findings, implications for family discussions and whether the result will influence current management, research eligibility or follow-up intensity. When SAA is positive in an uncertain parkinsonian syndrome, a useful formulation is: “This result supports the presence of misfolded α-synuclein biology, which is commonly seen in Parkinson’s disease and related Lewy body disorders, but the exact clinical diagnosis still depends on your symptoms, examination, imaging and evolution over time”. The clinician should then specify whether NfL, MRI, cognition, autonomic testing or genetics point toward PD, DLB, MSA, mixed pathology, or another disorder with α-synuclein co-pathology. When SAA is negative in a patient who otherwise meets clinical criteria for PD, the result should not be interpreted as invalidating the patient’s diagnosis or symptoms. A balanced explanation is: “The test did not detect α-synuclein seeding in this sample. This lowers the probability of typical Lewy body Parkinson’s disease, but false-negative or biologically synuclein-negative situations are recognized, especially in some genetic or normosmic presentations and in early or anatomically restricted disease”. The practical response is to review pre-test probability, assay type and matrix, olfaction, genetics, DAT imaging and longitudinal clinical evolution rather than abruptly withdrawing a clinically useful diagnosis. When SAA is positive in isolated RBD, hyposmia or another prodromal context, communication should avoid deterministic language. Patients can be told that the results indicate a higher biological risk of developing a synucleinopathy, but it does not predict with certainty whether, when or which syndrome will emerge. Counseling should focus on safety measures for RBD, longitudinal monitoring, research opportunities, management of anxiety and clear documentation that the individual does not necessarily have manifest PD at the time of testing. NfL results require a similarly cautious vocabulary. A high NfL can be described as “a signal of increased nerve-cell or axonal injury” that raises concern for a faster or atypical neurodegenerative process in the appropriate clinical context. It should not be labeled as a positive test for MSA or PSP. The discussion should explicitly mention confounders, such as age, renal function, body mass index, diabetes, peripheral neuropathy, vascular disease, and recent neurological injury, especially when the result is close to a proposed threshold. Challenges and future directions Remaining diagnostic gaps Despite high overall accuracy, several diagnostic challenges persist. Differentiation between PD and DLB using standard α-syn SAA alone remains difficult, with similar amplification kinetics [36]. While clinical features, imaging, and course distinguish these conditions, molecular biomarkers confirming one versus the other are lacking. Quantitative parameters and longitudinal kinetics may eventually provide discrimination [86]. Distinguishing MSA from Lewy body disease requires either optimized MSA-specific SAA protocols [34, 37] or integration with NfL [43] as standard protocols show near-zero MSA sensitivity [36]. Even with modified assays, specificity for type 2 MSA seeds was only 68–87% [34], indicating continued risk of false positives. Development of robust MSA-specific assays remains a priority. Detection of copathologies presents both opportunities and challenges. While α-syn SAA positivity in 9.5% of AD patients [32] and 8–36% of PSP/CBS cases [67, 68] usefully identifies mixed pathology, determining the dominant pathology driving clinical phenotype requires integration of multiple biomarkers. Amyloid/tau biomarkers, NfL, and potentially tau SAA must be interpreted collectively [102, 107]. Technical optimization needs Current SAA methods remain primarily qualitative (positive/negative) despite promising quantitative parameters. Further development of quantification approaches like the LFP score [86], seeded oligomer concentrations [87], and kinetic measures (Nrep, lag phase, Vmax) [64, 88] could enable correlation with disease severity and progression monitoring. Standardization of which parameters to report and how to interpret them requires consensus. For peripheral bio-specimens, technical challenges limit current utility. Olfactory mucosa sampling showed inhibitory effects on aggregation [50], while urine and saliva required substantial optimization for acceptable performance [42]. Sample preparation protocols, handling procedures, and storage conditions need optimization for each matrix [41]. The goal is accessible bio-specimens (blood, saliva) performing comparably to CSF. Harmonization across laboratories is essential before routine clinical use. Despite good reproducibility in expert hands [46, 66], the fair agreement (κ = 0.246) in the four-laboratory DLB study [33] demonstrates current protocol heterogeneity. International working groups should establish consensus protocols, reference materials, quality control procedures, and proficiency testing programs. Integration with emerging technologies Novel platforms may address current limitations. The iRS platform detecting α-synuclein misfolding achieved AUC 0.90 in combined datasets [131], offering complementary information to amplification-based methods by directly measuring conformational changes without requiring competent seed amplification. Multiplex proteomics panels simultaneously measuring 124 proteins including NfL, GFAP, and various neurodegenerative markers achieved AUC > 0.89 for distinguishing clinical diagnoses and SAA status [130]. Integration of biomarkers with advanced imaging shows promise. Tau PET combined with NfL and MRI planimetry improved PSP diagnosis [57], while quantitative susceptibility mapping of iron deposition plus plasma NfL enhanced MSA/PSP differentiation from PD [78]. Machine learning approaches integrating plasma biomarkers, clinical parameters, and radiomic features from MRI achieved AUC > 0.87 for PD versus atypical parkinsonism [82]. Development of biomarker-driven diagnostic criteria represents a paradigm shift from purely clinical definitions. Proposed frameworks incorporate α-syn SAA positivity as a biological anchor for PD diagnosis [67], analogous to amyloid PET/CSF for Alzheimer's disease. Such criteria would enable biology-based patient stratification in clinical trials, identification of pure versus mixed pathology cases, and ultimately precision medicine approaches targeting confirmed molecular abnormalities.