Skin biopsy for phosphorylated α-synuclein in MCI or dementia due to Lewy body disease in a convenience cohort from a subspecialty behavioral neurology practice

Abstract

Background:

Immunohistochemically detected phosphorylated α-synuclein (p-α-syn) in cutaneous nerves has been found in >95% of patients meeting consensus criteria for dementia with Lewy bodies (DLB); however, its performance in prodromal stages (mild cognitive impairment-Lewy body, MCI-LB) and clinical practice is unknown.

Objectives:

We evaluated skin biopsy for p-α-syn in a convenience cohort of patients with suspected MCI-LB or DLB in a tertiary care setting.

Methods:

53 patients underwent commercially available skin biopsy for skin nerve p-α-syn as part of a clinical evaluation for suspected MCI-LB/DLB. Consensus MCI-LB/DLB criteria blinded to skin biopsy result were applied through retrospective chart review; those not definitively meeting criteria were considered questionable synucleinopathy (QS).

Results:

Median age at skin biopsy was 70 years (range 43–87) with 81% men. Thirty-six patients (68%) had positive skin biopsies. 32/43 (74%) patients meeting clinical consensus criteria for DLB/MCI-LB had a positive skin biopsy, with higher frequency in MCI-LB (87%) compared to DLB (60%). Four of ten (40%) QS had positive biopsies. One patient with a positive skin biopsy had no evidence of CNS α-synuclein on autopsy.

Conclusion:

Positive skin biopsy for p-α-syn occurs across the spectrum of cognitive presentations of Lewy body disease including prodromal DLB, albeit at a lower frequency in this retrospective review of atypical cases compared to typical clinical presentations. A positive skin biopsy may not reflect CNS synucleinopathy as the etiology for a patient’s symptoms in some cases. Longitudinal studies of skin biopsies for p-α-syn in suspected MCI-LB/DLB with autopsy confirmation are needed.

INTRODUCTION

Dementia with Lewy bodies (DLB) and its prodromal form (mild cognitive impairment-Lewy body, MCI-LB) demonstrate substantial clinical heterogeneity and overlap with other neurologic diseases which leads to diagnostic uncertainty. Given the variability in presentation of MCI-LB, psychiatric or delirium-onset manifestations have been proposed, and are especially challenging to diagnose as patients often initially lack the core features of DLB: parkinsonism, REM sleep behavior disorder (RBD), recurrent fully-formed visual hallucinations, and cognitive fluctuations.¹ Commonly employed biomarkers for DLB such as the cingulate island sign on brain ¹⁸F-FDG-PET or presynaptic dopaminergic deficiency measured by ¹²³I-ioflupane SPECT have somewhat variable sensitivity and specificity for DLB diagnosis depending on image acquisition techniques and analytical methods.^2–4 In MCI-LB, with¹²³I-ioflupane SPECT has a sensitivity of 54.2% for combined possible and probable MCI-LB while a higher medial temporal to substantia nigra ratio on FDG-PET is >90% sensitive and >80% specific in differentiating MCI-LB from MCI-AD and healthy controls.^{5, 6} Clinical diagnostic criteria are available to diagnose DLB and MCI-LB, but demonstrate a variable sensitivity of 90–91% and specificity of 73–94% for DLB and MCI-LB, respectively.^{1, 7–9} Reliable biomarkers of α-synuclein pathology remain a significant unmet need to advance diagnostic certainty in MCI-LB/DLB, especially in atypical or complex presentations.

There are two major emerging biomarkers of pathogenic α-synuclein for suspected synucleinopathies: (1) the presence of phosphorylated α-synuclein (p-α-syn) within skin nerve fibers as demonstrated by immunohistochemistry and (2) oligomeric α-synuclein within CSF or other biosamples able to aggregate in vitro using synuclein seed amplification assays (SAA).¹⁰ Early autopsy-based studies showed p-α-syn within abdominal dermal nervous tissue in 40% of DLB cases and 70% of PD cases while more recent studies have reported p-α-syn in skin biopsies of 100% of patients with clinically diagnosed DLB.^11–13 In contrast, p-α-syn in skin biopsies of normal controls is nearly universally absent, including in elderly individuals who may be suspected to have incidental Lewy body pathology suggesting both a high sensitivity and specificity of p-α-syn in DLB.^{14, 15} In a large multi-center study utilizing a commercially available skin biopsy, presence of p-α-syn within intra nerve fibers was found in 96% of those with DLB and 3% of those who did not meet clinical criteria for a synucleinopathy.¹⁶ However, p-α-syn has not been evaluated in prodromal DLB and secondary analyses of a recent study suggests a high frequency of positive test results in cases where the ultimate diagnosis was not clear.¹⁶ Therefore, we explored the real-world performance of skin biopsy for p-α-syn in an academic behavioral neurology practice, specifically focusing on clinically suspected MCI-LB and DLB.

METHODS

Cohort Identification

Patients seen in the Behavioral Neurology Division at Mayo Clinic Rochester and who underwent clinical skin biopsy for α-synuclein testing were identified by searching the medical record database for the lab test code for p-α-syn, with dates of skin procedure ranging between September 2022 and October 2024.

Chart Review and Clinical Variable Abstraction

Clinical data were abstracted through retrospective chart review including presenting symptom(s), neurologic examination findings, imaging results, and laboratory testing. Clinical data closest to the time of skin biopsy were recorded. The presence of core DLB features was determined by explicit documentation and impression of the treating neurologist; when it was not explicitly stated, the determination of whether core features were present or absent was based on available documentation attempting to mirror the descriptions laid out in the most recent DLB criteria.⁹ For RBD, patients had probable RBD (pRBD) with a history of recurrent dream enactment behaviors (≥ 3 discrete episodes) in the absence of polysomnography (PSG), while those who underwent PSG that demonstrated REM sleep without atonia were considered to have PSG-proven RBD. Patients with a history of dream enactment but no evidence for REM sleep without atonia on PSG were considered to not have RBD. The most recent consensus clinical diagnostic criteria for MCI-LB, DLB, Parkinson’s disease (PD), PD-MCI, PD dementia (PDD), and multiple system atrophy (MSA) were applied to each patient by MJR and two subspeciality certified behavioral neurologists (SJM, HB), blinded to skin biopsy or other biomarker result.^{1, 9, 17–20} As all patients included had some clinical suspicion for MCI-LB or DLB by their treating neurologist, a diagnosis of questionable synucleinopathy (QS) was applied if there was consensus amongst the blinded reviewers that the patients did not definitively meet criteria for MCI-LB/DLB or had an alternative explanation for their core clinical feature (e.g. dream enactment behavior in the setting of untreated moderate or severe sleep disordered breathing that could cause “pseudo RBD”, one or two lifetime episodes of parasomnia behavior, or unchanged mild abnormality on longitudinal ¹²³I-ioflupane SPECT).²¹ Finally, longitudinal outcomes were abstracted from medical record review when available, including the treating neurologist’s final impression on the patient’s most likely diagnosis that took into account the result of the skin biopsy and any other clinical information; however, this information was not utilized for the initial application of consensus clinical diagnostic criteria.

Additional specific variables of interest collected but were not included in the determination of the consensus clinical diagnosis included cerebrospinal fluid Alzheimer’s disease biomarkers, autoimmune encephalopathy evaluations, and CSF α-synuclein seed amplification assay (SAA, SAAmplify^™-αSYN test, Amprion) when these were performed. The presence of visually determined cingulate island sign on ¹⁸F-FDG-PET was recorded.² We also performed exploratory analysis with our multi-label unsupervised machine learning algorithm on FDG-PET scans (StateViewer) which utilizes a K nearest neighbor model to identify similar FDG-PET scans of patients with known neurodegenerative disease.²²This produces a log odds ratio (OR) for the likelihood of a specific diagnosis (in this case the log OR for DLB) based on brain metabolism pattern. An abnormal ¹²³I-ioflupane SPECT was determined by visual and semiquantitative analysis using DaTQUANT^™ (GE Healthcare, Chicago, IL). The lowest putamen striatum binding ratio (SBR) z-score value was recorded with a z-score off < −1.0 considered abnormal for categorical determination of abnormal ¹²³I-ioflupane SPECT.⁴

Skin biopsy

All patients underwent clinical skin biopsy according to Syn-One^® manufacturer instructions (CND Life Sciences, Scottsdale, AZ) with 3 mm punch biopsies at the left posterior cervical (PC; 3 cm lateral to C7 vertebrae), left distal thigh (DT; 10 cm above lateral knee), and left distal leg (DL; 10 cm above lateral malleolus).²³ Data were extracted from the clinical diagnostic reports prepared by CND Life Sciences including presence or absence of p-α-syn deposition and skin biopsy location(s) if present.

Statistics

Data were analyzed using GraphPad Prism v10.0.0 (Dogmatics, Boston, MA). Descriptive statistics are reported either as median (range) or mean (standard deviation) for continuous variables and number (percentage) for categorical variables. Primary outcomes of interest were skin biopsy results by whether patients met blinded consensus criteria of a synucleinopathy (MCI-LB or DLB). Exploratory analyses were performed in subgroups defined by skin biopsy result. Exploratory comparisons of objective biomarkers of LBD (visually determined cingulate island sign and StateViewer log OR for DLB on FDG-PET, DatQUANT z-score) were also compared by skin biopsy result. In individuals who underwent both skin biopsy and CSF SAA for synuclein, primarily descriptive statistics were utilized apart from Cohen’s κ to assess agreement between skin biopsy and CSF SAA.²⁴ Group differences for continuous variables were determined by student’s t-test (two groups) or one-way ANOVA (three groups) with Tukey’s multiple comparison test used when ANOVA was significant. Categorical variables were compared with Fisher’s exact test or χ² test. Linear regression was used to assess relationships between continuous variables using the ordinary least squares method. Regression p values were those calculated under null hypothesis that the line slope β₁ = 0. No outlier exclusion, data imputation for missing values, or correction for multiple comparisons was performed. Significance was set as α < 0.05 using two-sided testing.

Standard Protocol Approvals, Registrations, and Patient Consents

The Mayo Clinic Institutional Review Board approved this study, and participating patients (or their legally authorized representatives) provided written consent to use their medical information for research purposes.

Data Sharing

All relevant data have been shared and published in this article.

RESULTS

Demographic and Clinical Variables

Fifty-three patients were included with a median age of 70 years old (range 43–87) at time of skin biopsy. Clinical consensus criteria at the time of skin biopsy were as follows: possible DLB (n = 6), probable DLB (n = 13), possible Parkinsons’s disease dementia (PDD) (n = 1), possible MCI-LB (n = 8), probable MCI-LB (n = 14), PD-MCI (n = 1), and questionable synucleinopathy (QS, n = 10). PDD and PD-MCI were grouped with DLB and MCI-LB, respectively, since these entities likely represent a spectrum of the same underlying pathology. Full demographic and clinical data can be found in Table 1. Male sex was less frequent in the DLB group (60%) than QS (90%) and MCI-LB (96%) (p = 0.009). Bedside cognitive testing scores were lower in patients with DLB than both MCI-LB and QS patients (p <0.001). The total number of core DLB clinical features, visual hallucinations, parkinsonism, and pRBD were more common in DLB and MCI-LB than QS (p <0.001, p = 0.008, p = 0.033, p = 0.049, respectively). Of the 28 patients (53%) who underwent ¹²³I-ioflupane SPECT, minimum putamen SBR z-score was lower in MCI-LB compared with DLB (p = 0.040). Thirty-six of 53 patients (68%) had longitudinal data available for secondary analysis with median follow up time of 16 months (range 4 to 31 months).

Table 1.

Comparison of clinical characteristics of study cohort by clinical consensus criteria

	QS	MCI-LB#	DLB&	Total	P value
Number of patients	10	23	20	53
Age in years, median (range)	70 (53–83)	69 (62–83)	74 (43–87)	70 (43–87)	0.475
Sex, male	9 (90%)	22 (96%)	12 (60%)	43 (81%)	0.009
Kokmen STMS	33 (4)	33 (3)	27 (5)	31 (5)	<0.001
MoCA equivalent mean score to STMS30	22	22	17	21
18 FDG-PET brain
Cingulate island sign	2 (20%)	12 (53%)	7 (35%)	21 (40%)	0.192
StateViewer OR of DLB	0.24 (0.89)	1.15 (1.51)	0.83 (1.37)	0.87 (1.38)	0.253
123 I-ioflupane SPECT brain
Abnormal by visual inspection	1/6 (17%)	6/11 (55%)	3/11 (27%)	10/28 (36%)	0.224
Putamen SBR z-score < −1	1/5 (20%)	7/11 (64%)	1/9 (11%)	9/25 (36%)	0.037
Minimum putamen SBR z-score	0.00 (1.44)	−1.69 (1.42)	−0.02 (1.41)	−0.75 (1.60)	0.026
Positive AD biomarker (CSF)	0/7 (0%)	1/13 (8%)	7/12 (59%)	8/32 (25%)	0.003
Core DLB features
pRBD	3 (30%)	17 (74%)	10 (50%)	30 (57%)	0.049
PSG-RBD	0/5 (0%)	7/20 (35%)	5/14 (36%)	12/39 (31%)	0.280
Parkinsonism	1 (10%)	11 (48%)	12 (60%)	24 (45%)	0.033
Recurrent visual hallucinations	1 (10%)	4 (17%)	11 (55%)	16 (30%)	0.008
Cognitive fluctuations	3 (30%)	8 (35%)	10 (50%)	21 (40%)	0.469
Sum of core DLB features (range 0–4)	0.5 (0.7)	1.3 (0.8)	1.9 (0.9)	1.4 (1.0)	<0.001
Skin biopsy
Positive result	4 (40%)	20 (87%)	12 (60%)	36 (68%)	0.019
Number of positive sites*	1.3 (0.5)	1.5 (0.7)	1.3 (0.7)	1.4 (0.6)	0.806
PC site positivity*	2/4 (50%)	17/20 (85%)	10/12 (83%)	29/36 (81%)	0.260
DT site positivity*	1/4 (25%)	7/20 (35%)	3/12 (25%)	11/36 (31%)	0.811
DL site positivity*	2/4 (50%)	5/20 (25%)	3/12 (25%)	10/26 (28%)	0.575

Positive result=any single skin biopsy site with phosphorylated alpha synuclein

Abbreviations: QS, questionable synucleinopathy, did not meet published clinical consensus criteria for a synucleinopathy at time of skin biopsy although had symptoms potentially suggestive of a synucleinopathy or had an alternative explanation for symptoms. MCI-LB, Mild cognitive impairment-Lewy body, both possible and probable MCI-LB and one participant with Parkinson’s disease-Mild Cognitive Impairment. DLB, Dementia with Lewy bodies, both possible and probable DLB and one participant with Parkinson’s disease dementia. STMS, short test of mental status. MoCA, Montreal Cognitive Assessment, FDG PET, fluorodeoxyglucose positron emission tomography. OR, odds ratio. SPECT, Single Photon Emission Computed Tomography. SBR, striatum binding ratio. AD, Alzheimer’s disease. pRBD, probable REM behavior disorder. PSG-RBD, REM behavior disorder diagnosed by polysomnogram. PC, posterior cervical. DT, distal thigh. DL, distal leg.

^*Only for patients with biopsies positive for phosphorylated α-synuclein at any site

^&One patient diagnosed with Parkinson’s disease dementia was included in the DLB group

^#One patient diagnosed with PD-MCI was included in the MCI-LB group

Skin Biopsy Results

Skin biopsy was positive for p-α-syn in one or more of the three biopsy sites in 20/23 (87%) of the MCI-LB group, 12/20 (60%) of the DLB group, and 4/10 (40%) of the QS group with an overall positivity rate of 32/43 (74%) in those meeting consensus criteria for DLB or MCI-LB (p = 0.019). There was no difference in p-α-syn positivity in those with possible MCI-LB or DLB vs those with probable MCI-LB or DLB (73% vs 75%, p > 0.999). The posterior cervical site was the most common site of p-α-syn positivity with similar site positivity across QS and the MCI-LB/DLB groups (Table 1, Supplementary Figure 1).

Comparisons of clinical and biomarker characteristics by skin biopsy result (n = 36, positive, n = 17 negative) are shown in Table 2. There was a trend for a higher proportion of patients meeting clinical criteria for synucleinopathy in those with a positive skin biopsy vs. those having negative skin biopsy (32/36 (89%) vs 11/17 (65%), p = 0.058), but otherwise there were minimal differences between groups. Visually determined cingulate island sign on FDG-PET was more common in those with a positive skin biopsy (50% vs 18%, p = 0.036). Further, there was a positive relationship between the StateViewer determined log OR of DLB metabolic pattern on FDG-PET and the number of positive skin biopsy sites, likely driven by a small group of individuals with all 3 biopsy sites being positive (p = 0.004) (Supplementary Figure 2). CSF average total nucleated cell (TNC) count was lower in those with positive skin biopsy (1 vs 4 TNC, p = 0.002), although this was not clinically significant.

Table 2.

Comparisons of clinical data by skin biopsy result

	Skin biopsy result		P value
	Negative	Positive
Number of patients	17	36
Age, years	67 (11)	72 (9)	0.098
Sex, male	13 (77%)	30 (83%)	0.709
Kokmen STMS	30 (5)	31 (5)	0.459
MoCA equivalent mean score to STMS30	20	21
18 FDG-PET Brain
Cingulate island sign	3 (18%)	18 (50%)	0.036
StateViewer Log OR of DLB	0.61 (0.89)	0.97 (1.53)	0.420
123 I-ioflupane SPECT brain
Abnormal by visual inspection	3/9 (33%)	7/19 (37%)	>0.999
Putamen SBR z-score < −1	2/8 (25%)	7/17 (41%)	0.661
Minimum putamen SBR z-score	−0.55 (0.92)	−0.85 (1.86)	0.670
CSF
Total nucleated cell count [n]	4 (3) [12]	1 (1) [22]	0.002
Protein [n]	70 (31) [13]	71 (33) [22]	0.890
p-tau181 [n]	24.3 (24.4) [9]	19.7 (12.7) [21]	0.498
Total tau [n]	250 (217) [9]	203 (107) [21]	0.428
Ab42 [n]	1184 (595) [9]	1089 (460) [21]	0.636
p-tau181/Ab42 [n]	0.026 (0.033) [9]	0.021 (0.017) [21]	0.581
Positive AD biomarker	2/9 (22%)	6/23 (26%)	>0.999
Core DLB features
pRBD	7 (41%)	23 (64%)	0.119
PSG-RBD	5/11 (46%)	7/28 (25%)	0.262
Parkinsonism	6 (35%)	18 (50%)	0.384
Recurrent visual hallucinations	7 (41%)	9 (25%)	0.337
Cognitive fluctuations	6 (35%)	15 (42%)	0.768
Sum of features (range 0–4)	1.4 (1.0)	1.4 (0.9)	0.936
COVID-19
History of infection prior to biopsy	4 (24%)	11 (31%)	0.748
Synucleinopathy by clinical criteria	11 (65%)	32 (89%)	0.058

Abbreviations: STMS, short test of mental status. Montreal Cognitive Assessment, FDG PET, fluorodeoxyglucose positron emission tomography. OR, odds ratio. SPECT, Single Photon Emission Computed Tomography. SBR, striatum binding ratio. AD, Alzheimer’s disease. pRBD, probable REM behavior disorder. PSG-RBD, REM behavior disorder diagnosed by polysomnogram.

Correlation Between CSF SAA and P-α-syn Skin Biopsy

Eleven patients underwent CSF α-synuclein SAA and skin biopsy as part of their diagnostic evaluation. Of those who did not meet clinical consensus criteria for synucleinopathy, 1 of 4 (25%) had a positive SAA, compared with 3 of 7 (43%) for those meeting criteria for MCI-LB/DLB. Seven of the 11 (64%) patients had concordant results on the skin biopsy for p-α-syn and CSF SAA (shown as grey rows in Table 3), resulting in fair agreement (Cohen’s κ = 0.290; 95% CI: −0.233 to 0.813).

Table 3.

Comparison of CSF α-synuclein seed amplification assay vs skin biopsy for phosphorylated α-synuclein in 11 patients

Clinical consensus criteria#	Skin biopsy	CSF SAA	Final diagnosis§
QS	−	+	MCI-LB
QS	+	−	MCI-LB
QS	−	−	MCI-LB
QS	−	−	bvFTD
Possible MCI-LB	+	−	CBS from CBD*
Possible MCI-LB	+	+	MCI-LB + AD
Possible MCI-LB	+	+	MCI-LB
Probable MCI-LB	+	+	MCI-LB
Possible DLB	−	−	DLB
Probable DLB	−	−	Seronegative AE
Possible PDD	+	−	GFAP-IgG astrocytopathy

Positive skin biopsy was defined by at least one of three sites staining for phosphorylated a-synuclein within intra- nerve fibers. Grey rows indicate concordant results between the two tests.

Abbreviations: QS, questionable synucleinopathy. MCI-LB, mild cognitive impairment-Lewy body. DLB, Dementia with Lewy bodies. PDD, Parkinson’s disease dementia. SAA, seed amplification assay. bvFTD, behavioral variant frontotemporal dementia. CBS, corticobasal syndrome. AD, Alzheimer’s disease. CBD, corticobasal degeneration. AE, autoimmune encephalitis. GFAP, glial acidic fibrillary protein.

^#Whether patients met clinical consensus criteria for MCI-LB/DLB by blinded consensus chart review at time skin biopsy was obtained

^§Via longitudinal outcome data based on final clinical impression of treating neurologist after integration of all available data, including both skin biopsy and CSF SAA results.

^*Confirmed on autopsy.

Longitudinal Discrepant Outcomes

In the QS group, two of four patients with a positive skin biopsy were not thought to have a synucleinopathy over longitudinal follow up. One had a resolution of spells, parasomnia behaviors, and cognitive complaints with treatment of OSA and anxiety and was ultimately thought to have a functional neurocognitive disorder with pseudo-RBD while the other case had a single lifetime episode of dream enactment but normalization of cognitive symptoms after treatment of depression. Two cases meeting criteria for MCI-LB/DLB with a negative skin biopsy were still thought to have DLB or MCI-LB at longitudinal follow up.

Three cases initially meeting criteria for MCI-LB/DLB were ultimately diagnosed with a non-synucleinopathy by their neurologist based on longitudinal follow-up. One patient met criteria for PDD but was found to have GFAP-IgG astrocytopathy, had negative CSF synuclein SAA, and improved with immunosuppression. Another met criteria for probable MCI-LB with a prominent neuropsychiatric syndrome and was subsequently diagnosed with DPPX-IgG autoimmune encephalitis with some improvement in cognition after IVIG and rituximab. Finally, a male in his 60s with a history of adult-onset bipolar disorder requiring ECT was evaluated for several years of visuospatial difficulties, executive dysfunction, worsening impulsivity, and mild cognitive fluctuations with exam demonstrating very mild parkinsonism and left sided apraxia. He did not have visual hallucinations nor develop dream enactment behavior. Initial FDG-PET brain imaging demonstrated occipitoparietal hypometabolism. He demonstrated marked cognitive improvement with donepezil. Skin biopsy was positive for p-α-syn in the posterior cervical and distal thigh, and CSF SAA was negative. Based on clinical features, imaging biomarkers and skin biopsy result he was diagnosed with corticobasal syndrome due to suspected Lewy body disease. Ultimately, his autopsy demonstrated corticobasal degeneration (CBD) without additional degenerative co-pathologies. Specifically, there was no evidence for CNS synuclein (Figure 1).

Figure 1. Histopathologic slides of a patient with cortical basal degeneration who met clinical consensus criteria for MCI-LB with a positive skin biopsy without evidence of α-synuclein pathology on autopsy.

(A) Schematic diagram showing location of histologic sections in subsequent panels. (B) Right parietal cortex stained for tau (AT8) with astrocytic plaques (thin arrow), neuronal cytoplasmic inclusions (thick arrow), and neuropil threads (arrow head). (C) CA3 region of the left hippocampus stained for 4R tau (RD4) with neuronal cytoplasmic inclusions (thick arrow). (D-H) Absent α-synuclein (p-S129) staining in the left amygdala (D), left anterior cingulate (E), locus coeruleus (F), dorsal motor nucleus of the vagus (G), and olfactory bulb (H). All scale bars shown are 10 μm.

DISCUSSION

Our single center experience of patients undergoing clinical skin biopsy to assess for phosphorylated α-synuclein as part of their diagnostic evaluation for suspected Lewy body disease suggests a skin biopsy can be positive across the cognitive continuum of LBD including prodromal DLB, although we found lower positive rate in DLB than previously reported.¹⁶ Our data raise important considerations about the potential for false positive results that could influence provisional antemortem diagnosis and management of patients with cognitive disorders attributed to synucleinopathy. Given the complexity and heterogenous presentations of evolving LBD, which often include a background of psychiatric disease, nonspecific symptoms, or discrepancies between self-reported cognitive symptoms and objective cognitive performance, an accurate and reliable biomarker to distinguish an underlying synucleinopathy etiology is of considerable importance.^{1, 25, 26}

For contextual interpretation of our results, published data on clinically available skin biopsy for p-α-syn, particularly in those with DLB, is restricted to cross sectional analysis of clinically diagnosed participants without longitudinal follow-up data. More importantly, there is not published data with autopsy-proven cases. In this context, our data demonstrate that p-α-syn is present during both the prodromal and clinically overt DLB stages. We found that p-α-syn was present in 87% of patients who met criteria for MCI-LB, which has not been previously been described. However, we observed a much lower positivity rate in DLB of 60% compared with other reports of 96–100% positivity in DLB patients with typical clinical presentations.^{11–13, 16}

Several factors could explain the differences between our and previous studies. First, Gibbons et al. applied rigorous inclusion criteria of patients recruited from multiple centers including community-based practices in their primary analysis.¹⁶ Thus, patients in that study were likely to represent classic phenotypical presentations of DLB and this test may have particular relevance in classical presentations. In contrast, our study is a clinical convenience cohort from a tertiary care subspecialty cognitive neurology clinic with skin biopsies being performed at the discretion of the treating neurologist who may have been more likely to perform the skin biopsy in cases where diagnostic uncertainty was higher or symptoms were less convincing. Additionally, our DLB cohort comprised a higher percentage of females (40%) and lower rate of probable RBD (36%) than most large DLB cohorts, supporting the notion our DLB patients may be overrepresented by atypical presentations.²⁷ Our DLB cohort also had a high frequency of positive AD biomarkers which is common in DLB patients, influences the clinical presentation of DLB and may also explain some of the more atypical features of our DLB cohort.^{28, 29} A prior autopsy study of CSF SAA demonstrated a low sensitivity of CSF SAA for amygdala-predominant Lewy bodies, which is associated with more advanced AD pathology.³⁰ Given the limited autopsy confirmed studies with p−α-syn, whether skin biopsy is similarly insensitive to more anatomically restricted distributions of Lewy body pathology is unknown, but may also explain lower skin biopsy positivity in our DLB subgroup. Furthermore, female DLB patients demonstrate different temporal evolution of core clinical features than male DLB patients, which may contribute to diagnostic uncertainty and may partially explain demographic differences and biopsy positivity rates seen between MCI-LB and DLB subgroups in our study compared with other cohorts of DLB patients.³¹ Finally, given our cohort comprises a convenience cohort of patients with complex and atypical presentations without autopsy confirmation in all but 1 case, misdiagnosis is possible.

Other prior skin biopsy studies of DLB patients have included a high proportion of individuals with substantial autonomic dysfunction, which could also potentially explain differences in positive skin biopsy rates. Autonomic symptoms were often relatively minor compared with other symptoms in our cohort, a reflection of the atypical nature of the cohort study and the selection bias of our cognitive clinic.^{12, 13} The “brain vs. body first” hypothesis of synuclein progression posits that α-syn deposition begins in the periphery and ascends to the CNS, thus those with more autonomic dysfunction may be more likely to demonstrate p−α-syn on skin biopsy.^{32, 33} Given the some of the atypical features of our DLB cohort it is conceivable that our cohort could also comprise more “brain first” DLB patients who may be less likely to have positive skin biopsies. Importantly, there was no difference in abnormal biopsy frequency when comparing possible vs probable MCI-LB/DLB, suggesting misdiagnosis by lumping possible and probable MCI-LB/DLB cases in our analysis did not influence our results.

We had a high frequency of positive p-α-syn biopsies in patients who had a questionable synucleinopathy (40%). This frequency is similar to the Gibbons et al. secondary analysis that showed 32/57 (56%) of reclassified cases to “unknown”diagnosis, in which they could not rule out another contributor to symptoms or had abnormal Montreal cognitive assessment score, had a positive skin biopsy.¹⁶ Given our case of corticobasal syndrome due to autopsy-proven corticobasal degeneration with a positive skin biopsy but without CNS α-synuclein, two biopsy-positive QS cases with resolution of symptoms after treatment of mood disorders or sleep apnea, and two cases of positive skin biopsies who were ultimately determined to have antibody-positive autoimmune encephalopathy, false positive results with this test are at least possible and longitudinal autopsy data are needed to better understand this test’s specificity. Positive results in these selected cases may represent an evolving body-first progression of synuclein deposition or may have identified incidental Lewy body pathology associated with aging.^{15, 32} However, at least as it pertains to our autopsy case who had cortical dysfunction with an absence of cortical or even subcortical synuclein, a positive skin biopsy could not have explained this patient’s symptoms. Since specific proteinopathy-directed treatments are being actively investigated in several clinical trials, clarifying such limitations of antemortem biomarkers is crucial.

A small number in our cohort underwent both skin biopsy and CSF SAA testing. We found fair agreement between the skin biopsy and CSF SAA (Cohen’s κ = 0.290), which is comparable to one study which showed immunofluorescence for p-α-syn in skin biopsies of iRBD patients had only fair agreement with CSF synuclein SAA (κ= 0.3; p < 0.01) but much lower than another study of all synucleinopathies.^{10, 34} These studies also suggest immunofluorescence for p-α-syn is more sensitive and specific than CSF SAA across the synucleinopathy spectrum. In our series, 3/11 (27%) patients had a positive skin biopsy with negative CSF SAA compared with only 1/11 (9%) who had positive CSF SAA and negative skin biopsy, suggesting p-α-syn may be more frequently positive than CSF SAA and differences in discrepant test positivity between skin biopsy and CSF could be explained by patients following a “body first” trajectory. Given our small number of participants with both skin biopsy and CSF SAA no definitive conclusions can be drawn from our cohort and additional data from larger series of pathologically proven synucleinopathy are necessary to determine the comparative diagnostic performance of p-α-syn skin biopsy and CSF SAA.

Patients with positive skin biopsy were more likely to have a cingulate island sign on FDG-PET (50% vs. 18%). There was also positive association between number of positive p-α-syn sites and higher log OR for DLB brain metabolism pattern as determined by StateViewer. There were no differences in abnormal visual or semiquantitative dopaminergic dysfunction measured by¹²³I-ioflupane SPECT based on skin biopsy result. This is likely due to the bias of cognitively impaired patients in our cohort. However, this also supports agreement in objective markers for LBD and suggests considering a combination of biomarkers, particularly in atypical cases, to improve diagnostic certainty. Finally, there was no difference in the frequency of AD copathology based on skin biopsy result which has important clinical implications given DLB and AD frequently co-occur and influence the evolution of clinical symptoms in DLB.^{29, 35}

Our study has several limitations. Our cohort is a relatively small retrospective convenience sample from a tertiary referral center which is likely biased towards atypical presentations of neurologic disease which limits the generalizability of our findings to more classic presentations of MCI-LB or DLB in community neurology settings. Further, skin biopsies were performed at the discretion of the treating neurologist who had some clinical suspicion for underlying LBD which limits its generalizability to those presenting with undifferentiated cognitive symptoms. Additionally, it is likely that individuals with more complex or undifferentiated presentations were more likely to undergo biopsy than cases with high diagnostic certainty. Consensus diagnoses were made based on documented signs and symptoms in the medical record which could have influenced our results and given the atypical features of our DLB cohort in particular the possibility of misdiagnosis is possible. We did not have consistently reported reliable data on hyposmia, which is strongly associated with CNS synucleinopathy. As skin biopsies for p-α-syn have only recently been implemented in our institution, we had relatively short duration longitudinal follow-up data, which is especially important in those patients with potential false positive or negative results to better understand the significance of these findings. Further longitudinal studies with autopsy confirmation are necessary to further characterize the accuracy of skin biopsies for p-α-syn.

CONCLUSION

Skin biopsy for phosphorylated α-synuclein can be positive at any point along the Lewy body disease spectrum in patients with cognitive manifestations, including those with prodromal DLB. Our single autopsy case of skin biopsy positivity for p-α-syn without CNS Lewy body pathology indicates that a positive skin biopsy does not necessarily indicate a CNS synucleinopathy to explain clinical symptoms. Further longitudinal clinical studies of various phenotypic presentations of MCI-LB/DLB with autopsy confirmation are necessary to understand the diagnostic utility of skin biopsies for p- α-syn.