Performance of ¹²³I‐ioflupane SPECT striatal binding in dementia with Lewy bodies

Abstract

INTRODUCTION

We aimed to determine whether ¹²³I‐ioflupane single‐photon emission computed tomography (SPECT) striatal binding ratio (SBR) correlated with parkinsonian motor symptoms in dementia with Lewy bodies (DLB) and if SBR predicts worsening of parkinsonism over time.

METHODS

A retrospective cohort study of the U.S. Dementia with Lewy Bodies Consortium dataset including individuals with DLB with baseline ¹²³I‐ioflupane SPECT analyzed with DaTQUANT and baseline and 24‐month Movement Disorder Society Unified Parkinson's Disease Rating Scale–Part III (MDS‐UPDRS‐III). A subset had cerebrospinal fluid α‐synuclein seed amplification assay (SAA) evaluation.

RESULTS

Baseline mean SBRs were significant predictors of baseline and 24‐month MDS‐UPDRS‐III scores, although they did not predict meaningful worsening over time. SAA positivity was associated with lower SBRs; Z score cut‐off values are provided.

DISCUSSION

In suspected DLB, ¹²³I‐ioflupane SPECT, at diagnosis, could be used to confirm underlying dopamine deficiency; it does not predict meaningful worsening of motor parkinsonism. More severe dopamine deficiency increases confidence in presence of synucleinopathy.

Highlights

• ¹²³I‐ioflupane single‐photon emission computed tomography (SPECT) can confirm underlying dopamine deficiency.

• Striatal binding ratio (SBR) Z scores predicted 24‐month Unified Parkinson's Disease Rating Scale–Part III (UPDRS‐III) scores.

• SBR Z scores are not predictive of subsequent meaningful worsening of parkinsonism.

• More severe dopamine dysfunction on SPECT is associated with presence of seed amplification assay (SAA).

• SBR Z score cut‐offs that indicate cerebrospinal fluid SAA positivity are provided.

1. BACKGROUND

The criteria for diagnosis of dementia with Lewy bodies (DLB) remains predominantly clinical. ¹ Parkinsonian motor symptoms are one of the four core clinical features of DLB ¹ and are seen in 85% of patients. ² Since 2007 [¹²³I] N‐x‐flouropropyl‐2b‐carbomethoxy‐3b‐(4‐iodophenyl) nortropane (¹²³I‐ioflupane, “DaTscan” ³ ) single‐photon emission computed tomography (SPECT) scans have been used to aid in diagnosis by confirming dopamine deficiency as the underlying cause of motor parkinsonism in DLB patients. ^{4 123}I‐ioflupane is a radioligand that binds to the dopamine transporter (DAT), a protein that is present on the presynaptic dopaminergic neurons of the substantia nigra and other dopaminergic pathways. ³ With SPECT, the uptake of the radioligand on dopaminergic neuronal processes in the striatum can be visualized, ³ thereby helping to discriminate patients with neurodegenerative parkinsonism from those with other forms of clinical parkinsonism. ⁵ , ⁶ Further study of the clinical applications of ¹²³I‐ioflupane SPECT found a specificity of 90% and a sensitivity of 77% in the differentiation of DLB from non‐DLB dementia, often Alzheimer's disease (AD). ^{4 123}I‐ioflupane SPECT has also been compared to the gold standard of pathologic criteria on autopsy and was found to be both sensitive (80%) and specific (92%). ⁷ The utility of DAT imaging as either a screening or confirmatory test, or as a marker of either disease progression or response to therapy in DLB, remains unclear.

¹²³I‐ioflupane SPECT is most often interpreted by visual inspection. ⁸ Software, such as DaTQUANT, developed by GE Healthcare in 2013, provides semi‐quantitative measurements of dopamine uptake in the basal ganglia relative to standard reference regions. ⁹ These software have been found to improve diagnostic accuracy and reproducibility of ¹²³I‐ioflupane SPECT interpretation. ¹⁰ , ¹¹ , ¹² A benefit of the DaTQUANT program is its ability to compare individual patient images to a normal age‐matched population. ⁹ The software generates an array of data (Figure 1) obtained from the initial ¹²³I‐ioflupane SPECT images, including striatal binding ratios (SBRs) from which Z scores are derived based on age‐matched controls. ⁹

FIGURE 1.

Screenshots of DaTQUANT software. A, Examples of scans with striatal and reference region outlines. B, High average SBR Z score = 1.71 and (C) low average SBR Z score = –3.755. SBR, striatal binding ratio

The Movement Disorder Society Unified Parkinson's Disease Rating Scale (MDS‐UPDRS) is a four‐part rating scale which addresses a multitude of clinical features seen in Parkinson's disease. ¹³ Part III of the MDS‐UPDRS (MDS‐UPDRS‐III) is a physician‐scored rating scale of parkinsonian findings on physical examination. ¹³ Because of the frequent parkinsonian features observed in DLB patients, the MDS‐UPDRS‐III has been used in DLB research for motor symptom monitoring ¹⁴ and as an outcome measure in multiple studies since its development in the early 2000s. ¹⁵ , ¹⁶ Multiple groups have attempted to determine a minimal clinically important difference (MCID) for the MDS‐UPDRS as well as its individual components to track disease progression. ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ For MDS‐UPDRS‐III, specifically in PD, a MCID of 5 points is commonly accepted; ¹⁷ , ²¹ there is no agreed‐upon MCID specific to DLB.

The primary aim of our study was to determine the relationship between ¹²³I‐ioflupane SPECT SBR Z scores measured via DaTQUANT analysis and motor parkinsonism as measured by the MDS‐UPDRS‐III at baseline and 24 months in a well‐phenotyped cohort. A secondary aim was to determine whether ¹²³I‐ioflupane SPECT SBR Z scores predict meaningful worsening of motor parkinsonian symptoms in DLB at 24 months as measured by the MDS‐UPDRS‐III. The relationship between striatal binding and MDS‐UPDRS‐III has been analyzed in patients with PD ²² , ²³ and DLB, ²⁴ , ²⁵ showing that dopamine deficiency on ¹²³I‐ioflupane SPECT is associated with severity of motor symptoms. However, to the best of our knowledge, no evaluation of the ability of ¹²³I‐ioflupane SPECT to predict progression of parkinsonism has been performed in the DLB population. Last, we explored the association between cerebrospinal fluid (CSF) α‐synuclein (α‐syn) seed amplification assay (SAA) and SBRs, as the relationship between CSF SAA and ¹²³I‐ioflupane SPECT semi‐quantitative analysis is not well studied in DLB.

2. METHODS

2.1. Study design

This is a retrospective cohort study using the US Dementia with Lewy Bodies Consortium (DLBC) dataset. This dataset includes participant information collected from multiple participating study sites in the United States. The sites involved include the Cleveland Clinic (Cleveland, Ohio and Las Vegas, Nevada), Rush University, VA‐Puget Sound Health Care System/University of Washington, University of California San Diego, University of Pennsylvania, University of Pittsburgh, Thomas Jefferson University, University of North Carolina, and Florida Atlantic University. Participants undergo clinical, imaging, and biofluid collection and evaluation. ²⁶ , ²⁷ Clinical data from the DLBC dataset is maintained in the Parkinson's Disease Biomarkers Program Data Management Resource (PDMP‐DMR). ²⁸ Data obtained from 2017 to November 2023 were used for this analysis. Details on data collection and storage have been published previously. ²⁷ , ²⁹ , ³⁰

2.2. Sample group

The sample group includes individuals with either probable DLB based on 2017 McKeith Criteria ¹ or high‐likelihood DLB individuals with mild cognitive impairment (MCI) based on prodromal DLB research criteria published by McKeith et al. in 2020. ³¹ All participants in the DLBC were screened for inclusion. Participants with Parkinson's disease dementia (PDD) were excluded from this analysis. Participants without baseline ¹²³I‐ioflupane SPECT analyzed with DaTQUANT or baseline MDS‐UPDRS‐III examination were also excluded.

RESEARCH‐IN‐CONTEXT

1. Systematic review: The authors reviewed the current literature using PubMed. There have been publications on the relationship between striatal binding and parkinsonian symptoms; however, few are specific to dementia with Lewy bodies (DLB). There have been no publications on the relationship between seed amplification assay (SAA) and striatal binding. Relevant references are appropriately cited.

2. Interpretation: Our findings support prior discoveries that ¹²³I‐Ioflupane single‐photon emission computed tomography can confirm dopamine deficiency in the context of clinical parkinsonism. We show that striatal binding ratio (SBR) Z scores obtained at baseline can predict 24‐month Movement Disorder Society Unified Parkinson's Disease Rating Scale Part III scores; however, not meaningful worsening of scores. Our findings suggest the presence of SAA in DLB indicates dopamine dysfunction; this can aid in diagnosis of NSD.

3. Future directions: Future studies should aim to determine whether SBRs obtained in pre‐parkinsonian DLB patients could predict development of parkinsonism in the future, which could aid in earlier diagnosis. Our findings are limited by small sample size and would benefit from repetition in larger cohorts.

2.3. Physical examination

MDS‐UPDRS‐III total  score was used for semi‐quantitative assessment of motor symptoms observed on physical exam. These examinations were performed by the enrolling physician at their respective institution. Examiners were required to complete formalized training on how to perform and grade the examination through the Movement Disorders Society. ³² As part of the MDS‐UPDRS, examiners indicated whether and how recently the participant took dopaminergic medications and whether the participant was “on” or “off” during the examination. “On” and “off” refer to the functional state the examiner determines the participant to be in during their examination in the presence of dopaminergic medication. ¹⁵ “On” is the expected functional state when the participant is having a good response (i.e., improvement in motor parkinsonism, including reduction of rigidity, bradykinesia, and/or tremor), whereas “off” is the expected functional state in which the patient is having a poor response. ¹⁵

2.4. ¹²³I‐ioflupane SPECT

¹²³I‐ioflupane SPECT data were processed using the DaTQUANT software, which computes an SBR for each basal ganglion. ⁹ The software spatially normalizes the SPECT images to a template which has specified regions for the left and right side of the striatum as well as a reference occipital region (Figure 1). ⁹ Raw ¹²³I‐ioflupane SPECT projection data were imported into DaTQUANT and reconstructed with a consistent set of parameters (OSEM, 2 iterations, 10 subsets, Butterworth 0.6 10 filter). ³³ SBR is computed as a ratio of regional mean image pixel values; derivation of these values to a built‐in age‐matched normal database yields the Z scores. ³⁴ In addition to the whole striatum, sub‐regions including the whole putamen, anterior putamen, posterior putamen, and caudate were also studied.

2.5. CSF α‐syn SAA testing

A subset of 45 participants had CSF tested with Amprion for α‐syn by SAA developed by Concha‐Marambio et al. ³⁵ The “protein misfolding cyclic amplification” protocol used has been outlined in detail previously. ³⁵

2.6. Statistical analysis

Demographics and participant characteristics were summarized by means with standard deviation for continuous variables and count with percentage for categorical variables. Pearson correlations were calculated between MDS‐UPDRS‐III scores (at baseline, 24 months, and change between baseline and 24 months) and baseline mean SBR Z scores.

Multivariable linear regression models were used to predict MDS‐UPDRS‐III scores at baseline and 24 months. Baseline mean SBR Z scores (striatum, putamen, posterior putamen, and caudate) were independent variables in separate models. Covariates included age, sex, race, diagnosis, dopaminergic medication use, antidopaminergic medication use, and clinical state at time of MDS‐UPDRS‐III exam. In the model predicting 24 month MDS‐UPDRS‐III, baseline MDS‐UPDRS‐III score was also included as a covariate.

Multivariable logistic regression models were used to predict meaningful worsening in MDS‐UPDRS‐III scores from baseline to 24 months. Meaningful worsening in MDS‐UPDRS‐III was defined as an increase of at least 5 points in MDS‐UPDRS‐III scores from baseline to 24 month exam. ¹⁷ Baseline mean SBR Z scores (striatum, putamen, posterior putamen, and caudate) were independent variables in separate models. Covariates included age, sex, diagnosis, dopaminergic medication use, and clinical state at baseline. Race and antidopaminergic medication use were excluded as covariates due to model convergence issue.

A subgroup analysis of the 45 participants who had available CSF SAA results was performed. The demographics and participant characteristics were stratified by CSF SAA results and were summarized by means with standard deviation for continuous variables and count with percentage for categorical variables. In the SAA‐positive group and SAA‐negative group, Pearson correlations were calculated between baseline MDS‐UPDRS‐III scores and mean SBR Z scores separately. Multivariable logistic regression models were used to predict SAA results (positive or negative). Baseline mean SBR Z scores (striatum, putamen, posterior putamen, and caudate) were independent variables in separate models. Covariates included age, sex, diagnosis, and baseline MDS‐UPDRS‐III score. Receiver operating characteristic (ROC) curve analysis was conducted to examine the ability of baseline mean SBR Z scores (striatum, putamen, posterior putamen, and caudate) to predict presence or absence of CSF α‐syn via SAA.

Statistical analysis was completed using SAS Enterprise Guide, version 8.2 (SAS Institute Inc.). P value < 0.05 was considered statistically significant.

3. RESULTS

Fifty‐four participants remained after excluding five participants for a diagnosis of PDD and 129 participants for not having baseline ¹²³I‐ioflupane SPECT analyzed via DaTQUANT and MDS‐UPDRS‐III examinations at baseline visit at the time of cut‐off for this analysis. Table 1 summarizes the demographics and participant characteristics stratified by whether participants had MDS‐UPDRS‐III examinations at 24 month follow‐up. Participants had a mean age of 68.9 ± 7.5 years, the majority were men (85.2%) with a diagnosis of probable DLB (83.3%), with the remainder having a diagnosis of MCI with Lewy bodies (MCI‐LB). Most participants were not on a dopaminergic medication at baseline (66.7%) or 24 month (76%) evaluation. Most participants were not on an antidopaminergic medication at baseline (94.4%) or 24 month (88%) evaluation. Of the eighteen participants on dopaminergic medications at baseline MDS‐UPDRS‐III examination, thirteen were considered “on” and five were considered “off” by the examining clinician. Of the six participants on dopaminergic medications during the 24 month MDS‐UPDRS‐III examination, three participants were considered “on” and two were considered “off,” and one was undocumented by the examining clinician. The average time since last levodopa dose for our cohort was 278 minutes with times ranging from 2 minutes to 1579 minutes. The average baseline MDS‐UPDRS‐III score was 30.8 ± 15.4. The average 24 month MDS‐UPDRS‐III score was 34.2 ± 20.8. Right and left SBR Z scores were not significantly different based on paired t test, striatum (P = 0.15), putamen (P = 0.08), posterior putamen (P = 0.06), and caudate (P = 0.72). The right and left SBR Z scores were averaged for each participant to simplify further analysis.

TABLE 1.

Demographics and participant characteristics of the study sample (N = 54).

	Total (N = 54)	Participants without MDS‐UPDRS‐III at 24 ‐month follow‐up (N = 29)	Participants with MDS‐UPDRS‐III at 24 month follow‐up (N = 25)	P value
Age at baseline, mean ± SD	68.9 ± 7.5	69.0 ± 8.0	68.9 ± 7.0	0.97a
Sex, n (%)				0.45 c
Female	8 (14.8)	3 (10.3)	5 (20.0)
Male	46 (85.2)	26 (89.7)	20 (80.0)
Race, n (%)				0.85 c
White	51 (94.4)	27 (93.1)	24 (96.0)
African—Black	1 (1.9)	1 (3.4)	0 (0.00)
American—Black	1 (1.9)	1 (3.4)	0 (0.00)
American Indian/Alaska Native	1 (1.9)	0 (0.00)	1 (4.0)
Diagnosis, n (%)				0.065 c
High likelihood DLB with MCI	9 (16.7)	2 (6.9)	7 (28.0)
Probable DLB	45 (83.3)	27 (93.1)	18 (72.0)
SAA results, n (%) *				0.11 b
DETECTED	29 (64.4)	18 (75.0)	11 (52.4)
NOT DETECTED	16 (35.6)	6 (25.0)	10 (47.6)
Baseline mean striatum SBR Z score, mean ± SD	−2.2 ± 1.6	−2.4 ± 1.8	−1.9 ± 1.5	0.27 a
Baseline mean putamen SBR Z score, mean ± SD	−2.3 ± 1.7	−2.5 ± 1.8	−2.0 ± 1.5	0.25 a
Baseline mean posterior putamen SBR Z score, mean ± SD	−2.1 ± 1.5	−2.3 ± 1.6	−1.9 ± 1.4	0.34 a
Baseline mean caudate SBR Z score, mean ± SD	−1.8 ± 1.5	−2.0 ± 1.6	−1.6 ± 1.3	0.29 a
Baseline MDS‐UPDRS‐III Score,  mean ± SD	30.8 ± 15.4	33.4 ± 17.5	27.7 ± 12.1	0.18 a
24 month MDS‐UPDRS‐III score, mean ± SD *	34.2 ± 20.8	—	34.2 ± 20.8
Meaningfully worsen in MDS‐UPDRS‐III, n (%) *	15 (60.0)	—	15 (60.0)
Baseline dopaminergic medication, n (%)				0.44 b
No	36 (66.7)	18 (62.1)	18 (72.0)
Yes	18 (33.3)	11 (37.9)	7 (28.0)
Baseline antidopaminergic medication, n (%)				0.99 c
No	51 (94.4)	27 (93.1)	24 (96.0)
Yes	3 (5.6)	2 (6.9)	1 (4.0)
Baseline clinical state for those on dopaminergic medication, n (%)				0.99 c
OFF	5 (27.8)	3 (27.3)	2 (28.6)
ON	13 (72.2)	8 (72.7)	5 (71.4)
24 month dopaminergic medication, n (%)
No			19 (76.0)
Yes			6 (24.0)
24 month antidopaminergic medication, n (%)
No			22 (88.0)
Yes			3 (12.0)
24 month clinical state for those on dopaminergic medication, n (%)
NA			1 (16.7)
OFF			2 (33.3)
ON			3 (50.0)

^*Data not available for all subjects.

 Missing values: SAA results = 9; 24 month MDS‐UPDRS‐III score = 29; meaningfully worsen in UPDRS = 29. Statistics presented as mean ± SD, N (column %). P values: ^a = Analysis of variance; ^b = Pearson chi‐squared test; ^c = Fisher exact test

Abbreviations: DLB, dementia with Lewy bodies; MCI, mild cognitive impairment; MDS‐UPDRS‐III, Movement Disorder Society Unified Parkinson's Disease Rating Scale–Part III; SAA, seed amplification assay; SBR, striatal binding ratio; SD, standard deviation.

Baseline MDS‐UPDRS‐III scores and 24 month MDS‐UPDRS‐III scores were significantly correlated with all baseline SBR Z scores. However, the change in MDS‐UPDRS‐III score from baseline to 24 months was not significantly correlated with baseline SBR Z scores (Figure 2 and Figures S1‐S3 and, Table S1 in supporting information).

FIGURE 2.

Plot of linear correlation between (A) baseline MDS‐UPDRS‐III scores and baseline mean caudate SBR Z scores (N = 54; r = –0.37; P = 0.006), (B) 24 month MDS‐UPDRS‐III scores and baseline mean caudate SBR Z scores (N = 25; r = –0.51; P = 0.009), (C) change in MDS‐UPDRS‐III score from baseline to 24 months and baseline mean caudate SBR Z scores (N = 25; r = –0.34; P = 0.092). Red triangles indicate positive CSF SAA cases; blue circles indicate negative CSF SAA cases; green stars indicate cases with missing CSF SAA results. See supporting information for plot of linear correlations for striatum, putamen, and posterior putamen. CSF, cerebrospinal fluid; MDS‐UPDRS‐III, Movement Disorder Society Unified Parkinson's Disease Rating Scale–Part III; SAA, seed amplification assay; SBR, striatal binding ratio

In multivariable linear regression models predicting MDS‐UPDRS‐III scores at baseline, striatum (estimate [standard error]: –2.65 [1.29], P = 0.046) and caudate (–3.31 [1.45], P = 0.028) SBR Z scores were significant predictors and, while there was a trend, the putamen (–2.43 [1.28], P = 0.064) and posterior putamen (–2.39 [1.46], P = 0.11) did not reach statistical significance. Age, sex, race, diagnosis, baseline clinical state, dopaminergic medication use, and antidopaminergic medication use were not significantly related to the baseline MDS‐UPDRS‐III score (Table S2 in supporting information).

Twenty‐five (46.3%) participants had both a baseline and 24 month MDS‐UPDRS‐III exam. These participants were demographically and clinically similar to the full sample (Table 1). In multivariable linear regression, we found that baseline SBR Z scores were significant predictors of MDS‐UPDRS‐III scores at 24 month visits, including striatum (–6.57 [2.64], P = 0.025), putamen (–6.14 [2.64], P = 0.035), posterior putamen (–6.52 [2.85], P = 0.037), and caudate (–7.55 [2.89], P = 0.020). Of the covariates, sex, baseline MDS‐UPDRS‐III scores, and 24 month clinical state were identified as significant predictors of MDS‐UPDRS‐III scores at 24 months. Females had worse scores than males on average; higher baseline MDS‐UPDRS‐III score predicted worse 24 month MDS‐UPDRS‐III score; and 24‐month clinical state of “on” predicted a better 24 month MDS‐UPDRS‐III score (Table S3 in supporting information).

Of the 25 participants with baseline and 24 month MDS‐UPDRS‐III data, 15 (60.0%) had a clinically important worsening defined as an increase in MDS‐UPDRS‐III score of ≥ 5 points at 24 months. ¹⁷ Baseline SBR Z scores were not predictive of meaningful worsening in MDS‐UPDRS‐III scores (Table S4 in supporting information).

A subgroup analysis of participants who had available CSF SAA results was performed. Table 2 summarizes the demographics and participant characteristics stratified by SAA results. Of the 54 participants analyzed, 29 were positive for CSF SAA, 16 were negative, and 9 had missing data. The positive cases had significantly lower SBR Z scores (P < 0.001 for all regions, Figure 3). The positive cases had higher baseline MDS‐UPDRS‐III scores on average (mean [standard deviation]: 34.9 [16.2] vs. 24.3 [12.4], P = 0.075) and higher 24 month MDS‐UPDRS‐III scores (37.9 [19.5] vs. 26.2 [15.0], P = 0.27) compared to negative cases. The positive cases were also more likely to be on dopaminergic medications at baseline exam compared to negative cases (41.4% vs. 31.3%, P = 0.24). Of the nine participants diagnosed with MCI‐LB, three (33.3%) had positive CSF SAA, four (44.4%) had negative CSF SAA, and two (22.2%) had missing data. Of the 45 participants with a diagnosis of probable DLB, 26 (57.8%) had positive CSF SAA, whereas 12 (26.7%) had negative CSF SAA, and 7 (15.6%) had missing data (Table 2).

FIGURE 3.

Box plot of baseline SBR Z scores by SAA results demonstrating the significantly lower mean SBR Z score in the striatum, putamen, posterior putamen, and caudate in α‐syn–positive participants compared to α‐syn negative participants. α‐syn, alpha synuclein; SAA, seed amplification assay; SBR, striatal binding ratio

TABLE 2.

Demographics and participant characteristics of the study sample (N = 54), stratified by SAA results.

	Total (N = 54)	Not detected (N = 16)	Detected (N = 29)	Missing (N = 9)	P value
Age at baseline, mean ± SD	68.9 ± 7.5	67.3 ± 8.0	68.5 ± 6.3	73.2 ± 9.2	0.15 a
Sex, n (%)					0.66 c
Female	8 (14.8)	3 (18.8)	3 (10.3)	2 (22.2)
Male	46 (85.2)	13 (81.3)	26 (89.7)	7 (77.8)
Race, n (%)					0.59 c
White	51 (94.4)	16 (100.0)	27 (93.1)	8 (88.9)
African—Black	1 (1.9)	0 (0.00)	1 (3.4)	0 (0.00)
American—Black	1 (1.9)	0 (0.00)	0 (0.00)	1 (11.1)
American Indian/Alaska Native	1 (1.9)	0 (0.00)	1 (3.4)	0 (0.00)
Diagnosis, n (%)					0.43 c
High likelihood DLB with MCI	9 (16.7)	4 (25.0)	3 (10.3)	2 (22.2)
Probable DLB	45 (83.3)	12 (75.0)	26 (89.7)	7 (77.8)
Baseline mean striatum SBR Z score, mean ± SD	−2.2 ± 1.6	−0.66 ± 0.78	−3.1 ± 1.3	−1.9 ± 1.8	<0.001 a
Baseline mean putamen SBR Z score, mean ± SD	−2.3 ± 1.7	−0.71 ± 0.81	−3.2 ± 1.3	−2.0 ± 1.9	<0.001 a
Baseline mean posterior putamen SBR Z score, mean ± SD	−2.1 ± 1.5	−0.84 ± 0.90	−2.9 ± 1.1	−1.9 ± 1.7	<0.001 a
Baseline mean caudate SBR Z score, mean ± SD	−1.8 ± 1.5	−0.50 ± 0.80	−2.6 ± 1.2	−1.5 ± 1.5	<0.001 a
Baseline MDS‐UPDRS‐III score, mean ± SD	30.8 ± 15.4	24.3 ± 12.4	34.9 ± 16.2	28.8 ± 14.1	0.075 a
24 month MDS‐UPDRS‐III score, mean ± SD *	34.2 ± 20.8	26.2 ± 15.0	37.9 ± 19.5	43.8 ± 33.4	0.27 a
Meaningful worsening in UPDRS, n (%) *	15 (60.0)	6 (60.0)	6 (54.5)	3 (75.0)	0.87 c
Baseline clinical state, n (%)					0.95 c
NA	31 (57.4)	10 (62.5)	15 (51.7)	6 (66.7)
OFF	8 (14.8)	2 (12.5)	5 (17.2)	1 (11.1)
ON	15 (27.8)	4 (25.0)	9 (31.0)	2 (22.2)
Baseline dopaminergic medication, n (%)					0.24 b
No	36 (66.7)	11 (68.8)	17 (58.6)	8 (88.9)
Yes	18 (33.3)	5 (31.3)	12 (41.4)	1 (11.1)
Baseline antidopaminergic medication, n (%)					0.99 c
No	51 (94.4)	15 (93.8)	27 (93.1)	9 (100.0)
Yes	3 (5.6)	1 (6.3)	2 (6.9)	0 (0.00)
24 month clinical state, n (%)					0.97 c
NA	41 (75.9)	12 (75.0)	21 (72.4)	8 (88.9)
OFF	3 (5.6)	1 (6.3)	2 (6.9)	0 (0.00)
ON	10 (18.5)	3 (18.8)	6 (20.7)	1 (11.1)
24 month dopaminergic medication, n (%)					0.71 c
No	48 (88.9)	14 (87.5)	25 (86.2)	9 (100.0)
Yes	6 (11.1)	2 (12.5)	4 (13.8)	0 (0.00)
24 month antidopaminergic medication, n (%)					0.99 c
No	51 (94.4)	15 (93.8)	27 (93.1)	9 (100.0)
Yes	3 (5.6)	1 (6.3)	2 (6.9)	0 (0.00)

^*Data not available for all subjects. Missing values: 24 month MDS‐UPDRS‐III score = 29; meaningfully worsen in UPDRS = 29. Statistics presented as mean ± SD, N (column %).

Abbreviations: DLB, dementia with Lewy bodies; MCI, mild cognitive impairment; MDS‐UPDRS‐III, Movement Disorder Society Unified Parkinson's Disease Rating Scale–Part III; SAA, seed amplification assay; SBR, striatal binding ratio; SD, standard deviation.

P values: ^a = Analysis of variance.

P values: ^b = Pearson chi‐squared test.

P values: ^c = Fisher exact test.

For SAA‐negative participants’ SBR Z scores for all regions and baseline MDS‐UPDRS‐III were not significantly correlated (Table S5 in supporting information). For SAA‐positive participants, caudate SBR Z scores and baseline MDS‐UPDRS‐III scores were significantly correlated (r = –0.38, P = 0.045), but not striatum, putamen, or posterior putamen (Table S5). There was no significant difference in striatum asymmetry between the SAA‐positive and SAA‐negative participants (P = 0.29). Multivariable logistic regression models were used to predict SAA results (positive vs. negative; Table S6 in supporting information). ROC curve analysis was conducted to examine the ability of baseline mean SBR Z scores (striatum, putamen, posterior putamen, and caudate) to predict presence or absence of CSF α‐syn by SAA. Based on the ROC curve analysis (Figure S4 in supporting information), optimum SBR Z score cut‐off values for predicting positive SAA results were chosen based on sensitivities and specificities; these are outlined in Table 3. There is currently no agreed‐upon SBR “cut‐off” value or location used to determine whether the result of a ¹²³I‐ioflupane SPECT is considered “positive” or “negative”; clinically, positivity versus negativity is determined by visual inspection alone. Table 3 provides options of cut‐off values and locations based on the ability of SBR Z score to predict presence of absence of CSF α‐syn by SAA.

TABLE 3.

Optimal SBR Z scores cut‐off values with their corresponding sensitivities and specificities based on ROC curve analysis examining the ability of baseline mean SBR Z scores (striatum, putamen, posterior putamen, and caudate) to predict presence or absence of CSF α‐syn by SAA.

Location	Cut‐off SBR Z score	Sensitivity	Specificity
Striatum	−2.075	86.21%	100%
Putamen	−1.99	86.21%	100%
	−1.89	89.66%	93.75%
Posterior putamen	−2.115	82.76%	100%
	−1.97	86.21%	93.75%
Caudate	−1.82	82.76%	100%
	−1.585	86.21%	93.75%

Abbreviations: α‐syn, alpha synuclein; CSF, cerebrospinal fluid; ROC, receiver operating characteristic; SAA, seed amplification assay; SBR, striatal binding ratio.

4. DISCUSSION

Despite the prevalence of non‐tremor predominant motor parkinsonism in DLB, our study revealed a significant relationship between ¹²³I‐ioflupane SPECT SBR Z scores measured by DaTQUANT and motor parkinsonism as measured by MDS‐UPDRS‐III in DLB patients at the time of study enrollment. This is consistent with the negative correlation between MDS‐UPDRS‐III and SBR seen in PD ²² , ²³ , ³⁶ , ³⁷ , ³⁸ and DLB. ²⁴ , ²⁵ We also found that baseline SBR Z scores were significantly correlated with 24 month MDS‐UPDRS‐III scores and that they predicted 24 month MDS‐UPDRS‐III scores when adjusting for key demographics, clinical characteristics, and baseline MDS‐UPDRS‐III score. However, baseline SBR Z scores were not predictive of meaningful worsening (based on MCID of ≥ 5 points) on MDS‐UPDRS‐III from baseline to 24 months. A recent systematic review analyzed 20 studies that assessed the clinical utility of ¹²³I‐ioflupane SPECT in patients with suspected parkinsonian syndromes. ³⁹ They found that ¹²³I‐ioflupane SPECT was associated with a change in management in half of the tested population and modification of diagnosis in one third of the tested population. ³⁹ Of these studies, only two clearly included DLB as the indication for the scan. ³⁹ , ⁴⁰ , ⁴¹ In one, ¹²³I‐ioflupane SPECT was found to prompt a change in clinical management in 65% of DLB cases ⁴⁰ and in the other, scans suggested an alternative diagnosis in 54% of patients with possible DLB. ⁴¹ Neither study used SBRs in the interpretation of the ¹²³I‐ioflupane SPECT scans.

In McKeith et al.’s 2020 publication of research criteria for prodromal DLB, ^{31 123}I‐ioflupane SPECT was cited as a proposed biomarker for MCI‐LB based on a publication in 2018 which found visually rated ¹²³I‐ioflupane SPECT to be 54.3% sensitive and 89% specific in identifying possible or probable clinically diagnosed MCI‐LB. ⁴² Multiple studies have identified that those with rapid eye movement sleep behavior disorder (RBD) will have decreased DAT binding in the striatum, predicting the development of α‐synucleinopathies. ⁴³ , ⁴⁴ , ⁴⁵ , ⁴⁶ , ⁴⁷ The Parkinson Associated Risk Syndrome study has also identified that those with hyposmia who have deficient dopamine uptake on ¹²³I‐ioflupane SPECT convert to clinical PD at a higher rate than those with normal dopamine uptake. ⁴⁸ , ⁴⁹ The findings of these studies are provocative given they propose ¹²³I‐ioflupane SPECT as a potential early biomarker for α‐synucleinopathies. Based on the predictive findings we observed, it is possible that SBR Z scores could be a useful biomarker to predict development of parkinsonism in patients who do not yet exhibit motor parkinsonism. However, most participants in our study had some degree of motor parkinsonism at their baseline visit, with an average MDS‐UPDRS‐III score of 30.8 ± 15.4; therefore, we were unable to address this question. Further analysis of patients prior to motor parkinsonism onset, such as those with isolated RBD, ⁵⁰ is needed. This is particularly important in the context of newly proposed definitions of α‐synucleinopathies such as neuronal α‐syn disease (NSD). ⁵¹

SBR semi‐quantitative analysis has high sensitivity and specificity in discriminating ¹²³I‐ioflupane SPECT. ²³ DaTQUANT software was used for SBR analysis because it is routinely used by clinicians, has a long publication history, and has a built‐in normal database for Z score reporting. Alternative analysis methods have been published recently which improve spatial normalization and statistical power though they are not widely available. ⁵² , ⁵³ DaTQUANT includes iterative reconstruction from raw projection data, which substantially improves quantitative harmonization for multi‐center studies such as this one. Although SPECT acquisition was performed in a consistent fashion (low energy collimators, 120 projections, 128 × 128 matrix), differences between scanners at participating sites remain. The most significant difference between scanners is image reconstruction software and settings. Performing image reconstruction consistently in DaTQUANT from the raw data minimized these differences. Small sample size per site prevented including site as a variable in statistical analysis, though the effect of scanner differences is expected to be small.

Due to DaTQUANT's widespread use, several groups have proposed SBR “cut‐off” thresholds for ruling in or ruling out a diagnosis of PD or DLB, yet there is some disagreement on this matter. ⁵⁴ Historically, a striatal or putaminal SBR Z score of –2 or –2.5 was arbitrarily used to identify patients with PD. ⁵⁵ , ⁵⁶ In 2021, a group identified the posterior putamen SBR (of the more affected side) to yield the highest diagnostic accuracy in Lewy body diseases (LBDs), making no distinction between PD and DLB, with a cut‐off posterior putamen SBR Z score of –1.8. ⁵⁷ In 2023, Lanfranchi et al. published their finding that a posterior putamen Z score cut‐off of –1.27 has a sensitivity of 97% and a specificity of 94% in differentiating PD and essential tremor, arguing for a less conservative cut‐off value without loss of specificity. ⁵⁴ This study also compared scans of DLB and AD patients and found a whole putamen Z score of –0.96 has a sensitivity of 66% and a specificity of 85% in differentiating DLB from AD. ⁵⁴ In regard to DLB specifically, one group performed ¹²³I‐ioflupane SPECT scans on patients with RBD and followed them for 5 years; they identified a cut‐off striatal Z score of –2.55 predicts development of DLB, with a sensitivity of 93.9% and a specificity of 72%. ⁴⁵ Notably, these studies were all based on clinical diagnoses of LBDs and none used α‐syn testing or pathologic confirmation. ⁴⁵ , ⁵⁴ , ⁵⁷ Given the lack of consensus opinion on this matter, we chose to evaluate mean SBR Z scores for the striatum, whole putamen, posterior putamen, and the caudate. In our cohort, there was no significant difference between left and right SBR Z scores, which is expected in DLB in which motor parkinsonism is more frequently symmetric. ⁵⁸ We found that the caudate had the strongest correlation with MDS‐UPDRS‐III score, which is expected given caudate changes in ¹²³I‐ioflupane SPECT indicate more advanced disease, ⁵⁹ and those with advanced disease likely have higher MDS‐UPDRS‐III scores.

Simuni et al. recently proposed a biological definition of NSD based on the presence of CSF α‐syn via SAA. ⁵¹ They recommend using the whole putamen SBR Z score of the most affected side to document evidence of dopaminergic dysfunction. ⁵¹ Their paper did not provide a Z score cut‐off to rule in or rule out dopaminergic dysfunction. ⁵¹ To the best of our knowledge, our study is the first to compare ¹²³I‐ioflupane SPECT SBRs to SAA results in DLB. Notably, 35.5% of the 45 individuals with CSF SAA results had negative CSF SAA compared to 64.4% with positive CSF SAA. The SAA findings in this cohort have recently been published. ²⁶ Those with positive CSF SAA had significantly lower SBR Z scores in all regions compared to those with negative CSF SAA, which indicates that those with α‐synucleinopathies and a clinical diagnosis of DLB are more likely to have ¹²³I‐ioflupane SPECT evidence of dopamine dysfunction. Applying the Lanfranchi cut‐off value of posterior putamen SBR Z score < –0.96⁵⁴ to our cohort of 45 participants with SAA results and baseline SBR measurements, we found that 28 of 29 of the SAA‐positive cases (sensitivity 93.6%) would meet these criteria; however, 7 of the 16 SAA‐negative cases (specificity 69.6%) would also meet this criteria. Therefore, if using ¹²³I‐ioflupane SPECT SBR Z scores to rule in or out NSD, an alternative cut‐off value may be needed. In Table 3 we provide options of SBR Z score cut‐offs with high sensitivity and specificity in identifying SAA‐positive cases. The Lanfranchi cut‐off ⁵⁴ is still very relevant as it applies to Parkinsonian diseases ⁶⁰ but may not be sensitive enough to rule in or out disease that is primarily a result of α‐syn aggregation as tested by SAA. It is worth noting that α‐syn aggregation is not the only cause of dopaminergic neuron deficiency, ⁶⁰ meaning it would be inappropriate to use ¹²³I‐ioflupane SPECT to rule in or rule out synucleinopathies, arguing for the use of α‐syn biomarkers such as CSF SAAs as suggested by Simuni et al. ⁵¹

The greatest limitation of our study is the small sample size, particularly at the 24 month follow‐up as only 25 participants were eligible for follow‐up at the time of analysis. Participants who followed up were less likely to have probable DLB and had higher SBR Z scores. To mitigate the effects of selection bias, models were adjusted for potential confounders, including diagnosis. Given the limited sample size, estimates may be biased toward the null, although guidelines suggest two cases per independent variable may provide sufficient power. ⁶¹ It is known that medications can influence the MDS‐UPDRS‐III score, ¹⁸ and while the majority of our participants were not on medications at the time of examination, there was a portion examined while on dopaminergic medication and while in the “on” state, which would affect their scoring on the MDS‐UPDRS‐III. We appreciate that those considered to be in the “on” state is not an ideal representation of the individual's true clinical dopamine deficiency; however, our cohort was too small to exclude all these participants. In regard to the anti‐dopaminergic medications, fortunately the only medications used in this cohort were quetiapine and clozapine, both of which are less likely to affect the MDS‐UPDRS‐III score compared to classic anti‐psychotics. ⁶² , ⁶³ Finally, while we have CSF SAA results on this cohort, we are still awaiting pathological confirmation of underlying diagnoses.

We conclude that in DLB, ¹²³I‐ioflupane SPECT at time of diagnosis can be used to confirm underlying dopamine deficiency in the context of clinical parkinsonism, which is consistent with prior findings regarding the utility of ¹²³I‐ioflupane SPECT early in DLB to aid in diagnosis. ⁴ We also conclude that ¹²³I‐ioflupane SPECT at time of diagnosis does not add value beyond MDS‐UPDRS‐III score at time of diagnosis to predict clinical progression over 2 years. Further studies on patients prior to presentation of parkinsonian motor symptoms need to be completed to understand if ¹²³I‐ioflupane SPECT can be predictive of future development of parkinsonism. Additionally, we identified that those with presence of SAA and a clinical diagnosis of DLB are more likely to have ¹²³I‐ioflupane SPECT evidence of dopamine dysfunction compared to those without presence of SAA. Therefore, analyzing a larger cohort and evaluating only those with SAA positivity may lead to different conclusions on ¹²³I‐ioflupane SPECT's ability to predict disease progression, making it a possible marker of disease progression for future therapeutics.

CONFLICT OF INTEREST STATEMENT

Emily F. Maly has nothing to disclose. Frank P. DiFilippo has nothing to disclose. Brittany Lapin has nothing to disclose. Yadi Li has nothing to disclose. Sarah Berman has nothing to disclose. Andrea C. Bozoki works as a site PI for an EIP Pharma clinical trial, a Site PI for aCognition Therapeutics clinical trial, and is a member of the data and safety monitoring board for AviadoBio. Jori E. Fleisher receives research support from NIH (R01AG079128, U19AG078105, U01NS100610), Parkinson's Foundation, CurePSP, and private philanthropic support. She has served as a site investigator for clinical trials sponsored by Cognition Therapeutics and EIP Pharma. James E. Galvin has nothing to disclose. David J. Irwin receives research grants from NIH and clinical trial funding from Prevail, Passage Bio, Alector, and Denali. Carol F. Lippa has nothing to disclose. Irene Litvan's research is supported by the National Institutes of Health grants: 5U01NS112010/807745, U01NS100610, R25NS098999; U19 AG063911‐1 and 1R21NS114764‐01A1; 2 P30 AG062429‐06; the Michael J Fox Foundation, Parkinson's Foundation, Roche, AbbVie, Lundbeck, EIP Pharma, Alterity, Novartis, and UCB.  She is a member of the Scientific Advisory Board for the Rossy PSP Program at the University of Toronto, Aprinoia, Amydis, and the Food and Drug Administration (FDA) Peripheral and Central Nervous System Drugs Advisory Committee. She receives her salary from the University of California San Diego and as chief editor of Frontiers in Neurology. Debby W. Tsuang has nothing to disclose. Cyrus P. Zabetian has nothing to disclose. Angela S. Taylor is employed by the Lewy Body Dementia Association. Lynn M. Bekris has nothing to disclose. Oscar L. Lopez is a consultant for Novo Nordisk and is a member of the data and safety monitoring board for Acumen. Douglas Galasko works as a consultant for GE Healthcare, Eisai, and Cognition Therapeutics. James B. Leverenz receives grant funding from GE Healthcare (DaTscan Ligand and DaTQUANT™ software) and Lilly Pharmaceuticals (honorarium for speaking engagement). Author disclosures are available in the supporting information.

CONSENT STATEMENT

All contributing centers obtained informed consent from participants and maintained individual institutional review board approval.

Supporting information

Supporting Information

Supporting Information

Maly EF, DiFilippo FP, Lapin B, et al. Performance of ¹²³I‐ioflupane SPECT striatal binding in dementia with Lewy bodies. Alzheimer's Dement. 2025;21:e70596. 10.1002/alz.70596