Quantitative Susceptibility Mapping and Iron Deposition in Autopsy‐Confirmed Progressive Supranuclear Palsy

Abstract

Background

Quantitative susceptibility mapping (QSM) is a promising neuroimaging biomarker for detecting abnormal iron deposition in progressive supranuclear palsy (PSP); however, the QSM profile and its accuracy in autopsy‐confirmed PSP remain unexplored.

Objective

The aim of the study was to elucidate and validate the QSM profiles in autopsy‐confirmed PSP.

Methods

Eight patients with a clinically diagnosed PSP syndrome (PSPS) underwent antemortem 3 T magnetic resonance imaging (MRI) scanning and subsequent pathological diagnosis, confirming 6 PSP pathology and 2 multiple system atrophy (MSA) pathology cases. QSM analysis and Prussian blue staining were performed across subcortical regions.

Results

PSP cases showed higher susceptibility than age‐matched controls in the pallidum, subthalamic nucleus, substantia nigra, and cerebellar dentate (false discovery rate [FDR]‐corrected P < 0.05). PSP had higher susceptibility in the subthalamic nucleus and lower susceptibility in the putamen than the clinically atypical MSA (P = 0.08). Subcortical susceptibilities correlated with Prussian blue–stained areas (r = 0.63, P < 0.001).

Conclusions

QSM could be a reliable measure to evaluate abnormal iron accumulation patterns in PSP.

Progressive supranuclear palsy (PSP) is a neurodegenerative disease typically characterized by postural instability, vertical supranuclear gaze palsy, axial rigidity, and bradykinesia. ¹ , ² The clinical spectrum of PSP encompasses various atypical presentations, including corticobasal syndrome, those characterized by prominent speech and language disturbances, and cases presenting with pure akinesia with gait freezing or parkinsonism. Given this clinical heterogeneity, neuroimaging biomarkers, including magnetic resonance imaging (MRI) and positron emission tomography (PET), have played an indispensable role in the study and diagnosis of PSP. ³ , ⁴

Quantitative susceptibility mapping (QSM) is an advanced MRI technique that visualizes magnetic susceptibility distribution in vivo and detects magnetic substances such as iron. ⁵ The estimated magnetic susceptibility is proportional to the subcortical iron concentration, ⁶ and a meta‐analysis suggested that QSM is more robust than the other iron‐sensitive MR technique in detecting abnormal iron deposition in Parkinson's disease. ⁷ Multiple QSM studies have demonstrated abnormal iron depositions in the subcortical structures in PSP, including the pallidum, substantia nigra, and red nucleus. ⁸ , ⁹ , ¹⁰ Our group recently revealed substantial differences in the spatial pattern of iron deposition across PSP clinical variants. ¹¹

However, these studies have predominantly relied on clinically diagnosed PSP (PSP syndrome [PSPS]) patients, which poses certain limitations in neurodegenerative disease research. Indeed, accurate clinical diagnosis is challenging due to the heterogeneity of clinical presentations and overlapping symptomatology. Autopsy confirmation is critical in PSP, given the considerable discrepancy between clinical and pathologic diagnoses. ¹² Other neurodegenerative disorders, particularly multiple system atrophy (MSA), can clinically mimic PSP, ¹³ , ¹⁴ , ¹⁵ leading to substantial diagnostic uncertainty. The reliance on clinical diagnoses in previous QSM studies potentially confounds the interpretation of iron deposition patterns and limits the generalizability of findings. In addition, the utility of QSM in differentiating between different underlying pathologies in PSPS patients remains unexplored, and its accuracy as an iron measurement tool has not been validated in autopsy‐confirmed PSP patients.

To address these critical limitations and establish the pathological correlates of QSM findings, we aimed to characterize QSM profiles in autopsy‐confirmed PSP and clinically atypical MSA (ca‐MSA) masquerading as PSP ¹⁴ to validate its accuracy as a measure of iron through comparison with histological iron detection using Prussian blue staining.

Methods

Participants

Nine clinically diagnosed PSP patients were prospectively recruited by the Mayo Clinic, Neurodegenerative Research Group (NRG), between May 2019 and October 2023, and subsequently died and underwent an autopsy performed by a neuropathologist (D.W.D.) at Mayo Clinic, Jacksonville. All patients underwent comprehensive neurological examinations, including the Montreal Cognitive Assessment (MoCA), ¹⁶ the Frontal Assessment Battery (FAB), ¹⁷ the Movement Disorders Society–sponsored revision of the Unified Parkinson's Disease Rating Scale Part III (MDS‐UPDRS III), ¹⁸ the PSP rating scale, ¹⁹ the PSP Saccadic Impairment Scale (PSIS), ²⁰ and a 3 T MRI scan with standardized protocol. Pathological examination revealed 7 patients with PSP pathology (PSP‐PSPS) and 2 cases of MSA pathology (ca‐MSA). After excluding 1 PSP‐PSPS case due to poor image quality, 8 patients, 6 PSP‐PSPS and 2 ca‐MSA, were included. The clinical diagnoses of 6 PSP‐PSPS included PSP‐Richardson syndrome (n = 4), PSP‐parkinsonism (n = 1), and PSP‐corticobasal syndrome (n = 1). The 2 ca‐MSA cases were diagnosed with PSP‐Richardson syndrome and PSP‐parkinsonism. Five patients had multiple visits with valid QSM scans. For comparison, 10 age‐matched cognitively ²¹ and motorically ²² normal subjects were included as controls.

MRI Analysis

All subjects underwent a standardized MRI protocol using a 3 T scanner (Magnetom Prisma, Siemens AG, Healthineers, Germany), which included a 3‐dimensional (3D) magnetization‐prepared rapid acquisition gradient echo (MPRAGE) sequence and a 3D 5‐echo gradient‐echo sequence. The detailed imaging protocol and QSM processing were the same as in previous studies. ²³ , ²⁴ , ²⁵ The QSM images were reconstructed using in‐house brain masking and STI suite ²⁵ , ²⁶ , ²⁷ and then rigidly registered to MPRAGE images. Brain atlases in the template space were nonlinearly registered to each MPRAGE space.

Magnetic susceptibility values were obtained in 7 subcortical regions of interest (ROIs), including the caudate, putamen, pallidum, subthalamic nucleus, substantia nigra, red nucleus, and cerebellar dentate, using brain atlases, as described previously. ²⁸ , ²⁹ , ³⁰ Magnetic susceptibility in the left and right hemispheres was averaged to analyze group comparison, and the magnetic susceptibility in the left hemisphere was used to compare with Prussian blue.

For the 5 patients who had multiple visits with QSM scans, we used the first visit for comparative analysis with controls to evaluate earlier diagnostic ability and the last visit for correlation analysis with Prussian blue to evaluate accuracy in the shorter time between QSM scan and autopsy.

Pathological Assessment

All patients underwent standardized assessment of the left hemisphere, including macroscopic observation, histological evaluation of hematoxylin and eosin‐stained sections and immunohistochemistry. ¹⁵ , ³¹ Immunohistochemistry was performed on a Dako autostainer using α‐synuclein (NACP antibody, rabbit polyclonal; 1:3000; formic acid pretreatment) and phosphorylated tau (CP13 antibody, mouse monoclonal; 1:1000; from the late Dr. Peter Davies; Feinstein Institute for Medical Research). PSP was diagnosed based on neurofibrillary tangles or pretangles and tufted astrocytes on phosphorylated tau‐stained sections. ² , ³² Neuropathologic diagnosis of MSA was established based on widespread glial cytoplasmic inclusions on α‐synuclein‐stained sections. ³³ , ³⁴

Iron deposition was assessed by Prussian blue staining and quantified using digital pathology techniques. ³⁵ Prussian blue staining was performed in the same 7 brain regions as the MRI analysis. The stained slides were digitally scanned at 20× magnification using an Aperio ScanScope AT2 (Leica Biosystems, Buffalo Grove, IL, USA). The scanned images were viewed and annotated using ImageScope software (version 12.4.3; Leica Biosystems, Deer Park, IL, USA). Regions of interest were outlined for each brain region. A custom color deconvolution algorithm was employed to separate and measure the blue staining intensity (Fig. S1). This analysis yielded a percentage load value, representing the proportion of the stained area within the total examined region. All image analyses were performed by investigators who were blinded to the participants’ clinical status.

Statistical Analysis

Demographics and clinical scores at the time of first QSM scans were compared between groups using a Fisher exact test for categorical values and a Wilcoxon rank‐sum test for continuous values.

Magnetic susceptibilities in the 7 subcortical regions were compared between PSP‐PSPS and controls using the Wilcoxon rank‐sum test and between PSP‐PSPS and ca‐MSA to evaluate the differentiating ability between PSP and MSA pathologies. The comparison was also performed between all PSPS (PSP‐PSPS plus ca‐MSA) and controls. False discovery rate (FDR) correction ³⁶ was applied to adjust for the multiple comparisons. The rank‐biserial correlation coefficient was calculated to assess effect size between groups, with > 0.5 considered large. Stained area was compared between PSP‐PSPS and ca‐MSA. Association between magnetic susceptibility and stained area in all ROIs was evaluated using Spearman's correlation.

Statistical analyses were performed using MATLAB 2022a Statistics and Machine Learning Toolbox (The MathWorks Inc., Natick, MA, USA).

Results

Supplementary Table S1 summarizes demographic characteristics and clinical scores. There were no differences in sex, age, or education between PSP‐PSPS and controls. Similarly, PSP‐PSPS and ca‐MSA groups showed no significant differences in age, sex, education, MoCA, FAB, PSP rating scale, UPDRS Part III, or PSIS. For the scans used in the autopsy comparison, the median interval between the last MRI and death was 0.8 years (interquartile range [IQR]: 0.35–1.35) across the cohort. Regarding clinical presentations of 2 ca‐MSA cases, 1 presented with early falls and vertical supranuclear gaze palsy, with urinary retention attributed to benign prostate hyperplasia, leading to a PSP diagnosis. The other showed early falls, partial levodopa responsiveness, and mild ocular motor abnormalities, with urinary incontinence attributed to prior transurethral resection of the prostate and autonomic reflex testing showing a peripheral pattern, resulting in a PSP‐parkinsonism diagnosis. ¹⁴

Figure 1 shows subcortical susceptibility comparisons. PSP‐PSPS showed significantly higher susceptibility compared to controls in the pallidum, subthalamic nucleus, substantia nigra, and cerebellar dentate nucleus (FDR‐corrected P < 0.05, Fig. 1A). Large effect size (≥0.5) was observed across all 7 regions examined. In the comparison between PSP‐PSPS and ca‐MSA (Fig. 1B), ca‐MSA demonstrated relatively higher susceptibility in the putamen, whereas PSP‐PSPS had relatively higher susceptibility in the subthalamic nucleus (uncorrected P = 0.08). In the comparison between all PSPS (PSP‐PSPS plus ca‐MSA) and controls, PSPS showed higher susceptibility in the putamen, pallidum, substantia nigra, and cerebellar dentate nucleus (FDR‐corrected P < 0.05, Fig. S2). Averaged QSM images in each group at the level of the basal ganglia were consistent with the regional analysis, showing higher signals in the pallidum for PSP‐PSPS and in the putamen for ca‐MSA (Fig. 1C). These features were also observed in T2*‐weighted images (Fig. S3).

Fig. 1.

Quantitative susceptibility mapping (QSM) analysis on progressive supranuclear palsy (PSP)‐PSP syndrome (PSPS) and clinically atypical multiple system atrophy (ca‐MSA). (A) Magnetic susceptibility in PSP‐PSPS and controls in the 7 subcortical regions. Asterisks indicate a significant difference, with false discovery rate (FDR)‐corrected P < 0.05 (*) and P < 0.01 (**). (B) Magnetic susceptibility in PSP‐PSPS and ca‐MSA in the 7 subcortical regions. The plus marks indicate a trend (uncorrected P < 0.1). (C) Averaged QSM maps in PSP‐PSPS (n = 6), ca‐MSA (n = 2), and controls (n = 10) at the level of the basal ganglia in the template space. Arrows indicate regions of higher magnetic susceptibility in comparison to controls. (D) Difference and ratio between susceptibilities in the putamen and pallidum. Asterisks indicate a significant difference, with P < 0.05 (*) and P < 0.001 (***).

The difference or ratio between pallidum and putamen also demonstrated large differential ability (Fig. 1D). The difference between putamen and pallidum effectively differentiated PSP‐PSPS from the others (ca‐MSA and controls) (P < 0.001). The ratio of these regions (putamen/pallidum) distinguished ca‐MSA from the others (PSP‐PSPS plus controls) (P < 0.05). No significant differences were observed in the stained iron area between PSP‐PSPS and ca‐MSA (Fig. S4).

Figure 2A illustrates the relationship between magnetic susceptibility and stained area across all 8 patients, demonstrating a significant positive correlation (r = 0.63, P < 0.001). In the subanalysis in each region, positive correlations were observed in the putamen and pallidum (r > 0.7, uncorrected P < 0.05, Fig. 2B).

Fig. 2.

Relationship between magnetic susceptibility (ppm) and area stained with Prussian blue (%). (A) Overall relationship, including all subcortical regions. Magnetic susceptibility was significantly correlated with the stained area (r = 0.63, P < 0.001, y = 0.091x + 0.043). Different colors represent different regions. (B) Relationships in each region. The significant correlations (*) were observed in the putamen and pallidum (P < 0.05).

Discussion

In this study, we investigated magnetic susceptibility changes in autopsy‐confirmed PSP patients and validated the utility of QSM as an iron measurement tool by comparison with Prussian blue staining. PSP‐PSPS demonstrated higher magnetic susceptibility in the pallidum, subthalamic nucleus, substantia nigra, and cerebellar dentate nucleus, generally consistent with previous studies of clinically diagnosed PSP. ⁸ , ⁹ , ¹⁰ , ¹¹ Furthermore, we identified distinct susceptibility differences between PSP‐PSPS and ca‐MSA in the putamen, pallidum, and subthalamic nucleus. Composite measures, including the ratio or difference between pallidum and putamen susceptibilities, enhanced these distinctions. Although previous studies mainly included clinically diagnosed PSP patients, ⁸ , ⁹ , ¹⁰ , ¹¹ this study focused on the autopsy‐confirmed PSP. This autopsy confirmation is essential in the PSP study, given the considerable discrepancy between clinical and pathologic diagnoses. Although the small sample size limited definitive conclusions, our findings suggest the potential utility of QSM in predicting underlying pathology in patients presenting with PSP‐like symptoms.

QSM profiles were slightly different between all PSPS (PSP‐PSPS and ca‐MSA) and PSP‐PSPS. The putamen showed higher susceptibility in PSPS than in controls, but not in PSP‐PSPS alone. Conversely, the subthalamic nucleus did not show higher susceptibility in PSPS than in controls, but it did in PSP‐PSPS alone. These results imply that magnetic susceptibility may be overestimated in the putamen and underestimated in the subthalamic nucleus due to the presence of ca‐MSA in the cohort of clinically diagnosed PSP. This finding emphasizes the importance of autopsy confirmation and provides insight into interpreting QSM studies on PSP.

Magnetic susceptibility generally showed positive correlations with iron concentration measured by Prussian blue in the examined regions. When all regions were included, a strong positive correlation was observed. For each region, although the small sample size limits firm conclusions, 5 of the 7 regions showed moderate‐to‐strong positive correlations (r ≥ 0.5), 2 of which were significant. These findings align with previous validation studies ⁶ , ³⁷ and further support the validity of QSM in PSP. Our results contribute to a better understanding of iron homeostasis in PSP and validate QSM as a valuable tool for studying and diagnosing this condition.

In contrast to QSM, no significant differences in the stained iron area were observed between PSP‐PSPS and ca‐MSA. One possible reason is that QSM measured iron throughout the tissue, whereas Prussian blue targeted the limited area. Although we confirmed the general relationship between QSM and iron staining, further studies with larger cohorts are needed to clarify this discrepancy.

The main limitation of this study is small sample size due to rare autopsy‐confirmed cases. Although there are only 2 ca‐MSA cases, the results of the statistical analysis, including P‐values and effect size, should be interpreted with caution. Another limitation would be the potential spatial discrepancy between MRI and histopathological annotations despite careful ROI selection. Nevertheless, this study has strength in its novel application of QSM assessment in autopsy‐confirmed PSP patients. We characterized susceptibility profiles of pathologically diagnosed PSP patients, validated QSM as an iron measurement tool, and demonstrated its ability to differentiate between PSP and MSA pathologies with similar clinical manifestations. Currently, QSM is not widely used in clinical practice due to requirements for additional processing and specialized interpretation. However, accumulating evidence of QSM's diagnostic utility, ⁵ including findings from this study and others, ¹¹ , ²⁴ , ³⁵ may facilitate its future clinical implementation.

In conclusion, we investigated magnetic susceptibility differences in autopsy‐confirmed PSP patients and validated them using Prussian blue staining. Our findings support the utility of QSM as a valid technique for measuring disease‐specific iron accumulation patterns in atypical parkinsonian disorders.

Author Roles

(1) Research project: A. Conception, B. Organization, C. Execution; (2) Statistical analysis: A. Design, B. Execution, C. Review and critique; (3) Manuscript preparation: A. Writing of the first draft, B. Review and critique.

R.S.: 1A, 1C, 2A, 2B, 3A

H.S.: 1A, 1C, 2A, 2C, 3A

F.A.: 1C, 2C, 3B

D.W.D.: 1B, 1C, 2C, 3B

J.L.W.: 1A, 1B, 1C, 2C, 3B

K.A.J.: 1A, 1B, 1C, 2C, 3B

Disclosures

Ethical Compliance Statement: This study was approved by the Mayo Clinic Institutional Review Board, and informed consent was obtained from all participants. We confirm that we have read the journal's position on issues involved in ethical publication and affirm that this work is consistent with those guidelines.

Funding sources and Conflict of Interest: This study was funded by NIH grant R01‐NS89757 (MPIs: Whitwell and Josephs). The authors declare that there are no conflicts of interest relevant to this work.

Financial disclosure for the previous 12 months: R.S. and H.S. have no financial disclosure to report. F.A., D.W.D., J.L.W., and K.A.J. receive research funding from National Institutes of Health. K.A.J. is an associate editor of the Annals of Clinical and Translational Neurology and an editorial board member of Acta Neuropathologica and Journal of Neurology.

Supporting information

Figure S1. Color deconvolution analysis of Prussian blue staining. Representative images showing the original Prussian blue–stained section (left) and its corresponding iron‐specific component after color deconvolution (right). The color deconvolution algorithm separates the blue signal specific to iron deposits from background staining, enabling quantitative analysis of iron content. Scale bar = 20 μm.

Figure S2. Magnetic susceptibility in all progressive supranuclear palsy syndrome (PSPS) cases (PSP‐PSPS plus clinically atypical multiple system atrophy [ca‐MSA]) and controls in the 7 subcortical regions. Asterisks indicate a significant difference with false discovery rate (FDR)‐corrected P < 0.05.

Figure S3. T2*‐weighted images (last echo gradient‐echo images) of 2 patients with PSP‐PSPS and 2 patients with clinically atypical multiple system atrophy (ca‐MSA). Hypointense signals were observed in the posterior putamen of ca‐MSA patients (yellow arrows).

Figure S4. Area stained with Prussian blue in PSP‐PSPS and clinically atypical multiple system atrophy (ca‐MSA) in the 7 subcortical regions.

Table S1. Demographics and clinical findings across PSP‐PSPS, ca‐MSA, and control groups. Results are presented as median (first and third quartiles) for continuous values. MoCA, Montreal Cognitive Assessment; FAB, Frontal Assessment Battery; MDS‐UPDRS III, Movement Disorder Society–sponsored revision of the Unified Parkinson's Disease Rating Scale Parts III; PSIS, PSP Saccadic Impairment Scale.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.