Autopsy validation of progressive supranuclear palsy-predominant speech/language disorder criteria

Abstract

Background:

Progressive supranuclear palsy (PSP) may present as a speech/language disorder (PSP-SL).

Objectives:

We assessed pathological correlates of PSP-SL patients who retained the suggestive of PSP-SL (s.o. PSP-SL) diagnosis versus those who progressed to possible/probable (poss./prob.) PSP.

Methods:

Thirty-four prospectively recruited s.o. PSP-SL patients completed comprehensive speech/language and neurological assessments longitudinally, died and underwent autopsy.

Results:

Twelve patients (35%) evolved to poss./prob PSP while 22 (65%) remained as s.o. PSP-SL. Pathological diagnoses differed across the groups (p=0.025). S.o. PSP-SL patients had four different neuropathologies [corticobasal degeneration (59%), PSP (13%), Pick’s disease (14%) and frontotemporal lobar degeneration with TDP-43 (14%)] while all poss./prob PSP patients had a 4R-tauopathy [PSP (67%) and corticobasal degeneration (33%)]. Development of poss./prob. PSP increased the chance of having PSP pathology by 2.38 times.

Conclusion:

PSP-SL is associated with heterogenous pathologies. Evolution of PSP-SL into poss./prob. PSP is more predictive of underlying PSP pathology than s.o. PSP-SL.

Introduction

Progressive supranuclear palsy (PSP) with predominant speech/language disorder (PSP-SL) is a clinical variant of PSP¹. PSP-SL patients present with progressive apraxia of speech (PAOS), a motor speech disorder affecting planning and programming of speech, or agrammatic aphasia, a language disorder affecting grammar production. According to the Movement Disorder Society clinical criteria for PSP (MDS-PSP)², patients with PAOS and/or agrammatic aphasia fulfill criteria for suggestive of PSP-SL (s.o. PSP-SL) and if they develop vertical supranuclear gaze palsy or slowing of vertical saccades they meet criteria for possible PSP-SL (poss. PSP-SL).

Most patients with PAOS develop a Parkinson-plus syndrome^{3, 4}, characterized by bradykinesia, rigidity, postural instability, ocular motor impairment and ideomotor apraxia^{3, 5, 6}, and approximately 50% eventually meet poss. PSP-SL criteria during follow-up⁷. Furthermore, some of these patients develop falls within 3 years and akinesia and meet probable PSP (prob. PSP) criteria⁷. At autopsy, PAOS patients most commonly have an underlying 4R tauopathy, either corticobasal degeneration (CBD) or PSP, although cases have been reported with Pick’s disease (PiD) or frontotemporal lobar degeneration with TDP-43 (FTLD-TDP)^{1, 8}. The relationship between underlying pathology and diagnostic certainty of the MDS-PSP criteria, and specifically whether clinical diagnosis of poss./prob. PSP increases specificity of underlying PSP pathology, is unknown.

In this study we evaluated 34 s.o. PSP-SL patients who were followed longitudinally for a median of 9.2 years, died and had brain autopsy. We assessed and compared underlying pathological diagnoses in patients that remained s.o. PSP-SL during follow-up and those who progressed to poss./prob. PSP. We hypothesized that patients who developed poss./prob. PSP would more likely have PSP pathology.

Methods

The Neurodegenerative Research Group (NRG) at Mayo Clinic, Rochester, consecutively recruited patients presenting with PAOS and/or agrammatic aphasia from the Neurology Department between 7/1/2010– 4/1/2021. All patients met MDS-PSP criteria for s.o. PSP-SL². Patients were followed longitudinally with annual speech/language, neurological, and neuropsychological evaluations⁹. To be included in this study patients must have had a brain autopsy (n=34). The MDS-PSP criteria were applied at final visit before death, operationalized as published⁷. Briefly, MDS-PSP ocular motor impairment level (O1–3) was determined using the PSP Saccadic Impairment Scale (PSIS)¹⁰; postural stability (P1–3) and akinesia (A1–3) using sub-scores from the PSP Rating Scale (PSP-RS)¹¹ and the Movement Disorders Society Sponsored revision of the Unified Parkinson’s Disorder Rating Scale part III (MDS-UPDRS III)¹². At final visit patients were categorized into those that remained as s.o. PSP-SL and those that progressed to either a possible or probable PSP diagnosis (poss./prob. PSP). Other clinical outcome measures assessed included the Montreal Cognitive Assessment battery (MoCA)¹³ for general cognition; Frontal Assessment Battery (FAB)¹⁴ for executive function; Frontal Behavioral Inventory (FBI)¹⁵ for abnormal behaviors (cut-point ≥27 indicate frontal lobe dementia¹⁵); Western Aphasia Battery-Aphasia Quotient (WAB-AQ)¹⁶ for aphasia severity; WAB-Praxis subtest for limb apraxia; and the Motor Speech Disorders (MSD) scale¹⁷ for motor speech disorder severity (AOS and/or dysarthria). Features of corticobasal syndrome (CBS)¹⁸ were assessed. Of the 34 patients in this autopsy study, 32 were included in a previous longitudinal clinical/imaging study characterizing PAOS across pathologies⁸; the MDS-PSP criteria were not assessed in that study. One patient was included in a case-report⁴. The Mayo Clinic IRB approved this study. All patients gave consent to participate.

Neuropathological evaluations were performed in accordance with current diagnostic protocol, as previously described^{19, 20}. Pathological diagnoses were rendered according to established criteria: PSP was diagnosed based upon presence of characteristic neuronal (pretangles and tangles) and glial lesions (tufted astrocytes and oligodendroglial coiled bodies) in vulnerable cortical and subcortical regions²¹; CBD by the presence of cortical and subcortical neuronal and glial lesions (astrocytic plaques) and thread-like processes in gray and white matter²²; FTLD-TDP type A had TDP-43 immunoreactive neuronal cytoplasmic inclusions, dystrophic neurites and intranuclear inclusions, while type B had predominantly neuronal cytoplasmic inclusions²³. PiD showed argyrophilic and tau-immunoreactive Pick bodies, negative on Gallyas silver stain²⁴.

Statistical analyses were done using SPSS 27.0, with T-test or Mann Whitney-U depending on distribution. Binary logistic regression was used to investigate relationships between pathology and MDS-PSP clinical diagnosis.

Raw data for this study is available upon request to the corresponding author.

Results

Of the 34 patients, 22 (65%) maintained a clinical diagnosis of s.o. PSP-SL at last visit before death, while 12 (35%) met criteria for poss./prob. PSP (10 for probable PSP with predominant parkinsonism [PSP-P], one for probable PSP Richardson’s syndrome [PSP-RS] and one for possible PSP-SL). The groups did not differ in demographics at last visit except on disease duration and age (Table 1). The s.o. PSP-SL group had shorter time from onset to death and time from onset to last visit than poss./prob. PSP. No difference was observed in time from last visit to death. The poss./prob. PSP patients were older at last visit and death. The poss./prob. PSP patients were more impaired on PSIS, UPDRS I and III, FAB, WAB praxis and MSD at last visit compared to s.o. PSP-SL (Table 1). The frequency of CBS features did not differ across groups with the majority in both groups meeting criteria for probable CBD¹⁸. AOS subtype²⁵ differed with prosodic subtype being more frequent in poss./prob. PSP and phonetic subtype more frequent in s.o. PSP-SL (Table 1).

Table 1.

Demographics and neurological data at last visit before death for the cohort.

Variable	s.o. PSP-SL (n=22)	poss./prob. PSP (n=12)	p-value
Gender (% male)	9 (40.9)	7 (58.3)	0.331
Handedness (% right)	18 (81.8)	9 (75.0)	0.456
Education (years)	14.50 (12.00, 18.00)	16.00 (15.00, 17.50)	0.423
Age at onset (years)	61.97 (53.78, 67.66)	69.25 (56.75, 74.09)	0.123
Time from onset to last visit	4.82 (3.60, 6.87)	9.37 (8.15, 10.06)	<0.001
Age at last visit (years)	67.78 (59.40, 72.07)	78.81 (66.96, 82.90)	0.005
Time from last visit to death	2.06 (0.73, 3.81)	1.18 (0.78, 2.68)	0.261
Age at death	69.43 (60.41, 74.20)	80.00(68.18, 84.46)	0.009
Disease duration (onset to death)	7.38 (5.08, 9.76)	10.61 (9.52, 11.74)	0.003
Clinical data at last visit
PSP Rating Scale	34.00 (15.00, 67.00)	53.50 (45.00, 60.75)	0.126
PSIS	1.00 (0.00, 1.00)	2.50 (2.00, 3.00)	<0.001
MDS-UPDRS I	9.00 (5.50, 16.50)	18.00 (11.00, 21.00)	0.037
MDS-UPDRS II	16.00 (4.00, 38.00)	36.00 (33.00, 41.00)	0.135
MDS-UPDRS III	22.00 (10.00, 52.25)	68.00 (58.75, 77.25)	0.003
MoCA	21.00 (13.50, 24.50)	18.00 (10.75, 24.00)	0.334
FAB	14.00 (8.50, 14.75)	6.50 (1.75, 9.75)	0.007
FBI	26 (21, 37)	24 (15, 27)	0.240
FBI > 27	10/21 (48%)	3 (25%)	0.278
WAB- Praxis	49.00 (41.75, 54.50)	30.00 (23.00, 43.00)	0.019
WAB- AQ	81.95 (74.85, 95.10)	92.30 (90.60, 95.35)	0.158*
MSD	3.00 (2.50, 6.50)	1.00 (1.00, 4.00)	0.022
Phonetic AOS/prosodic AOS	17 (85%)/3 (15%)†	4 (33%)/8 (67%)	0.003
CBS features at last visit
Limb rigidity or akinesia	16 (73%)	12 (100%)	0.069
Limb dystonia	4 (18%)‡	1 (8%)‡	0.635
Limb myoclonus	1 (5%)‡	1 (8%)‡	0.999
Orobuccal or limb apraxia	19 (86%)	12 (100%)	0.537
Cortical sensory deficit	NA	NA	NA
Alien limb phenomena	4 (18%)¶	0 (0%)	0.273
Met criteria probable sporadic CBD	20 (91%) α	12 (100%)	0.529

Values are shown as median (q1, q3) and % = percentage of patients. Significant comparisons are bolded. s.o. PSP-SL = suggestive progressive supranuclear palsy with predominant speech/language disorder; poss./prob. PSP = possible/probable progressive supranuclear palsy; PSIS = progressive supranuclear palsy saccadic impairment scale; MDS-UPDRS (I, II and III) = Movement Disorders Society Unified Parkinson’s Disorder Rating Scale (I, II and III); MoCA = Montreal Cognitive Assessment; FAB = Frontal Assessment Battery; FBI = Frontal Behavioral Inventory; WAB = Western Aphasia Battery; WAB-AQ = Western Aphasia Battery- Aphasia Quotient; MSD = Motor speech disorder rating; NA = not able to provide data.

^*WAB-AQ was only available in 53% of the cohort. Patients became untestable due to severity of apraxia of speech and mutism.

^†Two patients did not have AOS and were diagnosed with progressive agrammatic aphasia.

^‡All patients with these clinical features had CBD pathology

^¶Three of the four patients with alien limb phenomena had CBD pathology and one had PSP pathology

^αOne patient with CBD pathology did not meet probable CBD criteria only because they were under age 50

Figure 1 shows the distribution of pathological diagnoses across groups. In s.o. PSP-SL, three (14%) had PSP, 13 (59%) CBD, three (13%) FTLD-TDP (two type-A and one type-B) and three PiD (Figure 1a). In poss./prob. PSP, 8 (67%) had PSP and 4 CBD (%33) (Figure 1b). Stated another way, of the 11 patients with PSP, 8 (73%) were poss./prob. PSP before death and 3 (27%) were s.o. PSP-SL before death (Figure 1c). the 23 patients with pathologies other than PSP (CBD, FTLD-TDP and PiD), 19 (83%) were s.o. PSP-SL and 4 (17%) were poss./prob. PSP (Figure 1d). Adjusting for time from disease onset to last visit, the development of poss./prob. PSP increased the chances of PSP pathology by 2.38X compared to other pathologies (the unstandardized beta weight for the outcome (B)= 2.377, standard error= 1.059, Wald= 5.033, p=0.025).

FIGURE 1:

Pie charts illustrating the relationship between clinical diagnosis and underlying pathology. (AandB) The frequency of pathologies underlyings.o. PSP-SL and poss./prob. PSP are shown. (CandD) The frequencies of clinical diagnoses in patients with PSP pathology or non-PSP pathologyare shown. CBD, corticobasal degeneration; FTLD-TDP, frontotemporal lobar degeneration with TDP-43 inclusions; PiD, Pick’s disease; poss./prob.,possible/probable; PSP, progressive supranuclear palsy; s.o. PSP-SL, suggestive progressive supranuclear palsy with predominant speech/languagedisorder.

Discussion

This study of a large autopsy cohort of speech/language patients shows that patients who develop possible or probable PSP according to the MDS-PSP criteria are more likely to have PSP pathology compared to those who remain s.o. PSP-SL. The s.o. PSP-SL group had heterogeneous pathologies, including 4R and non-4R tauopathies.

These findings have clinical implications. When a patient is diagnosed with s.o. PSP-SL, at least four pathological diagnoses are possible. MDS-PSP criteria were previously shown to have high specificity for probable 4R tauopathies²⁶. Our findings agree since all patients with poss./prob PSP had a 4R tauopathy. The s.o criteria which has been shown to be sensitive²⁷ was demonstrated not to be specific for a 4R tauopathy. Over the disease course, if the patient evolves into poss./prob. PSP, he/she is 2.38X more likely to have PSP pathology. In other words, a patient remaining as s.o. PSP-SL over years is less likely to have PSP pathology, supporting the findings of another study²⁸. The increased likelihood of PSP pathology with increased diagnostic certainty not only provides physicians the ability to give prognostic estimates and counseling but also supports the clinical utility of MDS-PSP criteria, which were shown to have higher sensitivity compared to the Litvan et al. 1996 criteria^{28, 29}. Genetic testing could help diagnostically in the s.o. PSP-SL patients since, as we previously reported⁸, two FTLD-TDP cases had a progranulin mutation.

The patients who developed poss./prob. PSP had longer disease duration and were older at death compared to those who remained s.o. PSP-SL. This finding was not due to pathological diversity of the s.o. PSP-SL group as these differences persisted after the FTLD-TDP and PiD cases were removed from the analysis. The finding could have been driven by a difference in frequencies of CBD and PSP. In our previous clinicopathological study of this cohort, patients with CBD had more rapid clinical decline than those with PSP⁸, while one study reported similar survival between patients clinically diagnosed with PSP and CBS³⁰. The disease duration of 7.38 years in the s.o. PSP-SL group is similar to the median survival reported in CBD of 7.9 years, although in that cohort survival was not worsened by PSP-related features³¹. The longer disease duration may explain why the poss./prob. PSP group performed worse clinically at the last visit before death on several tests. There was a tendency for the s.o. PSP-SL group to have worse aphasia, although differences were not significant, possibly because many patients were aphemic or anarthric and untestable. We found that prosodic AOS was more common in the patients that progressed to poss./prob PSP. This concurs with our previous finding that prosodic AOS is associated with older age and PSP pathology⁸.

As previously described^{3, 5}, patients with PAOS commonly evolve into an atypical parkinsonian disorder characterized by features of both PSP and CBS. Indeed, most s.o. PSP-SL and poss./prob. PSP patients in this study also met probable CBD criteria at last study visit, with limb rigidity/akinesia and orobuccal/limb apraxia being the most common CBS features. Motor function and ideomotor apraxia were worse in the poss./prob. PSP group, although we did not find ideomotor apraxia severity to differ between those with CBD versus PSP pathology. Limb dystonia and myoclonus were rarely observed but when present always occurred in CBD cases. Many of the poss./prob. PSP patients met probable PSP-P criteria due to the development of akinesia and ocular motor impairment; postural instability and falls were common but rarely occurred within the first three years⁷. Many s.o. PSP-SL patients would also meet criteria for s.o. PSP-P and s.o. PSP-CBS, although using MAX criteria 2 the final diagnosis in all cases was s.o. PSP-SL. Behavioral dyscontrol also developed later on in many patients in both groups, although none met behavioral variant frontotemporal dementia criteria³².

The strengths of the study were that all patients had standardized clinical assessments, were evaluated, and diagnosed at baseline by speech-language pathologists and followed longitudinally until approximately 1-year before death which allowed us to capture progression of PSP symptoms.

In summary, s.o. PSP-SL is clinically and pathologically heterogeneous. However, those who develop poss./prob. PSP diagnoses are more likely to have PSP pathology and have longer disease duration which is clinically relevant when discussing prognosis with patients.