Abstract

Although magnetic resonance imaging is a standard investigation in neurodegenerative disease, sensitive and specific markers for the underlying histopathological diagnosis are largely lacking. This report presents evidence to indicate that corticobasal degeneration and progressive supranuclear palsy, in particular, might be identifiable at a single subject level with diffusion tensor imaging. Patients with clinical diagnoses of Alzheimer’s disease, semantic dementia and non-fluent primary progressive aphasia (n = 9 each) were contrasted with control subjects (n = 26) with the diffusion tensor imaging measures: fractional anisotropy, axial and radial diffusivity. At 1 year follow-up, all participants with non-fluent primary progressive aphasia had evolved either corticobasal degeneration (n = 5) or progressive supranuclear palsy (n = 4). The corticobasal degeneration/progressive supranuclear palsy set showed white matter abnormalities involving the entire cerebrum. Individual maps were similar to the group level results, even in the most minimally impaired patients. Fractional anisotropy was consistently the most sensitive metric. In Alzheimer’s disease and semantic dementia, by contrast, group level and individual analyses revealed limited areas of abnormality centred on the posterior cingulate and rostral temporal lobes, respectively. In both groups radial diffusivity was the most sensitive metric. Scrutiny of the standard scores for each group’s most sensitive metric revealed that, although the values for every patient with corticobasal degeneration or progressive supranuclear palsy fell outside 95% of the normal mean, none of the other two groups’ members had values outside this range. Further underscoring the hypothesis that this finding relates specifically to a diffuse pathological process in the white matter of the tauopathies, and is not merely a function of disease severity, a grey matter analysis consisting of group level voxel-based morphometry revealed only focal areas of atrophy in all three groups. Consistent with past reports for the respective clinical syndromes, these were centred on the left frontal operculum and caudate nucleus in non-fluent primary progressive aphasia (the corticobasal degeneration/progressive supranuclear palsy set), anterior temporal lobes in semantic dementia, and hippocampus and posterior cingulate gyrus in Alzheimer’s disease. Detection of this extensive white matter lesion in corticobasal degeneration and progressive supranuclear palsy—a pathologically proven feature of these conditions—in single subjects with diffusion tensor imaging appears to have strong diagnostic marker potential for these diseases.

Introduction

The pathological underpinning of a clinical dementia syndrome is a probabilistic diagnosis in life. In some circumstances, for instance where all indicators point to a typical clinical presentation of Alzheimer’s disease, the likelihood of Alzheimer’s pathology is exceedingly high (Alladi et al., 2007). Some clinical dementia syndromes are, however, far more heterogeneous in terms of their pathological substrate, and for these more specific in vivo imaging markers of pathology are desirable. Primary progressive aphasia is arguably the most difficult of these as it can be associated with ubiquitin-staining pathology [usually TAR DNA-binding protein 43 (TDP-43)-positive], tauopathies (such as corticobasal degeneration or progressive supranuclear palsy) or Alzheimer’s pathology (Gorno-Tempini et al., 2011).

The advent of amyloid ligand imaging with PET has enabled in vivo visualization of Alzheimer’s pathology. On the other hand, the high prevalence of incidental Alzheimer’s pathology in later life, plus emerging evidence that this scan modality becomes positive for amyloid years before onset of symptoms, mean that its most definitive clinical role would be in ruling out Alzheimer’s disease—i.e. a ‘positive’ amyloid scan does not prove a causal role for Alzheimer’s pathology in a clinical dementia syndrome (Caso et al., 2012). Differentiating between different types of dementia-causing non-Alzheimer’s pathologies with imaging is even more difficult as there are at present no specific ligands for inclusions such as tau, TDP-43 or alpha synuclein etc.

A further challenge to developing predictors of non-Alzheimer’s pathologies is identification of robust markers for single subject diagnosis. Although a vast number of imaging studies have been published on the atrophy or metabolic profiles of clinical dementia syndromes, very few have examined discriminability of differing pathologies in clinically ambiguous presentations. Furthermore, studies taking this further important step have almost all reported group-level findings without single subject validation. This report presents the first evidence to indicate that the tauopathies—corticobasal degeneration and progressive supranuclear palsy—might be identifiable at a single-subject level with diffusion tensor imaging.

Materials and methods

Participants

The study cohort comprised four groups: nine participants with an initial clinical diagnosis of non-fluent primary progressive aphasia who at 12-months follow-up had evolved a pattern of either corticobasal degeneration (five patients) or progressive supranuclear palsy (four patients, one of whom has since died and was confirmed at necropsy to have progressive supranuclear palsy pathology); nine patients with a clinical presentation of typical Alzheimer’s disease; nine patients with a diagnosis of semantic dementia; and 26 healthy, age- and education-matched control participants. All patients were recruited from the memory clinics held at Addenbrooke’s Hospital, University of Cambridge, UK. Written informed consent was obtained from the participants and, where appropriate, their next of kin. The study was approved by the regional Ethics Committee.

The diagnoses of neurodegenerative syndromes were made in accordance with accepted criteria for diagnosis of Alzheimer’s disease (Dubois et al., 2007), semantic dementia (Hodges and Patterson, 2007), non-fluent primary progressive aphasia (Neary et al., 1998; Mesulam, 2001), corticobasal degeneration (Litvan et al., 1997) and progressive supranuclear palsy (Litvan et al., 1996). Due to the presumed similar underlying pathology in corticobasal degeneration and progressive supranuclear palsy (Boeve et al., 2003; Scaravilli et al., 2005), scans for these two subgroups were combined to form a corticobasal degeneration/progressive supranuclear palsy set for the group level analyses. Prospective participants were excluded if there was evidence of significant leukoaraiosis in their T2-weighted imaging (excluded three patients). All control participants were free of cognitive and psychiatric illnesses as evidenced by their performance in clinical and neuropsychological evaluations.

Neuropsychological battery

All patients underwent a detailed neuropsychological assessment prior to imaging. The tests included global measures: Mini-Mental State Examination (Folstein et al., 1975) and the revised version of Addenbrooke’s Cognitive Examination (ACE-R) (Mioshi et al., 2006); forward and backward digit span subtests of the Wechsler Memory Scale (Wechsler, 1997); letter and category fluency subtests of the ACE-R; the Camel and Cactus Test of non-verbal semantic abilities (Bozeat et al., 2000); the 64-item naming subtest of the Cambridge semantic memory battery (Hodges and Patterson, 2007); cube analysis from the Visual and Object Space Perception battery (VOSP) (Warrington and James, 1991); Delis-Kaplan Executive Function System (D-KEFS) Trail making test (Delis et al., 2001); and copy and delayed recall of the Rey-Osterrieth complex figure. We also designed a novel apraxia assessment battery for assessment of orobuccal and limb apraxia for both transitive and intransitive gestures.

Predictive Analytics SoftWare version 18 was used for analysis of the neuropsychological data. For the parametric subset of the markers, one-way ANOVA was used to compare the performance in different groups and Gabriel’s test was used for post hoc comparisons. For variables with a skewed distribution, the Kruskal-Wallis test was used for multiple group comparisons and the Mann Whitney U test with Bonferroni correction to explore the underlying group differences when the Kruskal-Wallis test result was significant. In all neuropsychological tests, a two tailed P-value < 0.05 was considered significant.

Imaging

Study participants were scanned within an average of 1.1 months [standard deviation (SD) 0.6 months] from cognitive assessment. In the corticobasal degeneration/progressive supranuclear palsy set, the scans were obtained while the patients still had a clinical diagnosis of non-fluent primary progressive aphasia with no definite clinical evidence of conversion to either corticobasal degeneration or progressive supranuclear palsy. All MRI scans were performed on a Siemens Trio 3 T system (Siemens Medical Systems).

Diffusion tensor imaging

Whole-brain diffusion experiments were sensitized along 63 non-colinear directions and one b-value of 1000 s/mm2. Full details of the data acquisition protocol have been published previously (Acosta-Cabronero et al., 2011). Diffusion tensor imaging maps for axial and radial diffusivity and fractional anisotropy were generated, skeletonized and tested using standard tract-based spatial statistics (TBSS v1.2) (Smith et al., 2006) as described elsewhere (Acosta-Cabronero et al., 2011). For group-level comparisons, patient-cohort diffusion tensor imaging maps were contrasted against those from control participants and the resulting statistical maps were thresholded at the stringent P-value of 0.01, corrected for multiple comparisons. Tract-based spatial statistics was also utilized to detect white matter abnormalities in individual subjects by testing diffusion tensor imaging data from single patients against the control group. The permutation-based method of tract-based spatial statistics enables the identification of clusters in which diffusivity (or anisotropy) from a single subject ranks worse than all control participants. The stringency of the statistical threshold is therefore dictated by the size of the control group. In other words, the most stringent P-value would be equal to one divided by the total number of possible permutations, which in turn is defined by the number of control participants. Controls (n = 26) allowed for thresholding the statistical maps at P < 0.04, uncorrected for multiple comparisons.

For each metric, we also calculated the proportion of abnormal voxels at the centre of white matter tracts. Group-level results represent the number of abnormal voxels—inferred from the corrected statistical maps thresholded at 0.01—divided by the total number of voxels included in the analysis. For individual patients, the threshold used to infer the extent of abnormality was 0.04 uncorrected. Note that in order to avoid the arbitrary limitations imposed by statistical tests, we also computed standard scores for each diffusion tensor imaging metric at individual and group levels.

Volumetric T1 imaging

T1-weighted anatomical images were acquired using 3D MP-RAGE with the following imaging parameters: repetition time/echo time/inversion time/flip angle = 2300 ms/2.86 ms/900 ms/9°, 144 slices, 192 × 192 matrix dimensions and 1.25 × 1.25 × 1.25 mm3 voxel size. Full details of the acquisition protocol can be found elsewhere (Acosta-Cabronero et al., 2011). All obtained T1 volumes were preprocessed by bias correction and skull-stripping (Acosta-Cabronero et al., 2008) before group level analysis. We used two sample t-test implemented in statistical parametric mapping version 5 (Ashburner and Friston, 2005) for voxel-based morphometry of the patient groups’ grey matter volumes against controls. The statistical maps were thresholded at a stringent level of P < 0.01, corrected for multiple comparisons (false discovery rate = 0.01). In all calculations, age and total intracranial volumes [derived from the sum of grey matter, white matter and CSF (Pengas et al., 2009)] were fed into the statistical models as nuisance covariates.

Results

Demographics and neuropsychology

Table 1 summarizes the demographic and neuropsychological data for each participating group. There was no statistical difference between the groups for any of the demographic parameters. The three patient groups showed comparable performances in the general neuropsychological tests (Mini-Mental State Examination and ACE-R) that were significantly worse than controls. At a group level, while patients with Alzheimer’s disease were impaired at non-verbal memory and visuospatial tasks, patients with semantic dementia were particularly poor at naming and the test of non-verbal semantic knowledge (Camel and Cactus Test). The corticobasal degeneration/progressive supranuclear palsy group, on the other hand, had relatively preserved semantic knowledge but was impaired at tests of executive function and visuospatial tasks. Also it was the only group with a significant impairment in the assessments of orobuccal and limb praxis.

Table 1

Between-group comparison of demographic and general neuropsychological markers

 CBD/PSP Alzheimer’s disease Semantic dementia Control Omnibus significance (P-value) 
Median (range) (n = 9) Median (range) (n = 9) Median (range) (n = 9) Median (range) (n = 26) 
Demographics      
Age at test (years) 64 (58–73) 66 (61–74) 65 (51–73) 69 (57–79) NS 
Symptom duration (years) 3ċ0 (2–5) 4.5 (3–7) 4.0 (2–9) NA NS 
Education (years) 12 (10–19) 11 (10–16) 14 (10–19) 13 (10–18) NS 
Sex 2 male, 7 female 4 male, 5 female 5 male, 4 female 11 male, 15 female NA 
Handedness 9 right 8 right 8 right 24 right NA 
Neuropsychological tests      
MMSE (30) 21 (15–28)*** 19 (12–23)*** 22 (15–27)*** 30 (28–30) < 0.001 
ACE-R (100) 57 (28–89)*** 55 (52–69)*** 53 (22–84)*** 97 (88–99) < 0.001 
Forward digit span 4 (3–6)***,§§ 5 (3–6)* 7 (4–8) 6 (5–8) < 0.001 
Backward digit span 2 (1–4)***,§§ 3 (1–5)** 4 (2–6) 5 (4–7) < 0.001 
TRAIL-A 148 (47–307)***,§§ 95 (42–294)* 43 (35–98) 32 (22–62) < 0.001 
Letter fluency (wpm) 5 (2–9)*** 8 (3–16)* 6 (1–8)*** 17 (4–21) < 0.001 
Category fluency (wpm) 9 (5–16)*** 13 (5–15)** 4 (2–8)***,† 19 (13–31) < 0.001 
Camel and Cactus Test (64) 54 (38–56)* 44 (27–56)* 31 (0–45)*** 61 (53–63) < 0.001 
64 item naming (64) 57 (49–64),§§§ 54 (27–56),§§§ 25 (2–49)*** 63 (58–64) < 0.001 
VOSP (Cube analysis) (10)a 7 (3–10)** 6 (2–10)** 10 (9–10) 10 (6–10) < 0.001 
Rey copy (36) 19 (11–29)***,§§ 17 (5–34)***,§§§ 34 (30–36) 35 (27–36) < 0.001 
Rey recall (36)a 14.0 (6–23)**,† 3.5 (0–6.5)*** 17.0 (2–25) **,† 21.5 (7–29) < 0.001 
Apraxia battery(14)a 9 (3–14)*** 14 (12–14)*,# 13 (8–14)*,## 14 (13–14) < 0.001 
 CBD/PSP Alzheimer’s disease Semantic dementia Control Omnibus significance (P-value) 
Median (range) (n = 9) Median (range) (n = 9) Median (range) (n = 9) Median (range) (n = 26) 
Demographics      
Age at test (years) 64 (58–73) 66 (61–74) 65 (51–73) 69 (57–79) NS 
Symptom duration (years) 3ċ0 (2–5) 4.5 (3–7) 4.0 (2–9) NA NS 
Education (years) 12 (10–19) 11 (10–16) 14 (10–19) 13 (10–18) NS 
Sex 2 male, 7 female 4 male, 5 female 5 male, 4 female 11 male, 15 female NA 
Handedness 9 right 8 right 8 right 24 right NA 
Neuropsychological tests      
MMSE (30) 21 (15–28)*** 19 (12–23)*** 22 (15–27)*** 30 (28–30) < 0.001 
ACE-R (100) 57 (28–89)*** 55 (52–69)*** 53 (22–84)*** 97 (88–99) < 0.001 
Forward digit span 4 (3–6)***,§§ 5 (3–6)* 7 (4–8) 6 (5–8) < 0.001 
Backward digit span 2 (1–4)***,§§ 3 (1–5)** 4 (2–6) 5 (4–7) < 0.001 
TRAIL-A 148 (47–307)***,§§ 95 (42–294)* 43 (35–98) 32 (22–62) < 0.001 
Letter fluency (wpm) 5 (2–9)*** 8 (3–16)* 6 (1–8)*** 17 (4–21) < 0.001 
Category fluency (wpm) 9 (5–16)*** 13 (5–15)** 4 (2–8)***,† 19 (13–31) < 0.001 
Camel and Cactus Test (64) 54 (38–56)* 44 (27–56)* 31 (0–45)*** 61 (53–63) < 0.001 
64 item naming (64) 57 (49–64),§§§ 54 (27–56),§§§ 25 (2–49)*** 63 (58–64) < 0.001 
VOSP (Cube analysis) (10)a 7 (3–10)** 6 (2–10)** 10 (9–10) 10 (6–10) < 0.001 
Rey copy (36) 19 (11–29)***,§§ 17 (5–34)***,§§§ 34 (30–36) 35 (27–36) < 0.001 
Rey recall (36)a 14.0 (6–23)**,† 3.5 (0–6.5)*** 17.0 (2–25) **,† 21.5 (7–29) < 0.001 
Apraxia battery(14)a 9 (3–14)*** 14 (12–14)*,# 13 (8–14)*,## 14 (13–14) < 0.001 

The numbers in the parentheses represent the maximum scores for the tests; symptom duration was based on caregiver accounts of when the patients’ cognitive decline first began to emerge.

*P < 0.05 compared to controls; **P < 0.01 compared to controls; ***P < 0.001 compared to controls; #P < 0.01 compared to corticobasal degeneration; ##P < 0.001 compared to corticobasal degeneration; P < 0.01 compared to Alzheimer’s disease; §P < 0.05 compared to semantic dementia; §§P < 0.01 compared to semantic dementia; §§§P < 0.001 compared to semantic dementia.

CBD = corticobasal degeneration; MMSE = Mini-Mental State Examination; NS = non-significant; NA = not applicable; PSP = progressive supranuclear palsy; VOSP = Visual Object and Space Perception Battery; wpm = words per minute.

aNon-parametric test.

Imaging analysis

Corticobasal degeneration/progressive supranuclear palsy

As previously mentioned, the study cohort included four and five cases, respectively, whose later clinical profiles indicated progressive supranuclear palsy and corticobasal degeneration; diffusion tensor imaging maps for these subgroups were combined for group level comparisons. The corticobasal degeneration/progressive supranuclear palsy group (n = 9) showed extensive fractional anisotropy and radial diffusivity white matter abnormalities essentially involving the entire cerebrum but with more intense predilection for anterior regions (Fig. 1). Fractional anisotropy appeared to be the most sensitive metric (Table 2) in detecting group level abnormalities, closely followed by radial diffusivity; axial diffusivity revealed relatively restricted, mainly anterior, areas of tract degeneration.

Figure 1

Group level (P < 0.01 corrected) demonstration of white matter abnormalities for diffusivity and anisotropy metrics (top three rows) and grey matter atrophy (bottom row). AD = Alzheimer disease; CBD = corticobasal degeneration; GM = grey matter; PSP = progressive supranuclear palsy; SD = semantic dementia.

Figure 1

Group level (P < 0.01 corrected) demonstration of white matter abnormalities for diffusivity and anisotropy metrics (top three rows) and grey matter atrophy (bottom row). AD = Alzheimer disease; CBD = corticobasal degeneration; GM = grey matter; PSP = progressive supranuclear palsy; SD = semantic dementia.

Table 2

Individual and group level proportions of abnormal voxels and standard scores for values of different metrics at the centre of white matter tracts

Patients ACE-R score FA abnormal voxels (%) RD abnormal voxels (%) λ1 abnormal voxels (%) FA standard score RD standard score λ1 standard score 
PSP1 79 36.71 28.35 4.13 −2.97 2.79 0.94 
CBD1 77 53.34 46.64 22.70 −4.05 3.98 1.76 
PSP2 89 31.18 19.63 2.72 −2.37 1.96 0.38 
CBD2 71 35.48 35.35 26.01 −3.03 2.84 2.65 
PSP3 60 14.54 15.12 3.40 −1.99 1.97 1.06 
CBD3 47 34.08 33.87 19.63 −3.10 3.52 2.18 
PSP4 47 38.24 25.59 5.03 −2.66 2.24 0.12 
CBD4 28 26.77 25.06 6.15 −2.02 2.26 0.47 
CBD5 35 28.15 25.14 11.04 −2.26 2.46 2.49 
CBD/PSP 37.98 34.31 9.62 −2.71 2.67 1.34 
AD1 57 3.65 3.91 0.1 −0.46 0.61 0.18 
AD2 56 3.16 1.49 0.45 −0.65 0.32 0.12 
AD3 63 2.43 4.29 2.81 −0.51 0.93 0.74 
AD4 60 2.28 6.38 3.46 −0.49 0.93 0.26 
AD5 67 0.12 1.29 6.70 1.1 −0.12 1.21 
AD6 69 0.11 0.47 0.41 0.08 0.16 0.02 
AD7 52 3.10 5.98 3.25 1.02 −0.54 −1.47 
AD8 55 0.47 0.82 0.26 0.02 −0.31 −0.71 
AD9 61 1.18 1.42 0.72 −0.77 0.49 0.03 
Alzheimer’s disease 0.81 4.34 1.98 −0.07 0.27 0.04 
SD1 52 2.47 4.79 2.93 0.47 −0.35 −0.82 
SD2 22 3.27 3.95 1.20 −0.76 0.74 0.03 
SD3 43 1.25 1.55 0.74 −0.31 −0.15 −0.91 
SD4 84 0.57 2.83 1.36 −0.16 0.29 0.50 
SD5 53 1.67 2.26 0.51 −0.80 0.17 −1.03 
SD6 61 3.57 9.21 4.93 −0.76 1.00 0.91 
SD7 70 5.05 9.88 4.06 −0.84 1.16 0.79 
SD8 55 8.75 9.39 3.12 −1.78 1.92 0.88 
SD9 60 4.60 6.41 2.54 −1.00 1.03 0.94 
Semantic dementia 0.78 3.03 0.59 −0.66 0.65 0.14 
Patients ACE-R score FA abnormal voxels (%) RD abnormal voxels (%) λ1 abnormal voxels (%) FA standard score RD standard score λ1 standard score 
PSP1 79 36.71 28.35 4.13 −2.97 2.79 0.94 
CBD1 77 53.34 46.64 22.70 −4.05 3.98 1.76 
PSP2 89 31.18 19.63 2.72 −2.37 1.96 0.38 
CBD2 71 35.48 35.35 26.01 −3.03 2.84 2.65 
PSP3 60 14.54 15.12 3.40 −1.99 1.97 1.06 
CBD3 47 34.08 33.87 19.63 −3.10 3.52 2.18 
PSP4 47 38.24 25.59 5.03 −2.66 2.24 0.12 
CBD4 28 26.77 25.06 6.15 −2.02 2.26 0.47 
CBD5 35 28.15 25.14 11.04 −2.26 2.46 2.49 
CBD/PSP 37.98 34.31 9.62 −2.71 2.67 1.34 
AD1 57 3.65 3.91 0.1 −0.46 0.61 0.18 
AD2 56 3.16 1.49 0.45 −0.65 0.32 0.12 
AD3 63 2.43 4.29 2.81 −0.51 0.93 0.74 
AD4 60 2.28 6.38 3.46 −0.49 0.93 0.26 
AD5 67 0.12 1.29 6.70 1.1 −0.12 1.21 
AD6 69 0.11 0.47 0.41 0.08 0.16 0.02 
AD7 52 3.10 5.98 3.25 1.02 −0.54 −1.47 
AD8 55 0.47 0.82 0.26 0.02 −0.31 −0.71 
AD9 61 1.18 1.42 0.72 −0.77 0.49 0.03 
Alzheimer’s disease 0.81 4.34 1.98 −0.07 0.27 0.04 
SD1 52 2.47 4.79 2.93 0.47 −0.35 −0.82 
SD2 22 3.27 3.95 1.20 −0.76 0.74 0.03 
SD3 43 1.25 1.55 0.74 −0.31 −0.15 −0.91 
SD4 84 0.57 2.83 1.36 −0.16 0.29 0.50 
SD5 53 1.67 2.26 0.51 −0.80 0.17 −1.03 
SD6 61 3.57 9.21 4.93 −0.76 1.00 0.91 
SD7 70 5.05 9.88 4.06 −0.84 1.16 0.79 
SD8 55 8.75 9.39 3.12 −1.78 1.92 0.88 
SD9 60 4.60 6.41 2.54 −1.00 1.03 0.94 
Semantic dementia 0.78 3.03 0.59 −0.66 0.65 0.14 

Note the qualitatively distinct pattern of widespread involvement of white matter tracts in the corticobasal degeneration/progressive supranuclear palsy group.

ACE–R = Addenbrooke’s cognitive examination–revised; CBD = corticobasal degeneration; FA = fractional anisotropy; λ1 = axial diffusivity; PSP = progressive supranuclear palsy; RD = radial diffusivity.

At an individual level, the patterns of abnormality looked remarkably similar to the group level results with no obvious differences between those with a corticobasal degeneration versus those with a progressive supranuclear palsy phenotype (Fig. 2 and Supplementary Fig. 1). Despite some variation in the distribution and extent of degeneration, all scans revealed a qualitatively distinct pattern of widespread involvement of white matter tracts with an anterior bias. Fractional anisotropy was consistently the most sensitive metric although, again, it was followed closely by radial diffusivity. Compared to the other groups, abnormalities accounted for a significant proportion of white matter (Table 2). Moreover, the standard scores for both radial diffusivity and fractional anisotropy were outside the 95% confidence interval of normal mean in all individuals and at a group level (Table 2).

Figure 2

Single-subject (P < 0.04 uncorrected) demonstration of white matter abnormalities for diffusivity and anisotropy metrics. AD = Alzheimer disease; CBD = corticobasal degeneration; PSP = progressive supranuclear palsy; SD = semantic dementia.

Figure 2

Single-subject (P < 0.04 uncorrected) demonstration of white matter abnormalities for diffusivity and anisotropy metrics. AD = Alzheimer disease; CBD = corticobasal degeneration; PSP = progressive supranuclear palsy; SD = semantic dementia.

Group level comparison of grey matter density using voxel-based morphometry revealed a small cluster of cortical atrophy centred on the left frontal operculum with a less significant involvement of the left sided premotor area and caudate nucleus (Fig. 1).

Alzheimer’s disease

Group level comparisons revealed limited areas of reduced fractional anisotropy and increased axial diffusivity and radial diffusivity in a number of posterior association and projection tracts (Fig. 1). As demonstrated in Table 2, radial diffusivity was the most sensitive metric in detecting group level abnormalities followed by axial diffusivity and fractional anisotropy. For none of the metrics, however, did the proportion of abnormal voxels exceed 5% of the white matter skeleton. Individual scans showed less confluent abnormalities in a similar distribution (Fig. 2 and Supplementary Fig. 1). Consistent with the group level findings, different diffusion tensor imaging metrics showed differential sensitivity, with radial diffusivity the most consistent in detecting abnormalities at an individual level. It is worth emphasizing, however, that for the majority of the patients, the proportion of abnormal tracts remained <5%. Moreover, scrutiny of standard scores in different diffusion tensor imaging metrics, both at a group level and individually, revealed values within the normal range (Table 2).

Voxel-based morphometry of the group revealed bilateral atrophy of the hippocampi and posterior cingulate gyri with some minor patchy changes in postero-lateral temporo-parietal areas (Fig. 1).

Semantic dementia

For the semantic dementia group, areas of reduced fractional anisotropy and increased axial and radial diffusivities were mainly centred on the anterior temporal and, less significantly, orbito-frontal areas (Fig. 1). Fractional anisotropy and radial diffusivity maps revealed larger areas of abnormality when compared with axial diffusivity (Table 2). Similarly, individual diffusion tensor imaging comparisons revealed a consistent asymmetrical (mainly left sided) involvement of anterior temporal tracts (Fig. 2 and Supplementary Fig. 1). Amongst the different metrics, radial diffusivity was uniformly the most sensitive in all patients with semantic dementia (Table 2).

This group’s voxel-based morphometry analysis identified severe grey matter atrophy in both anterior temporal lobes that was more significant on the left where it extended to posterior temporal areas. There was also some minor ventral frontal involvement (Fig. 1).

Discussion

The results demonstrate that a qualitatively distinct pattern of diffusion tensor behaviour can be identified at a single subject level for the presumed tauopathies, corticobasal degeneration/progressive supranuclear palsy. Single subject analysis in Alzheimer’s disease yielded fairly minimal changes in radial and axial diffusion that, consistent with previous group level analysis (Acosta-Cabronero et al., 2010; Bosch et al., 2012), was most apparent in the white matter of the posterior cingulate and adjacent posterior parieto-temporal regions. In semantic dementia, changes were focused on the rostral temporal lobe and principally involved an increase in radial diffusion, again consistent with previous group level results (Acosta-Cabronero et al., 2011).

The most striking finding, though, was that individuals with non-fluent primary progressive aphasia had profound and diffuse changes throughout essentially all of the white matter. This predominantly involved increased radial diffusion and therefore decreased fractional anisotropy. Because, at follow-up, these individuals had developed signs of either corticobasal degeneration or progressive supranuclear palsy, it is proposed that this diffuse white matter lesion is a diagnostic marker for this class of tauopathies. Although this has been shown in only nine such patients in the present study, it is important to stress that the findings were seen in each and every patient in the group, and in none of the 18 patients without putative corticobasal degeneration/progressive supranuclear palsy; furthermore, there was no overlap at all between the corticobasal degeneration/progressive supranuclear palsy group and the other two patient groups in terms of the diffusion tensor imaging finding. The probability of these results occurring in this way by chance (i.e. all nine of the patients with corticobasal degeneration/progressive supranuclear palsy, and none of the other patients, having this diffusion tensor imaging profile) is of the order of one in 10 million. In agreement with the findings of the present study, previous group level comparisons investigating the patterns of white matter degeneration in progressive supranuclear palsy and corticobasal degeneration have indicated extensive white matter abnormalities in these two conditions (Erbetta et al., 2009; Knake et al., 2010; Saini et al., 2011).

With the exception of one pathologically verified patient, the diagnosis of corticobasal degeneration/progressive supranuclear palsy in this study was clinical and further pathological verification will be necessary to confirm the specificity of these findings in future work, particularly as features of a corticobasal syndrome can also be associated with Alzheimer’s pathology (Chand et al., 2006). That said, previous clinicopathological studies have found that progressive non-fluent aphasia occurring before or after the motor features of corticobasal degeneration or progressive supranuclear palsy in the same patient is highly predictive for these pathological entities (Kertesz et al., 2005; Shelley et al., 2009). The specificity of the findings is also reinforced by the observation that the diffuse white matter lesion of the patients with corticobasal degeneration or progressive supranuclear palsy was not seen in any patients with semantic dementia. Previous clinicopathological studies indicate that ubiquitin-positive TDP-43 is the typical pathology in semantic dementia whereas tauopathies are relatively rare (Rohrer et al., 2011). The individuals with semantic dementia also showed a specific profile, with changes most significant in the rostral temporal lobe (spilling over into the ventral frontal region in some cases) that were predominantly driven by increased radial diffusion. This pattern, while specific at a single-subject level, is not as groundbreaking as the corticobasal degeneration/progressive supranuclear palsy finding because atrophy in the rostral temporal lobe on standard structural imaging also has predictive value for pathology in semantic dementia (Pereira et al., 2009). This could alter, however, if the diffusion tensor imaging pattern in the rare case of semantic dementia associated with tau pathology had a qualitatively different diffusion tensor imaging profile to those with TDP-43, because previous structural imaging work reported such cases to be identical in terms of atrophy alone (Pereira et al., 2009).

Returning to the key finding of this study, a crucial question is why patients with corticobasal degeneration or progressive supranuclear palsy have such diffuse changes in white matter. The answer seems unlikely to relate to grey matter atrophy or axonal loss. If it did, in comparison to the other groups, one would expect these patients to have a severe and global dementia with wide spread cortical atrophy, but this was clearly not the case. At a group level, patients with corticobasal degeneration or progressive supranuclear palsy were equivalent to the other two patient groups on the global measures of Mini-Mental State Examination and ACE-R and showed a rather restricted pattern of cortical atrophy that was in agreement with the previous group studies of non-fluent primary progressive aphasia (Whitwell et al., 2004; Josephs et al., 2006; Wilson et al., 2010). Moreover, as demonstrated in Table 2, there was no distinctive difference in the extent of white matter abnormality between those individuals with progressive supranuclear palsy or corticobasal degeneration with a very mild degree of cognitive impairment—as evidenced by their global cognitive scores—and those who had very low scores in ACE-R. Therefore, we speculate that this diffuse white matter change relates to the known glial pathology in these diseases (Forman et al., 2002), which, nonetheless, could interfere with action potential propagation. The diffuse distribution of the diffusion tensor imaging changes could, therefore, offer a credible explanation for the profound slowing of cognitive processing that is a hallmark of corticobasal degeneration or progressive supranuclear palsy (Dubois et al., 1988).

A likely related finding is that previous studies have identified profound levels of insoluble tau in the white matter of patients with progressive supranuclear palsy (Zhukareva et al., 2006) and corticobasal degeneration (Forman et al., 2002). In the present study there was no discernable difference in the diffusion tensor imaging pattern of those who evolved progressive supranuclear palsy versus corticobasal degeneration. This finding has some resonance with the known close relationship between these two pathologies (Boeve et al., 2003; Scaravilli et al., 2005); for instance, both are associated with the same predominantly four-repeat tau isoform. Interestingly, Pick’s disease, in which there is predominantly three-repeat tau (Cairns et al., 2007), has also been shown to exhibit prominent white matter tau pathology (Zhukareva et al., 2002). It will therefore be interesting for future studies to examine whether Pick’s disease, and indeed other tauopathies, yield similar or different diffusion changes to corticobasal degeneration/progressive supranuclear palsy. The present results offer strong evidence. however, that the corticobasal degeneration/progressive supranuclear palsy result is not a non-specific feature of any tauopathy because the results in Alzheimer’s disease, the most common tau disorder, were completely dissimilar. This finding accords with the known cortical predilection, and relative sparing of subcortical areas, for tau pathology in Alzheimer’s disease (Zhukareva et al., 2006) that stands in sharp contrast to the subcortical, mainly glial, pathology seen in corticobasal degeneration/progressive supranuclear palsy. One final point to emphasize on the underlying mechanism of the diffusion tensor imaging lesion is that we do not believe that this relates to tau pathology per se, but rather to its relationship to surrounding tissue and its resulting impact on diffusivity of water in these specific diseases.

The findings of the present study indicate a clear potential for utilization of diffusion tensor imaging data as a state-specific diagnostic biomarker for corticobasal degeneration/progressive supranuclear palsy. The clinical importance of this finding becomes more apparent when one considers that none of the available imaging, serological or CSF biomarkers are capable of making a positive prediction for these conditions (Toledo et al., 2012). Although one could argue that the average duration of symptoms in the corticobasal degeneration/progressive supranuclear palsy group was 3 years, delayed presentation due to the insidious onset of the symptoms is a well-recognized feature of all neurodegenerative diseases. The fact that six of nine patients with non-fluent primary progressive aphasia were recruited following their first presentation to the clinic further corroborates this claim. Moreover, the corticobasal degeneration/progressive supranuclear palsy group had the shortest symptom duration amongst the groups and this argues against widespread white matter abnormalities being a mere consequence of advanced disease.

There are some important caveats to note with respect to future studies. First is that the approach employed here requires a relatively large control database for comparison. The second is that diffusion tensor imaging acquisition can be realized in a multitude of different forms with respect to strength of magnetic field, resolution, directions and b-values, and these technical details may prove to be critically important. The current results were acquired at 3 T with 63 non-colinear directions; it cannot be assumed without empirical validation that less detailed acquisitions or lower field strengths would yield similar results. Moreover, it is worth emphasizing that we excluded those participants whose structural MRI revealed a significant degree of white matter disease (leukoariosis). It will be of interest to assess—in future studies—whether corticobasal degeneration/progressive supranuclear palsy can be differentiated from other diseases in cases with concomitant leukoariosis or whether the findings are only valid once leukoariosis is excluded. A final point that must be stressed is that the present study explicitly selected cases with non-fluent primary progressive aphasia that developed clinical corticobasal degeneration or progressive supranuclear palsy because this combination is known to be highly predictive of corticobasal degeneration/progressive supranuclear palsy pathology. The study, therefore, is not directly comparable to diffusion tensor imaging studies of non-fluent primary progressive aphasia as a clinical syndrome (Schwindt et al., 2011). In the latter, one might expect heterogeneous pathology (tau, TDP-43 and even Alzheimer’s pathology) and therefore less extensive white matter changes depending on the proportion of tau cases that were included. A corollary of this issue is that the current results raise serious concerns about using diffusion tensor imaging to map white matter tract degeneration in non-fluent primary progressive aphasia; if a cohort contains a large proportion of tau cases, changes might be seen that do not relate to axonal loss. As discussed above, the hypothesis that the white matter lesion in the cases with corticobasal degeneration/progressive supranuclear palsy relates to glial, rather than axonal pathology is not only supported by past pathological studies; its lack of relationship to axonal loss was highlighted by the present finding that it was independent of disease severity (Table 2). This observation, nevertheless, emphasizes its diagnostic potential because the lesion was equally evident in even the most minimally impaired subject.

In summary, this study suggests that patients with putative tauopathies, corticobasal degeneration and progressive supranuclear palsy, have a distinct white matter diffusion tensor imaging profile that can be visualized at the level of the individual. This finding arose from studying patients with non-fluent primary progressive aphasia in whom corticobasal degeneration/progressive supranuclear palsy is one known pathological substrate. Further pathological confirmation of these findings will take some time but, in the interim, a next important step will be to determine whether the findings can be replicated in corticobasal degeneration/progressive supranuclear palsy presenting as a movement disorder, and in contrast to Parkinson’s disease. To conclude, it is interesting to note that when unexpected profuse white matter tau pathology was first discovered in progressive supranuclear palsy by Zhukareva et al. (2006), the authors speculated that it could have diagnostic implications if their discovery could be exploited by neuroimaging. The present findings suggest that diffusion tensor imaging may have now achieved this.

Funding

This work was supported by a Donald Forrester Trust research grant and NIHR Cambridge Biomedical Research Centre.

Abbreviations

    Abbreviations
  • ACE-R

    Addenbrooke’s Cognitive Examination-Revised

References

Acosta-Cabronero
J
Patterson
K
Fryer
TD
Hodges
JR
Pengas
G
Williams
GB
, et al.  . 
Atrophy, hypometabolism and white matter abnormalities in semantic dementia tell a coherent story
Brain
 , 
2011
, vol. 
134
 
Pt 7
(pg. 
2025
-
35
)
Acosta-Cabronero
J
Williams
GB
Pengas
G
Nestor
PJ
Absolute diffusivities define the landscape of white matter degeneration in Alzheimer's disease
Brain
 , 
2010
, vol. 
133
 
Pt 2
(pg. 
529
-
39
)
Acosta-Cabronero
J
Williams
GB
Pereira
JM
Pengas
G
Nestor
PJ
The impact of skull-stripping and radio-frequency bias correction on grey-matter segmentation for voxel-based morphometry
Neuroimage
 , 
2008
, vol. 
39
 (pg. 
1654
-
65
)
Alladi
S
Xuereb
J
Bak
T
Nestor
P
Knibb
J
Patterson
K
, et al.  . 
Focal cortical presentations of Alzheimer's disease
Brain
 , 
2007
, vol. 
130
 
Pt 10
(pg. 
2636
-
45
)
Ashburner
J
Friston
KJ
Unified segmentation
Neuroimage
 , 
2005
, vol. 
26
 (pg. 
839
-
51
)
Boeve
BF
Lang
AE
Litvan
I
Corticobasal degeneration and its relationship to progressive supranuclear palsy and frontotemporal dementia
Ann Neurol
 , 
2003
, vol. 
54
 
Suppl 5
(pg. 
S15
-
9
)
Bosch
B
Arenaza-Urquijo
EM
Rami
L
Sala-Llonch
R
Junque
C
Sole-Padulles
C
, et al.  . 
Multiple DTI index analysis in normal aging, amnestic MCI and AD. Relationship with neuropsychological performance
Neurobiol Aging
 , 
2012
, vol. 
33
 (pg. 
61
-
74
)
Bozeat
S
Lambon Ralph
MA
Patterson
K
Garrard
P
Hodges
JR
Non-verbal semantic impairment in semantic dementia
Neuropsychologia
 , 
2000
, vol. 
38
 (pg. 
1207
-
15
)
Cairns
NJ
Bigio
EH
Mackenzie
IR
Neumann
M
Lee
VM
Hatanpaa
KJ
, et al.  . 
Neuropathologic diagnostic and nosologic criteria for frontotemporal lobar degeneration: consensus of the Consortium for Frontotemporal Lobar Degeneration
Acta Neuropathol
 , 
2007
, vol. 
114
 (pg. 
5
-
22
)
Caso
F
Gesierich
B
Henry
M
Sidhu
M
Lamarre
A
Babiak
M
, et al.  . 
Nonfluent/agrammatic PPA with in-vivo cortical amyloidosis and Pick's disease pathology
Behav Neurol
 , 
2012
, vol. 
26
 (pg. 
95
-
106
)
Chand
P
Grafman
J
Dickson
D
Ishizawa
K
Litvan
I
Alzheimer's disease presenting as corticobasal syndrome
Mov Disord
 , 
2006
, vol. 
21
 (pg. 
2018
-
22
)
Delis
D
Kaplan
E
Kramer
J
Delis-Kaplan executive function system
 , 
2001
San Antonio, TX
Harcourt Brace & Company
Dubois
B
Feldman
HH
Jacova
C
Dekosky
ST
Barberger-Gateau
P
Cummings
J
, et al.  . 
Research criteria for the diagnosis of Alzheimer's disease: revising the NINCDS-ADRDA criteria
Lancet Neurol
 , 
2007
, vol. 
6
 (pg. 
734
-
46
)
Dubois
B
Pillon
B
Legault
F
Agid
Y
Lhermitte
F
Slowing of cognitive processing in progressive supranuclear palsy. A comparison with Parkinson's disease
Arch Neurol
 , 
1988
, vol. 
45
 (pg. 
1194
-
9
)
Erbetta
A
Mandelli
ML
Savoiardo
M
Grisoli
M
Bizzi
A
Soliveri
P
, et al.  . 
Diffusion tensor imaging shows different topographic involvement of the thalamus in progressive supranuclear palsy and corticobasal degeneration
AJNR Am J Neuroradiol
 , 
2009
, vol. 
30
 (pg. 
1482
-
7
)
Folstein
MF
Folstein
SE
McHugh
PR
“Mini-mental state”. A practical method for grading the cognitive state of patients for the clinician
J Psychiatr Res
 , 
1975
, vol. 
12
 (pg. 
189
-
98
)
Forman
MS
Zhukareva
V
Bergeron
C
Chin
SS
Grossman
M
Clark
C
, et al.  . 
Signature tau neuropathology in gray and white matter of corticobasal degeneration
Am J Pathol
 , 
2002
, vol. 
160
 (pg. 
2045
-
53
)
Gorno-Tempini
ML
Hillis
AE
Weintraub
S
Kertesz
A
Mendez
M
Cappa
SF
, et al.  . 
Classification of primary progressive aphasia and its variants
Neurology
 , 
2011
, vol. 
76
 (pg. 
1006
-
14
)
Hodges
JR
Patterson
K
Semantic dementia: a unique clinicopathological syndrome
Lancet Neurol
 , 
2007
, vol. 
6
 (pg. 
1004
-
14
)
Josephs
KA
Duffy
JR
Strand
EA
Whitwell
JL
Layton
KF
Parisi
JE
, et al.  . 
Clinicopathological and imaging correlates of progressive aphasia and apraxia of speech
Brain
 , 
2006
, vol. 
129
 
Pt 6
(pg. 
1385
-
98
)
Kertesz
A
McMonagle
P
Blair
M
Davidson
W
Munoz
DG
The evolution and pathology of frontotemporal dementia
Brain
 , 
2005
, vol. 
128
 
Pt 9
(pg. 
1996
-
2005
)
Knake
S
Belke
M
Menzler
K
Pilatus
U
Eggert
KM
Oertel
WH
, et al.  . 
In vivo demonstration of microstructural brain pathology in progressive supranuclear palsy: a DTI study using TBSS
Mov Disord
 , 
2010
, vol. 
25
 (pg. 
1232
-
8
)
Litvan
I
Agid
Y
Goetz
C
Jankovic
J
Wenning
GK
Brandel
JP
, et al.  . 
Accuracy of the clinical diagnosis of corticobasal degeneration: a clinicopathologic study
Neurology
 , 
1997
, vol. 
48
 (pg. 
119
-
25
)
Litvan
I
Agid
Y
Jankovic
J
Goetz
C
Brandel
JP
Lai
EC
, et al.  . 
Accuracy of clinical criteria for the diagnosis of progressive supranuclear palsy (Steele-Richardson-Olszewski syndrome)
Neurology
 , 
1996
, vol. 
46
 (pg. 
922
-
30
)
Mesulam
MM
Primary progressive aphasia
Ann Neurol
 , 
2001
, vol. 
49
 (pg. 
425
-
32
)
Mioshi
E
Dawson
K
Mitchell
J
Arnold
R
Hodges
JR
The Addenbrooke's Cognitive Examination Revised (ACE-R): a brief cognitive test battery for dementia screening
Int J Geriatr Psychiatry
 , 
2006
, vol. 
21
 (pg. 
1078
-
85
)
Neary
D
Snowden
JS
Gustafson
L
Passant
U
Stuss
D
Black
S
, et al.  . 
Frontotemporal lobar degeneration: a consensus on clinical diagnostic criteria
Neurology
 , 
1998
, vol. 
51
 (pg. 
1546
-
54
)
Pengas
G
Pereira
JM
Williams
GB
Nestor
PJ
Comparative reliability of total intracranial volume estimation methods and the influence of atrophy in a longitudinal semantic dementia cohort
J Neuroimaging
 , 
2009
, vol. 
19
 (pg. 
37
-
46
)
Pereira
JM
Williams
GB
Acosta-Cabronero
J
Pengas
G
Spillantini
MG
Xuereb
JH
, et al.  . 
Atrophy patterns in histologic vs clinical groupings of frontotemporal lobar degeneration
Neurology
 , 
2009
, vol. 
72
 (pg. 
1653
-
60
)
Rohrer
JD
Lashley
T
Schott
JM
Warren
JE
Mead
S
Isaacs
AM
, et al.  . 
Clinical and neuroanatomical signatures of tissue pathology in frontotemporal lobar degeneration
Brain
 , 
2011
, vol. 
134
 
Pt 9
(pg. 
2565
-
81
)
Saini
J
Bagepally
BS
Sandhya
M
Pasha
SA
Yadav
R
Pal
PK
In vivo evaluation of white matter pathology in patients of progressive supranuclear palsy using TBSS
Neuroradiology
 , 
2011
, vol. 
54
 (pg. 
771
-
80
)
Scaravilli
T
Tolosa
E
Ferrer
I
Progressive supranuclear palsy and corticobasal degeneration: lumping versus splitting
Mov Disord
 , 
2005
, vol. 
20
 
Suppl 12
(pg. 
S21
-
8
)
Schwindt
GC
Graham
NL
Rochon
E
Tang-Wai
DF
Lobaugh
NJ
Chow
TW
, et al.  . 
Whole–brain white matter disruption in semantic and nonfluent variants of primary progressive aphasia
Hum Brain Mapp
 , 
2013
, vol. 
34
 (pg. 
973
-
84
)
Shelley
BP
Hodges
JR
Kipps
CM
Xuereb
JH
Bak
TH
Is the pathology of corticobasal syndrome predictable in life?
Mov Disord
 , 
2009
, vol. 
24
 (pg. 
1593
-
9
)
Smith
SM
Jenkinson
M
Johansen-Berg
H
Rueckert
D
Nichols
TE
Mackay
CE
, et al.  . 
Tract-based spatial statistics: voxelwise analysis of multi-subject diffusion data
Neuroimage
 , 
2006
, vol. 
31
 (pg. 
1487
-
505
)
Toledo
JB
Brettschneider
J
Grossman
M
Arnold
SE
Hu
WT
Xie
SX
, et al.  . 
CSF biomarkers cutoffs: the importance of coincident neuropathological diseases
Acta Neuropathol
 , 
2012
, vol. 
124
 (pg. 
23
-
35
)
Warrington
E
James
M
Visual object and space perception battery
 , 
1991
Bury St Edmunds
Thames Valley Test Company
Wechsler
D
Wechsler Memory Scale (WMS–III)
 , 
1997
San Antonio, TX
The Psychological Corporation
Whitwell
JL
Anderson
VM
Scahill
RI
Rossor
MN
Fox
NC
Longitudinal patterns of regional change on volumetric MRI in frontotemporal lobar degeneration
Dement Geriatr Cogn Disord
 , 
2004
, vol. 
17
 (pg. 
307
-
10
)
Wilson
SM
Henry
ML
Besbris
M
Ogar
JM
Dronkers
NF
Jarrold
W
, et al.  . 
Connected speech production in three variants of primary progressive aphasia
Brain
 , 
2010
, vol. 
133
 
Pt 7
(pg. 
2069
-
88
)
Zhukareva
V
Joyce
S
Schuck
T
Van Deerlin
V
Hurtig
H
Albin
R
, et al.  . 
Unexpected abundance of pathological tau in progressive supranuclear palsy white matter
Ann Neurol
 , 
2006
, vol. 
60
 (pg. 
335
-
45
)
Zhukareva
V
Mann
D
Pickering-Brown
S
Uryu
K
Shuck
T
Shah
K
, et al.  . 
Sporadic Pick's disease: a tauopathy characterized by a spectrum of pathological tau isoforms in gray and white matter
Ann Neurol
 , 
2002
, vol. 
51
 (pg. 
730
-
9
)