Imaging amyloid deposition in Lewy body diseases

Study design.

Eight subjects with DLB, 7 with PDD, and 11 with PD were recruited from the Massachusetts General Hospital (MGH) Movement and Memory Units for the study (table 1). All subjects with PD fulfilled criteria for a diagnosis of idiopathic PD according to the UK Parkinson’s Disease Society Brain Bank Research Center (UKPDSBRC) clinical diagnostic criteria,⁹ had Hoehn and Yahr (H&Y) stages I through IV, and were cognitively normal, with Blessed Dementia Scale (BDS) scores < 5. Subjects with PDD met UKPDSBRC criteria for idiopathic PD, were H&Y stages I through IV and, in addition, met Diagnostic and Statistical Manual of Mental Disorders, 3rd Edition, Revised, criteria for dementia, with BDS scores ≥ 5. Although the recent consensus criteria for PDD¹ were unavailable at the study’s inception, all subjects with PDD would have met these criteria at the time of enrollment. Subjects with DLB met consensus criteria for probable DLB of the Dementia with Lewy Bodies Consortium,² with the presence of at least two of the following: parkinsonism, hallucinations, and fluctuations in cognition. The BDS scores in DLB were ≥5. Six of 7 DLB, 11 of 11 PD, and 6 of 7 PDD subjects were treated with dopaminergic drugs, such as l-dopa. Acquired data from these parkinsonian subjects were compared with those of a separately collected cohort of 15 AD and 37 NC subjects who were participants in longitudinal studies of normal aging and dementia at MGH and Brigham and Women’s Hospital. The diagnosis of NC required normal neurologic examination results, a Clinical Dementia Rating (CDR) scale¹⁰ score of 0, and normal cognition, with a BDS score < 3 or Mini-Mental State Examination (MMSE) score > 27. All subjects with AD met National Institute of Neurological and Communicative Disorders and Stroke–Alzheimer’s Disease and Related Disorders Association criteria for probable AD¹¹ and had BDS scores ≥ 5 or MMSE scores < 23. All participants gave written informed consent according to the protocols approved by the Partners Healthcare Inc. Institutional Review Board.

Clinical evaluation.

Clinical evaluation of motor function involved the H&Y staging of PD and United Parkinson’s Disease Rating Scale (UPDRS) part III¹² testing in the on state; the mean lag from dose of dopaminergic drug to examination in 21 of 23 DLB, PDD, and PD subjects was not different between groups (p > 0.37). Additional measures included the CDR, the CDR Sum of Boxes, the Mayo fluctuations screen,¹³ the Neuropsychiatric Inventory Questionnaire,¹⁴ including the hallucinations subscale, and the Geriatric Depression Scale. Neuropsychological testing included the MMSE, BDS, digits forward, digits backward, Trails A and B, Logical Memory IA and IIA, Free and Cued Selective Reminding Tests (FRSRT and FCSRT, respectively),¹⁵ Benton visual form discrimination test, Wechsler Adult Intelligence Scale–Revised, letter fluency, two-item category fluency (animals and vegetables), 30-item Boston Naming Test, and the American version of the National Adult Reading Test verbal IQ assessment. The majority of these tests correspond with the Uniform Data Set of cognitive tests, as developed by the AD Centers. The mean interval between neuropsychological testing and PiB-PET was 0.07 ± 2.12 months.

PET imaging acquisition.

N-Methyl-[11C]2-(4-methylaminophenyl)-6-hydroxybenzothiazole (PiB) was prepared at Massachusetts General Hospital as described previously.¹⁶ Subjects were positioned in either of two PET cameras for dynamic acquisition, a Siemens/CTI ECAT HR scanner (three-dimensional mode; 63 image planes; 15.2-cm axial field of view; 5.6-mm transaxial resolution; 2.4-mm slice interval; 69 frames: 12 × 15 seconds, 57 × 60 seconds; Knoxville, TN) or a GE PC4096 scanner (two-dimensional mode; 15 image planes; 10.0-cm axial field of view; 7.0-mm transaxial resolution; 6.0-mm slice interval; 39 frames: 8 × 15 seconds, 4 × 60 seconds, 27 × 120 seconds; Milwaukee, WI). After a transmission scan, 8.5 to 15 mCi 11C-PiB was injected as a bolus and followed immediately by a 60-minute dynamic acquisition. PET data were reconstructed with ordered set expectation maximization and corrected for attenuation, and each frame was evaluated to verify adequate count statistics and absence of head motion.

PET image analysis.

Global and regional cortical PiB retention was calculated using the Logan graphical analysis method,^17,18 with cerebellar cortex as the reference tissue input function, to evaluate specific PiB retention expressed as the distribution volume ratio (DVR) as in previous PiB studies.^19–22 Parametric images of DVR over the late (40- to 60-minute) epoch were assessed visually and classified as PiB negative and positive without knowledge of clinical classification, by identifying the presence of either specific cortical PiB retention (at least 50 cortical voxels with DVR >1.3; PiB positive) or the presence of only nonspecific white matter uptake (much less than 50 cortical voxels with DVR >1.3; PiB negative).^19,21 Subjects were classified as PiB positive if even a relatively restricted portion of cortex had evidence of specific binding; hence, aggregate binding in such individuals averaged over large areas of nonspecific binding remained low.

To compare PiB retention in regions of interest (ROIs) among groups, we transformed each subject’s PiB-PET data set into a standard space and calculated PiB DVR in anatomically defined ROIs, as described previously.²² The standard space used was the single-subject Automated Anatomic Labeling (AAL),²³ and we chose Statistical Parametric Mapping (SPM2)²⁴ to determine the warping transformation that carried each subject’s PiB-PET data into AAL space. The resulting warping transformation was applied to the individual dynamic frames of PiB data. Activity was measured in each individual PiB data set in the following ROI aggregates: global, occipital, frontal lateral, anterior cingulate, lateral temporal, medial temporal, parietal (lateral and precuneus)/posterior cingulate, and striatum. The components of each aggregate are tabulated in appendix e-1 on the Neurology® Web site at www.neurology.org. Time–activity curves were derived in each ROI aggregate in each subject, and the DVR was calculated according to standard methods,^17,18 using the cerebellar gray as reference. We used analysis of covariance (ANCOVA) and ratios to derive regional retention in the ROI aggregates relative to global retention. We tested the following hypotheses relating regionally specific amyloid burden to cognitive test performance: parietal (lateral and precuneus)/posterior cingulate DVR and performance on the Benton visual form discrimination test; occipital DVR and performance on the Benton visual form discrimination test; frontal lateral DVR and performance on digits backward; frontal lateral DVR and performance on FCSRT; anterior cingulate DVR and performance on digits backward; anterior cingulate DVR and performance on FCSRT; medial temporal DVR and performance on FRSRT; occipital DVR and hallucinations. In a separate study, using overlapping subject groups, intercamera reliability was found to be satisfactory. The mean DVR in the global ROI did not differ between the two PET cameras among groups with or without dementia (age and multiple test adjusted tests, p > 0.30).

Clinical features.

Demographics of the cohorts are shown in table 1. MMSE scores were comparable between the DLB and PDD groups but were lower in the DLB group than in the AD group (p < 0.017, analysis of variance). The DLB and PDD groups had similar H&Y scores. However, the H&Y score was lower in the PD group than in the PDD group (p < 0.009). The DLB and PDD groups had more severe motor dysfunction than did the PD group as measured by the UPDRS III (DLB vs PD, p < 0.03; PDD vs PD, p < 0.006). The DLB and PDD groups performed similarly on the battery of cognitive tests. They had more cognitive fluctuations than did the PD group (DLB vs PD, p = 0.004; PDD vs PD, p = 0.008) and marginally more hallucinations than did the PD group (DLB vs PD, p = 0.055; PDD vs PD, p = 0.063). All NC subjects scored normally on all neuropsychologic tests.

Global amyloid burden.

Cortical amyloid burden differed among the AD, DLB, PDD, PD, and NC groups (p < 0.0001, nonparametric permutation test; figures 1 and 2). The DLB group had greater mean PiB retention than did the PDD group (p < 0.05, Tukey adjusted post hoc test), despite similar degrees of cognitive impairment (table 1). This difference remained significant after adjusting for nonsignificant differences in performance on the MMSE or BDS, with ANCOVA. The mean amyloid burden in the DLB group was comparable to levels associated with AD (difference not significant [NS]). PiB retention was higher in the DLB group than in the PD or NC group (p < 0.05). In the PDD group, PiB retention was lower than in the AD group (p < 0.05); the mean amyloid burden in the PDD group was not statistically different from that in the PD or NC group. Cortical amyloid burden in the PD group was similar to that in the NC group (NS). These findings persisted after adjusting for age with ANCOVA. There was neither a significant interaction of gender with diagnosis nor a gender main effect on amyloid burden (factorial ANCOVA with crossed diagnosis and gender factors and an age covariate).

We visually assessed the parametric images to differentiate cases with specific cortical PiB retention (PiB positive) from cases with only nonspecific white matter uptake (PiB negative). All 15 AD subjects, 7 of 8 DLB subjects, and all 7 PDD subjects were PiB positive. Seven of 11 PD subjects and 19 of 37 NC subjects were PiB positive. PiB-positive and PiB-negative PD subjects overlapped in demographic, clinical, and neuropsychological features, and there were no significant differences in mean age, performance on the MMSE, or Hoehn and Yahr stage. PiB-positive and PiB-negative NC subjects similarly overlapped in demographic, clinical, and neuropsychological features, including age and MMSE performance.

Regional amyloid burden.

The specific PiB retention measured in ROIs reflected global PiB retention (absolute, unadjusted DVRs are shown in table 2). We therefore also analyzed differences between diagnostic groups in regional PiB retention, adjusting for global PiB retention via ANCOVA. Adjusted amyloid burden in the frontal lateral, lateral temporal, medial temporal, parietal (lateral and precuneus)/posterior cingulate, and anterior cingulate ROIs did not differentiate the DLB, PDD, or PD groups from the AD group (for all comparisons, NS). In contrast, adjusted occipital PiB retention was lower in the AD group than in the DLB (p = 0.013, Tukey adjusted post hoc test), PDD (p = 0.0002), or PD (p = 0.0004) group. The AD group had lower adjusted occipital PiB than the NC group as well (p < 0.0001). Adjusted PiB burden in the ROIs did not differentiate the DLB, PDD, and PD groups from one another.

Association of global amyloid with clinical variables in Lewy body diseases.

Across the DLB, PDD, and PD groups pooled, the greater the amyloid burden was, as measured by the global cortical PiB DVR, the lower the MMSE score was (r = −0.50, p = 0.01). However, this effect was dominated by differences between groups rather than by relations within groups, such that the pooled within-group relation was not significant (within group partial r = −0.1, p = 0.65).

Association of regional amyloid with clinical, cognitive, and motor variables in Lewy body diseases.

We tested eight hypotheses relating regionally specific amyloid burden with specific tests of cognitive performance across the DLB, PDD, and PD groups. Regional PiB retention was adjusted for global retention. We found that regional amyloid burden in the parietal (lateral and precuneus)/posterior cingulate region—but not in the occipital region—showed a within-group relation to decline in the Benton visual form discrimination test (whether we adjusted for global PiB retention as a simple ratio [p = 0.011; false discovery rate p = 0.047 correcting for multiple ROI–cognitive test comparisons] or via ANCOVA [p = 0.005]). In contrast, in AD and NC subjects, amyloid burden in this region did not correlate with Benton performance. Regional amyloid burden in the DLB, PDD, and PD groups did not relate significantly to the other hypothesized focal cognitive impairments, whether blind to diagnostic group or within diagnostic group. In particular, frontal executive tasks and cued recall (digits backward and FCSRT) did not relate to PiB retention in the frontal lateral or anterior cingulate regions; memory impairment (FRSRT) was not associated with increased PiB retention in the medial temporal lobe region; and hallucinations were not associated with occipital PiB retention.

There was a significant interaction of diagnostic group (DLB, PDD, PD) with the relation of striatal PiB retention to UPDRS III score (whether we adjusted for global PiB retention as a simple ratio [p = 0.039] or via ANCOVA [p = 0.046]). In the case of the DLB and PDD groups but not the PD group, the greater the amyloid burden in the striatum was, the lower (better) the UPDRS III score was (DLB, r = −0.87 p = 0.01; PDD, r = −0.90 p = 0.005).