Measurements of the Amygdala and Hippocampus in Pathologically Confirmed Alzheimer Disease and Frontotemporal Lobar Degeneration
Total Page:16
File Type:pdf, Size:1020Kb
ORIGINAL CONTRIBUTION Measurements of the Amygdala and Hippocampus in Pathologically Confirmed Alzheimer Disease and Frontotemporal Lobar Degeneration Josephine Barnes, MA; Jennifer L. Whitwell, PhD; Chris Frost, MA, DipStat; Keith A. Josephs, MST, MD; Martin Rossor, MD, FRCP; Nick C. Fox, MD, FRCP Background: Differentiating between Alzheimer dis- Results: Geometric mean amygdala and hippocampal ease (AD) and frontotemporal lobar degeneration (FTLD) volumes were, respectively, 15.0% (95% confidence in- can be difficult, particularly in the earliest stages of the terval [CI], 4.2%-24.5%) and 16.4% (95% CI, 5.9%- diseases. Patterns of atrophy on magnetic resonance im- 25.6%) lower in the AD than in the control group. In aging may help distinguish these diseases and aid diag- FTLD, the equivalent differences were 43.1% (95% CI, nosis. 31.9%-52.6%) in the amygdala and 36.1% (95% CI, 27.5%- 43.7%) in the hippocampus. Volumes were signifi- Objective: To assess the diagnostic utility of magnetic cantly lower in the FTLD than in the AD group (PϽ.01 resonance imaging–derived amygdala and hippocampal in both regions). Within the FTLD clinical subgroups, volumes from patients with pathologically proved AD and there was evidence of a difference in pattern of atrophy FTLD. with greater asymmetry (left smaller than right) in se- mantic dementia compared with frontal variant FTLD Design: Cross-sectional volumetric magnetic reso- (PϽ.001). On average, the left hippocampus was 14% nance imaging study of the hippocampus and amyg- smaller in semantic dementia than in frontal variant FTLD, dala. whereas the right hippocampus was 37% larger. On av- erage, the left amygdala was 39% smaller in semantic de- Setting: Specialist cognitive disorders clinic. mentia than in frontal variant FTLD, whereas the right amygdala was only 1% smaller. Subjects: Thirty-seven subjects, including 10 patients with pathologically proved AD, 17 patients with patho- Conclusions: Hippocampal atrophy is not specific to AD logically proved FTLD, and 10 age-matched control sub- or FTLD. However, severe or asymmetrical amygdala at- jects. rophy should suggest FTLD. Atrophy patterns follow clini- cal syndromes rather than pathology. Main Outcome Measures: Hippocampal and amyg- dala volumes. Arch Neurol. 2006;63:1434-1439 Author Affiliations: Dementia N THE ELDERLY POPULATION,ALZ- cially in the early stages, has led to inter- Research Centre, University heimer disease (AD) is the most est in imaging studies to help differenti- 1 6 College London, Institute of common cause of dementia. ate FTLD from AD. Neurology (Ms Barnes; However, in patients presenting Prominent hippocampal atrophy has Drs Whitwell, Rossor, and Fox; with dementia before age 65 been demonstrated to be the signature pat- and Mr Frost), Medical years, the prevalence of frontotemporal lo- tern of atrophy on magnetic resonance im- I 7-12 Statistics Unit, London School bar degeneration (FTLD) approximates aging (MRI) studies in AD, whereas of Hygiene and Tropical that of AD.2 Differentiating AD from FTLD more anterior temporal losses, including Medicine (Mr Frost), Department of Clinical is important because the prognoses and severe amygdala atrophy, have been sug- 13-15 Neurology, National Hospital management of these diseases are differ- gestive of FTLD. These findings sup- for Neurology and ent. A diagnosis of AD is suggested if port the clinical presentations because hip- Neurosurgery (Drs Rossor and memory decline is an early clinical fea- pocampal damage is linked to loss of Fox), and Division of ture, whereas prominent behavioral and episodic memory,16 and amygdala dam- Neuroscience and Psychological language dysfunction suggests a diagno- age has been associated with emotional and Medicine, Faculty of Medicine, sis of FTLD. However, recent clinicopatho- behavioral dyscontrol.15,17 Imperial College London logical studies have identified cases of Pathology remains the gold standard in (Dr Rossor), London, England; and Departments of Diagnostic FTLD with prominent episodic memory the diagnosis of AD and FTLD. Although 3 13,14,18-22 Radiology (Dr Whitwell) and loss, and a frontal presentation of AD has several MRI studies have exam- Neurology (Dr Josephs), been described.4,5 The overlap in clinical ined the ability of volumetric measures to Mayo Clinic, Rochester, Minn. features and diagnostic uncertainty, espe- differentiate clinically diagnosed FTLD (REPRINTED) ARCH NEUROL / VOL 63, OCT 2006 WWW.ARCHNEUROL.COM 1434 ©2006 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 10/02/2021 from AD, clinical diagnostic accuracy is uncertain in the absence of histopathological confirmation, and poten- Table 1. Subject Demographics* tial exists for circularity in study design. Therefore, this study assessed the value of hippocampal and amygdala Subject Groups measurements in separating patients with AD and FTLD Control AD FTLD from control subjects and from one another in a patho- Subject Characteristics (n = 10) (n = 10) (n = 17) logically defined cohort of AD and FTLD cases. Sex, No. M/F 6/4 7/3 11/6 Age, y 56 (11) 57 (9) 56 (10) METHODS No. with sporadic/familial disease NA 9/1 12/5 MMSE score (maximum of 30) 29 (1) 15 (6) 22 (6) Disease duration at time NA 3.1 (0.7)† 3.5 (2.0) SUBJECTS of imaging study, y Time from imaging to death, y‡ NA 3.7 (2.3) 4.5 (2.1) Patients were recruited from the Cognitive Disorders Clinic at the National Hospital for Neurology and Neurosurgery. Ten pa- Abbreviations: AD, Alzheimer disease; FTLD, frontotemporal lobar tients with pathologically confirmed AD and 17 patients with dysfunction; MMSE, Mini-Mental State Examination; NA, not applicable. pathologically confirmed FTLD (5 with Pick disease, 8 with ubiq- *Unless otherwise indicated, data are expressed as mean (SD). †Disease duration was not available in 1 AD patient. uitin-only immunoreactive neuronal inclusions [FTLD-U], 3 ‡Includes only 6 patients with AD and 15 with FTLD who came to tau positive with tau exon 10ϩ16 mutations [tau exon 10ϩ16], postmortem examination. and 1 with dementia lacking distinctive histology [DLDH]) were included in this study. Pathological diagnosis was based on the 23 most recent consensus criteria. Pathological confirmation by STATISTICAL ANALYSIS biopsy rather than postmortem findings was available in 4 of the patients with AD and 2 of the patients with FTLD. We analyzed the data using STATA version 8 (StataCorp, Col- We reviewed medical records and recorded the clinical pre- lege Station, Tex) and SAS (SAS Institute Inc, Cary, NC) statis- sentation, disease duration, and Mini-Mental State Examina- tical software. Volumes were adjusted for head size differences tion score24 at the time of the MRI study. A family history of ϩ using the total intracranial volume, derived according to a pre- dementia was present in all cases of tau exon 10 16, 2 cases viously described protocol.30 Standardization was carried out sepa- of FTLD-U, and 1 case of AD. We used established clinical cri- rately for the amygdala and hippocampus. For both structures, teria to classify the patients with FTLD into frontal variant FTLD this assumed a linear relationship between log-transformed val- (fvFTD) (n=12), semantic dementia (SD) (n=4), and progres- 25 ues and log–total intracranial volume, with the slopes of the as- sive nonfluent aphasia (n=1). None of the patients with FTLD sociations estimated from left and right measurements in an ex- had any clinical evidence of motor neuron disease. Ten healthy panded control group to improve precision.31 age-matched controls without evidence or symptoms of cog- To minimize the number of hypothesis tests performed, we nitive decline were also included in the study. The demo- carried out a structured analysis using linear mixed models with graphic characteristics of all subjects are reported in Table 1. interaction terms and disease group–specific unstructured co- variance matrices. Such an approach extends analysis of vari- IMAGE ANALYSIS ance by removing the requirement for homogeneity of vari- ances. Such models were used to make comparisons between Image Acquisition geometric mean volumes by (1) disease group (FTLD vs AD vs control group in the primary analysis), (2) amygdala vs hip- The MRI studies were performed on a 1.5-T imaging system pocampus, and (3) structural laterality. We used contrasts of (GE Signa Unit; General Electric Medical Systems, Milwau- mean levels to estimate effect sizes and Wald tests to examine kee, Wis) using a spoiled gradient-echo technique (256ϫ128 the statistical significance of interaction terms and main ef- matrix; field of view, 24ϫ19.2 cm; repetition time, 35 milli- fects. We also used models of this type to compare geometric seconds; echo time, 5 milliseconds; number of excitations, 1; mean volumes by FTLD subtype. Because unstructured covari- and flip angle, 35°; yielding 124 contiguous 1.5-mm-thick sec- ance models do not facilitate comparison of variances be- tions). If an individual had more than 1 diagnostic imaging study, tween disease groups and regions, direct-product covariance the earliest one was chosen for inclusion in the study. matrices (amygdala/hippocampusϫleft/right side) were used to compare variances between (1) disease groups, (2) struc- Manual Segmentation tures (amydgala and hippocampus), and (3) left and right sides. A likelihood ratio test was used to compare the fit of a model The MIDAS software package was used for all imaging analy- where the disease group–specific amygdala and hippocampal sis and allows tracing of regions of interest in 2 orthogonal variances were assumed constant with those of one where they views.26 Before tracing, all images were registered to a stan- were allowed to be different. A similar approach was used to dard template.27 In addition, each image was reflected across compare left- and right-sided variances. In a similar compari- the midsagittal plane, and traces were performed on the right son of variances by disease groups, we assumed left- and right- side of the presented image.