C9ORF72 and the FTD-ALS spectrum: A systematic review of neuroimaging studies

Objective To perform a systematic review of the literature on the neuroimaging investigation of frontotemporal dementia (FTD) and amyotrophic lateral sclerosis (ALS) associated with C9ORF72 mutation. Methods The search was performed on PubMed and LILACS with the following terms: C9ORF72, MRI, SPECT, PET, ALS, FTD. No filters were added. Results Twenty articles were selected. Most studies found consistent involvement of frontotemporal regions in C9ORF72 carriers, including prefrontal cortex, and also cingulate, subcortical regions, especially the thalami, and posterior regions such as the parietal and occipital lobes. Functional connectivity was also explored and impaired sensorimotor connectivity in striatum and thalami was found in behavioral variant FTD C9ORF72 carriers. Some papers have reported an absence of significant abnormalities on brain imaging. Conclusion The inclusion of patients at different stages of the disease, differences in neuroimaging methods across studies, and distinct clinical phenotypes associated with C9ORF72 may account for the heterogeneity of results.


INTRODUCTION
F rontotemporal dementia (FTD) and amyotrophic lateral sclerosis (ALS) share common clinical, pathological and genetic features.FTD encompasses a heterogeneous group of clinical presentations, with variable phenotypes including behavioral changes and deficits in language and other cognitive func-tions. 1,2On the other hand, besides motor symptoms, ALS is also characterized by cognitive impairment and behavioral disorders, overlapping with the cognitive profile of FTD. 3 Indeed, the association between dementia and ALS has been recognized since the nineteenth century and almost 50% of ALS patients are Prado et al.
believed to have cognitive impairment and up to 15% of these fulfill criteria for FTD. 3,4Conversely, motor neuron disease can appear during the course of FTD in up to 15% of patients. 5Therefore there is a clinical and pathophysiological continuum between FTD and ALS.
The recent discovery that an expanded hexanucleotide (GGGGCC) repeat insertion in a noncoding promoter region of open-reading frame 72 (C9ORF72) is a cause of familial FTD and ALS opened a promising window for the understanding of the FTD-ALS spectrum. 6,7The neurobiological functions of C9ORF72 and the pathophysiological mechanisms by which it participates in neurodegenerative processes are unknown. 8The C9ORF72 genotype may account for 10-50% of familial cases of behavioral variant FTD (bvFTD). 1,8Conversely, up to 41% of familial ALS and 5% of sporadic ALS cases may have C9ORF72 mutation. 9Co-morbid FTD is more common in ALS patients with the C9ORF72 genotype, and these patients may have faster disease progression and more pronounced cognitive and behavioral disorders. 9,10ince its discovery, there has been an intense research effort to investigate the clinical phenotypes associated with C9ORF72 repeat expansion.More specifically, neuroimaging methods have been employed to investigate neuroanatomical features of FTD and/or ALS patients with C9ORF72 mutation.Brain imaging may provide clinical markers for both the diagnosis and/or the follow-up of these patients, and may also shed light on the pathophysiological mechanisms of neurodegeneration associated with C9ORF72 repeat expansion.In the current paper, we aimed to review the literature on neuroimaging studies of FTD and/or ALS patients with C9ORF72 mutation.

METHODS
We conducted a systematic review of the literature according to a predetermined protocol as described elsewhere. 11The search aimed to identify original papers reporting neuroimaging data in FTD and/or ALS patients with C9ORF72 repeat expansion.
The search was performed in July 26 th 2015 on two electronic databases: PubMed and LILACS.The following terms (alone and in combination) were employed for the search on PubMed: C9ORF72, MRI, SPECT, PET, ALS, FTD.The same keywords were entered for the search on the LILACS database.We did not employ language or chronological filters in the search.
Titles and abstracts of the papers retrieved in the initial search were screened according to the following eligibility criteria: [1] original research, [2] case series, cohort or cross-sectional design, and [3] imaging methods (MRI, PET and/or SPECT).Abstracts with insufficient information, individual case reports and review articles were not included in the final selection.Disagreements on eligibility were resolved through discussion among the authors.

RESULTS
Table 1 presents findings reported in the selected studies, including the number of patients, neuroimaging technique, and main results.
The initial search resulted in 110 and 69 papers retrieved on PubMed and LILACS, respectively.After this initial screening, papers were selected according to the aforementioned inclusion criteria and duplicate articles removed.The final selection comprised twenty articles (Figure 1).
Selected publications are presented below in three parts: Part I, comprising studies which included FTD patients only; Part II, which describes studies limited to ALS patients; and Part III, which presents studies that included ALS, FTD and FTD-ALS patients.
Part I: FTD Patients.A series of studies assessed the pattern of brain atrophy in FTD patients with C9ORF72 10,12-21 using mainly MRI volumetric analysis.
A widespread, symmetrical pattern of brain atrophy was reported in FTD-C9ORF72 patients compared with healthy controls. 12,13The more atrophic compromised areas were the anterior brain regions, including temporal lobes and all the main subregions of the prefrontal cortex (dorsolateral, orbitofrontal and medial regions).Atrophy in posterior regions (parietal and occipital regions) was also observed in C9ORF72 carriers. 13,14However, these findings were not replicated in a series of C9ORF72 FTD patients in which brain atrophy was assessed using a visual rating scale, and which failed to find significant differences in atrophy patterns between carriers and healthy controls in prefrontal regions (orbitofrontal cortex, anterior cingulate) or temporal regions. 16A recent study reported that carriers of C9ORF72 repeat expansion exhibited significant atrophy in specific brain regions in the pre-symptomatic phase of FTD (before the onset of clinical symptoms). 19Compared to healthy controls, C9ORF72 carriers had marked atrophy in subcortical (thalamus, e.g.) and cortical regions (including frontal, temporal and parietal regions) 20-25 years prior to expected disease onset. 19nvestigating white matter tract changes in different genetic groups of bvFTD 21 compared with healthy controls, C9ORF72 carriers had altered diffusivity in the corpus callosum and cingulum bundle.However, these  data are limited by the small size of the sample (only four bvFTD carriers).Some studies compared neuroimaging features of C9ORF72-bvFTD with sporadic bvFTD and other mutations.C9ORF72-bvFTD patients had less gray matter loss than sporadic bvFTD in the anterior cingulate, orbitofrontal cortex, anterior temporal lobe and insula. 16nother study reported that the majority of subjects with mutation in the microtubule associated protein tau gene (MAPT) and C9ORF72 subjects had symmetric frontal atrophy, while most subjects with mutation in the progranulin gene (GRN) had asymmetric atrophy. 139ORF72 carriers had greater atrophy in posterior (parietal and occipital) lobes in comparison with MAPT and sporadic bvFTD groups, while patients with MAPT mutations had greater impairment in temporal poles, compared with the C9ORF72 group.13 Patients with GRN mutation had more loss in parietal lobes than C9ORF72 carriers.13 In the same study, by applying a multinomial logistic regression model based on atrophic patterns, it was possible to classify FTD patients with different genotypes with 93% accuracy, suggesting that neuroimaging may be useful to distinguish C9ORF72-FTD patients from patients with other mutations at a single-subject level.13 Only one study investigated white matter patterns across bvFTD patients grouped according to genetic status.21 There were no differences between C9ORF72 and sporadic bvFTD cases, but MAPT patients had abnormal fractional anisotropy in the anterior region of the left temporal lobe, compared with the C9ORF72 group.
The integrity of the intrinsic connectivity network in bvFTD was explored in a group of 14 bvFTD C9ORF72 carriers and 14 bvFTD non-carriers. 12These groups were compared against healthy controls.Patients with C9ORF72 did not exhibit differences in the default mode network compared to controls.Conversely, bvFTD noncarriers exhibited a different pattern, presenting both impaired (in striatum and thalamus) and enhanced (in precuneus and posterior cingulate) connectivity compared with controls. 12In the same study, it was reported that C9ORF72 carriers had impaired sensorimotor connectivity in striatum and thalami, compared with bvFTD non-carriers.There was no difference in the salience network connectivity between carriers and non-carriers. 12he progression of brain atrophy in bvFTD patients with different genetic status was assessed in a longitudinal study. 18GRN patients had greater rates of atrophy than sporadic, MAPT and C9ORF72 groups.Sporadic bvFTD patients had greater rates of gray matter loss in anterior cingulate than C9ORF72 carriers, while the latter had greater rates of atrophy in cerebellum and occipital lobes, compared with MAPT carriers. 18Another study found that C9ORF72-bvFTD patients had increased rates of brain atrophy and ventricular expansion compared with healthy controls. 14n summary, widespread brain atrophy was reported in FTD C9ORF72 patients, mostly in anterior brain regions, but also with possible damage in posterior cortical areas.Brain atrophy may be identified before disease onset.However, the absence of significant changes in FTD C9ORF72 carriers has also been reported.
Part II: ALS Patients.Five articles investigated neuroimaging features of ALS patients with C9ORF72 expansion.Four of these studies employed the MRI technique 9,[22][23][24] while the remainder used FDG-PET. 25LS patients with C9ORF72 mutation had greater atrophy in prefrontal regions, including frontal gyri and the anterior cingulate, compared to those with sporadic ALS.9,22 The right precentral gyrus was also affected in one study.9 Mild hypometabolism in the thalamus and posterior cingulate was found on PET-FDG in ALS carriers compared with non-carriers.25 Compared with ALS non-carriers, C9ORF72 carriers had more cortical and subcortical involvement, affecting both cortical (fusiform, supramarginal, and orbitofrontal cortex and Broca's area) and subcortical regions (thalamus).23 Interestingly, in the same study, white matter abnormalities in ALS non-carriers were relatively limited to corticospinal and cerebellar pathways, while carriers had more widespread involvement.These data suggested that non-motor changes (e.g. cogntive impairment) in ALS could be largely driven by C9ORF72 repeat expansion. In short, ALS C9ORF72 carriers had greater atrophy, with predominance in prefrontal regions, compared to sporadic ALS patients.Mild hypometabolism in the thalamus and posterior cingulate, more widespread abnormalities of white matter, and greater basal ganglia involvement has also been demonstrated in ALS carriers compared with non-carriers.
Part III: FTD, ALS and FTD-ALS patients.The imaging patterns of patients with ALS, FTD or FTD-ALS according to their genetic status were compared in a series of studies. 5,10,20,22,26,27Most of the papers employed structural brain MRI.
In a group of eighteen patients with C9ORF72 repeat expansion (fourteen bvFTD, three with FTD/ALS and one with ALS), gray matter loss was found in cortical areas including frontotemporal regions, 26 in a similar pattern to that reported by others. 10,20,22Most studies reported symmetrical patterns of brain atrophy, except for patients presenting with predominant language deficit.Some patients may have parietal cortical atrophy and thalamic involvement. 10,26These studies are limited by the absence of direct comparisons between bvFTD and FTD-ALS.
A group of patients with C9ORF72 expansion (15 bvFTD, 11 FTD-ALS and 5 ALS) was compared against 48 sporadic non-carrier patients (48 bvFTD, 19 FTD-ALS and 6 ALS). 27The authors found that bvFTD-C9 patients had more parietal and bilateral thalamic atrophy and less medial frontal atrophy compared to sporadic bvFTD patients.FTD-ALS C9ORF72 patients had more dorsal frontal and bilateral posterior cortical atrophy and less damage to the temporal pole than sporadic FTD-ALS patients. 27onversely, some studies reported that C9ORF72 carriers may not have brain atrophy. 22,26These findings were expanded by a recent study, which demonstrated that almost 18% of bvFTD cases with C9ORF72 mutation had no abnormalities on PET/SPECT. 5n a study that investigated the metabolic patterns of C9ORF72 carriers on PET-FDG, ALS carriers of C9ORF72 had more pronounced hypometabolism in cortical (cingulate cortex, and frontotemporal regions) and subcortical structures (caudate and thalami) compared with sporadic ALS patients. 28In the same study, ALS patients with C9ORF72 expansion had impaired metabolism in the left temporal cortex, compared with the ALS-FTD group. 28Accordingly, ALS C9ORF72 patients may have a more severe clinical picture and more widespread central nervous system involvement than sporadic ALS patients, regardless of the association with bvFTD.
Taken together, C9ORF72 carriers had symmetrical gray matter loss in cortical regions, except for patients with predominant language deficit, who demonstrated asymmetrical cortical involvement.ALS C9ORF72 patients had more widespread central nervous system involvement than sporadic ALS and/or FTD groups.Some studies have reported an absence of abnormalities on structural and functional neuroimaging.Prado et al.

DISCUSSION
For many years, neuroimaging was of limited applicability in the everyday evaluation of neurodegenerative disorders.For instance, the exclusion of focal lesions or hydrocephalus as causes of cognitive deficits was the main utility of imaging exploration in patients suffering from cognitive disorders.This picture has changed, with modern imaging techniques which provide useful and specific markers for the diagnosis and the follow-up of neurodegenerative diseases, such as ALS and FTD. 29In this paper we systematically reviewed neuroimaging data in FTD and/or ALS patients with C9ORF72 repeat expansion.
3][14][15] Subcortical regions, especially thalami, may also be affected in C9ORF72 carriers. 10,19,22It is of note that some studies reported that patients with C9ORF72 mutation may not have abnormalities on structural and functional brain imaging. 5,16,22,26These disparate patterns may be due to a number of different reasons.The inclusion of patients at different stages of disease and differences in neuroimaging methods across studies may account for the variability of results.One factor that may partially account for these disparate findings is that different phenotypes are associated with C9ORF72 and heterogeneity may occur even among patients with the same clinical phenotype. 22][32][33] Repeated expansion in C9ORF72 may also contribute to Alzheimer's disease. 33,34In summary, although FTD and/ or ALS are the most common phenotypes of C9ORF72 repeat expansion, other clinical presentations may occur, with different neuroimaging patterns.It remains unclear why some patients with the C9ORF72 expansion have minimal atrophy on neuroimaging studies.The possible pathways by which C9ORF72 mutation participates in the pathophysiological process associated with different neurodegenerative diseases also remain elusive.
From a clinical perspective, the variability of clinical findings associated with C9ORF72 limits the interpretation of neuroimaging features at an individual level.A single-center study reported the utility of a multinomial regression model to accurately identify C9ORF72 patients based on patterns of brain atrophy 13 at singlesubject level.However, this strategy seems limited to research centers with advanced expertise in neuroimaging techniques.Moreover, C9ORF72 carriers may have no structural abnormalities on brain MRI. 16,22,26Therefore, atrophic features in brain MRI are of limited value for identification of C9ORF72 carriers in clinical practice.
On the other hand, neuroimaging assessment may be useful for the follow-up of patients with C9ORF72 repeat expansion and for suggesting prognostic aspects.FTD and/or ALS patients with C9ORF72 mutation may have faster disease progression and shorter survival than noncarriers, 10,16 even though this is not consistent across studies. 27In this scenario, neuroimaging can identify markers of disease progression, such as the rate of brain atrophy and ventricular expansion. 14These markers could help track disease changes and guide clinical management, especially in the prospect of disease-modifying drugs that will target the pathophysiological process of neurodegenerative disorders.
New modern neuroimaging techniques may provide useful biomarkers for the diagnosis and follow-up of C9ORF72 carriers.Disruption of functional connectivity may be seen in the absence of brain atrophy and could be regarded as an early marker of disease. 12Only one study to date explored functional connectivity in C9ORF72 carriers, and found that there is a convergent, large-scale, disrupted network among different patterns of brain atrophy. 12The investigation of functional connectivity may enhance our understanding about the neural networks compromised by C9ORF72 mutation, thus providing valuable information for the comprehension of the pathophysiology of the FTD-ALS spectrum.
Techniques exploring the integrity of the white matter tract may also be of clinical value in the assessment of patients with C9ORF72 repeat expansion.Degeneration of the corticospinal tract is a hallmark of ALS, and disruption of this tract can differentiate ALS patients from bvFTD and ALS-FTD patients. 23,35Further studies are needed to describe putative white matter changes associated withC9ORF72 mutation.
Besides its value for diagnostic purposes, neuroimaging is also important for the understanding of the neural basis of cognitive and behavioral disorders observed in the FTD-ALS spectrum.ALS and bvFTD patients with C9ORF72 mutation have a greater frequency of psychiatric disorders, especially psychotic symptoms, such as delusions, paranoid ideation and hallucinations. 5,16,22,36ndeed, almost 40% of FTD patients with C9ORF72 repeat expansion presented psychotic symptoms. 22aranoid or irrational thinking were also frequent in the same study. 22The neuropsychological profile of bvFTD patients with C9ORF72 expansion is similar to non-carrier bvFTD patients, 16 with comparable performance in memory, language and executive skills.Deficits in executive functions are the most common observed feature in ALS-FTD patients 3 and can also be associated with prefrontal dysfunction.
Taken together, these data emphasize the complex interaction between C9ORF72 mutation and clinical presentations of neurodegenerative diseases, especially the FTD-ALS spectrum.The discovery of the C9ORF72 repeat expansion has opened a window for the understanding of the continuum between FTD and ALS.The next advances in neuroimaging investigation may provide valuable markers for the diagnosis and follow-up of these patients, and may also clarify the common pathophysiological pathways between ALS and FTD, with possible clinical outcomes.

111Figure 1 .
Figure 1.Flowchart depicting selection of items for systematic review on PubMed and Lilacs databases using the terms C9ORF72, ALS, FTD, MRI, SPECT and PET.

Table 1 .
Synthesis of articles included in the present review.

Title Authors Journal Year Population Methods Results
DTIThe mean of the brain volume was lower in the C9ORF72 carriers, with decreased grey matter in prefrontal cortex and cerebellar vermis.The DTI showed increased radial diffusivity and decreased fractional anisotropy bilaterally in anterior thalamic radiations, uncinate fasciculus, anterior cingulum and anterior corpus callosum, right posterior corpus callosum, posterior inferior longitudinal fasciculus and superior longitudinal fasciculus.