Brain metabolic correlates of apathy in amyotrophic lateral sclerosis: An 18F‐FDG‐positron emission tomography stud

The aim of this study was to evaluate brain metabolic correlates of apathy in amyotrophic lateral sclerosis (ALS).


INTRODUC TI ON
Amyotrophic lateral sclerosis (ALS) is a neurodegenerative disease affecting upper and lower motor neurons. Death usually occurs within 2 to 5 years, mainly due to respiratory failure [1]. According to population-based studies, approximately 50% of ALS patients show cognitive and/or behavioural impairment along the frontotemporal degeneration spectrum at diagnosis [2,3]. Apathy has been included among features characterizing behavioural dysfunction since the first diagnostic criteria were established for ALS-related frontotemporal syndromes [4]. Apathy has assumed a central role in the recently revised criteria, which state that the presence of apathy by itself allows a diagnosis of behavioural impairment associated with ALS [5]. Apathy is a feature shared among many neurological and psychiatric disorders. It has been defined in the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition [6] as being characterized by "diminished motivation and reduced goal-directed behaviour, accompanied by decreased emotional responsiveness." The diagnostic criteria for apathy were revised in 2018 as follows [7]: the patient must present a quantitative reduction of goal-directed activity as compared with her/his previous level of functioning; symptoms must persist for at least 4 weeks, and affect at least two of the three apathy dimensions (behaviour/cognition; emotion; social interaction); apathy should lead to functional impairment, and should not be fully ascribable to other factors (e.g., effects of substances or major changes in the patient's environment).
In order to determine if behavioural symptoms of ALS patients represent a change attributable to the neurodegenerative process, premorbid status must be assessed. At the ALS Centre of Turin, Italy, the neuropsychological assessment of ALS patients includes the evaluation of behavioural dysfunction based on direct observation, patient's history, and the Frontal Systems Behaviour Scale (FrSBe) [8]. The FrSBe evaluates three domains (apathy, disinhibition and executive dysfunction) and provides "before" and "after" ratings, referring respectively to the premorbid condition and the time the scale is performed.
As 18 F-2-fluoro-2-deoxy-D-glucose ( 18 F-FDG) positron emission tomography (PET) is a marker of neuronal integrity in vivo [9], in the present study we evaluated brain metabolic correlates, assessed through 18 F-FDG-PET, of the apathy subscore on the FrSBe in a series of patients with ALS. Since we hypothesized that both the "after" apathy score and the change between "before" and "after" conditions could be relevant in characterizing ALS-related behavioural dysfunction, we aimed to evaluate the associations of both of these scores with brain metabolism.

MATERIAL S AND ME THODS Patients
A total of 165 patients diagnosed with definite, probable or probable laboratory-supported ALS according to El Escorial Revised Diagnostic Criteria [10] at the ALS Centre of Turin in the period 2009 to 2015 were included in this study. They were enrolled at diagnosis or, less frequently, during the first follow-up visit (usually 2 months later). Patients with a history of neurological disorders affecting cognition (major stroke, severe head injuries, mental retardation), alcohol and drug dependence, psychiatric diseases (including mood disorders), or use of high-dose psychoactive medications were not enrolled, nor were patients whose native language was not Italian.
Respiratory failure was excluded through clinical assessment, peripheral blood oxygen saturation, and, when necessary, spirometry and arterial blood gases analysis, within 4 weeks before or after the enrolment. Patients underwent 18 F-FDG-PET and neuropsychological assessment including the FrSBe. The complete test battery has been reported elsewhere [3]. Neuropsychological evaluation and 18 F-FDG-PET were performed within 1 month of each other.

Acquisition of 18 F-FDG-PET images
The 18 F-FDG-PET was performed according to published guidelines [11]. Patients fasted at least 6 h before the examination. Blood glucose was <7.2 mmol/L in all cases before the procedure. After a 20-min rest, approximately 185 MBq of 18 F-FDG was injected.
The acquisition started 60 min after the injection. PET/computed hypometabolism in bilateral DLPFC and DMPFC, and left ACC and PMC, and relative cerebellar and pontine hypermetabolism.

Conclusion:
No studies on brain 18 F-2-fluoro-2-deoxy-D-glucose positron emission tomography correlates of apathy have been performed in ALS. We found that FrSBe "after" apathy subscore correlated with metabolic changes in brain regions known as neuroanatomical correlates of apathy. Furthermore, our findings support the relevance of the gap between premorbid and morbid conditions to detect behavioural changes due to the neurodegenerative process underlying ALS.

K E Y W O R D S
18 F-FDG-PET, amyotrophic lateral sclerosis, apathy tomography (CT) scans were performed with a Discovery ST-E System (General Electric, Boston, MA, USA). Brain CT (thickness of 3.75 mm, 140 kVolt, 60-80 mAs) and PET scan (1 field of view of 30 transaxial cm) were sequentially acquired, the former being used for attenuation correction of PET data. PET images were reconstructed with four iterations and 28 subsets with an initial voxel size of 2.34 × 2.34 × 2.00 mm, and data were collected in 128 × 128 matrices.

Behavioural assessment
The FrSBe [8] is a 46-item scale, including a total score and three subscores: apathy (14 items); disinhibition (15 items); and executive dysfunction (17 items). Items are rated on a five-point scale: 1, almost never; 2, seldom; 3, sometimes; 4, frequently; 5, almost always. The FrSBe contains "before" and "after" ratings, referring respectively to the premorbid condition and the time the scale is performed (in our series, at diagnosis). We used the Family version evaluated by a close relative, since reports from caregivers are extremely important given the possible loss of insight of patients [5]. The higher the FrSBe score, the more severe the behavioural impairment. Scores ≥65 are interpreted as pathological according to the FrSBe manual for each section and the total score of the scale [8]. We considered the "after" apathy subscore as a measure of behavioural impairment at diagnosis. The "before-after" change was estimated in two different ways. Firstly, it was measured as the difference, or "gap" between "before" and "after" apathy subscores, calculated as follows: "after" apathy subscore -"before" apathy subscore. Secondly, it was estimated through assessment of the apathetic/non-apathetic status based on the cut-off of 65 points to evaluate eventual change of status between "before" and "after" conditions. Thus, we could subdivide apathetic patients (i.e., "after" apathy subscore ≥65) into two groups: patients with a premorbid score already in the pathological range (i.e. "before" apathy subscore ≥65) and patients with a premorbid score within the normal range (i.e. "before" apathy subscore <65). Both methods were considered as possible proxies of behavioural changes attributable to the neurodegenerative process. In order to identify a possible threshold above which to consider a "before-after" gap as significant, we examined a comparable neurological group as the reference group, as suggested by the manual for the scale [8].
We considered 517 incident ALS patients from the Piemonte and Valle d'Aosta Register for ALS [12], who underwent a neuropsychological assessment, including the FrSBe, at diagnosis, between 2009 and 2015. We excluded 22 patients, who displayed a negative gap between "before" and "after" conditions, possibly due to misinterpretation of the scale by the rater. We also excluded those patients who underwent PET (n = 165, the present study sample).

Statistical analysis
Comparisons between means were made using the Student's t-test or analysis of variance; comparisons between categorical variables were made using the chi-squared test and Fisher's test when applicable.
SPM12 implemented in Matlab R2018b (MathWorks, Natick, MA, United States) was used for image normalization. A customized brain 18 F-FDG-PET template [13] was utilized for spatial normalization. Intensity normalization was performed using the 0.8 default SPM value of grey matter threshold, and images were subsequently smoothed with a 10-mm filter and submitted to statistical analysis.
First, we aimed to evaluate the correlations between brain metabolism and both "after" apathy subscore and the "before-after" gap of the apathy subscore of the FrSBe, performing two multiple regression analyses in the whole sample (n = 165). Subsequently, we focused on patients with an "after" apathy subscore ≥65, that is, patients with scores considered as pathological at diagnosis (n = 84), to evaluate whether a further characterization of such patients based on the "before-after" change was worthwhile. We divided this group into two subgroups to compare them: patients with a "before" apathy subscore ≥65 (i.e., already in the pathological range) versus patients with a "before" apathy subscore <65 (i.e. within the normal range). Then, we divided the same group of patients with the "after" apathy subscore ≥65 into the following two subgroups to compare them: patients showing a "before-after" gap <22 versus patients with a "before-after" gap ≥22.
Comparisons were performed through the two-sample t-test model of SPM12.
In all analyses we did not include age, sex and education as covariates, since the FrSBe scores were already corrected for these variables. Furthermore, we did not include a measure of global cognitive status (i.e., classification according to the diagnostic criteria for ALS-frontotemporal spectrum disorder) [5] or executive dysfunction as covariates, since they were highly correlated with apathy subscores (r = 0.77, p < 0.001). We included the FrSBe "after" subscore related to disinhibition as a covariate in all the analyses, since it was only marginally correlated with the "after" apathy subscore (r = 0.57; p < 0.001). Details regarding the pitfalls of including highly correlated variables as covariates in multiple regression models are reported elsewhere [14].
For all the analyses the height threshold was set at p < 0.005 uncorrected (p < 0.05 FWE-corrected at cluster level) and only clusters containing >125 contiguous voxels were considered significant. Brodmann areas were identified at a 0-to 2-mm range from the Talairach coordinates of the SPM output isocentres corrected by Talairach Client (http://www.talai rach.org/index.html).

Protocol approvals
The study was approved by the ethical committee, "Comitato Etico Interaziendale Azienda Ospedaliero-Universitaria Città della Salute e della Scienza di Torino". The study was performed in accordance with the ethical standards as laid down in the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards.
Patients provided written informed consent.

Demographic and clinical data
We compared the demographic and clinical data of patients who underwent 18 F-FDG-PET (n = 165) with those of the reference population-based series (n = 330). The comparison is summarized in Table S1. No significant difference was found for sex distribution, education, and site of onset (bulbar/spinal). Otherwise, in patients who underwent 18 F-FDG-PET, age was slightly older and ALS Functional Rating Scale-Revised (ALSFRS-R) score slightly higher, probably due to the greater difficulty experienced by elderly people and patients with worse disability in reaching the PET centre.
In the group of patients with an "after" apathy subscore ≥65, that is, those patients with scores considered as pathological at diagnosis (n = 84), we compared demographic and clinical data of those with a "before" apathy subscore ≥65 versus those with a "before" apathy subscore <65, and patients with a before-after gap <22 versus patients with a before-after gap ≥22. In both comparisons, we did not find any difference in terms of sex distribution, site of onset (bulbar/spinal), age at assessment, education, or ALSFRS-R at assessment. These data are summarized in Table S2.

Data obtained from 18 F-FDG-PET
Correlation between the "after" apathy subscore and brain metabolism in the whole sample (n = 165) The "after" apathy subscore negatively correlated with metabolism Correlation between "before-after" gap and brain metabolism in the whole sample (n = 165) The "before-after" gap negatively correlated with metabolism in bilateral DLPFC and DMPFC, left VLPFC, left ACC, bilateral PMC (Table 2, Figure 1b), and positively correlated with clusters including the cerebellum and pons (Figure 2b).

TA B L E 1
Clusters of negative correlation between Frontal Systems Behaviour Scale "after" apathy subscore and whole-brain metabolism in the whole sample Comparison among patients with the "after" apathy subscore ≥65 (n = 84) In patients with an "after" apathy subscore ≥65, we found no difference between those with a "before" subscore ≥65 (n = 26) and those with a "before" subscore <65 (n = 58).
In patients with "before-after" gap ≥22 (n = 40) as compared to patients with "before-after" gap <22 (n = 44), clusters of relative hypometabolism were found in bilateral DLPFC and DMPFC, left ACC, and left PMC (Table 3, Figure 3a), while clusters of relative hypermetabolism were found in cerebellum and pons (Figure 3b).

DISCUSS ION
To our knowledge, no other studies on brain 18 F-FDG-PET correlates of apathy have been performed in ALS patients. Furthermore, F I G U R E 1 (a) Clusters of negative correlation between Frontal Systems Behaviour Scale (FrSBe) "after" apathy subscore and wholebrain metabolism in the whole sample (n = 165) are projected on brain surface. (b) Clusters of negative correlation between FrSBe apathy "before-after" gap and whole brain metabolism in the whole sample (n = 165) are projected on brain surface F I G U R E 2 (a) Clusters of positive correlation between Frontal Systems Behaviour Scale (FrSBe) "after" apathy subscore and whole-brain metabolism in the whole sample (n = 165) are represented on a brain magnetic resonance imaging (MRI) template. (b) Clusters of positive correlation between FrSBe apathy "before-after" gap and whole brain metabolism in the whole sample (n = 165) are represented on a brain MRI template we aimed to evaluate the relationship between cerebral metabolism and behavioural changes, defined as the difference between "before" and "after" apathy subscores on the FrSBe scale. We found that the higher the apathy subscore at diagnosis, the lower the metabolism in brain regions known to be involved in apathy circuitry (the DLPFC, DMPFC, VLPFC, PMC, ACC, and insula).
Similarly, the metabolism of largely overlapping regions tended to decrease as the "before-after" gap increased, suggesting the possible metabolic correlates of behavioural changes due to the neurodegenerative process. Since motor impairment remains the core feature of ALS and bulbar onset is significantly associated with cognitive impairment, we ran further analyses to control for the possible impact of motor disability and site of onset on our results, adding the ALSFRS-R total score and spinal/bulbar onset as covariates in the multiple regression analyses. They provided substantially unchanged results (data not shown).
Many structural magnetic resonance imaging (MRI) studies of apathy in frontotemporal dementia (FTD) have been conducted.

TA B L E 2
Clusters of negative correlation between Frontal Systems Behaviour Scale apathy "before-after" gap and brain metabolism in the whole sample

TA B L E 3 Clusters of relative hypometabolism in patients with Frontal Systems
Behaviour Scale (FrSBe) apathy "before-after" gap ≥22 as compared to patients with "before-after" gap <22, in the sample of patients with FrSBe "after" apathy subscore ≥65 In a voxel-based morphometry study including patients with behavioural variant FTD (bvFTD, n = 48) and primary progressive aphasia (n = 14), FrSBe apathy subscore was significantly correlated with atrophy of the right DLPFC, with trends towards significance in the left DLPFC, right ACC, right lateral orbitofrontal cortex, right temporoparietal junction, and right putamen [15]. A voxel-based morphometry and diffusion tensor imaging MRI study [16] evaluated the grey and white matter correlates of apathy across the three components of initiation, planning and motivation as measured by the Philadelphia Apathy Computerized Test, in a sample of 18 bvFTD patients. DLPFC atrophy was predominantly related to the cognitive component (planning) and to deficits in set-shifting, task setting and abstraction. ACC atrophy was linked to the initiation component deficit. The PMC was found to play an important role in energization and intentional movement planning. These data suggest that the components of apathy underlie partially distinct circuits.
A more recent study [17] applied principal component analysis to identify clusters of behavioural changes based on the Frontal Behaviour Inventory subscores in 102 non-demented ALS patients.
The apathetic profile was correlated with thinning of the bilateral orbitofrontal cortex. tive motivation state across all components, given its role in the F I G U R E 3 (a) Clusters of relative hypometabolism in patients with Frontal Systems Behaviour Scale (FrSBe) apathy "before-after" gap ≥22 as compared to patients with "before-after" gap <22 are projected on brain surface. (b) Clusters of relative hypermetabolism in patients with FrSBe apathy "before-after" gap ≥22 as compared to patients with "before-after" gap <22 are represented on a brain MRI template perception of emotionally significant stimuli, integration of interoceptive inputs and close connections with prefrontal structures.
In our study we identified clusters of negative correlation between apathy subscores and glucose metabolism in regions including the DLPFC, DMPFC, VLPFC, PMC, ACC, and insula, largely overlapping with cortical regions previously shown to be related to different apathy components in FTD [21]. Clusters of positive correlation included the cerebellum and the pons. Notably, cerebellar and brainstem metabolism tends to increase as ALS-related cognitive impairment worsens [22]. The cerebellum is known to be involved in cognitive and behavioural processes. Cerebellar damage can lead to cerebellar cognitive affective syndrome (Schmahmann's syndrome) [23]. Data from neuroimaging and neuromodulation/neurostimulation studies suggest that cerebellar compensatory reorganization might be involved in neurodegenerative diseases affecting cognition, for example, Alzheimer's disease and FTD [24]. Such compensatory cerebellar changes are expected to be more prominent as clinical cognitive and behavioural impairment become more severe [25]. A possible explanation for the finding of a positive correlation between cerebellar metabolism and both the "after" apathy score and the "before-after" gap is the involvement of the cerebellum in compensatory mechanisms. These might be prevalent in earlier stages and represent an adaptive mechanism to overcome frontal cognitive impairment, with effect dissipation over time. This point strengthens the view of ALS as a disease involving multiple neural systems and networks.
Clusters of negative and positive correlation between apathy subscores and brain metabolism were substantially overlapping for the "after" apathy subscore and the "before-after" gap. This finding underlines the importance of the "before-after" gap in the clinical use of the scale, since it could represent a proxy for the behavioural change attributable to the degenerative process. In agreement with the FrSBe manual [8], we examined a comparable, reference, population-based series [12] to identify a possible cut-off value for the gap to attribute a behavioural change to the neurodegenerative process.
We propose that the threshold between the third and fourth quartile be considered as a possible cut-off value. The results of group comparisons support the hypothesis that the entity of the "before-after gap" might be more relevant than the change in category based on the cut-off value of 65 to attribute a behavioural change to the neurodegenerative process of ALS. Therefore, we suggest the "before-after gap" be considered along with the classification based on the cut-off value of 65 points in the clinical assessment of apathy through the FrSBe. However, we cannot exclude the possibility that the different sample sizes of the two groups in the comparison between apathetic patients with a "before" apathy subscore ≥65 (n = 26) versus apathetic patients with "before" apathy subscore <65 (n = 58), might have had a minimal effect on the results. Otherwise, in the comparison between apathetic patients with a before-after gap of <22 and apathetic patients with a before-after gap ≥22, the two groups were similar in size (n = 44 and n = 40, respectively).
A possible limitation of the present study is the fact that MRI scans were not available for all patients, which precluded partial volume effect correction for cortical atrophy. Nevertheless, studies employing voxel-based atrophy correction of resting glucose metabolism showed that metabolic measurements were relatively independent of brain atrophy [26]. Another possible limitation is that we did not characterize brain metabolic changes associated with different components of apathy.
In conclusion, to our knowledge, no other studies on brain 18 F-FDG-PET correlates of apathy have been performed in ALS patients. We found that FrSBe "after" apathy subscore correlated with metabolic changes in brain regions known as neuroanatomical correlates of apathy. Furthermore, our data suggest the relevance of the gap between the premorbid and morbid conditions to detect behavioural changes attributable to the neurodegenerative process underlying ALS.