Brain ¹⁸F-FDG PET Metabolic Abnormalities in Patients with Long-Lasting Macrophagic Myofascitis

Patients

The study population included consecutive symptomatic patients with histopathologic features of MMF at muscle biopsy. Patients with a history of cerebral disease were excluded. All patients underwent both ¹⁸F-FDG PET/CT of the brain and neuropsychological testing (routinely performed in the Garches–Necker–Mondor–Hendaye reference center for rare neuromuscular diseases) as standard care. The Institutional Review Board (Comité de Protection des Personnes Ile-de-France VI, Hôpital Pitié-Salpêtrière, Paris, France) took into account the retrospective nature of this study, approved the protocol (on December 18, 2013), and waived the need for patient informed consent.

Control Group

Healthy subjects from the NCT00484523 clinical trial, similar in age and sex, were included as a control group. They were free from neurologic or psychiatric disease and cognitive complaints, had a normal brain MRI, and underwent a PET scan and then a standardized neurologic examination (including the Mini-Mental State Examination) to confirm the lack of a cognitive disorder.

Neuropsychological Assessment

MMF patients underwent a comprehensive battery of neuropsychological tests exploring specific domains of the MMF-associated cognitive disorder (Table 1). The assessment of executive function included tests exploring working memory (backward digit span and Zazzo cancellation tests), flexibility (Trail Making Test [TMT]), inhibition (Stroop test), and planning (Rey–Osterrieth test). Attention was also explored with the Zazzo cancellation test. Long-term visual memory was assessed with the delayed-recall portion of the Rey–Osterrieth test. Immediate verbal memory was assessed with the forward digit span test, and episodic verbal memory was assessed with the Grober–Buschke long-term free-recall (recovery capacity) and cued-recall (storage capacity) tests. Semantic (animal) and letter (P) fluency tasks were used to examine quantitative (word generation) and qualitative (category clustering and switching) aspects of verbal fluency. Finally, the interhemispheric connection was tested with dichotic listening tests for words and sentences in succession.

Apart from the evaluation of dichotic listening, for which raw performance scores were expressed as the mean number of words or sentences relative to those of the control group, all other raw performance scores were converted to mean z scores by comparison with a control population to express a pathologic threshold at −1.65 SDs of the normal average, as previously described (7). For the TMT and the Stroop interference test (expressed as time), mean z scores of less than 0 and greater than or equal to 0 indicated good performance and poor performance, respectively. For all other tests, a mean z score or a mean number of words or sentences of greater than or equal to 0 indicated good performance, whereas a mean z score or a mean number of words or sentences of less than 0 indicated poor performance.

At the end of the neuropsychological assessment, patients were classified into 4 general categories: those with no significant cognitive impairment (subgroup 1), those with frontal subcortical (FSC) dysfunction (subgroup 2), those with Papez circuit dysfunction (subgroup 3), and those with callosal disconnection (subgroup 4). Patients considered to be cognitively healthy (subgroup 1) were those who performed over the pathologic threshold on all tests (1.65 SDs). Patients with FSC dysfunction (subgroup 2) were those who had pathologic performance on executive function tests (TMT, backward digit span test, Grober–Buschke long-term free-recall test, Rey–Osterrieth complex figure test, Zazzo cancellation test, Stroop test, and verbal fluency). They had good performance on episodic memory and dichotic listening tests. Patients identified as having Papez circuit dysfunction (subgroup 3) had executive dysfunction similar to that of patients in subgroup 2 but had poorer performance on the episodic memory test. They had pathologic results on the 3 total-recall tasks and the Grober–Buschke delayed total-recall test, which assesses storage and consolidation capacities. Finally, patients with callosal disconnection (subgroup 4) potentially had the same cognitive problems as patients in subgroups 2 and 3. They had pathologic results on the dichotic listening test, which revealed left-ear extinction and, therefore, callosal disconnection.

¹⁸F-FDG PET/CT Acquisition

Brain imaging of MMF patients was performed with a Gemini GXL PET/CT camera (Philips) after intravenous injection of ¹⁸F-FDG at 2 MBq/kg. Patients were required to fast for at least 6 h before undergoing the scan and to have normal blood glucose levels, and they were maintained in a neurosensory resting state for 10 min before and 30 min after the injection. A low-dose helical CT scan was performed for anatomic correlation and attenuation correction (x-ray tube tension, 120 kV; current, 80–100 mAs; rotation time, 0.5 s; pitch, 0.938; slice thickness, 2 mm), and then a 1-step emission scan with a duration of 15 min was performed. Images were reconstructed with a line-of-response–row-action maximum-likelihood algorithm (2 iterations; 28 subsets; postprocessing filter, 5.1 mm) and CT attenuation correction (matrix size, 128 × 128; voxel size, 2 × 2 × 2 mm).

Brain imaging of the control group had previously been performed with a Discovery ST PET/CT camera (GE Healthcare). The acquisition parameters were similar (1 step of 15 min at 30 min after intravenous injection of 150 MBq of ¹⁸F-FDG), but the reconstruction algorithm was slightly different—that is, ordered-subset expectation maximization (5 iterations; 32 subsets) with CT attenuation correction (matrix size, 256 × 256; voxel size, 0.8 × 0.8 × 0.8 mm).

Statistical Parametric Mapping Analysis

All ¹⁸F-FDG PET brain image volumes were spatially normalized onto the Montreal Neurologic Institute template (McGill University) with a 12-parameter affine transformation followed by nonlinear transformations. The dimensions of the resulting voxels were 2 × 2 × 2 mm. The images were smoothed with a gaussian filter (full width at half maximum, 10 mm) to blur individual variations in anatomy and to increase the signal-to-noise ratio. Spatial preprocessing and statistical analysis were performed with the statistical parametric mapping (SMP12) software implemented in MATLAB version R2014a (The MathWorks, Inc.). Through the use of analysis of covariance, cerebral ¹⁸F-FDG PET images of all patients and images of patients in each subgroup were compared with images of subjects in the control group to identify an MMF glucose metabolism pattern. The results were collected at a P value of less than 0.005 at the voxel level for clusters k ≥ 200 contiguous voxels (corrected for cluster volume); adjustment for age and sex was made because these nuisance variables are known to have an influence on regional brain metabolism. All significant results were listed with the individual K value, which represented the number of significant voxels in a particular cluster (cluster size), brain regions, side (hemisphere), Brodmann areas involved, and the peak T value (defined through our statistical model as the SD of the measure of cerebral glucose metabolism in comparison to the reference population).