Dexmedetomidine inhibits unstable motor network in patients with primary motor area gliomas

Participants characteristics

The study prospectively recruited 21 patients (12 males, age = 42±12 (mean±SD) years) who were diagnosed as brain frontal-parietal gliomas between September 2017 and December 2018. The study was approved by the appropriate Institutional Review Board (IRB), and written informed consent was obtained from all subjects. Sixteen patients received surgical treatments. Inclusion criterion: patients with supratentorial mass lesion; aged 18 to 75; ASA I-II; MRI compatible. Exclusion criteria:unable to comprehend and cooperate with the behavioral and fMRI examination; impaired mental status; taking sedative drugs in the past 24 hours; taking pain reliever in the past 24 hours; drug and/or alcohol abuse; pregnant and/or lactation period woman. All patients accomplished the task-fMRI and resting-state fMRI (rs-fMRI) examinations. The patients were divided into three groups according to the spatial location of tumor and M1: the tumor was located in SMA, PMd and M1 (pre-M1 group, 9 patients), in postcentral gyrus (post-M1 group, 6 patients) and midline postcentral gyrus (control group, 6 patients).

The work-flow of experiment and sedation protocol

The work-flow of experiment before and after sedation is showed in Figure 1A. Subjects fasted for at least 6h from solids and 2h from liquids before sedation. During the study and the recovery period, an anesthesiologist monitored electrocardiogram, blood pressure, pulse oximetry (SpO₂), and breathing frequency (Monitor: MAGLIFE C PLUS; Schiller Medical; Germany). After the first several tests, dexmedetomidine (Jiangsu Hengrui Medicine Co., Ltd., China) was infused through an intravenous catheter placed into a vein of the right hand or forearm. Dexmedetomidine was administered as a 1μg/kg loading bolus over 10min, followed by a 0.7μg/kg/h infusion. The sedation states were judged by the anesthesiologist who adjusted the drug administration as the study needed. The patients’ sedation level was evaluated by Observer’s Assessment of Alertness/Sedation (OAA/S) scale that was developed to measure the level of alertness in subjects who are sedated [18]. The OAA/S is scored from 1 to 5, indicating deep sleep to fully alert (5 = alert, 4 = lethargic, 3 = aroused by voice, 2 = aroused by shaking, 1 = deep sleep). The titrated sedative doses were guided by OAA/S, targeting a score ≤ 3. Once achieved a score of 3, the second rs-fMRI and task-fMRI was performed. At last the second muscle power test was performed and then the drug infusion was stopped. Throughout the study, the subjects breathed spontaneously. The vital signs were stable, and no study was interrupted because of the drug administration.

fMRI data acquisition

MRI data were performed at a Siemens Verio 3.0 Tesla MRI scanner (Siemens Healthineers, Erlangen, Germany). A three-plane localizer image was initially obtained to identify anatomic landmarks and to allow positioning of the transverse sections parallel to the anterior commissure–posterior commissure line. The head of each participant was snugly fixed by foam pads to reduce head movements and scanner noise. The structural images were obtained in a sagittal orientation employing a magnetization prepared rapid gradient echo (MPRAGE) sequence over the whole brain: 192 slices, TI/TR/TE = 900/2,300/3.25msec, flip angle: 8°, slice thickness 1.0 mm, field of view = 256 × 256 mm², matrix = 256 × 256.

The brain task-fMRI (Figure 1B) was performed during alternating left and right four finger-to-thumb opposition movements by using T2*-weighted blood oxygen level-dependent based Echo-planar MR imaging (36 slices, repetition time/echo time = 3,000/30 ms, flip angle = 90°, field of view = 192 × 192 mm², matrix=64 × 64, slice thickness 3.0 mm, 0.75mm intersection gap, in plane resolution: 3mm × 3mm). The task instructions were auditory-cued using a digital audiotape. All subjects were visually monitored during the experiment to ensure compliance with the protocol. After drug was infused, if the patient stopped moving because of drowsiness, he/she would be verbally reminded to continue immediately through speaker. If there were more than 2 stops in one task block, the task would be aborted and repeated.

Rs-fMRI was also performed before and after dexmedetomidine was administered, with blood oxygen level–dependent (BOLD) gradient-echo T2*-weighted echo-planar imaging (32 slices, repetition time/echo time = 2,220/30 ms, flip angle = 90°, field of view = 192 × 192 mm², matrix= 64 × 64, slice thickness 3 mm, 1mm intersection gap, in plane resolution: 3mm × 3mm). During the echo-planar imaging data acquisition, subjects were instructed to keep relaxed with their eyes closed and remain motionless as much as possible. Each scan lasted for 9 min 6 secs and 246 volumes were obtained.

Regions-of-interest (ROI) creation in the motor network of rs-fMRI

Ten ROIs were selected to detect the functional connectivity (FC) of the hand motor network. We focused on the dynamic changes in the organization of the motor execution network controlling for the movement of the contralesional hand. First, because the M1 may be damaged or significantly shifted from the original site, we define the ROI of M1 individually. We used the voxels of task-evoked activation to create an individual mask of M1, and then the mask was normalized using the non-linear transformation to MNI space. For supplementary motor area (SMA), because in all cases there was activation during task-fMRI, for the sake of simplicity we used an individual spherical region (radius = 5.0 mm) centered at the peak activation point of SMA. The ipsilesional dorsal premotor area (PMd) was damaged in pre-M1 group, so in all cases we only used contralesional PMd spherical ROI with 5.0 mm radius and defined it in MNI space (left -22, -13, 57; right 28, -10, 54) [15, 19]. For other ROIs, we also used MNI coordinates from previous reports including bilateral thalamus (ventral lateral nucleus), superior cerebellum (lobule VI), and dentate nucleus (Table 1) [19].

Data pre-processing and analysis of task-fMRI and rs-fMRI

The task-fMRI images were preprocessed using DPABI standard pipeline including slice timing, realignment, structural images co-registration to functional images, segment (DARTEL) and smoothing [20]. Statistical analyses of functional images were performed in MATLAB 2015b (Mathworks, Natick, MA, USA) with SPM12 (Wellcome Centre of Human Neuroimaging, London, UK) using the general linear model. For each subject, images were corrected for subject motion with the first volume of each study used as a reference. Volumes affected by excessive motion (10 mm displacements) were discarded. The EPI images were co-registered with anatomic images. Tumor segmentation was performed based on the high resolution T1 anatomical dataset. The resulting images were smoothed with a Gaussian spatial filter to a final smoothness of 5 mm. The results of individual analyses were thresholded at the p < 0.05 level of significance corrected for family wise error (FWE). For less conservative assessment, if there was no activation or cluster size < 10, the uncorrected threshold of p < 0.001 was used. The functional image was then overlaid on a high resolution T1 structural image in order to have the anatomic localization of the functional foci.

The rs-fMRI data were also preprocessed using DPABI standard pipeline. The first 10 volumes were discarded to allow for magnetization equilibrium effects and the adaptation of the subjects to the circumstances, leaving 236 volumes for further analysis. The processing steps included slice timing, realignment and structural images co-registered to functional images. Structural images were normalized with new segment and DARTEL. The nuisance covariates were regressed out with Friston 24 (6 head motion parameters), CSF and white matter. Then the functional images were normalized, detrended, smoothed (Gaussian kernel full width half maximum = 4 mm), and temporal band pass filter (0.01–0.1 Hz) was used to reduce low-frequency drift and high-frequency physiological noise. Rs-fMRI time courses of each ROI were extracted, and ROI-to-ROI FC values were calculated using Pearson correlation and transformed to Fisher z.

Maximum grip strength test

The maximum grip strength was measured before and after sedation using a digital dynamometer (Zhongshan Camry Electronic Co. Ltd, Guangdong, China) as motor ability. Lateralization index (LI) of maximum grip strength was defined as (C–I)/(C+I) where C and I mean the contralateral and ipsilateral maximum grip strength.

Surgery strategy and clinical results

Participant sociodemographic and medical variables (Table 2) were archived in a purpose-made database for later analysis. Surgery was performed by one senior neurosurgeon with intraoperative neurophysiological monitoring (appendix). All the participants were evaluated on preoperative day 1, postoperative day 7, and at follow-up after 1 month regarding their motor function.

Statistical analysis

Lateralization index (LI) of M1 was defined as (C–I)/(C+I) where C and I mean the contralateral and ipsilateral sensorimotor area to the ipsilesional hemisphere. Magnitude LI (LI-M) using the maximum T-value and volumetric LI (LI-V) using the number of activated voxels within the sensorimotor area was calculated. Statistical analyses were conducted using SPSS (IBM Corp., Chicago, IL, USA). The significance level was set at P < .05. The false discovery rate (FDR) correction was applied to control false positives from multiple comparisons. FDR adjusted P value (p[FDR]) less than 0.05 was considered significant.

Clinical characteristics and grouping of patients

The imaging and clinical data of the 21 patients are summarized in Table 2. All the patients underwent fMRI examinations before and after sedation successfully without any observed side effects. 16 patients received surgical treatments. After the operation, new motor deficits of upper-limb occurred in five of six patients in the pre-M1 group, and none in other groups (P < .01). Probability maps of lesion distribution of the three groups were shown in Figure 2A.

Behavioral test

We found no significant difference of the motion counts among the three groups and no significant difference of the motion counts or maximum grip strength laterality between pre- and post-administration of dexmedetomidine within each group.

Task-fMRI: sedation alters lateralization of hand motor areas

Figure 2B showed typical activation maps before and after sedation. In pre-M1 group, we found no statistical significance of either activation volume or activation magnitude between the lesional and healthy hemispheres before sedation; however, after sedation, both activation volume (P < .02) and activation magnitude (P < .02) in the lesional hemisphere was significantly lower than that in the healthy hemisphere, which were not observed in other groups. The LI-M and LI-V of M1 were strongly correlated both before (r = 0.901; P < .001) and after sedation (r = 0.908; P < .001). Both LI-V (P < .03) and LI-M (P < .02) increased significantly after sedation in pre-M1 group (Figure 2C). Surprisingly, in post-M1 group and control group, we found no significant differences for either LI-V (P = .173 versus .173) or LI-M (P = .344 versus .917) after sedation.

Rs-fMRI: sedation alters FCs within the motor network

Before sedation, the FCs distribution of post-M1 (mean, 0.41; 95%CI, 0.39-0.44) were lower than that of pre-M1 group (mean, 0.44; 95%CI, 0.42-0.47; P < .05) and control group (mean, 0.47; 95%CI, 0.44-0.51; P < .05); however, we found no significant difference of FCs between pre-M1 group and control group (P = .376). After sedation, FCs distribution of pre-M1 group (mean 0.24; 95%CI, 0.22-0.26) was significantly lower than that of post-M1 (mean 0.31, 95%CI, 0.28-0.34, P < .0001) and control group (mean 0.47, 95%CI, 0.44-0.50, P < .0001); FCs distribution of post-M1 group were significantly lower than that of control group (P < .0001). Distribution of FCs in pre-M1 group (P < .0001) and post-M1 group (P < .0001) shifted to the lower FCs obviously after sedation (Figure 3A, 3B), which was not observed in control group (Figure 3C, P = 0.90).

Compared with pre-sedation, 27 of 45 FCs within motor areas decreased significantly after sedation in pre-M1 group (p[FDR] < .05, Figure 3D). However, none of these FCs decreased significantly in post-M1 group and control group (Figure 3E, 3F). Compared with pre-sedation, FCs of the ipsilesional M1 with the M1-H, PMd, SMA, Cere-L, Cere-H, Den-L and Den-H were all significantly decreased after sedation in pre-M1 group (p[FDR] < .05, Figure 4D); significantly decreased FC was observed only between M1-L and Den-L in post-M1 group (p[FDR] < .05, Figure 4E); none of the FCs decreased significantly in control group (Figure 4F).

We set FC threshold at 0.3 where maximal differences were observed among the three groups (Figure 5). 34 of 45, 26 of 45 and 4 of 45 FCs were under threshold after sedation in pre-M1, post-M1, and control group respectively (Figure 4A–4C).