Combined Whole-brain Structural and Functional MRI for the Prediction of Neurological Recovery After Cardiac Arrest

Use of Brain Structural and Functional Connectomes for the Prediction of Neurological Recovery in Coma Patients After Cardiac Arrest

To assess the performance of a predictive model resulting from the analysis of sMRI/fMRI/contrast-enhanced MRI-derived personalized connectomic data, as compared with standard predictors (clinical examination, electrophysiology, serum biomarker, standard neuroimaging) collected ≥ 72h from sedation withdrawal and in normothermia condition, to predict anoxoischemic coma neurological outcome at 6 months.

There is a major need for timely, reliable and generalizable methods to predict outcomes in anoxo-ischemic coma patients. Standard predictors of poor outcome after cardiac arrest (CA) include clinical, electrophysiological and serum biomarkers data. All have substantial limitations in terms of reliability and generalizability. By providing whole-brain structural and functional connectivity maps, or connectomes, advanced MRI techniques have precisely revealed the brain network damages induced by CA. Because these individualized connectomic profiles contains critical information about consciousness recovery potential after CA, it can be hypothesized that these whole-brain quantitative data can be used to elaborate highly performant predictive algorithms for anoxo-ischemic coma patients.

Regarding advanced structural MRI (sMRI), two recent studies, including one from the investigators group, have shown a high sensitivity and specificity of these advanced techniques (diffusion tensor imaging -DTI; voxel-based morphometry -VBM) for predicting poor neurological recovery in anoxo-ischemic patients. However, these two studies collected data using poorly defined time window for MRI, across lengthy data collection periods (> 8 years) and did not apply a strict protocol of withdrawal or limitation-of-care decision to control from misclassification of outcome due to so-called selffulfilling prophecies. Concerning functional MRI (fMRI), a recent study from the investigators group reports that the strength of frontoparietal functional connectivity differs between anoxo-ischemic coma patients who recover and those who eventually score an unfavorable outcome at 3 months.

Furthermore, converging evidence suggest that task-based fMRI can be used to detect active, command-following modulation of cortical activity and, hence, consciousness in behaviorally unresponsive patients. This task-based fMRI pattern named Cognitive Motor Dissociation (CMD) show promise of radically improving good outcome neuroprognostication after CA. Finally, aiming to maximize the performance of MRI-derived predictive models, the investigators group have recently reported in a " proof-of-concept " study that a combined sMRI/fMRI connectomes and contrast-enhanced MRI data analysis, synergistically outperform alternative predictive models based on sMRI or fMRI data in isolation.

As recommended in recent guidelines for the management of anoxo-ischemic coma patients[7-9], a standard multimodal prognostication procedure will be followed, including the collection of standard predictors after at least 72h from complete withdrawal of sedation in normothermia condition: i) clinical examination and behavioral data (Day 1, 3 and 7 after inclusion): Glasgow Coma Scale - GCS, Full Outline of UnResponsiveness - FOUR, Coma Recovery Scale Revised - CRSR; and standardized brainstem reflex testing (FOUR, Glasgow-Liège score); ii) severity stratification scoring (Day 1 after inclusion): Cardiac Arrest Hospital Prognosis - CAHP, Out-of-Hospital Cardiac Arrest - OHCA, iii) laboratory findings (Day 1, 3 and 7 after inclusion) : NSE blood level (Day 1); iv) electrophysiological assessments: standard EEG using ACNS classification (once between Day 1 and Day 15 after inclusion); v) standard neuroimaging (once between Day 1 and Day 15 after inclusion): standard brain CT or MRI data (T1, T2*, SWI, DWI, FLAIR).

In addition to standard clinical neuroprognostication procedure, an advanced whole-brain sMRI/fMRI/contrast-enhanced MRI scan will be acquired at least after 72h from complete withdrawal of sedation in normothermia condition (between Day 1 and Day 7 after inclusion).

sMRI/fMRI/ contrast-enhanced MRI data will be collected during the same scanning plot that will be used for standard MRI (T1, T2*, SWI, DWI, FLAIR) will encompass (total acquisition time = 45 min for all centers, except for Toulouse center = 60 min):

• Structural MRI (total acquisition time = 30 min): i) gray matter: 3D T1-weithed data will be computed to assess whole brain cortical thickness and deep gray matter quantitative volumetry, ii) white matter, whole-brain DTI will be acquired to measure whole brain with matter fractional anisotropy (WWM-FA) and mean average diffusion coefficient (WB-aDC). A normalization procedure will be applied (healthy controls data from each neuroimaging facility).
• Functional MRI (total acquisition time = 10 min for all centers, except for Toulouse center = 25 min): i) passive-task: multislice T2*-weighted for resting-state fMRI analysis images, acquisition time = 10 min; ii) active tasks: fMRI will be used to probe for volitional thought without selfexpression output (motor imagery and motor action), acquisition time = 15 min (only for Toulouse center).
• Contrast-enhanced MRI (acquisition time = 5 min) for blood-brain barrier permeability assessment. As exploratory goals and seeking to: i) study the potential changes over time of advanced brain sMRI/fMRI/contrast-enhanced MRI data, a second identical advanced sMRI/fMRI/contrastenhanced MRI will be performed minimum 7 days (with an allowance of + 3 days) after the first MRI assessment in patients enrolled in Toulouse center (N = 30); ii) To investigate the usefulness for patient's neuroprognostication of novel brain injury fluid-derived biomarkers (ref), three peripheral blood samples will be collected two times (each sample blood volume= 5 ml), a first one immediately at patient's inclusion and second one 7 days later (N = 30, only for Toulouse center).

To gauge the clinical significance of this findings, the investigators plan to use largely validated neurological functional score (mRS, CPCs). Additionally, the investigator plan to explore as secondary evaluation criteria patients' level of consciousness (CRS-R) and the restauration of the pre-arrest health-related quality of life (HRQOL). These assessments will be performed at hospital discharge (mRS, CPC) and at 3 (mRS, CPC) and 6 months (mRS, CPC, CRS-R, HRQOL) after CA by specifically trained investigators, during the planned follow-up visit. Patient's medical care will be not be influenced by patient's study participation because the treating teams will be fully blinded to advanced sMRI/fMRI/contrast-enhanced MRI data (Figure 3). Patient's management will be performed in agreement with international guidelines. A strict and homogenous WLST protocol will be used in all the recruiting centers.

Inclusion Criteria:

• Adult patients (male or female ≥ 18 years).
• Coma, as indicated by a Glasgow Coma Scale (GCS) ≤ 8 (motor score ≤ 2) immediately after CA resuscitation and before sedation onset.
• Persisting unconsciousness, defined as the inability to obey verbal commands, after at least 72 hours from complete withdrawal of sedation in normothermia conditions.
• Written informed consent from patient's legal representative.
• Affiliation or beneficiary to the French social security system.

Exclusion Criteria:

• Brain death.
• Coma explained by other cause than CA.
• Likely poor neurological outcome based on early predictors, following ERC-ESCIM 2021 recommendations. In a comatose patient with GCS motor score ≤ 3 at ≥ 72 h from ROSC, in the absence of confounders, the identification of at least two of the following: bilaterally absent pupillary light and corneal reflexes at ≥ 72h, bilaterally absent N20 SSEP ≥ 24h; neuron-specific enolase (NSE) > 60 μg/l at 48h and/or 72h, status myoclonus ≤72h.
• Decision of WLST previous to patient recruitment, based on early predictors of poor neurological outcome, age, co-morbidity, general organ function and patient's preferences.
• Life expectancy shorter than 6 months based on pre-morbid conditions.
• Former neurological functional disability (mRS > 2 before CA).
• MRI contraindication: medical material not MRI compatible, claustrophobia
• Known hypersensitivity to gadoteric acid, meglumin or any drug containing gadolinium
• Severe kidney failure defined as a KDIGO score > 3 (glomerular filtration rate < 30 ml/min/1.73 m2 or renal replacement therapy).
• Hemodynamic shock or severe respiratory failure precluding patient's transport and MRI scanning.
• Pregnancy or nursing woman.
• Patient under juridical protection.