Multimodal prediction of 3- and 12-month outcomes in ICU-patients with acute disorders of consciousness

Abstract

Background: In intensive care unit (ICU) patients with coma and other disorders of consciousness (DoC), outcome prediction is key to decision-making regarding prognostication, neurorehabilitation, and management of family expectations. Current prediction algorithms are largely based on chronic DoC, while multimodal data from acute DoC are scarce. Therefore, CONNECT-ME (Consciousness in neurocritical care cohort study using EEG and fMRI,NCT02644265) investigates ICU-patients with acute DoC due to traumatic and non-traumatic brain injuries, utilizing EEG (resting-state and passive paradigms), fMRI (resting-state) and systematic clinical examinations. Methods: We previously presented results for a subset of patients (n=87) concerning prediction of consciousness levels at ICU discharge. Now, we report 3- and 12-month outcomes in an extended cohort (n=123). Favourable outcome was defined as modified Rankin Scale ≤3, Cerebral Performance Category ≤2, and Glasgow Outcome Scale-Extended ≥4. EEG-features included visual-grading, automated spectral categorization, and Support Vector Machine (SVM) consciousness classifier. fMRI-features included functional connectivity measures from six resting-state networks. Random-Forest and SVM machine learning were applied to EEG- and fMRI-features to predict outcomes. Here, Random-Forest results are presented as area under the curve (AUC) of receiver operating curves or accuracy. Cox proportional regression with in-hospital death as competing risk was used to assess independent clinical predictors of time to favourable outcome. Results: Between April-2016 and July-2021, we enrolled 123 patients (mean age 51 years, 42% women). Of 82 (66%) ICU-survivors, 3- and 12-month outcomes were available for 79 (96%) and 77 (94%), respectively. EEG-features predicted both 3-month (AUC 0.79[0.77-0.82] and 12-month (0.74[0.71-0.77]) outcomes. fMRI-features appeared to predict 3-month outcome (accuracy 0.69-0.78) both alone and when combined with some EEG-features (accuracies 0.73-0.84), but not 12-month outcome (larger sample sizes needed). Independent clinical predictors of time to favourable outcome were younger age (Hazards-Ratio 1.04[95% CI 1.02-1.06]), TBI (1.94[1.04-3.61]), command-following abilities at admission (2.70[1.40-5.23]), initial brain-imaging without severe pathology (2.42[1.12-5.22]), improving consciousness in the ICU (5.76[2.41-15.51]), and favourable visual-graded EEG (2.47[1.46-4.19]). Conclusion: For the first time, our results indicate that EEG- and fMRI-features and readily available clinical data reliably predict short-term outcome of patients with acute DoC, and EEG also predicts 12-month outcome after ICU discharge.

Exploring the Significance of Cognitive Motor Dissociation on Patient Outcome in Acute Disorders of Consciousness

“…Besides EEG (and ERPs) and fMRI, the prognostic value of other emerging techniques such as functional near-infrared spectroscopy or transcranial magnetic stimulation combined with EEG (TMS-EEG) will be investigated in the future in the acute clinical setting. 29,30 Finally, an effort has been undertaken to investigate the added value of multimodal assessment, such as done in the study of Amiri et al 31 However, they failed to prove the added value of the combination of EEG and fMRI data and using machine learning to predict 12month outcome while EEGs alone-as well as thorough repeated neurological assessment alone-were good models to predict outcome at that time point. Mortality rates were not addressed systematically, but we found similar death rates for patients with CMD (33%) compared with non-CMD patients (30.5%).…”