Prediction for Coma Recovery in ICU With Comaweb: a Validation Study

The general objective of the comaScore project is to provide an external validation of the accuracy of the comaScore, a score derived from magnetic resonance imaging (MRI), to predict 1 year outcome of patients unresponsive to simple orders after traumatic brain injury (TBI), aneurysmal subarachnoid hemorrhages (aSAH) and cardiac arrest (CA) in the day 7 - day 45 period post brain injury. As recent studies reported late awakeners cases, even in cardiac arrest and, in contrary, that around 10% of patients with acute brain injury remain with permanent disorders of consciousness (DOC), the need of reliable prognosis tool at the early phase, while the patient is still in the ICU, is critical. ComaScore, based on the quantitative analysis of diffusion tensor imaging, was developed from a derivation cohort of 506 patients. It is much more performing than existing tools (IMPACT, OHCA) in this respect.

This study aims to prospectively validate the predictive accuracy of comaScore to predict unfavorable outcome at 1-year after the first insult, in an independent sample (external validation). Unfavorable outcome is defined as a Glascow Outcome Scale Extended (GOSE) of 1 to 3. This scale ranges from 1 to 8. Scores of 1 to 3 correspond to death, vegetative state, minimally conscious state or bedridden.