Associations Between Neuroimaging Findings and Extubation Failure in Traumatic Brain Injury

Tuesday, March 24, 2026

8:45 AM - 9:45 AM Central Time

Location: Connections Central - RST 06

Ali Abu-Alya, DO (he/him/his)

University of Washington, Washington

Disclosure(s): No relevant financial relationship(s) to disclose.
AG

Aaron Gallagher, MD

Neurology Resident
University of Washington

Disclosure information not submitted.
EB

Eliza Baird-Daniel, MD

Neurosurgery Resident
University of Washington

Disclosure information not submitted.
CF

Christine Fong, N/A

Senior Computer Specialist
University of Washington

Disclosure information not submitted.
AL

Abhijit Lele, MBBS, MD, MSCR, FNCS, CPHQ

Neurocritical Care Physician
University of Washington

Disclosure information not submitted.
RB

Robert Bonow, MD

Neurocritical Care Physician
University of Washington

Disclosure information not submitted.
JT

James Town, MD

Critical Care Physician
University of Washington

Disclosure information not submitted.
SW

Sarah Wahlster, MD

Neurocritical Care Physician
University of Washington

Disclosure information not submitted.

Introduction: The ability to predict successful extubation in patients with acute brain injury remains inadequate, reflected by high extubation failure rates in this population and imperfect predictive models. We examined associations between neuroimaging findings and extubation failure in patients with moderate to severe traumatic brain injury (TBI).

Methods: In this single center retrospective study (2021-2025), we included patients with TBI and Glasgow Coma Scale (GCS) ≤12 who received invasive mechanical ventilation (IMV) for ≥24 hours and underwent an extubation trial. Patients with cervical spinal cord injury and major chest trauma were excluded. The primary outcome was extubation failure within 5 days of extubation. Head computed tomography (CT) scans on admission and prior to extubation were reviewed. Using multivariable logistical regressions adjusting for age, GCS, and prior cardiopulmonary disease, we assessed associations between extubation failure and multiple pre-defined neuroimaging findings, including validated TBI scoring systems and structural lesions of neuroanatomical structures involved in airway protection.

Results: Of 181 patients included (mean age 45 years, 26% female, 92% admission GCS 3-8, 42% underwent neurosurgical intervention, median IMV 2.2 days [IQR 1.5-4.7]), 32 (17.68%) experienced extubation failure. Marshall score 3/4/6 (OR 2.9, 95% CI 1.1-7.1), Rotterdam score ≥3 (OR 2.6, 95% CI 1.1-7.1), Helsinki score ≥6 (OR 3.5, 95% CI 1.4-8.4), edema score ≥3 (OR 4.0, 95% CI 1.7-10) and presence of cisternal compression (OR 3.1, 95% CI 1.4-7.4) on pre-extubation CT (obtained within a median of 1.5 days (IQR 0.9-3) prior to extubation) were associated with extubation failure, however, no associations were found for the same findings on the admission CT.

Conclusions: Neuroimaging findings on pre-extubation CT but not admission CT were associated with extubation failure. The predictive value of neuroimaging findings within extubation failure warrants further evaluation.