{
  "abstract": "Aims and Objectives Over one million patients with a traumatic brain injury (TBI) present to UK emergency departments (EDs) every year, highlighting the need for accurate prognostic tools. To the best of our knowledge, we created the largest acute TBI MRI dataset currently available. We examined a range of imaging methods, from routine MRI reporting to lesion annotation and quantitative analysis, to assess their added prognostic value beyond ED clinical assessment.Method and Design Data were collected 2006-2019 across four European prospective cohorts (Cambridge, CENTER-TBI, Trondheim, Turku). Adults (≥16 years) presenting to the ED with TBI who underwent CT within 24 hours and MRI within one month were included. Each scan was reported and annotated by two independent reviewers blinded to outcome. Quantitative analysis comprised volumetric T1 and diffusion tensor imaging. Six-month outcomes were assessed with the Extended Glasgow Outcome Scale: incomplete recovery = GOSE <8; unfavourable outcome = GOSE <5. Logistic regression models using clinical predictors—with and without imaging variables—were internally validated by bootstrapping.Results and Conclusion Of 787 patients (mean age 40, 68% male), 416 had uncomplicated mild, 172 complicated mild, and 199 moderate–severe TBI. Unfavourable outcomes occurred in 11% and incomplete recovery in 45%. Clinical predictors alone explained 27% (95% CI 14–41) and 38% (95% CI 32–44) of the variation in incomplete recovery and unfavourable outcomes, respectively. Adding quantitative MRI significantly improved prognostic performance to 40% (95% CI 28–53) and 48% (95% CI 41–55). Subgroup analysis showed the greatest incremental value of advanced imaging in moderate–severe TBI, with no meaningful benefit in mild TBI. More detailed image analysis may further improve prognostication in mild cases, but currently, comprehensive ED clinical assessment remains a stronger predictor in mild TBI than in moderate–severe injury.",
  "authors": [
    {
      "affiliations": [
        "University of Cambridge & Cambridge University Hospitals"
      ],
      "name": "Sophie Richter"
    },
    {
      "affiliations": [
        "University of Cambridge"
      ],
      "name": "Stefan Winzeck"
    },
    {
      "affiliations": [
        "University of Cambridge"
      ],
      "name": "Marta M Correia"
    },
    {
      "affiliations": [
        "University of Cambridge"
      ],
      "name": "Guy B Williams"
    },
    {
      "affiliations": [
        "Turku University Hospital & University of Turku"
      ],
      "name": "Olli Tenovuo"
    },
    {
      "affiliations": [
        "Norwegian University of Science and Technology & Trondheim University Hospital"
      ],
      "name": "Anne Vik"
    },
    {
      "affiliations": [
        "Norwegian University of Science and Technology & Trondheim University Hospital"
      ],
      "name": "Kent Goran Moen"
    },
    {
      "affiliations": [
        "University of Antwerp & Antwerp University Hospital"
      ],
      "name": "Andrew Maas"
    },
    {
      "affiliations": [
        "Leiden University Medical Center"
      ],
      "name": "Ewout Steyerberg"
    },
    {
      "affiliations": [
        "University of Cambridge & Cambridge University Hospitals"
      ],
      "name": "David K Menon"
    },
    {
      "affiliations": [
        "University of Cambridge & Cambridge University Hospitals"
      ],
      "name": "Virginia FJ Newcombe"
    }
  ],
  "title": "4122 Added value of advanced MRI for outcome prediction after traumatic brain injury presenting to the emergency department",
  "uid": "3365b6f9-af4c-5c1a-a484-fbfd74725a10"
}