Marshall classification of traumatic brain injury

The Marshall classification of traumatic brain injury is a CT scan-derived metric using only a few features and has been shown to predict outcomes in patients with traumatic brain injury (TBI). 

Usage

This system was first published in 1992 ¹, building on findings from a large cohort of head injury cases described in 1990 ², and is in common use (c. 2024 ⁷) for grading acute TBI based on CT findings.

The Rotterdam score is a more recent system that attempts to address some of the recognized limitations of the Marshall system, such as the struggle to classify patients with multiple types of injuries ^3-5. It has been suggested that although both systems are valuable in predicting an outcome, as the newer Rotterdam score incorporates additional variables (e.g. subarachnoid hemorrhage) and may be a preferable measure ⁶.

Classification

The Marshall system places patients into one of six categories (I to VI) of increasing severity based on findings on non-contrast CT scans of the brain. Higher categories have worse prognosis and survival. It is primarily concerned with two features: 

1. degree of swelling, as determined by

   • midline shift and/or

   • compression of basal cisterns

2. presence and size of contusions/hemorrhages referred to "high or mixed density lesions"

Categories

• diffuse injury I  (no visible pathology)

  • no visible intracranial pathology

• diffuse injury II

  • midline shift of 0 to 5 mm

  • basal cisterns remain visible

  • no high or mixed density lesions >25 cm³

• diffuse injury III (swelling)

  • midline shift of 0 to 5 mm

  • basal cisterns compressed or completely effaced

  • no high or mixed density lesions >25 cm³

• diffuse injury IV (shift)

  • midline shift >5 mm

  • no high or mixed density lesions >25 cm³

• evacuated mass lesion V

  • any lesion evacuated surgically

• non-evacuated mass lesion VI

  • high or mixed density lesions >25 cm³

  • not surgically evacuated