Marshall CT Classification as an Early Radiological Predictor of In-Hospital Mortality in Adult Surgical Traumatic Brain Injury: A Tertiary Indonesian Cohort Study

Abstract

Background: Computed tomography (CT) classification anchors surgical triage in traumatic brain injury (TBI), but recent Indonesian data on the prognostic value of the Marshall CT classification in surgically managed patients are scarce.

Objective: To quantify how strongly and independently the Marshall CT classification is associated with in-hospital mortality after cranial surgery for TBI.

Methods: This observational retrospective cohort study was conducted at Dr. Mohammad Hoesin Central General Hospital, Palembang. Consecutive adult patients operated on for TBI between January 2023 and December 2025 were enrolled. Two independent raters assigned Marshall categories from non-contrast head CT (κ = 0.87). The primary outcome was in-hospital mortality, analyzed with Spearman rank correlation, point-biserial correlation, penalized L2 logistic regression and receiver operating characteristic (ROC) analysis.

Results: Of 36 patients (75.0% male; median age 25 years), 7 died before discharge, a mortality of 19.4% (7/36; 95% confidence interval [CI] 9.8–35.0%). Higher Marshall categories carried higher mortality (Spearman ρ = 0.532, 95% CI 0.245–0.731; p = 0.001). With penalized adjustment the odds ratio (OR) was 3.12 per category (95% CI 1.63–4.41; p = 0.013). Marshall category gave an area under the ROC curve (AUC) of 0.852 (95% CI 0.724–0.960); admission Glasgow Coma Scale (GCS) score gave 0.778 (DeLong p = 0.298).

Conclusion: The admission Marshall category independently predicted in-hospital death after TBI surgery in this small Indonesian cohort; because the penalized model was exploratory, external validation is needed before the score is built into operative triage for early risk stratification.