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

Abstract

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

Methods: An observational retrospective cohort study was conducted at Dr. Mohammad Hoesin Central General Hospital, Palembang. Consecutive adult surgical TBI patients operated on between January 2023 and December 2025 were enrolled. Marshall CT categories were assigned from non-contrast head CT scans by two independent raters (κ = 0.87). The primary outcome was in-hospital mortality. We quantified the association between Marshall score and in-hospital mortality using Spearman rank correlation, point-biserial correlation, penalised L2 logistic regression, and ROC analysis.

Results: Thirty-six patients met the inclusion criteria (75.0% male; median age 25 years). In-hospital mortality was 19.4% (7/36; 95% CI 9.8–35.0%). Spearman correlation between Marshall score and mortality was ρ = 0.532 (95% CI 0.245–0.731, p = 0.001). Marshall category was an independent predictor in penalised logistic regression (adjusted OR 3.12, 95% CI 1.63–4.41; p = 0.013). ROC AUC was 0.852 (95% CI 0.724–0.960), numerically superior to admission GCS AUC of 0.778 (DeLong p = 0.298).

Conclusion: Marshall CT classification is an independent, point-of-CT predictor of in-hospital mortality after surgery for TBI in an Indonesian tertiary cohort. Integration into structured operative triage may enhance early risk stratification.