INTRODUCTION

Background

Traumatic brain injury (TBI) is one of the most severe forms of medical trauma, often resulting in adverse prognoses and significant neurological impairments that negatively impact a patient's quality of life and physiological well-being. TBI can occur as a result of various incidents—including blunt impacts or penetrating trauma from traffic accidents, criminal acts, or sports injuries—and frequently leads to impaired brain function in affected individuals [1]. According to the Global Burden of Disease report from 1999 to 2016, the global incidence and prevalence of TBI have risen, with age-standardized incidence and prevalence rates increasing by 3.6% and 8.4%, respectively [2]. Thus, TBI poses a significant challenge not only for affected individuals but also for global healthcare systems.

For diagnostic purposes and mortality prediction in TBI patients, clinicians routinely employ computed tomography (CT) scans. CT scans produce detailed cross-sectional images of the brain, enabling the identification of traumatic tissue damage.

However, with advancements in imaging technology, there is an increasing need for a more systematic approach to interpreting CT scan results in TBI patients. Several scoring systems based on noncontrast brain CT scans have been introduced. For example, CT scoring systems for TBI include the Marshall, Rotterdam, and Helsinki CT scores. These systems generate numerical scores based on CT findings to predict injury severity, potential complications, and overall prognosis. Therefore, an effective scoring system should be designed to clearly differentiate between patients likely to experience favorable versus unfavorable outcomes [3].

Objectives

Based on the above, we aim to analyze the effectiveness of various CT scan assessment methods for TBI to identify the most optimal approach for time-efficient diagnosis and mortality prediction. This analysis was conducted through a systematic review and bibliometric analysis of studies published between 2003 and 2024.

METHODS

Study design and setting

This research employed two methodologies: a systematic review following the 2020 PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines [4] and a bibliometric analysis.

Inclusion and exclusion criteria

Inclusion criteria were established to ensure that the studies selected for the systematic review were available in full text and published in English. Studies that were unavailable in full text, presented as narrative reviews, not in English, or that did not align with the study characteristics were excluded.

Data collection

Secondary data were gathered from databases including PubMed, Scopus, and Google Scholar. The search focused on studies published between 2003 and 2024. The keywords used for each database are listed in Table 1.

The collected data were exported to the Rayyan research tool, where 60 duplicate manuscripts were removed. Subsequently, 99 manuscripts were excluded based on titles and abstracts that did not align with the study objectives. An additional 15 manuscripts were not screened further due to not meeting the predetermined inclusion criteria, such as language incompatibility. Following manual screening, 15 manuscripts were selected for full-text analysis, with 9 manuscripts subsequently excluded due to ineligible research characteristics. Ultimately, 6 manuscripts met the inclusion standards for this review (Fig. 1) [5–10].

Bibliometric analysis

Unlike the systematic review, the bibliometric analysis utilized only one database—Scopus—for data collection, along with VOSviewer ver. 1.6.19 (Centre for Science and Technology Studies, Leiden University) for data visualization and analysis. A total of 72 documents were included. Extracted data comprised publication titles, authors, affiliations, keywords, publication years, and citation information. Two visualization methods were employed: network visualization and overlay visualization. The network visualization displays labels, circles, and connecting lines between topics, with circle size indicating the strength of correlation between research outputs. The overlay visualization differentiates objects by color according to the year of publication, ranging from blue (older/lowest score) to green and yellow (newer/highest score) [11].

RESULTS

The systematic review included six studies that underwent the full manuscript analysis process, as detailed in Table 2 [5–10].

Effectiveness of CT scoring systems

Based on the review of the six studies presented in Table 2 [5–10], effectiveness appears to be dominated by the Helsinki and Rotterdam CT scoring systems, which received comparable ratings. However, each study employed different assessment criteria, potentially accounting for the variations in results. Overall, the review indicates a close competition between the Rotterdam and Helsinki parameters; notably, two studies reporting area under the curve (AUC) and Nagelkerke pseudo-R² calculations showed that the Helsinki CT score excels [5,6].

Co-occurrence analysis

Fig. 2 presents a network visualization derived from a co-occurrence analysis of publication trends in CT scoring systems as mortality predictors in TBI. It illustrates the relationships among the main research topics and themes related to CT scoring systems published between 2003 and 2024. The analysis identified the most prominent topics as “Marshall CT score,” “Rotterdam CT score,” followed by “Helsinki computed tomography.”

Fig. 3 displays an overlay visualization from the co-occurrence analysis, comparing the efficacy of CT scoring systems as predictors of TBI mortality. The analysis revealed that studies focusing on the Rotterdam CT score peaked in 2017, while those referencing the Marshall CT score exhibited the highest total link strength and were more common in publications after 2020.

Output of publication trend tendencies

Fig. 4 illustrates the yearly trends in publications analyzing CT scoring systems for TBIs. The bibliometric analysis reveals a periodic escalation in the number of publications on this topic, reflecting fluctuations in research focus that may be associated with the increasing incidence of TBIs. However, the overall numbers remain relatively modest. The publication peak is observed in 2023, despite minor decreases during the periods 2016–2017 and 2020–2022.

DISCUSSION

In a previous review study, despite comparisons among the Helsinki, Neuroimaging Radiological Interpretation System (NIRIS), and Marshall parameters, the Helsinki CT score demonstrated superior performance compared to NIRIS. It should be noted that NIRIS was not included in the current review due to its distinct standardization and parameter design. Previous studies have suggested that NIRIS is more broadly applicable than the earlier-developed Marshall and Rotterdam systems. This indirectly supports the notion that the Helsinki CT score may be superior to both the Marshall and Rotterdam scoring systems [12]. The Helsinki CT score serves as a predictive tool for intensive care unit (ICU) patients, facilitating the estimation of appropriate therapeutic interventions based on prognosis and mortality probability [7]. Thus, given its design and detailed parameters, it is unsurprising that the Helsinki CT score outperforms NIRIS in outcome prediction. In one study, the Helsinki CT score outperformed both the Marshall and Rotterdam systems in predicting 6-month outcomes, thereby enhancing the performance of IMPACT (International Mission for Prognosis and Analysis of Clinical Trials in TBI), a well-known prognostic model for long-term outcomes [13]. A comparison of the three CT scoring systems, including their strengths and weaknesses, is briefly described in Table 3.

This review demonstrated that both the Rotterdam and Helsinki parameters are highly competent, as evidenced by similar AUC and Nagelkerke pseudo-R² values. Nevertheless, the Helsinki CT score is considered superior, with a Nagelkerke pseudo-R² of 0.203 and an AUC of 0.744 for 6-month mortality prediction [5,6]. This finding is further supported by studies reporting higher sensitivity and specificity for the Helsinki parameters over a 6-month prediction period. It is important to note that early-stage prediction is preferable; relying solely on detection during the critical phase may favor the Rotterdam parameters for predicting short-term mortality [8]. The studies by Mohammadifard et al. [9] and Goswami et al. [10] that reported the superiority of the Rotterdam CT score compared its sensitivity for mortality detection only against the Marshall CT score, excluding the Helsinki parameter. Rotterdam CT score's advantage over the Marshall score lies in its inclusion of assessments for subarachnoid hemorrhage, intraventricular hemorrhage [14], and epidural hematoma. The distinct CT findings emphasized by each scoring system contribute to their unique clinical applications. To elaborate, Table 4 [6,14] summarizes the identification capabilities of the different scoring systems.

The Helsinki parameters exhibit a notably favorable correlation between intracranial pathology and CT parameter scores. The Helsinki CT score offers a simplified and rapid assessment, even though examiners may sometimes struggle to distinguish whether an “intracerebral hematoma/contusion” is located within the parenchyma or the subarachnoid space [6]. Accurate prognostic information is vital not only for patient management but also for determining the most appropriate course of treatment in life-threatening conditions. In the context of TBI, the implementation of effective measures can significantly improve outcomes. An enhanced prognostic model that is both easy to apply and incorporates critical outcome predictors is essential—rendering the Helsinki parameters technically superior to the Rotterdam system. Based on recent protocols, the Helsinki score assigns distinct values for different types of brain lesions (subdural, intraparenchymal, and epidural hematomas) and evaluates the size and status of the suprasellar cisterns [15].

The Marshall CT score categorizes injuries into six groups, including class V (mass lesions amenable to surgical treatment) and class VI (mass lesions not suitable for surgical treatment). The Rotterdam CT score, similar to the Helsinki CT score, arranges cases by increasing severity [6]. Detailed parameters for each scoring system—Marshall, Rotterdam, and Helsinki—for TBI are presented in Table 5 [5,6].

Bibliometric analysis

The ranking based on the number of publications indicates that the “Marshall CT score” dominates—likely because it was published earlier, allowing for more extensive discussions and citations. Although the Helsinki CT scoring system is claimed to have higher sensitivity, it has not received equivalent research attention. Despite limited analytical results compared to the other systems, the Helsinki score is visually represented by a growing color gradient (from green to yellow) in the co-occurrence analysis from 2015 to the present, indicating increasing attention. Additionally, the Helsinki parameter appears to be the most effective choice for identifying prognosis and mortality in TBI patients. This bibliometric analysis highlights research gaps and suggests that further attention to the superiority of certain CT scoring parameters may enhance diagnostic effectiveness. The findings support the use of risk-stratification models based on tomography data in evolving epidemiological contexts, thereby improving decision-making, resource allocation, and risk communication among healthcare professionals, patients, and caregivers.

Limitations

Studies that include only a subset of the population may not be fully representative, and the limited availability of data can introduce bias into the analysis.

Conclusions

In summary, between 2003 and 2024, publication trends in CT scoring systems for TBI have predominantly focused on the Marshall CT score, Rotterdam CT score, and, subsequently, the Helsinki computed tomography score. The results of the review indicate that the Helsinki CT score offers the most accurate predictive performance for both mortality probability and long-term prognosis. Research related to the Helsinki system has been steadily growing since 2015.

ARTICLE INFORMATION

Author contributions

Conceptualization: AES; Formal analysis: AES, EP; Investigation: all authors. Methodology: all authors; Software: AES, EP; Supervision: RKP, YY; Writing–original draft: AES, EP; Writing–review & editing: all authors. All authors read and approved the final manuscript.

Conflicts of interest

The authors have no conflicts of interest to declare.

Funding

The authors received no financial support for this study.

Data availability

Data sharing is not applicable as no new data were created or analyzed in this study.

Fig. 1.

The 2020 PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) flow diagram.

Fig. 2.

Network visualization from a co-occurrence analysis, comparing publication trends of computed tomography scoring systems as mortality predictor in traumatic brain injury.

Fig. 3.

Overlay visualization from a co-occurrence analysis comparing the efficacy of computed tomography scoring systems as traumatic brain injury mortality predictors.

Fig. 4.

Trends over time in publications analyzing the use of computed tomography scoring systems.

REFERENCES

• 1. Prins M, Greco T, Alexander D, Giza CC. The pathophysiology of traumatic brain injury at a glance. Dis Model Mech 2013;6:1307–15.ArticlePubMedPMCPDF
• 2. GBD 2016 Traumatic Brain Injury and Spinal Cord Injury Collaborators. Global, regional, and national burden of traumatic brain injury and spinal cord injury, 1990-2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet Neurol 2019;18:56–87.ArticlePubMedPMC
• 3. Lefering R. Trauma scoring systems. Curr Opin Crit Care 2012;18:637–40.ArticlePubMed
• 4. Page MJ, McKenzie JE, Bossuyt PM, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 2021;372:n71.ArticlePubMedPMC
• 5. Rodrigues de Souza M, Aparecida Cortes M, Carlos Lucena da Silva G, et al. Evaluation of computed tomography scoring systems in the prediction of short-term mortality in traumatic brain injury patients from a low- to middle-income country. Neurotrauma Rep 2022;3:168–77.ArticlePubMedPMC
• 6. Thelin EP, Nelson DW, Vehvilainen J, et al. Evaluation of novel computerized tomography scoring systems in human traumatic brain injury: an observational, multicenter study. PLoS Med 2017;14:e1002368ArticlePubMedPMC
• 7. Vehvilainen J, Skrifvars M, Reinikainen M, et al. External validation of the NeuroImaging Radiological Interpretation System and Helsinki computed tomography score for mortality prediction in patients with traumatic brain injury treated in the intensive care unit: a Finnish intensive care consortium study. Acta Neurochir (Wien) 2022;164:2709–17.ArticlePubMedPMC
• 8. Biuki NM, Talari HR, Tabatabaei MH, et al. Comparison of the predictive value of the Helsinki, Rotterdam, and Stockholm CT scores in predicting 6-month outcomes in patients with blunt traumatic brain injuries. Chin J Traumatol 2023;26:357–62.ArticlePubMedPMC
• 9. Mohammadifard M, Ghaemi K, Hanif H, Sharifzadeh G, Haghparast M. Marshall and Rotterdam computed tomography scores in predicting early deaths after brain trauma. Eur J Transl Myol 2018;28:7542.ArticlePubMedPMCLink
• 10. Goswami B, Nanda V, Kataria S, Kataria D. Prediction of in-hospital mortality in patients with traumatic brain injury using the Rotterdam and Marshall CT scores: a retrospective study from Western India. Cureus 2023;15:e41548ArticlePubMedPMC
• 11. van Eck NJ, Waltman L. VOSviewer manual. Version 1.6.19. Univeristeit Leiden; 2023. PDF
• 12. Dewangan NK, Sharma A. Validation of the revised NeuroImaging Radiological Interpretation System for acute traumatic brain injury in adult and pediatric population. Indian J Neurotrauma 2021;18:32–7.Article
• 13. Raj R, Siironen J, Skrifvars MB, Hernesniemi J, Kivisaari R. Predicting outcome in traumatic brain injury: development of a novel computerized tomography classification system (Helsinki computerized tomography score). Neurosurgery 2014;75:632–47.ArticlePubMed
• 14. Khormali M, Soleimanipour S, Baigi V, et al. Comparing predictive utility of head computed tomography scan-based scoring systems for traumatic brain injury: a retrospective study. Brain Sci 2023;13:1145.ArticlePubMedPMC
• 15. Yao S, Song J, Li S, et al. Helsinki computed tomography scoring system can independently predict long-term outcome in traumatic brain injury. World Neurosurg 2017;101:528–33.ArticlePubMed

Figure & Data

References

Citations

Citations to this article as recorded by  

• Comparative predictive performance of computed tomography scoring systems in traumatic brain injury: a systematic review, Bayesian comparison, and meta-analysis
  Armin Khavandegar, Zahra Ramezani, Elaheh Khodadoust, Mahgol Sadat Hassan Zadeh Tabatabaei, Tahereh Maleki, Negin Safari Dehnavi, Mario Ganau, Lara Prisco, Ghazaleh Kheiri, Nasim Ramzi, Maral Moafi, Roberto Parisi, Sadra Kheiri, Soroush Mozaffari, Mustafa
  Neurosurgical Review.2026;[Epub]     CrossRef