Journal of Neurotrauma
XX:1–50
Mary Ann Liebert, Inc.
DOI: 10.1089/neu.2023.0553
Open camera or QR reader and
scan code to access this article
and other resources online.
REVIEW
Imaging Findings in Acute Traumatic Brain Injury:
a National Institute of Neurological Disorders
and Stroke Common Data Element-Based Pictorial
Review and Analysis of Over 4000 Admission Brain
Computed Tomography Scans from the Collaborative
European NeuroTrauma Effectiveness Research
in Traumatic Brain Injury (CENTER-TBI) Study
Thijs Vande Vyvere,1,2,*,** Dana Pisică,3,4,** Guido Wilms,5 Lene Claes,6 Pieter Van Dyck,1,2 Annemiek Snoeckx,1,2
Luc van den Hauwe,1 Pim Pullens,7 Jan Verheyden,6 Max Wintermark,8 Sven Dekeyzer,1,9
Christine L. Mac Donald,10,11 Andrew I.R. Maas,12,13,*** and Paul M. Parizel14,***;
and The CENTER-TBI Participants and Investigators****
Abstract
In 2010, the National Institute of Neurological Disorders and Stroke (NINDS) created a set of common data
elements (CDEs) to help standardize the assessment and reporting of imaging findings in traumatic brain
injury (TBI). However, as opposed to other standardized radiology reporting systems, a visual overview and
data to support the proposed standardized lexicon are lacking. We used over 4000 admission computed
tomography (CT) scans of patients with TBI from the Collaborative European NeuroTrauma Effectiveness
Research in Traumatic Brain Injury (CENTER-TBI) study to develop an extensive pictorial overview of the
NINDS TBI CDEs, with visual examples and background information on individual pathoanatomical lesion
1
Department of Radiology, 12Department of Neurosurgery, Antwerp University Hospital, Antwerp, Belgium.
2
Department of Molecular Imaging and Radiology (MIRA), 13Department of Translational Neuroscience, Faculty of Medicine and Health Science, University of Antwerp,
Antwerp, Belgium.
3
Department of Neurosurgery, 4Department of Public Health, Erasmus MC - University Medical Center Rotterdam, Rotterdam, the Netherlands.
5
Department of Radiology, University Hospitals Leuven, Leuven, Belgium.
6
icometrix, Research and Development, Leuven, Belgium
7
Department of Imaging, University Hospital Ghent; IBITech/MEDISIP, Engineering and Architecture, Ghent University; Ghent Institute for Functional and Metabolic
Imaging, Ghent University, Belgium.
8
Department of Neuroradiology, University of Texas MD Anderson Center, Houston, Texas, USA.
9
Department of Radiology, University Hospital Ghent, Belgium.
10
Department of Neurological Surgery, School of Medicine, Harborview Medical Center, Seattle, Washington, USA.
11
Department of Neurological Surgery, School of Medicine, University of Washington, Seattle, Washington, USA.
14
Department of Radiology, Royal Perth Hospital (RPH) and University of Western Australia (UWA), Perth, Australia; Western Australia National Imaging Facility (WA NIF)
node, Australia.
**Shared first authors.
***Shared last authors.
****The CENTER-TBI Participants and Investigators may be found at the end of this article.
*Address correspondence to: Thijs Vande Vyvere, PhD, Department of Radiology, Antwerp University Hospital, Drie Eikenstraat 655, Edegem 2650, Belgium E-mail:
thijs.vandevyvere@uza.be
ª Thijs Vande Vyvere et. al 2024; Published by Mary Ann Liebert, Inc. This Open Access article is distributed under the terms of the Creative Commons License (CC-BY)
(http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited.
1
,2 VANDE VYVERE ET AL.
types, up to the level of supplemental and emerging information (e.g., location and estimated volumes). We
documented the frequency of lesion occurrence, aiming to quantify the relative importance of different
CDEs for characterizing TBI, and performed a critical appraisal of our experience with the intent to inform
updating of the CDEs. In addition, we investigated the co-occurrence and clustering of lesion types and the
distribution of six CT classification systems. The median age of the 4087 patients in our dataset was 50 years
(interquartile range, 29-66; range, 0-96), including 238 patients under 18 years old (5.8%). Traumatic sub-
arachnoid hemorrhage (45.3%), skull fractures (37.4%), contusions (31.3%), and acute subdural hematoma
(28.9%) were the most frequently occurring CT findings in acute TBI. The ranking of these lesions was the
same in patients with mild TBI (baseline Glasgow Coma Scale [GCS] score 13-15) compared with those with
moderate-severe TBI (baseline GCS score 3-12), but the frequency of occurrence was up to three times
higher in moderate-severe TBI. In most TBI patients with CT abnormalities, there was co-occurrence and
clustering of different lesion types, with significant differences between mild and moderate-severe TBI
patients. More specifically, lesion patterns were more complex in moderate-severe TBI patients, with
more co-existing lesions and more frequent signs of mass effect. These patients also had higher and
more heterogeneous CT score distributions, associated with worse predicted outcomes. The critical app-
raisal of the NINDS CDEs was highly positive, but revealed that full assessment can be time consuming,
that some CDEs had very low frequencies, and identified a few redundancies and ambiguity in some def-
initions. Whilst primarily developed for research, implementation of CDE templates for use in clinical prac-
tice is advocated, but this will require development of an abbreviated version. In conclusion, with this study,
we provide an educational resource for clinicians and researchers to help assess, characterize, and report
the vast and complex spectrum of imaging findings in patients with TBI. Our data provides a comprehensive
overview of the contemporary landscape of TBI imaging pathology in Europe, and the findings can serve as
empirical evidence for updating the current NINDS radiologic CDEs to version 3.0.
Keywords: neuroimaging; NIH/NINDS Common Data Elements; structured reporting; traumatic brain injury
Introduction resonance imaging (MRI), however, is more sensitive in
Patients who suffer a traumatic brain injury (TBI) are ini- detecting subtle hemorrhagic lesions such as micro-
tially assessed clinically using the Glasgow Coma Scale bleeds, small traces of subarachnoid hemorrhage, as
(GCS), which broadly categorizes them into ‘‘mild well as certain non-hemorrhagic abnormalities such as
TBI’’ (GCS score 13-15), ‘‘moderate TBI’’ (GCS score edema, ischemia, and distinct forms of axonal injuries
9-12) or ‘‘severe TBI’’ (GCS score <9; Supplementary and contusions.3 As a consequence, it can better reveal
Table S1).1 Despite its utility in the clinical and research the full lesion burden of TBI.
setting,2 this classification is unidimensional and rela- Depending on the magnitude and direction of external
tively crude. The terms mild, moderate, and severe TBI forces to the head, a vast spectrum of findings can be
may lead to treatment bias (e.g., nihilism in severe TBI encountered on both standard and advanced neuro-
and disregard for the adverse long-term consequences imaging.4–9 Individual findings may vary according to
of mild TBI). Importantly, patients and their families location (e.g., temporal, frontal), anatomical compart-
object to the use of these terms. A more refined and ment (e.g., intra-axial, extra-axial), pattern (focal vs. dif-
multi-dimensional approach to characterization of TBI fuse), temporality (primary vs. secondary injury) and
is required. extent (e.g., small vs. large) and may be solitary or coex-
Neuroimaging studies play a pivotal role in the diagno- ist in various combinations with other findings. In the
sis, management, and clinical outcome prediction of clinical setting, the interpretation of neuroimaging stud-
brain-injured patients, and can contribute to a better ies is predominantly visual, qualitative, and subjective.
characterization of TBI.3 In the acute phase post-injury, Findings are reported in narrative form, with different
non-contrast computed tomography (CT) is the preferred physicians often using different terminology for the
imaging modality in adults, primarily due to its cost- same pathoanatomic lesion type (e.g., extradural col-
effectiveness, speed, widespread accessibility, and sensi- lection, epidural hematoma, extradural hematoma, etc.).
tivity in detecting a broad spectrum of injuries affecting Substantial observer differences have been reported,
the skull, brain, and blood vessels.3 CT is eminently even between expert neuroradiologists.10,11 Such report-
suited to identify pathological conditions that may req- ing inconsistencies are then perpetuated and potentially
uire surgery or other urgent interventions, such as enhanced when transferring imaging reports in study
intracranial pressure (ICP)–lowering therapies. Magnetic databases for research purposes, with negative
,IMAGING FINDINGS IN ACUTE TBI 3
implications for data sharing, harmonization, aggrega- entities) across TBI severities (mild TBI, defined by base-
tion, and interpretation across studies. line GCS score 13-15, and moderate-severe TBI defined
To address these issues, a multi-disciplinary task by baseline GCS score 3-12). This empirical evidence
force, led by the National Institutes of Health (NIH)- will provide both an educational resource for clinicians
National Institute of Neurological Disorders and Stroke and researchers, and a comprehensive overview of the
(NINDS), created in 2010 a set of common data ele- neuroimaging case-mix of contemporary TBI patients
ments (CDEs) with practical operational definitions in the European landscape. Further, it provides an empir-
and standardized terminology for reporting imaging ical evidence base to inform updating of the CDEs to
findings in TBI.12,13 These data elements were then version 3.0.
updated to version 2 (v.2) in 2013.14 A structured inter-
pretation scheme was proposed, with an outside-in ap-
proach (i.e., from the skull inward) of inspecting and Methods
reporting findings. Information on each individual le- Study population
sion or encountered abnormality can be reported by The CENTER-TBI study was a prospective, longitudinal,
tiers of increasing detail and complexity: 1) core infor- observational study (ClinicalTrials.gov, NCT02210221),
mation, relating to whether the lesion or abnormality in conducted between December 2014 and December 2017
question is present, absent or indeterminate; 2) in over 60 centers across Europe and Israel.21 The study
supplemental-highly recommended information, relat- recruited patients within 24 h of injury, with a clinical
ing to extent and anatomic location; and 3) emerging in- diagnosis of TBI, a clinical indication for CT imaging
formation, relating to in-depth measurements of based on the judgment of the treating team, and no severe
specific features, often performed with computer- pre-existing neurologic disorders that could confound
aided image analysis.12,13 We note that, in contrast to outcome assessment.21 The majority of participating
the clinical TBI CDEs, these designations relate to the centers were referral centers for neurotrauma. Conse-
detail of scoring, and not to the relative importance of quently, the population studied may not be readily gen-
radiologic CDEs. eralizable to community hospital settings. Further,
Standardized reporting of neuroimaging findings, pediatric patients are underrepresented, as most centers
using the CDE recommendations, substantially mini- mainly focused on adult TBI.
mizes inter- and intra-observer variability in research set-
tings, for both CT and MRI.15,16 Moreover, with adequate
training by experienced neuroradiologists, non-physician Central scan interpretation
researchers (e.g., neuropsychologists, neuroscientists, and reporting process
etc.) can also reliably interpret and report CDEs.16 As Interpretable admission CT datasets were forwarded to
such, TBI CDEs have a great potential to increase neuro- a neuroimaging repository and centrally assessed by
imaging reporting quality, in both clinical and research trained investigators using NINDS CDE-based standard-
settings. However, the CDE recommendations were pub- ized templates, under the supervision of an expert neuro-
lished strictly as a written guide, without the support of radiologist. The central review methodology has been
illustrated cases and data. This is in contrast to other described in detail elsewhere.16 The pathoanatomic
imaging classification systems, such as the widely used lesion types assessed are listed in Table 1. Apart from
BI-RADS (Breast Imaging-Reporting and Data System), the NINDS CDEs, the CENTER-TBI CDE reporting
or PI-RADS (Prostate Imaging Reporting and Data Sys- template included brain herniation and cortical sulcus ef-
tem), which are supported by visual examples to increase facement as distinct lesion types, and an extra item to re-
clarity, maximize adoption, and further decrease report- cord incidental neuroimaging findings, of potential
ing variability.17–20 interest in the context of TBI. Some types of incidental
The large-scale Collaborative European NeuroTrauma findings were preset choices (e.g., old stroke, prior TBI,
Effectiveness Research in Traumatic Brain Injury normal/abnormal prominent ventricles), but reviewers
(CENTER-TBI) study has created one of the largest neu- could also add free text remarks. Free text was mostly
roimaging repositories in the world and implemented used to give extra information regarding interpretation
scan interpretation and reporting according to the decisions or to add information when the readers felt
CDEs. Based on this experience we aim to: 1) present that certain options were lacking in the standard CDE
an extensive pictorial overview of TBI neuroimaging templates. As part of the structured reporting in
CDEs, with visual examples and background informa- CENTER-TBI, we also graded the admission CTs
tion; 2) report the frequencies of core, supplemental, according to the Rotterdam, Marshall and Helsinki clas-
and emerging CDEs of observed findings on more sification systems, and the Fisher, Morris-Marshall, and
than 4000 admission CTs of patients with acute TBI; Greene CT grading systems for quantification of trau-
and 3) explore clustering of core CDEs (pathoanatomic matic subarachnoid hemorrhage (tSAH).
, 4 VANDE VYVERE ET AL.
Table 1. Occurrence of Pathoanatomic Lesion Types on Admission CT in CENTER-TBI
All TBIs Mild TBI* Moderate-severe TBI* Classification
Pathoanatomic lesion type NIH/NINDS Primary/ Focal/
(TBI CDE core information) CDE ID N = 4087 N = 2744 N = 1193 v2 p value secondary diffuse
Skull fracture (%) C02463 1529 (37.4) 728 (26.5) 731 (61.3) < 0.001 Primary Focal/diffuse
Epidural hematoma (%) C02427 463 (11.3) 228 (8.3) 214 (17.9) < 0.001 Primary Focal
Extraaxial hematoma (%) C02430 24 (0.6) 8 (0.3) 13 (1.1) 0.003 Primary Focal/diffuse
Subdural hematoma, acute (%) C02472 1183 (28.9) 500 (18.2) 633 (53.1) < 0.001 Primary Focal/diffuse
Subdural hematoma, subacute C02476 22 (0.5) 16 (0.6) 5 (0.4) 0.681 Primary Focal/diffuse
or chronic (%)
Subdural hematoma, mixed C02480 86 (2.1) 43 (1.6) 39 (3.3) 0.001 Primary Focal/diffuse
density (%)
Traumatic subarachnoid C02469 1852 (45.3) 840 (30.6) 927 (77.7) < 0.001 Primary Focal/diffuse
hemorrhage (%)
Vascular dissection (%) C02489 2 (0.0) 1 (0.0) 0 (0.0) 1.000 Primary Focal
Traumatic aneurysm (%) C02484 1 (0.0) 1 (0.0) 0 (0.0) 1.000 Primary Focal
Venous sinus injury (%) C02491 0 (0.0) 0 (0.0) 0 (0.0) NA Primary Focal
Midline shift (%) C02455 470 (11.5) 107 (3.9) 341 (28.6) < 0.001 Secondary Focal
Cisternal compression (%) C02410 652 (16.0) 130 (4.7) 485 (40.7) < 0.001 Secondary Focal
Ventricular shift/effacement (%) C02435 591 (14.5) 129 (4.7) 431 (36.1) < 0.001 Secondary Focal
Contusion (%) C02414 1280 (31.3) 540 (19.7) 677 (56.7) < 0.001 Primary Focal/diffuse
Intracerebral hemorrhage (%) C02440 126 (3.1) 39 (1.4) 80 (6.7) < 0.001 Primary Focal
Intraventricular hemorrhage (%) C02446 491 (12.0) 121 (4.4) 347 (29.1) < 0.001 Primary/ Focal
secondary
Diffuse and traumatic axonal C02420/21 336 (8.2) 104 (3.8) 216 (18.1) < 0.001 Primary Focal/diffuse
injury (%)
Penetrating injury (%) C02459 18 (0.4) 5 (0.2) 12 (1.0) 0.001 Primary Focal/diffuse
Cervicomedullary/brainstem C02407 5 (0.1) 0 (0.0) 5 (0.4) 0.004 Primary/ Focal/diffuse
injury (%) secondary
Cerebral edema (%) C02424 56 (1.4) 4 (0.1) 49 (4.1) < 0.001 Secondary Focal/diffuse
Brain swelling/hyperemia (%) C02404 14 (0.3) 0 (0.0) 14 (1.2) < 0.001 Secondary Focal/diffuse
Ischemia (%) C02451 4 (0.1) 0 (0.0) 4 (0.3) 0.013 Secondary Focal/diffuse
Pathoanatomic lesion types currently not in the CDEs but collected extra in CENTER-TBI
Brain herniation (%) 390 (9.5) 85 (3.1) 285 (23.9) < 0.001 Primary/ Focal/diffuse
secondary
Cortical sulcus effacement (%) 459 (11.2) 113 (4.1) 319 (26.7) < 0.001 Secondary Focal/diffuse
Incidental findings (%) 622 (15.2) 451 (16.4) 155 (13.0) 0.007
The data contains no missing values.
*Mild TBI was defined by baseline Glasgow Coma Scale (GCS) score 13-15, regardless of presence of CT abnormalities. Moderate-severe TBI was
defined by baseline GCS score 3-12. The baseline GCS score used for classification was a derived variable calculated centrally for baseline risk adjustment.
It represents the post-stabilization GCS value, which was imputed when absent using IMPACT methodology: work back in time towards pre-hospital value
until a non-missing value is found.
CDE, common data element; CENTER-TBI, Collaborative European NeuroTrauma Effectiveness Research in Traumatic Brain Injury; CT, computed tomog-
raphy; NA, not applicable; NIH/NINDS, National Institutes of Health/National Institute of Neurological Disorders and Stroke; TBI, traumatic brain injury.
For pathoanatomic lesion types with size as a sup- by baseline GCS score 13-15, regardless of presence of
plemental CDE, the recommendations state reporting CT abnormalities) and moderate-severe TBI (defined by
volume or length, width, and maximal thickness. In baseline GCS score 3-12).
CENTER-TBI, these lesions were measured by volume,
calculated using the width · depth · length · 0.5 formula Occurrence of pathoanatomic lesion types. The abso-
(ABC/2; Supplementary Fig. S1). For subdural hemato- lute and relative frequencies of pathoanatomic lesion
mas in particular, the width was chosen on an axial types included in the TBI CDEs reporting scheme are
slice of average width, and not on the slice of largest reported for admission CTs of all patients and separately
width, to avoid overestimation of non-ellipsoid volumes. for patients with mild and moderate-severe TBI. Addi-
Data inconsistencies, such as discrepancies in timing and tionally, we reported the occurrence of pathoanatomic
naming of the scans, were addressed through diligent data lesion types for patients at the extremes of the GCS-
curation efforts. based severity spectrum: the subgroup of patients with
baseline GCS score 3 and the subgroup with baseline
Statistical analysis GCS score 15.
Characteristics of the study sample. Demographic and
injury characteristics are reported for the entire study Individual pathoanatomic lesion types. We then elab-
sample and separately for patients with mild (defined orated on each individual pathoanatomic lesion type,
