Written by Ketan Patel and Kyle Rankin
Spoon Feed
When eye exam is limited by trauma, we rely on orbital CT. While specificity is 94%, sensitivity is only 75% for open globe injuries (OGI), making surgical consultation necessary when suspicion remains.
Eye spy with my CT eye: sharp on specificity, blunt on sensitivity
Eye trauma is common in ED settings, and OGI is a true surgical emergency—missed cases can lead to permanent vision loss or other severe complications. CT is widely used because bedside ophthalmologic exams are often limited or impossible in trauma patients.
This paper is a prospectively registered systematic review and meta-analysis that pooled data from 8 studies, 827 patients (830 injured globes)—339 confirmed OGIs—to assess how well CT detects OGIs overall and by specific imaging signs. Across studies, the pooled sensitivity of CT for open globe injury was ~75% (95%CI 0.69–0.79). Specificity was ~94% (95%CI 0.91–0.96), meaning a positive CT is very likely true. Among individual signs, globe contour change/scleral irregularity had the best combination of sensitivity (~71%) and very high specificity (~98%); other signs (anterior chamber depth changes, vitreous hemorrhage, lens abnormality, intraocular air) were highly specific but not sensitive.
Limitations include the relatively small number of studies and heterogeneity in imaging protocols, need for specialized radiologists, and varying definitions of CT signs, which limits generalizability across settings and scanner technologies.
How does this change my practice?
These types of injuries are not uncommon in our trauma center. We’ll often have facial trauma present with significant soft tissue swelling that limits ocular exams, making direct visualization difficult. I will continue to rely on CT (understanding its limitations), but also draw on other techniques such as ultrasound as well, and bring in ophthalmology in cases with ongoing suspicion for this high-stakes diagnosis.
Source
Diagnostic test accuracy of CT for open globe injury: A systematic review and meta-analysis of overall impression and individual imaging signs. Injury. 2026 Jan 10;57(3):113041. doi: 10.1016/j.injury.2026.113041. Epub ahead of print. PMID: 41534455.
View JournalFeed Critical Appraisal
Critical Appraisal
Study Identification
Background
Study Question
Study Design & Conduct
Prospective / Retrospective: Retrospective
Multicenter: Yes
Unit of Allocation: Not applicable
Unit of Analysis: Globes
Randomization Method: Not applicable
Allocation Concealment: Not applicable
Blinding: CT readers were blinded to clinical findings
Follow-up Duration: Not applicable
Population
- Patients with suspected ocular trauma
- Underwent CT imaging of the globes
- Conference abstracts
- Non-English studies
- Review articles
Number Enrolled: Not reported
Number Analyzed: 830
Key Baseline Characteristics
Sex: Majority male (72% to 86.5%)
Disease Severity: Not reported
Care Setting Distribution: Trauma settings
Additional Baseline Characteristics
- Exclusion of patients with prior ocular trauma or surgery
Exposures / Interventions
Description: CT imaging of the globes
Definition / Dose: Not applicable
Timing: During trauma assessment
Classification Method: Not reported
Protocolized / Discretionary: Not reported
Description: Ophthalmological examination or surgical findings
Definition: Reference standard for open globe injury
Outcomes & Results
Primary Outcomes
Definition: Sensitivity and specificity for detecting open globe injury
Time Point: Not applicable
Measurement Method: Meta-analysis of included studies
Results: Sensitivity 75%, Specificity 94%
Secondary Outcomes
Definition: Sensitivity and specificity of signs like change in globe contour
Time Point: Not applicable
Measurement Method: Meta-analysis of included studies
Results: Sensitivity 71%, Specificity 98% for change in globe contour
Risk of Bias
Risk of Bias - QUADAS-2
- Patient Selection (Some concerns): Some studies had stringent exclusion criteria, potentially biasing results.
- Index Test (Low): CT readers were blinded to clinical findings in most studies.
- Reference Standard (Some concerns): Potential for confirmation bias as clinical findings may have been influenced by CT results.
- Flow and Timing (Low): Consistent application of reference standards across studies.
Transparency
COI Statement Present: TRUE
