Background: There is a subset of blunt trauma patients that present with symptoms suspicious for cervical spine injury or with unreliable clinical exams whose initial plain radiographs or cervical computed tomography (CT) scan are negative. Uncertainty remains, however, because no gold standard has been established for definitively clearing the cervical spine of injury in this patient cohort. Individual studies have detailed the use of magnetic resonance imaging (MRI) in this patient population without conclusive results.
Methods: Comprehensive database searches were conducted for prospective or retrospective diagnostic studies of blunt trauma patients who were entered into a cervical spine clearance protocol that included MRI. Inclusion criteria were minimum 30 patients with clinically suspicious or unevaluatable cervical spines, clinical follow-up as the gold standard, data reported to allow the collection of true positives, true negatives, false positives, and false negatives, MRI obtained within 72 hours of injury, and plain radiographs that disclosed nothing abnormal of the cervical spine with or without a CT scan that disclosed nothing abnormal. Log odds meta-analysis of the sensitivity, specificity, positive, and negative predictive value of MRI was performed.
Results: Five Level I diagnostic protocols, enrolling 464 patients receiving MRI, were included. There were zero false negatives in the five studies resulting in a negative predictive value of 100%. Log odds meta-analysis produced a 94.2% positive predictive value (95% confidence interval [CI] 75.0, 989), 97.2% sensitivity (95% CI 89.5, 99.3), and 98.5% specificity (95% CI 91.8, 99.7). Ninety-seven (97 of 464, 20.9%) patients had abnormalities identified by MRI that were not identified by plain radiographs with or without CT.
Conclusion: A magnetic resonance image that did not disclose anything abnormal can conclusively exclude cervical spine injury and is established as a gold standard for clearing the cervical spine in a clinically suspicious or unevaluatable blunt trauma patient. An accurate number of false positive MRI scans cannot be determined.