A retrospective comparison of CT and MRI in detecting pediatric cervical spine injury

Abstract

Although clinical criteria have been applied in the assessment of pediatric cervical spine trauma, no consensus has been established when imaging is required. With the increasing prevalence of computed tomography (CT) use in pediatric trauma and the concern for radiation in children, we sought to evaluate magnetic resonance imaging (MRI) and CT in detecting pediatric cervical spine injuries. We retrospectively queried a pediatric trauma database and identified pediatric patients who underwent both CT and MRI studies of the cervical spine and derived the statistical measures of each imaging modality to detect osseous and ligamentous/soft tissue injury. Eighty-four patients were identified with a mean age of 9.0 ± 5.8 years (56% male). Sixteen patients were identified with injury, 12 with soft tissue abnormalities on MRI (nine edema and six ligamentous), and 6 with osseous abnormalities on CTs (six osseous fractures and one discogenic injury). Of the six patients who presented with CT-identified osseous injuries, MRI detected all six fractures as well as an additional compression fracture. Using CT as the standard for osseous injury, MRI had a sensitivity of 100%, specificity of 97%, negative predictive value (NPV) of 75%, and positive predictive value (PPV) of 100%. Using MRI as the standard for soft tissue injury, CT had a sensitivity of 23%, specificity of 100%, NPV of 88%, and PPV of 100%. Further studies are required to investigate the use of MRI to detect osseous injuries.