The early work-up for isolated ligamentous injury of the cervical spine: does computed tomography scan have a role?

BACKGROUND: Helical computed tomography (HCT) scan is the preferred modality for diagnosing fractures of the cervical spine in blunt trauma. We hypothesize that HCT can be used as a screening tool for isolated ligamentous injury (LI) in blunt trauma.

METHODS: A prospective, consecutive series study design was used to include patients that could not have their cervical spine cleared clinically. All patients underwent HCT (occiput-T1) and plain radiographs (PR) with five views of the cervical spine. Patients with clinical or radiographic abnormalities without fracture underwent cervical magnetic resonance imaging (MRI). Demographic and outcome data were collected. The attending radiologist's interpretation was used for clinical management. Three neuroradiologists in a blinded fashion re-reviewed the studies (HCT, PR, and MRI) of the MRI subgroup.

RESULTS: One thousand five hundred seventy-seven patients met the study criteria. Two hundred seventy-eight had 416 cervical spine fractures. PR failed to identify 299 of 416 (72%) cervical spine fractures in 208 of 278 (74.8%) patients. Of the 1,299 (82%) patients who had no fracture, 85 (6.5%) required an MRI. The mean time from admission to MRI was 3 days for the LI subgroup. Of these, 21 of 85 (25%) had LI by MRI. Seven of 21 (33.3%) patients had an abnormal HCT versus 3 of 21 (14.3%) patients who had an abnormal PR. Four of 85 (4.7%) patients had spinal cord injury without radiographic abnormality. One (1.2%) patient required surgical stabilization of LI, as seen on all studies performed (PR, HCT, and MRI). Sensitivities for PR and HCT for LI were 16% and 32%, respectively. Negative predictive values for PR and HCT for LI were 74% and 78%, respectively. Measurements of interrater reliability for MRI, HCT, and PR had kappa values of 0.60, 0.14, and 0.41, respectively.

CONCLUSION: HCT is the most sensitive, specific, and cost-effective modality for screening the cervical spine bony injuries, but it is not an effective modality for screening for cervical LI. MRI is clearly superior to HCT for LI. The indications for MRI include abnormalities on HCT, neurologic deficits, cervical pain or tenderness on examination, or the inability to clear the cervical spine in the obtunded patient. With the current state of the art technology, we have redefined the definition of spinal cord injury without radiographic abnormality to include spinal cord injuries without boney injuries or LI.