Cervical Spine CT Can Miss Fractures in American Football Players When Protective Equipment is in Place: A Cadaver Study.

Abstract

Iatrogenic worsening of spinal injury can result in significant harm to American football players and complicate management when equipment is removed in the acute setting by inexperienced personnel. Spine imaging before removal of protective equipment mitigates this risk. There is no consensus regarding the ideal timing of equipment removal or whether current diagnostic imaging modalities are effective to detect such injuries without equipment removal. Prior data suggest that CT is a diagnostic modality for this purpose; however, radiologists' accuracy in detecting fractures in the presence of protective equipment requires additional study. (1) Does the introduction of American football equipment result in a significant reduction in sensitivity for cervical spine fracture detection? (2) Absent specific guidance as to parameters needed to establish diagnostic quality, can a radiologist determine whether such CTs are of diagnostic quality by subjectively relying on the ability to identify anatomic landmarks? A pendulum device was engineered to deliver a measured axial load to the crown of cadavers to produce a variety of cervical spine fractures in 13 cadaver specimens. The cadavers were then imaged using a standardized CT protocol first without and then with protective football equipment. The images were presented to three board-certified, fellowship-trained radiologists to (1) identify all fractures from the occiput to T1 and (2) subjectively assess the diagnostic quality of the resulting CTs. A sensitivity analysis was performed against a reference standard of fractures produced by the consensus of all radiologists in this study to determine whether there was any reduction in radiologists' ability to detect fractures once football equipment was in place. We found that CT scans obtained with football protective equipment in place resulted in lower sensitivity in diagnosing cervical spine injuries than CT scans obtained without pads. A total of 42 fractures were identified in the reference standard, allowing for a combined 126 possible fracture identifications between the three interpreters. Without football equipment, a combined 98 fractures were identified, whereas a combined 65 fractures were identified once the equipment was introduced. Overall, the sensitivity was reduced by 26% (52% [65 of 126] versus 78% [98 of 126] [95% CI 14.8% to 37.5%]; p < 0.001). Of the 78 total CT series imaged with football equipment, 92% (72 of 78) were considered to be of diagnostic quality. However, the study radiologists failed to identify 50% (53 of 105) of fractures present in those CT images. The sensitivity of cervical spine fracture detection using CT is diminished in the setting of protective American football equipment. Future studies in live subjects with cervical spine fracture may be warranted to support these conclusions. These findings contradict previous studies that determined CT to be a diagnostic imaging modality to image the cervical spine through equipment. Although the interpreting radiologists consistently deemed CTs performed in the presence of helmets and shoulder pads to have subjectively diagnostic quality, numerous fractures that had been detected in the absence of equipment were missed in their presence. Furthermore, this study established that subjective approval of the appearance of an imaging study based on the ability to recognize anatomic landmarks is insufficient to reliably determine the diagnostic quality of a CT study.