1656 - The Diagnostic Precision of Computed Tomography for Traumatic Cervical Spine Injury: An In Vitro Investigation

INTRODUCTION: Computed tomography (CT) is commonly used clinically to diagnose fractures of the cervical spine. Our lab recently performedbiomechanical tests applying dynamic axial compression and lateral bending to in vitro cervical spine specimens, simulating traumatic loads such as thoseexperienced in a rollover car crash [1,2]. These tests afforded us the opportunity to evaluate CT identification of fractures and fracture patterns compared to detailed dissection of the injured spine.METHODS: 35 three-vertebra human cadaver cervical spine specimens (13 C3-5, 7 C4-6, 7 C5-7, 8 C6-T1) potted in PMMA (Harry J. Bosworth Co.) wereused for the biomechanical tests. We applied dynamic axial compression at a rate of 0.5 m/s using a servohydraulic materials test system (model 8874,Instron Canton MA, USA) to specimens with three lateral eccentricities (lateral distance to the center of the spine, low 5% of the spine transverse diameter,middle 50%, high 150%) and two end conditions (19 constrained lateral translation and 16 unconstrained). Pre and post-injury images were acquired with high-resolution CT (Xtreme CT, Scanco Medical, Bru00fcttisellen, Switzerland, resolution 246 u03bcm). One spine clinician diagnosed vertebral fractures based on the CT images, blinded to specimen identification. The same spine surgeon subsequently diagnosed vertebral fractures through detailed dissection. In both injury assessments, each vertebra was divided into 34 anatomical structures, specifically transverse processes (right and left), pedicles (right and left), facet joints (right and left), lateral masses (right and left), laminae (right, middle and left), spinous processes (right, middle and left), vertebral body (9 parts reflecting right, middle, left and anterior, central and posterior regions), endplate (9 parts as in vertebral body) and uncinate processes (right and left). The extent of damage was graded into no damage, partial damage and complete damage groups. Discrepancies between the CT and dissection grading were evaluated in a secondary examination through CT consultation and physical inspection of the vertebra to confirm the presence/absence of injury. We defined a fracture found in both CT and dissection in the same anatomical structure as u201ccorrectly identified through CTu201d. A fracture identified on CT, but not on dissection nor confirmed after secondary examination was defined as u201cCT false positiveu201d. A fracture found during dissection but missed on CT assessment was categorized as u201cCT false negativeu201d.RESULTS: The incidence of injury to each anatomical structure and the results of the CT and dissection comparison are shown in Table 1.We found that the precision of CT was highest for fractures of the vertebral body (84%) and endplate (65%) in terms of correctly identifying fracture. On the other hand, CT accuracy was lowest in fractures of the lateral mass (25%) and pedicle (25%). The highest rate of CT false positive identification occurredfor the uncinate process and lateral mass.DISCUSSION: Traumatic injuries of the spine and spinal cord are common and potentially devastating lesions. Multi detector computed tomography is therecommended primary imaging modality in blunt spinal trauma patients [3,4]. CT has the high sensitivity (93.7%) for detecting cervical spine fractures byblunt trauma, but it is not perfect [5]. Although clinicians fear missing the occult spine fracture, there are few in vitro studies quantifying CT sensitivity for diagnosing spinal column injuries. In one study, axial CT detected only 54% of dislocations and subluxations in trauma victims [6]. Makino reported thedifference between postmortem CT and autopsy of 42 cervical spine injury cases and found that the percentage of CT-detected injuries that were missed at autopsy (35.0%, 14/40) was lower than the percentage of autopsy-detected injuries that were missed with CT (67.5%, 54/80) regarding intervertebral disc injury [7]. Stabler examined 10 cadavers with 28 posttraumatic lesions and found that two fractures were missed at the initial MR imaging reading [8]. These studies indicate the possibility of missing occult spine fractures on CT or MRI is larger than we thought. In this study, we clarified the discrepancy between CT and detailed dissection in an axial compression and lateral bending spine injury model. It represents a highly injured group but clinically relevant to cases of severe trauma, such as occurring in a motor vehicle crash.We focused on which part of the vertebral anatomy was most likely to Best, missed on CT. The discrepancy of CT and dissection injury assessment indicates the possibility that fractures of the lateral mass and the pedicle could be missed in CT diagnosis. The lateral mass and pedicle are important structures of the cervical spine because spine surgeons usually place implants in them to stabilize the spine [9]. In this axial compression lateral bending cervical spine fracture mode common to rollover accidents, and perhaps other loading modes, care should be taken in diagnosing lateral mass and pedicle fractures through CT, particularly if subsequent surgery will utilize this anatomy for implant stabilization.LIMITATION: It is possible to miss fracture in both CT and dissection, but it is rare due to the CT scan consultation and physical inspection.SIGNIFICANCE/CLINICAL RELEVANCE: Care should be taken in diagnosing lateral mass and pedicle fractures through CT of blunt spinal trauma patients.REFERENCES: [1] Van Toen C, et al. J. Biomech., vol.47, no.5, 1164-1172, 2013[2] Van Toen C, et al. Eur. Spine J., vol.24, no.1, 136u2013147 2015[3] Van Goethem JW, et al. Eur. Radiol., vol.15 no.3, 582-90, 2005[4] Bailitz J, et al. J. Trauma., vol.66, no.6, 1605-9, 2009[5] Mushahid R, et al. Injury, vol.44, no.11, 2013[6] Woodring JH, et al. J. Trauma, vol.33, no.5, 698u2013708, 1992[7] Makino Y, et al. Forensic Science International, vol.281, 44-51,2017[8] Stabler A, et al. Radiol., vol.221, no.2, 2001[9] Jones EL, et al. Spine, vol.22, no.9, 977-982, 1997

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.