Clinical role of radiography for thoracic spine fractures in daily practice in the MDCT era: a retrospective review of 255 trauma patients

Transcription

1 Jpn J Radiol (2012) 30: DOI /s ORIGINAL ARTICLE Clinical role of radiography for thoracic spine fractures in daily practice in the MDCT era: a retrospective review of 255 trauma patients Tsutomu Inaoka Kenjirou Ohashi Georges Y. El-Khoury Harnoor Singh Kevin S. Berbaum Received: 15 March 2012 / Accepted: 7 June 2012 / Published online: 6 July 2012 Ó Japan Radiological Society 2012 Abstract Purpose To retrospectively assess the diagnostic efficacy of radiography in detecting vertebral body fractures of the thoracic spine compared with MDCT, to assess the confounding factors reducing the diagnostic efficacy, and to investigate the outcomes of radiographically overlooked patients. Materials and methods Two hundred fifty-five patients suspected of thoracic spine fractures were enrolled. We assessed the diagnostic efficacy of radiography for the patients sub-grouped based on five confounding factors: chest abnormalities, head injuries, cervical spine fractures, upper extremity injuries, and age of 65 years or older. We investigated the outcomes of radiographically overlooked patients. Results Three hundred fifty-one vertebral body fractures were detected. The per-fracture sensitivities and specificities were 55 % and 94 % for vertebral body fractures and 41 and 99 % for unstable fractures. In patients with upper extremity injuries or aged 65 years or older, radiography was less sensitive in detecting the unstable fractures (P \ 0.05). Nineteen patients were overlooked by radiography; two had T. Inaoka (&) Department of Radiology, Toho University Sakura Medical Center, Shimoshizu, Sakura , Japan tsutomu.inaoka@med.toho-u.ac.jp T. Inaoka K. Ohashi G. Y. El-Khoury K. S. Berbaum Department of Radiology, University of Iowa Carver College of Medicine, Iowa, USA H. Singh Department of Family Practice, University of Iowa Carver College of Medicine, Iowa, USA neurological deficits and needed surgical fixation; 15 with no neurological deficit were conservatively treated with uneventful outcomes. Conclusion Radiography had low sensitivity but high specificity. In daily practice, primary use of MDCT is beneficial for patients with neurological deficit or upper extremity injuries or elderly patients. Keywords Thoracic spine Fracture Trauma Radiography MDCT Introduction Fractures of the thoracic spine have been discussed in the context of fractures of the lumbar spine [1 5]. However, the thoracic spine, particularly T1 T10, is significantly different from the lumbar spine, since it is supported by the ribcage with rigid bony and ligamentous structures. Actually, the resistance to violence of the thoracic spine is much stronger than that of the lumbar spine, whereas the spinal cord in the thoracic segment is less resistant to injury because of its sparse blood supply [2, 3, 5]. If a fracture occurs in the thoracic segment, therefore, it should be considered to be a serious injury. Some investigators had previously addressed the importance of paying special attention to fractures of the thoracic spine because of the high risk of serious cord injury [2, 3, 5 7]. In trauma centers, integration of whole-body computed tomography (CT) for early trauma care can increase the probability of survival in patients with polytrauma [8, 9]. In addition, reconstructed images from whole-body CT are enough to assess associated spinal injuries [10 12]. In daily practice, however, radiographic examinations play a major role in screening the patients suspected of having thoracic

2 618 Jpn J Radiol (2012) 30: spine fractures in addition to physical examinations. However, the thoracic spine abnormalities are often obscured on radiographs by normal or abnormal silhouettes of the surrounding structures such as the shoulder, ribs, diaphragm, mediastinum, and lung. Furthermore, cardiovascular or pulmonary injuries may mask the clinical significance of thoracic spine fractures for patient management [2, 7, 11]. As a result, substantial numbers of fractures in the thoracic spine are overlooked compared to those in the cervical or lumbar spine [5, 7, 8, 10]. Currently, multi-detector row CT (MDCT) has been increasingly used for skeletal trauma patients [13]. Recently, therefore, overutilization of CT has become a critical issue, since it causes an increase in medical costs and the administration of unnecessary radiation to the patients [14, 15]. Fractures of the cervical and lumbar spine have been studied [14 17]. To our knowledge, however, there are only a few studies that have evaluated the diagnostic efficacy of radiography for patients with fractures of the thoracic spine in the era of MDCT [10, 12]. Furthermore, there is no study assessing the confounding factors that reduce the diagnostic efficacy of radiography and that may thus influence patient management. The purposes of our study were to retrospectively assess the diagnostic efficacy of radiography in detecting vertebral body fractures of the thoracic spine compared with MDCT, to assess the confounding factors reducing the diagnostic efficacy of radiography, and to investigate the outcomes of radiographically overlooked patients with vertebral body fractures of the thoracic spine. Materials and methods Study population Institutional review board approval was obtained, and the study was conducted in compliance with the Health Insurance Portability and Accountability Act. The need for informed consent was waived. There were 836 consecutive patients who were suspected of having thoracic spine fractures during a 6-year period (2002 through 2008). A total of 255 patients met the following inclusion criteria: (1) patients who had a history of trauma (except for gunshot or penetrating injuries); (2) patients who came to the hospital less than 1 week after the traumatic event; (3) patients imaged by both radiography (anteroposterior and lateral views) and MDCT of the thoracic spine within an interval of 48 h; (4) patients with no previous spinal surgery. In this study, we enrolled the patients with minor trauma through multiple organ injuries to assess the confounding factors that may reduce the diagnostic performance of radiography of the thoracic spine. Imaging techniques Radiographic studies of the thoracic spine were performed with direct digital radiography (Fuji CR900; Fuji Medical Systems, Stamford, CT; Kodak DirectView CR900 system; Kodak, Rochester, NY; Siemens FD-X, Siemens Medical Solutions, Malvern, PA). Anterior posterior and lateral radiographs of the thoracic spine were obtained for all 255 patients. An additional swimmer s view was obtained in 109 patients (43 %). Four types of MDCT scanners were used: a four-detector row CT (Aquilion; Toshiba Medical Systems, Tustin, CA), a six-detector row CT (Emotion 6; Siemens, Malvern, PA), a 16-detector row CT (Sensation 16; Siemens, Malvern, PA), and a 64-detector row CT (Sensation 64; Siemens, Malvern, PA). A total of 1,887 thoracic vertebrae were studied in 255 studies (7.4 vertebrae on average per study). The scan range was determined based on the patients symptoms and the results of physical examination. Classification of thoracic spine fractures Radiographic and MDCT images of the thoracic spine were assessed using eight vertebral body fracture categories [1, 2]: wedge-compression fracture, two-column burst fracture, three-column burst fracture, anterior fracture dislocation, posterior fracture dislocation, Chance-type fracture, anterior dislocation without fracture, and posterior dislocation without fracture. We defined compression and two-column burst fractures as stable fractures and the other fractures as unstable fractures. Fracture dislocations or dislocations without fracture involving C7/T1 or T12/L1 were included in this analysis. Image review methods Radiographic and MDCT images were assessed for the presence, location, and types of vertebral body fractures in the thoracic spine independently by two experienced musculoskeletal radiologists. The MDCT reading was followed by a radiography reading in random order at least 6 weeks apart in order to avoid recall bias. Both two readers were blinded to the patients identification and clinical information except the history of acute trauma. Radiographic studies were evaluated on a four-monitor system in our intradepartmental picture-archiving and communication system network (Kodak System 4; Eastman Kodak, Rochester, NY). MDCT images were evaluated on a computer workstation (Vitrea 2, version 3.5; Vital images, Plymouth, MN). The reconstructed slice thickness was less than 2.0 mm. Multiplanar reformatted (MPR) and three-dimensional (3D) volume-rendered (VR) images were interactively reviewed by each reader.

3 Jpn J Radiol (2012) 30: Differentiation of acute from chronic fractures was made based on the presence of associated intravertebral and/or paravertebral hemorrhage or the absence of sclerotic changes at the fracture sites. In addition, disagreements between the two readers on the MDCT interpretation were resolved by consensus. Undetected fractures were determined when two readers both could not detect vertebral body fractures of the thoracic spine on the radiographs. Clinical assessment of patients Medical records including clinical, operative, and radiological reports were reviewed by two investigators. Patients age, gender, mechanism of injury, neurological findings, and associated injuries of the rest of the spine, head, neck, chest, abdomen, pelvis, and extremities were recorded. Neurological assessment was recorded based on the American Spinal Injury Association impairment scale (ASIA) (Grades A E) [18]. Chest abnormalities included clavicle fractures, scapula fractures, sternum fractures, ribs fractures, mediastinal injuries, aortic injuries, lung contusion, and pleural effusion. Abdominal injuries included injuries of the liver, spleen, pancreas, gall bladder, kidney, adrenal glands, and urinary bladder. Pelvic abnormalities included injuries of the intrapelvic organs, sacrum, and pelvic ring. These injuries were confirmed by MDCT and/ or operative findings. Extremity injuries included fractures, dislocations, and ligamentous injuries of the upper or lower extremities, which were confirmed by radiography, MDCT, or MRI. Simple lacerations were excluded. The managements and outcomes were recorded during the follow-up period (mean 189 days, range 0 1,770 days). Statistical analysis To test the diagnostic efficacy of radiography in detecting the vertebral body and unstable fractures of the thoracic spine, we calculated the per-fracture sensitivities and specificities of the two readers and the exact binominal 95 % confidence intervals (CIs) for each fracture using MDCT findings as the reference standard. We selected five confounding factors that might be thought to reduce the radiographic efficacy: chest abnormalities, head injuries, cervical spine fractures, upper extremity injuries, and age of 65 years or older. The threshold of age was determined based on the previous studies of cervical spine fractures in which the elderly population was 65 years and older [19, 20]. Pure neck injuries, abdominal abnormalities, and pelvic abnormalities were excluded because of less common association with thoracic spine fractures. Since lower extremity injuries and lumbar spine fractures were not thought to be related to obtain the optimal radiographic imaging of the thoracic spine or read the images, they were excluded. Regarding neurological deficits, we separately dealt with this factor since the patients with neurological deficits were commonly planned to be surgically treated. We assessed the diagnostic efficacy of radiography separately for each subgroup of patients divided by the presence or absence of the five confounding factors. Chisquare tests were used for the analyses. A statistically significant difference was indicated at the level of a P value \0.05. Results Clinical assessment The demographics of the study population are shown in Table 1. Neurological deficits were shown in 35 patients. Complete (ASIA grade A) and incomplete neurological deficits (ASIA grades B, C, and D) were found in 21 and 14 patients, respectively. The numbers of patients with head injuries, neck injuries, cervical spine injuries, lumbar spine injuries, chest abnormalities, abdominal abnormalities, pelvic abnormalities, and extremity injuries are shown Table 2. Diagnostic efficacy of radiography compared with MDCT Out of 255 patients, 201 patients showed 351 vertebral body fractures of the thoracic spine detected by MDCT. Disagreements regarding the MDCT interpretation by the two readers were seen in 21 patients (8 %, 21/255) and were solved by consensus. Out of the 351 vertebral body fractures, 78 (22 %) were unstable fractures (Fig. 1). The most common type of fracture was compression fracture (204; 58 %), followed by two-column burst fracture (69; 20 %), anterior fracture dislocation (31; 9 %), three-column burst fracture (25; 7 %), posterior fracture dislocation (11; 3 %), and Chancetype fracture (11; 3 %). No pure dislocation without fracture was found. Table 1 Demographics of the 255 patients included in this study Number of patients 255 Age range (mean) 6 90 years (44 years) Male/female 183/72 Mechanism of injury Motor vehicle accident 161 Fall 68 Others 26

4 620 Jpn J Radiol (2012) 30: Table 2 Distribution of associated injuries in 255 patients suspected of having thoracic spine fractures enrolled in this study Location of associated injury Number of patients Head 53 Neck 7 Cervical spine 26 Lumbar spine 53 Chest abnormalities 157 Clavicle 15 Scapula 12 Sternum 4 Ribs 95 Lung contusion 57 Pleural effusion 105 Mediastinum 6 Aorta 2 Abdominal abnormalities 8 Liver 4 Spleen 4 Pancreas 1 Kidney 1 Gastrointestinal tract 2 Adrenal gland 1 Pelvic abnormalities 12 Sacrum 5 Pelvic ring 9 Urinary bladder 1 Upper extremity 34 Lower extremity 12 Regarding diagnostic efficacy of radiography in detecting the vertebral body and unstable fractures of the thoracic spine compared with MDCT, the per-fracture sensitivities and specificities (95 % CIs) of the two readers are shown in Table 3. The results of the diagnostic efficacy of radiography for the detection of the vertebral body and unstable fractures for each group of patients with or without the five confounding factors are shown in Table 4. The numbers of patients with chest abnormalities, head injuries, cervical spine injuries, upper extremity injuries, and an age of 65 years or older were 157, 53, 26, 34, and 39, respectively. Outcomes of patients with vertebral body fractures of the thoracic spine overlooked by radiography Out of the 201 patients (351 vertebral body fractures of the thoracic spine) diagnosed by MDCT, 19 patients (9 %) [25 vertebral body fractures (7 %)] were radiographically read as negative by both readers. Out of the 19 patients, two were serious multitrauma patients and died shortly after admission from multiorgan injuries; therefore, their neurological Fig. 1 Distribution of vertebral body fractures of the thoracic spine. Note: Stable fractures; compression fractures and two-column burst fractures, unstable fractures; three-column burst fractures, Chancetype fractures, and fracture dislocations examinations were not reliable. Two patients presented with complete neurological deficits (ASIA grade A) at admission and underwent surgical thoracic spinal fixation, and the remaining 15 showed no neurological deficit and were conservatively treated with an uneventful recovery. The 25 radiographically overlooked vertebral body fractures consisted of 19 stable fractures (17, compression fractures; 2, two-column burst fractures) and 6 unstable fractures (3, three-column burst fractures; 2, anterior fracture dislocations; 1, posterior fracture dislocation) (Figs. 2, 3). Discussion Prior articles indicated that the diagnostic efficacy of radiography in detecting thoracic spine fractures is much lower than that of CT [8, 10, 12, 16]. In addition, the sensitivity and specificity of recent MDCT are nearly 100 % for thoracic spine fractures [12]. To our knowledge, there was only one article that tested the diagnostic efficacy of radiography in comparison with MDCT [12]. According to this report by Herzog [12], the sensitivity and specificity of radiography were 57 and 86 %, respectively, based on 12 patients, in whom 21 vertebral body fractures of the thoracic spine were detected. Our study showed a similar sensitivity but higher specificity of radiography using two experienced musculoskeletal radiologists who reviewed the studies of 255 trauma patients. For the detection of unstable fractures (three-column burst fractures, Chance-type fractures, and fracture dislocations) of the thoracic spine, radiography was significantly less sensitive for patients with upper extremity injuries and for patients whose age was 65 years or older (P \ 0.05). These results may be explained by the technical difficulty in obtaining optimal radiography among these patient

5 Jpn J Radiol (2012) 30: groups. In addition, concomitant degenerative changes or osteopenia in elderly patients may make the detection of the thoracic spine fractures difficult. Because of the significantly low sensitivity of radiography in these patient Table 3 Diagnostic efficacy of radiography of the thoracic spine in detecting vertebral body fractures and unstable fractures of the thoracic spine compared with MDCT Sensitivity (95 % CI) Specificity (95 % CI) Vertebral body fractures (%) 55 (51 %, 58 %) 94 (93 %, 95 %) Unstable fractures (%) 41 (35 %, 48 %) 99 (99 %, 100 %) Sensitivities and specificities were calculated for each fracture 95 % CI 95 % confidence interval, unstable fractures: three-column burst fractures, Chance-type fractures, and fracture dislocations subgroups, the primary use of MDCT may be beneficial for patients with upper extremity injuries or those aged 65 years or older. Van Beek [7] retrospectively evaluated the radiographic diagnosis (chest and spinal films) of the upper thoracic spine fractures in 23 multiple trauma patients and reported that 22 % of the fractures were missed since they were frequently associated with life-threatening injuries. In our results, 9 % (19/201) of the patients with vertebral body fractures of the thoracic spine were read as negative by radiography. Fractures involving the upper thoracic spine tended to be overlooked on the thoracic spine radiographs, even when swimmer s view was also considered. This can be explained by the fractures being obscured by the shoulder silhouettes [7]. Fractures involving the lower thoracic spine were missed by thoracic spine radiography Table 4 Diagnostic efficacy of radiography in detecting vertebral body and unstable fractures of the thoracic spine in patients with or without each confounding factor Sensitivities and specificities were calculated for each fracture 95 % CI 95 % confidence interval, unstable fractures: three-column burst fractures, Chance-type fractures, and fracture dislocations *1,*2 Significant differences (P \ 0.05) Sensitivity (95 % CI) Specificity (95 % CI) (a) Vertebral body fractures Chest abnormalities Present, % (n = 157) 52 (48 %, 56 %) 95 (94 %, 96 %) Absent, % (n = 98) 61 (55 %, 67 %) 93 (92 %, 95 %) Head injuries Present, % (n = 53) 59 (50 %, 68 %) 94 (92 %, 96 %) Absent, % (n = 202) 54 (50 %, 58 %) 94 (94 %, 95 %) Cervical spine fractures Present, % (n = 26) 57 (47 %, 67 %) 92 (89 %, 95 %) Absent, % (n = 229) 55 (51 %, 58 %) 95 (94 %, 95 %) Upper extremity injuries Present, % (n = 34) 49 (37 %, 60 %) 95 (92 %, 97 %) Absent, % (n = 221) 56 (52 %, 59 %) 94 (93 %, 95 %) 65 years 65 years or older, % (n = 39) 44 (33 %, 55 %) 95 (93 %, 97 %) Less than 65 years, % (n = 216) 56 (52 %, 60 %) 94 (93 %, 95 %) (b) Unstable fractures Chest abnormalities Present, % (n = 157) 41 (34 %, 48 %) 99 (99 %, 100 %) Absent, % (n = 98) 42 (26 %, 59 %) 100 (99 %, 100 %) Head injuries Present, % (n = 53) 33 (20 %, 47 %) 100 (99 %, 100 %) Absent, % (n = 202) 44 (36 %, 51 %) 99 (99 %, 100 %) Cervical spine fractures Present, % (n = 26) 22 (9 %, 40 %) 98 (96 %, 99 %) Absent, % (n = 229) 44 (37 %, 52 %) 99 (99 %, 100 %) Upper extremity injuries Present, % (n = 34) 6 (2 %, 14 %)* (98 %, 100 %) Absent, % (n = 221) 59 (51 %, 67 %)* 1 99 (99 %, 100 %) 65 years 65 years or older, % (n = 39) 9 (19 %, 24 %)* 2 99 (98 %, 100 %) Less than 65 years, % (n = 216) 47 (40 %, 54 %)* 2 99 (99 %, 100 %)

6 622 Jpn J Radiol (2012) 30: more than expected. On the radiographs of the thoracic spine, this region was obscured by the silhouette of the bilateral diaphragm. Lumbar spine radiographs were not considered in this study because we did not expect such a low diagnostic performance of thoracic spine radiography for the lower thoracic segment. With referring lumbar spine radiographs, it might be better. The association of chest Fig. 2 Distribution of 25 vertebral body fractures overlooked by radiography of the thoracic spine. Note: Stable fractures; compression fractures and two-column burst fractures, unstable fractures; threecolumn burst fractures and fracture dislocations. Fracture dislocation was counted as 0.5 at each spine level. For example, one T1/2 fracture dislocation was counted as 0.5 at T1 and 0.5 at T2 for counting purposes abnormalities might be thought to reduce the diagnostic performance of radiography, but there was no statistical significance in detecting fractures of the thoracic spine. So far, when optimal radiography of the thoracic spine is used, we believe that this association of chest abnormalities creates no problems when evaluating fractures of the thoracic spine. Out of the 19 patients with vertebral body fractures of the thoracic spine overlooked by radiography, 2 patients were serious multi-trauma patients who died shortly after admission from multiorgan injuries. Therefore, their thoracic spine fractures did not require further evaluation. Two patients presented with complete neurological deficits (ASIA grade A) at admission and needed further evaluation by MDCT for the operative planning. The remaining 15 patients presented with no neurological deficits at admission, and were conservatively treated and uneventfully recovered. In daily practice, therefore, we believe that radiography of the thoracic spine seems to be adequate for the initial imaging assessment of patients suspected of having thoracic spine injuries unless they present with neurological deficits or a deterioration in neurological status. The limitations of our study are mainly derived from the retrospective nature of the study design. The indications of MDCT following radiography were not standardized, and the scanning levels of MDCT for the thoracic spine were guided by clinical findings. Next, the clinical values of Fig. 3 A 67-year-old male with T1/2 anterior fracture dislocation (unstable fracture) from motor vehicle accident, which was radiographically overlooked. He presented with complete neurological deficit (ASIA grade A). Radiography of the thoracic spine: a anteroposterior view, b lateral view, c swimmer s view, and d sagittal reconstruction MDCT image. There appears to be superior mediastinal widening on anteroposterior view and anterior displacement of the trachea on lateral view, which could be due to mediastinal hematoma. However, the anterior fracture dislocation is not seen

7 Jpn J Radiol (2012) 30: radiography alone for thoracic spine fractures could not be assessed. The use of MDCT was suggested for detecting thoracic spine fractures in certain conditions. Further investigation regarding the contribution of MDCT to patient management and analysis of its cost-effectiveness are necessary to clearly define MDCT indications for patients who are suspected of having thoracic spine injuries. Finally, we determined two-column burst fractures to be stable fractures since we believed that two-column burst fractures were not as mechanically unstable in the thoracic segment as those in the cervical or lumbar segment. However, further study concerning the stability of burst fractures in the thoracic segment is warranted. In conclusion, the sensitivity of radiography of the thoracic spine was low in comparison with MDCT. In particular, regarding the upper thoracic segment, thoracic spine radiographs, even including the swimmer s view, missed unstable fractures. In addition, radiography of the thoracic spine was less sensitive in evaluating lower thoracic spine injuries than expected. In daily practice, however, we believe that the integration of radiography with physical and neurological examinations could certainly be adequate as the initial imaging tool for patients since radiography had very high specificity. However, the primary use of MDCT may be beneficial, particularly in patients complicated with upper extremity injuries and those aged 65 years or older, since radiography did not identify unstable fractures. In addition, for the patients presenting with neurological deficits, the primary use of MDCT should be considered because the sign of neurological deficits was very important for detecting unstable fractures. References 1. Denis F. The three column spine and its significance in the classification of acute thoracolumbar spinal injuries. Spine. 1983;8: Bohlman HH. Treatment of fractures and dislocations of the thoracic and lumbar spine. J Bone Jt Surg Am. 1985;67: El-Khoury GY, Whitten CG. Trauma to the upper thoracic spine: anatomy, biomechanics, and unique imaging features. AJR. 1993;160: Magerl F, Aebi M, Gertzbein SD, Harms J, Nazarian S. A comprehensive classification of thoracic and lumbar injuries. Eur Spine J. 1994;3: Vialle LR,Vialle E. Thoracic spine fractures. Injury 2005;36,Suppl 2:B65-SB Rogers LF, Thayer C, Weinberg PE, Kim KS. Acute injuries of the upper thoracic spine associated with paraplegia. AJR. 1980;134: van Beek EJ, Been HD, Ponsen KK, Maas M. Upper thoracic spinal fractures in trauma patients a diagnostic pitfall. Injury. 2000;31: Linsenmaier U, Krotz M, Hauser H, Rock C, Rieger J, Bohndorf K, et al. Whole-body computed tomography in polytrauma: techniques and management. Eur Radiol. 2002;12: Huber-Wagner S, Lefering R, Qvick LM, Korner M, Kay MV, Pfeifer KJ, et al. Effect of whole-body CT during trauma resuscitation on survival: a retrospective, multicentre study. Lancet. 2009;373: Wintermark M, Mouhsine E, Theumann N, Nordasini P, van Melle G, Leyvraz PF, et al. Thoracolumbar spine fractures in patients who have sustained severe trauma: depiction with multidetector row CT. Radiology. 2003;227: Roos JE, Hilfiker P, Platz A, Desbiolles L, Boehm T, Marincek B, et al. MDCT in emergency radiology: is a standardized chest or abdominal protocol sufficient for evaluation of thoracic and lumbar spine trauma? AJR. 2004;183: Herzog C, Ahle H, Mack MG, Maier B, Schwarz W, Zangos S, et al. Traumatic injuries of the pelvis and thoracic and lumbar spine: does thin-slice multidetector-row CT increase diagnostic accuracy? Eur Radiol. 2004;14: Geijer M, El-Khoury GY. MDCT in the evaluation of skeletal trauma: principles, protocols, and clinical applications. Emerg Radiol. 2006;13: Griffith B, Bolton C, Goyal N, Brown ML, Jain R. Screening cervical spine CT in a level I trauma center: overutilization? AJR. 2011;197: Kokabi N, Raper DM, Xing M, Giuffre BM. Application of imaging guidelines in patients with suspected cervical spine trauma: retrospective analysis and literature review. Emerg Radiol. 2011;18: Holmes JF, Akkinepalli R. Computed tomography versus plain radiography to screen for cervical spine injury: a meta-analysis. J Trauma. 2005;58: Dai LY, Wang XY, Jiang LS, Jiang SD, Xu HZ. Plain radiography versus computed tomography scans in diagnosis and management of thoracolumbar burst fractures. Spine. 2008;33:E Maynard FM Jr, Bracken MB, Creasey G, Ditunno JF Jr, Donovan WH, Ducker TB, et al. International standards for neurological and functional classification of spinal cord injury. Spinal Cord. 1997;35: Lomoschitz FM, Blackmore CC, Mirza SK, Mann FA. Cervical spine injuries in patients 65 years old and older: epidemiologic analysis regarding the effects of age and injury mechanism on distribution, type, and stability of injuries. AJR. 2002;178: Bud LD, Blackmore CC, Mann FA, Lomoschitz FM. Cervical spine fractures in patients 65 years and older: a clinical prediction rule for blunt trauma. Radiology. 2004;234:143 9.