Brief Report

Computed tomographic screening for thoracic and lumbar fractures: is spine reformatting necessary?

Eric A. Gross MD

Department of Emergency Medicine, Hennepin County Medical Center, Minneapolis, MN 55415, USA

Received 12 August 2008; revised 3 September 2008; accepted 3 September 2008

Abstract

Introduction: Patients who sustain traumatic vertebral fractures often have multiple other Associated injuries. Because of the mechanisms of injury, many of these patients routinely undergo chest computed tomographic (CCT) and/or abdominal/pelvic computed tomographic (APCT) scans to diagnose intrathoracic or Intra-abdominal injuries. These scans are routinely reformatted to provide more detailed imaging of the spine. Although the patient does not incur more radiation, the charges associated with this are significant. This study compared the sensitivity of these CT modalities in detecting thoracolumbar spine fractures.

Methods: A retrospective chart review identified blunt Trauma victims, admitted through the emergency department, with a discharge diagnosis of thoracic or lumbar spine fracture that received (1) a chest and T-spine CT, (2) an abdominal/pelvic and lumbar spine CT, or both. Final radiologic readings of these patients’ CT scans were obtained, and the sensitivities of the different imaging methods were compared. Discharge diagnosis of spine fracture was considered the gold standard.

Results: One hundred seventy-six APCT scans with reformatting and 175 CCT scans with reformatting were available for comparison. There were 9 of 176 false-negative APCT scans vs 3/176 false-negative lumbar spine CT scans. There were 14/175 false-negative CCT scans vs 2/175 false-negative thoracic spine CT scans. The differences in sensitivity were significant (P b .001) for both comparisons.

Conclusions: Reformatting of CCT and APCT scans gives improved sensitivity in the detection of thoracic and lumbar spine fractures in trauma patients. Future study looking at clinically significant fractures or those that change clinical management decisions may find that the reformatted images are not routinely needed as a screening tool.

(C) 2010

Introduction

Fractures of the thoracolumbar (TL) spine are seen in 4% to 5% of blunt trauma patients [1]. Because of the

Presented at the American College of Emergency Physicians’ Research Forum, October 2007.

E-mail address: [email protected].

mechanisms of injury underlying most of these fractures, many of these patients undergo chest computed tomographic (CCT) and/or abdominal/pelvic computed tomographic (APCT) scans to diagnose intrathoracic or intra-abdominal injuries. These scans are routinely reformatted to provide transverse, coronal, and sagittal images of the spine for more accurate screening for fractures.

Although the patient does not incur more radiation, the charges associated with this are significant. At one institution,

0735-6757/$ - see front matter (C) 2010 doi:10.1016/j.ajem.2008.09.013

74 E.A. Gross

Table 1 Comparison of CCT to TSCT

Chest CT Fracture

No fracture

the difference in charges between CCT with and without thoracic spine reconstruction (TSCT) is $2450 ($4150 vs

$1700). The difference in charges between APCT with and without lumbar spine reconstructions (LSCTs) is also $2450 ($6050 vs $3600). These charges are exclusive of any radiologist reading fees.

Whether these reformatted images are necessary as a screening tool is not known. To date, no studies have compared the sensitivity of chest or abdominal/pelvic scans with and without reformatting in detecting spine fractures. The objective of this study was to compare the accuracy of CCT to thoracic spine CT (TSCT) and APCT to lumbar spine CT (LSCT) in the detection of thoracic and lumbar spine fractures.

Methods

This was a retrospective study performed at an urban level I trauma center with an annual emergency department (ED) census of 97,000 patients. This study was approved by the institutional review board and granted a waiver of informed consent. Victims of blunt trauma admitted to the hospital through the ED with final discharge diagnoses of lumbar or Thoracic spine fractures were identified by an International Classification of Diseases, Ninth Revision (ICD-9), search. In that group, patients who received (1) a chest CT (CCT) and TSCT, (2) an APCT and LSCTs, or both were enrolled. If patients did not receive their CT scans at the time of initial ED evaluation, they were excluded. If patients did not receive both axial chest or abdominal-pelvic and the corresponding reconstructed sagittal and coronal images of the spine, they were excluded.

Final radiographic readings of all CT scans and reconstructions were obtained. Sensitivities in diagnosing spine fractures of CCT compared to TSCT and APCT compared to LSCT with reconstructions were compared. The discharge diagnosis of spine fracture was considered the gold standard for this study.

Data were analyzed using STATA 10.0. Descriptive statistics were used as appropriate. Sensitivity rates were compared using ?2 tests. A power analysis was done before initiation of the study. To show equivalence of CT with and without spine reconstructions within 10%, we needed 166 subjects per group. The following assumptions were made to arrive at this number: ? .01, ? .90, and 98% sensitivity of CT with reformatting.

Results

From December 1, 2004, to August 28, 2007, an ICD-9 search identified 338 patient charts for screening for the type of CT scans ordered. Seventy-five patients were excluded: 73 were misclassified by the ICD-9 search, 1 was a duplicate patient entry, and 1 had CT scans from an outside facility only.

TSCT	Fracture	82	14
	No fracture	2	77

In addition, comparison studies were missing in 82 instances of patients with appropriate diagnoses and a positive CT. These studies were not included in the data analysis.

In the remaining patients, 176 APCT scans with lumbar spine reformatting and 175 CCT scans with thoracic spine reformatting were available for comparison. Many patients had all 4 studies performed. There were 14 of 175 false- negative CCT scans vs 2 of 175 false-negative TSCT scans (see Table 1). There were 9 of 176 false-negative APCT scans vs 3 of 176 false-negative LSCT scans (see Table 2). The differences in sensitivity were significant (P b .001) for both comparisons.

Discussion

Previous research on the appropriate modality for screening for TL spine fractures has focused on CT vs Plain radiography. Computed tomography, even without reformatting, has been shown to be superior to radiographs in the detection of these fractures. One study found a sensitivity of 97% for CT of the chest/abdomen/pelvis vs 58% for standard radiographs of the TL spine [2]. Similarly, reformatted CT scans offer significant advantages to plain radiographs, offering sensitivity of 97% for TSCT (vs 68% for T-spine x-ray) and 95% for LSCT (vs 86% for lumbar spine x-ray) [3]. Clearly, data support the use of current generation, multidetector CT scan in the screening of trauma patients for TL spine fracture. The Eastern Association for the Surgery of Trauma‘s guidelines states that, “Reformatting of images allows a superior visualiza- tion of the spine and may be appropriate for areas of high concern” [4]. Whether routine reformatting is justified is not clear. This author was able to find only one study comparing CCT and APCT to reformatted images; how- ever, reformatted images were only obtained for suspicious areas. In addition, Image quality, not sensitivity as a screening tool for fractures, was the main outcome [5].

Despite similar sensitivities for fracture screening found

in the aforementioned studies, routine reformatting of chest and APCT scans showed improved detection of thoracic and lumbar spine fractures in this study.

Whether routine reformatting is necessary may be debatable. The cost associated with the routine reformatting in all patients is significant. In addition, Delay in diagnosis of TL fractures does not increase the incidence of neurologic injury [6]. Of note, most of the Missed fractures were either

Reformatting in thoracolumbar spine trauma 75

This study found that reformatting of CCT and APCT does increase sensitivity for the detection of TL spine fractures. Future study looking at clinically significant fractures or those that change clinical management decisions may find that the reformatted images are not routinely needed as a screening tool.

Table 2 Comparison of APCT to LSCT

APCT

Fracture

LSCT Fracture 106

No fracture 3

No fracture

9

58

spinous process or transverse process fractures. The clinical significance of these fractures is unclear. Neurologic deteriora- tion or need for operative intervention is unlikely, but management decisions may still be affected. Physical therapy, pain medication regimens, or treatment with orthoses for comfort may be required in these patients.

This study was performed retrospectively. As such, radiology reads of individual scans were not done indepen- dently. A radiologist reading both scans on same patient may influence final radiographic interpretations and bias results. However, one would expect improved sensitivities of the CCT and APCT if the reformatted spine images were seen contemporaneously. Alternatively, radiologists may not have focused intently on the bony windows of the CCT and APCT, knowing that dedicated spine CT scans were also obtained. Radiologists did dictate a “bone” section to most reads on the APCT and CCT scans, though. Therefore, the true sensitivity of CCT and APCT in this study is unclear. In addition, an ICD-9 search of medical records may miss some patients who should have been included.

References

1. Cooper C, Dunham CM, Rodriguez A. Falls and Major injuries are risk factors for thoracolumbar fractures: cognitive impairment and multiple injuries impede the detection of back pain and tenderness. J Trauma 1995;38:692-6.
2. Hauser CJ, Visvikis G, Hinrichs C, et al. Prospective validation of computed tomographic screening of the thoracolumbar spine in trauma. J Trauma 2003;55(2):228-35.
3. Sheridan R, Peralta R, Rhea J, et al. Reformatted visceral protocol helical computed tomographic scanning allows conventional radio- graphs of the thoracic and lumbar spine to be eliminated in the evaluation of blunt trauma patients. J Trauma 2003;55(4):665-9.
4. Diaz JJ, Cullinane DC, Altman DT, et al. Practice management guidelines for the screening of thoracolumbar spine fracture. J Trauma 2007;63(3):709-18.
5. Roos JE, Hilfiker P, Platz A, et al. MDCT in emergency radiology: is a standardized chest or abdominal protocol sufficient for evaluation of thoracic and lumbar spine trauma? AJR Am J Roentgenol 2004;183: 959-68.
6. Reid DC, Henderson R, Saboe L, et al. Etiology and clinical course of missed spine fractures. J Trauma 1987;27(9):980.