| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Diagnostic Imaging, Radboud University Nijmegen Medical Centre,
Internal number (Huispost) 667, Geert Groote plein 10, 6500 HB Nijmegen, The
Netherlands.
2 Department of Surgery, Radboud University Nijmegen Medical Centre, Nijmegen,
The Netherlands.
Received October 9, 2007;
accepted after revision December 2, 2007.
Address correspondence to M. Brink
(m.brink{at}rad.umcn.nl).
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. A prospective cohort study was conducted in 464 consecutive blunt trauma patients who met criteria indicative of severe blunt trauma (66% male; age range, 16-93 years; median injury severity score, 13). After clinical evaluation and conventional radiography of the chest and thoracic spine, all patients underwent routine thoracic MDCT with an IV contrast agent (routine MDCT algorithm). Within this routine MDCT group, a subgroup was prospectively defined with abnormal or inconclusive clinical or conventional radiography evaluation (selective MDCT group). Two investigators determined the type, extent, and clinical impact of additional injuries found on MDCT as compared to conventional radiography for both MDCT groups.
RESULTS. Of all 464 patients within the routine MDCT group, 164 patients underwent selective MDCT, which resulted in detection of additional diagnoses compared with conventional radiography in 97 (59%) patients. The routine MDCT algorithm detected additional diagnoses compared with conventional radiography in 201 of 464 patients (43%). Compared with the selective MDCT algorithm, this was an absolute increase of 104 of 464 (22%) extra patients, resulting in a change in patient management in 34 (7%; 95% CI, 5-9.7%), mostly because of additional findings of pulmonary and mediastinal injury.
CONCLUSION. Routine MDCT has relatively lower, though still substantial, added diagnostic value compared with selective MDCT of the chest.
Keywords: blunt trauma • CT • emergency radiology • thorax • wounds and injuries
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Regarding MDCT of the head, cervical spine, pelvis, and abdomen, several clinical prediction tools have been developed and validated [3-6]. Some recommendations exist concerning vascular injuries of the thorax [1, 7-9]. To our knowledge, little evidence exists on scanning indications for the complete chest, although chest injury has a high incidence and is an important contributor to mortality [10]. Nevertheless, routine MDCT of the chest in trauma patients has become common practice [11], although little is known about the added value of this practice compared with a more selective use of trauma MDCT.
The purpose of our study was to evaluate the added diagnostic value of a routine (low-threshold) thorax MDCT algorithm compared with a more selective (physical examination- and conventional radiography-driven) thorax MDCT algorithm in blunt trauma patients. We investigated a cohort of blunt trauma patients who fulfilled the criteria for our high-energy trauma protocol and consequently underwent a routine thoracoabdominal MDCT algorithm. For our study, we prospectively defined a subgroup of these patients on the basis of abnormal physical and conventional radiography parameters who would have received only MDCT if a selective MDCT algorithm was to be used. For both the routine and the selective algorithm, we determined the additional diagnostic value of MDCT compared with conventional radiography. We defined additional diagnostic value as the number of additional findings with impact on patient management compared with conventional radiography.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Imaging Studies
The diagnostic trauma protocol consisted of clinical evaluation; laboratory
investigation; conventional radiography; focused abdominal sonography for
trauma (FAST); and routine MDCT of the cervical spine, thorax, abdomen, and
pelvis. Conventional radiography of the thorax consisted of at least one
supine anteroposterior chest radiograph and a supine radiograph of the
thoracic spine in the lateral and anteroposterior directions. This was
performed on a Vertix 3D-III system (Siemens Medical Solutions), which is
located in the emergency department where the patients were resuscitated.
A routine MDCT protocol was executed on a Somatom Sensation 16-MDCT scanner (Siemens Medical Solutions) with automated tube current modulation located adjacent to the trauma bay. MDCT of the thorax was performed as a part of a thoracoabdominal scan from the acromio-clavicular joint to the lesser trochanter at a tube potential of 120 kV with a reference value of effective tube current-time product of 200 mAs and a mean dose-length product of 1,169 mGy. The detector configuration was 16 x 1.5 mm. A total of 100 mL of iodinated contrast material (iobitridol [Xenetix 300, Guerbet]) was injected. Reconstructed section thickness was 3 mm for lung, soft tissue, and bone reconstruction kernel, with an increment of 3, 3, and 1.5 mm, respectively. Sagittal and coronal reformatted images of the spine were obtained.
Interpretation of Clinical and Radiologic Findings
A surgical resident performed and interpreted the physical examinations
under supervision of a senior trauma surgeon according to advanced trauma life
support (ATLS) guidelines
[12]. The attending
radiologist or resident in radiology interpreted all conventional radiographs
on a PACS workstation at the time of the initial evalu ation of the patient.
Subsequently, and only for our research project, it was prospectively
determined whether patients met criteria for a selective MDCT. A selective
MDCT was defined as an indicated supplement to abnormal clinical or
radiographic evaluation. Before MDCT started, the attending surgical resident
explicitly noted if abnormalities on the physical examination and conventional
radiography fulfilled the criteria for a selective CT. At physical
examination, these criteria were the clinical suspicion of chest injury or
thoracolumbar vertebral injury: subcutaneous emphysema, asymmetric
auscultative findings, tenderness to palpation (in nonobtunded patients)
[13], or neurologic signs
suggestive of spinal injury
[14]. Criteria for selective
MDCT on conventional radiography were either the subjective impression of an
abnormal media stinum [7,
15], more than three rib
fractures [12], pulmonary
consolidation suspected to be lung con tusion or hemothorax
[12], intrapleural air or sub
cutaneous emphysema suspected to be pneumo thorax
[10], or thoracolumbar
vertebral fractures [16]. When
these criteria were not fulfilled, routine MDCT was considered a nonselective
MDCT (Fig. 1).
|
Thereafter, MDCT was performed in all patients. A resident in radiology interpreted all MDCT images under supervision of a board-certified radiologist who either was physically present at the initial reading of the MDCT images in the trauma bay or reviewed all CT images within several hours if patients were resuscitated during the night. On the basis of this inter pretation, the trauma team started or changed patient management as needed. Finally, in accordance with standard procedures for trauma patients in our hospital, an effort was made to follow every patient for at least 6 months, either in the outpatient clinics or by telephone.
Collection and Analysis of the Data
Every working day, two investigators attended briefings and resuscitations
of the included patients and reviewed all clinical charts and radiologic
reports of conventional radiography and MDCT. They assessed the type of
trauma-related abnormalities (signs of aortic injury, pneumothorax,
hemothorax, pulmonary contusion, lung laceration, diaphragmatic injury,
presence and number of thoracic cage fractures, and spinal fractures) on
conventional radiography and MDCT as reported by the attending radiologists.
In addition, they retrospectively assessed the extent of pulmonary injuries on
MDCT. Pneumothoraces and pulmonary contusions were classified according to
their location and size (unilateral or bilateral; minimal, moderate, or
severe). Hemothoraces were classified according to their location and volume
(unilateral or bilateral; less than 500 mL, 500-1,500 mL, or more than 1,500
mL). Disagreement between the investigators was resolved by consensus.
Using MDCT as the standard of reference, the number of patients with additional diagnoses on MDCT compared with conventional radiography was determined as a primary outcome measure. Subsequently, the extent of these additional diagnoses was assessed.
As a secondary study outcome measure, the number of patients with additional MDCT findings with an impact on patient management was identified. Impact on patient man agement was defined as the occurrence of one or more changes in treatment as a direct result of the MDCT findings. These changes included additional diagnostic workup, a change in intensity of care (care level upgrade), and immediate intervention directly and actually started by the attending trauma team. At least 6 months post trauma, all patients' charts and operational re ports were rereviewed to establish whether MDCT had missed any diagnoses that had manifested over time. In addition, for every patient the definitive injury severity score (ISS) [17] was collected from the regional trauma registry.
Furthermore, one investigator, who was not involved in treatment of the patients, manually measured the time to perform conventional radiography in a convenience sample of 47 patients. Because conventional radiography was performed in the emergency department, this time was measured from the start of patient positioning until completion of each view. Completion was achieved when the preview images were processed and ready for determination of adequacy. The time needed to perform addi tional radiographs in case of prior inadequate views was included in the measure ments. In addition, the time required to perform the MDCT protocol was recorded in a convenience sample of 57 patients. This was defined as the time between departure and return from the trauma bay to the adjacent CT room minus the time needed to obtain scans of other body regions.
Statistical Analysis
Descriptive statistics (mean and range) on age, sex, and ISS were used to
provide information on the cohort composition. We calculated the extra time
required to perform routine MDCT from the median of the time values as
measured in the convenience sample. For further analysis, study outcome
measures were assessed for all included patients (routine MDCT group) com
pared with the subgroup of patients with selective MDCT (selective MDCT
group). Taking into account this paired design of the study, the estimated
mean and its 95% CI of the difference in study outcomes between the selective
and routine MDCT groups were calculated
[18]. We performed the data
analysis using the statistical soft ware package SPSS for Microsoft Windows,
version 12.0.1 (Statistical Package for the Social Sciences).
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Routine MDCT Performance
Fifty-two percent of all routine MDCT patients (242/464) had trauma-related
injuries on chest MDCT. These injuries were predominantly pulmonary injuries,
rib fractures, and thoracic spine fractures
(Table 2). The routine MDCT
algorithm showed additional diagnoses compared with conventional radiography
in 43% of all patients (201/464). These additional diagnoses were
pneumothoraces (n = 85 patients), pulmonary contusions (n =
94), rib fractures (n = 57), scapular fractures (n = 28),
sternal fractures (n = 21), vertebral body fractures (n =
21), transverse process fractures (n = 24), aortic injury (n
= 2), esophageal injury (n = 1), and diaphragmatic injury (n
= 1). As displayed in Table 3,
these additional diagnoses induced 96 changes in management in 81 patients
(17%).
|
|
Of all included patients, 22 patients (5%) were lost to follow-up. Completed clinical follow-up showed that in one patient, MDCT of the spine appeared to be false-positive for thoracic spine fracture. However, the same MDCT showed additional pulmonary contusions and rib fractures. In the remaining 441 patients, routine MDCT of the chest missed no diagnoses that had an impact on patient management.
Selective MDCT Performance
A subgroup of 164 patients met the criteria for selective MDCT. In this
group, the median ISS was 22 and the incidence of trauma-related injuries was
74% (122/164) (Table 2). The
use of this MDCT algorithm would have shown additional diagnoses compared with
conventional radiography in 59% (97/164) of patients. In 29% of the patients
(47/164), this had impact on patient management
(Table 3).
Routine Versus Selective MDCT Algorithm
The percentage of chest injuries was lower in the routine MDCT group
compared with the selective MDCT subgroup (52% vs 74%)
(Table 2). In 104 patients, the
use of a routine MDCT algorithm resulted in more findings of additional
injuries compared with the selective MDCT algorithm. In the total patient
group, this is an absolute increase of 104 of 464 = 22% (95% CI, 19-26%).
These additional injuries were predominantly pulmonary lesions and thoracic
cage fractures (Table 4).
|
Figures 2A, 2B, 2C and 2D illustrates the severity of hemothoraces, pneumothoraces, and pulmonary contusions and the extent of rib fractures that were diagnosed on MDCT but missed on conventional radiography in both algorithms. As shown in this figure, these pulmonary injuries that would have been missed if the selective CT algorithm had been used were predominantly of minor severity.
|
|
|
|
In six patients, routine MDCT exclusively showed moderate to large hemopneumothoraces that needed chest tube insertion. Epidural anesthesia was started in three patients with multiple additional rib fractures. A cardiologist was consulted in three patients with additionally diagnosed sternal fractures because it was assumed that these lesions might be associated with blunt cardiac injury. However, in none of these patients could trauma-related cardiac injury be confirmed at follow-up. One additionally diagnosed vertebral body fracture was treated nonoperatively on an orthopedic ward. Most additional diagnoses with impact on patient management consisted of a combination of pneumothoraces, multiple rib fractures, or pulmonary contusions that led to an upgrade in intensity of care.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
However, our results also show that the use of a routine MDCT algorithm increases scanning frequency. This consequently increases examination time and costs. Moreover, the use of a routine algorithm increases radiation exposure (here presented in terms of dose-length product) as well. Although it is hard to estimate the detrimental effect of radiation from a single extra CT in this sample of patients of varying age, this is of concern because the trauma population is relatively young [19]. Finally, we found that the diagnoses exclusively revealed by the MDCT algorithm are relatively less extensive and are less frequently therapeutically relevant compared with the selective MDCT algorithm. Thus extension of scanning criteria automatically leads to a lower pretest probability of relevant diagnoses and to an absolute increase, but a relative decrease, in the additional diagnostic value of MDCT.
Several studies detail the absolute increase of additional findings of (clinically relevant) trauma-related diagnoses of helical CT and MDCT compared with conventional radiography. These studies cite numbers of additional diagnoses as high as our results [20, 21]. Most of these studies are retrospective surveys and incorporate bias in patient selection and record information.
In the majority of prospective studies on this topic, patient selection is roughly comparable to our selective MDCT group. These studies report comparable frequencies of additional visceral or spinal injuries. Hauser et al. [22] reported additional spinal injuries on CT compared with conventional radiography in 7% of 215 evaluable patients. Guerrero-Lopez et al. [23] reported additional findings with impact on patient management in 30% of 103 patients who underwent chest CT, and Trupka et al. [10] reported that CT detected major chest trauma complications that changed therapy and were missed on chest radiography in 41% of 104 patients. The higher incidence of additional findings with impact on patient management in the Trupka et al. study might be due to a more aggressive approach to additionally diagnosed pneumothoraces than in our practice. On the basis of the findings, most of these studies recommend performing CT in the type of population they studied. Moreover, the authors advocate that CT should also be performed in a less selective population.
Studies that more extensively examined the value of a routine algorithm in a representative but low-threshold trauma patient group are less prevalent in the literature. A study that compared single-detector CT indicated by abnormal radiographic or physical examination (n = 110) and mechanism-driven thoracic CT (n = 28) found additional relevant clinical findings in 20% and 5%, respectively, of all patients [24]. Another study group showed that conventional radiography of the chest did not reveal aortic injuries in 3% of patients who had MDCT after decelerating injury, irrespective of chest radiography findings [25]. The incidence of vascular injury and mean ISS score in this study group were much higher compared with our group.
The same study group showed that mechanism-driven MDCT shows additional diagnoses compared with conventional radiography of the chest in only 8% of 1,000 blunt trauma patients who have no obvious signs of injury [11]. In our nonselective routine MDCT group, this was substantially more (22%) (Table 4). Although detailed patient characteristics and figures concerning the thoracic spine are unclear in the last two studies, the contrast in outcomes with our data could be explained by differences in local practice and patient selection.
Our study adds to prior knowledge in the following ways: First, it quantifies the added value of primarily mechanism-driven routine MDCT of the entire chest compared with selective (physical examination- and conventional radiography-driven) MDCT in a representative and well-defined trauma patient group. That we detected substantially more clinically relevant additional diagnoses with the low-threshold routine MDCT algorithm advocates the routine use of thoracic MDCT in trauma patients.
Does this mean that the costs and time needed to perform conventional radiography of the chest and spine might be circumvented? In centers where patients can be resuscitated in the CT suite and a CT scanogram can be used to detect injuries that need immediate intervention, it might. Previous studies showed that in these infrastructures, conventional radiography of the thoracolumbar spine can be safely omitted [26, 27]. However, even if routine MDCT is performed, we believe that the usefulness of initial conventional radiography of the chest should not be underestimated. Chest conventional radiography can still serve as an initial quick test for large injuries at hundreds of times lower radiation dose than chest MDCT [28].
Second, this study shows a trend toward a decreased rate, severity, and clinical relevance of additional findings as scanning criteria expand. This indicates that it might be time to make a trade-off to prevent excessive and unnecessary imaging. First of all, the outcomes of this study can be useful in instigating that transition. Second, it might be possible to more specifically identify patients who are at high risk of having clinically relevant thoracic injuries using other (less strict) clinical factors than we used here. This might lead to the establishment of a clinical prediction rule and the consequent containment of costs, time consumption, and radiation exposure. In the meantime, other efforts can be made to minimize ionizing radiation exposure because most trauma patients receive high radiation doses not only at their primary evaluation but even more at their clinical follow-up [29]. Consequently, we should face the challenge to investigate alternative trauma CT protocols that use either low kilovoltage [30], automated tube current modulation with lower reference tube current-time products, or adapted beam collimation [31], or even use only digital scout views [27] or short-segment acquisitions without compromising accuracy for relevant traumatic lesions.
Our study has some limitations. First of all, estimation of performance of conventional radiography and MDCT was not done independently of clinical information. Radiologists were not blinded to this information. In our clinic, surgeons and radiologists work in close cooperation; the radiologist is present in the trauma bay during resuscitation. However, because of the nature of our study purpose and design, this was not considered a major problem.
Second, we eliminated hindsight bias as much as possible by insisting that clinicians and radiologists thoroughly assess conventional radiography before MDCT was executed. However, in the middle of the night, no investigator was present to protect and ensure the prospective nature of selective MDCT classification, and clinicians and radiologists were trusted with respect to their reports. This might have induced hindsight bias in interpretation of radiography and clinical evaluation. In a minority of the cases, this might have resulted in misjudgment of physical examination and conventional radiography performance and in misclassification of selective MDCT.
Third, our second study outcome, presence of impact on patient management, was determined in consensus. This outcome is sensitive to subjectivity and strongly depends on local treatment algorithms. For instance, we reported diagnostic workup to exclude blunt cardiac injury in patients with occult sternal fractures, although the relation between sternal fractures and cardiac contusion is not clear in the literature [32]. This practice might have resulted in an overestimation of clinically relevant additional diagnoses. On the other hand, the spectrum of impact of additional diagnoses was not exhaustive in our study because an outcome such as increasing diagnostic confidence in the presence or absence of injuries is a debatable outcome and very hard to quantify.
Finally, selection bias might have been introduced. Violation of the protocol occurred in 7% of eligible patients who had, according to the supervising traumatologists, a very low suspicion of trauma-related injuries and therefore MDCT was not performed. None of these patients suffered severe complications of thoracic trauma in clinical follow-up. This acceptably small selection bias possibly led to an overestimation of the added value of routine MDCT.
In conclusion, in this prospective study, we showed that performing routine MDCT of the thorax has a substantial added value compared with a physical examination- and conventional radiography-driven selective MDCT algorithm. This advocates the routine use of thoracic MDCT in trauma patients. However, this study also quantified that the virtually automatic use of MDCT increases the chance of finding no trauma-related injuries and exposing more patients to the potentially harmful effects of MDCT.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
