- Email: [email protected]
Multidetector-row CT of right hemidiaphragmatic rupture caused by blunt trauma: a review of 12 cases
Clinical Radiology (2005) 60, 1280–1289
Multidetector-row CT of right hemidiaphragmatic rupture caused by blunt trauma: a review of 12 cases O. Reesa, S.E. Mirvisb,*, K. Shanmuganathanb a
University of Wales College of Medicine, Cardiff, UK; and bDepartment of Radiology and Maryland Shock Trauma Center, University of Maryland School of Medicine, Baltimore, MD, USA
Received 23 June 2004; received in revised form 31 May 2005; accepted 22 June 2005
KEYWORDS Diaphragm (injuries); Diaphragm (CT); Computed tomography; Clinical effectiveness; Trauma; Thorax (injuries)
AIM: To determine the usefulness of multidetector-row CT (MDCT) with multiplanar reformatted (MPR) images in the sagittal and coronal plane in diagnosing acute right hemidiaphragmatic rupture. MATERIALS AND METHODS: Twelve patients were identified who received chest and abdominal MDCT after major blunt trauma diagnosed with right diaphragmatic injury. Sagittal and coronal reformations were performed in all cases. The images were retrospectively reviewed by two experienced radiologists for signs of right diaphragm injury, such as direct diaphragm discontinuity, the “collar sign”, the “dependent viscera sign”, and intra-thoracic location of herniated abdominal contents. RESULTS: Of the 12 cases of right hemidiaphragm rupture, diaphragm discontinuity was seen in seven (58%) cases, the collar sign in five (42%), the dependent viscera sign in four (33%), and transdiaphragmatic herniation of the right colon and fat in another. Two variants of the collar sign were apparent on high-quality sagittal and coronal reformations. The first, termed the “hump sign”, describes a rounded portion of liver herniating through the diaphragm forming a hump-shaped mass, and the second, termed the “band sign,” is a linear lucency across the liver along the torn edges of the hemidiaphragm. The hump sign occurred in 10 (83%) patients and the band sign in four (33%). CONCLUSION: MDCT is very useful in the diagnosis of right hemidiaphragm injury caused by blunt trauma when sagittal and coronal reformatted images are obtained, and should allow more frequent preoperative diagnosis. Q 2005 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
Introduction Traumatic diaphragm rupture (TDR) is a serious injury resulting from blunt abdominal trauma, the majority of which are caused by motor vehicle collisions and vehicles striking pedestrians, and occurs in 0.5–8% of patients undergoing surgical exploration.1–11 Hemidiaphragm rupture more commonly involves the left side and occurs in 56–86% of * Guarantor and correspondent: S.E. Mirvis, Department of Radiology, 22 South Greene Street, Baltimore, MD, USA. Tel.: C1 410 328 8845; fax: C1 410 328 0641. E-mail address: [email protected] (S.E. Mirvis).
cases. Right-sided tears occur in 11–39% of cases, and bilateral and other sites account for 2.4–13.0% of cases.1–11 A higher percentage of right-sided injuries are reported in surgical series, because historically this injury has been far easier to diagnose by direct inspection than by imaging methods.1,8 There are several theories concerning the mechanism of diaphragm rupture. These include avulsion of the attachments of the diaphragm or shearing of the stretched membrane after right or left lateral impact to the chest wall, rib fracture fragments directly penetrating the diaphragm, and a sudden increase in intra-abdominal pressure
0009-9260/$ - see front matter Q 2005 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.crad.2005.06.013
Multidetector-row CT of right hemidiaphragmatic rupture caused by blunt trauma
throughout the abdomen with the relatively weak, unprotected left diaphragm tearing from the force.10,12 Experimental studies also suggest that the right hemidiaphragm is mechanically stronger than the left, and requires a larger force to disrupt it.12,18 The right hemidiaphragm is also protected from abdominal impact by the energy-absorbing liver.12,18 Associated injuries are common in traumatic rupture of the diaphragm due to the relatively large force required to disrupt the diaphragm concurrently damaging adjacent organs. Associated major injuries are reported in 52–100% of patients.1,2,6,8,10–12 The most commonly damaged intra-abdominal organs are the liver, reported as being disrupted in 93% of patients with right-sided injury, and spleen in 24% of those with left-sided injury.12 In all cases of diaphragm tears, splenic injuries occur in 27– 63%.2,6,10,12 Other commonly associated abdominal injuries include pelvic and renal injuries and associated intra-thoracic injuries include haemopneumothorax and multiple rib fractures.1,2,6,8,10–12 The preoperative diagnosis of TDR is often difficult due to a lack of specific clinical signs and the presence of distracting or more obvious serious concurrent injuries. The use of chest radiography for initial screening of TDR has a relatively low sensitivity for diagnosis, reported as diagnostic in 46% of patients with rupture of the left side, and suspicious for TDR in another 18%.4 The use of helical computed tomography (CT) with sagittal and coronal reformations has increased the detection rate of TDR, with sensitivities and specificities for overall detection of right and left ruptures reported between 61–90% and 77–100%, respectively.13–18 However, the sensitivity for diagnosis of rightsided rupture is not as high as for the left side, with a CT sensitivity and specificity of left hemidiaphragm rupture reported as 78–100% and 100%, respectively, and the sensitivity and specificity of right hemidiaphragm rupture reported as 50–83% and 100%, respectively.12,16 The main diagnostic CT signs of rupture of the diaphragm include direct visualization of diaphragm discontinuity, abdominal visceral herniation above the diaphragm, a waist-like constriction of herniating viscera, the so-called “collar sign”, direct contact of the herniated liver with the chest wall without intervening lung, the “dependent viscera” sign, and thickening of the hemidiaphragm from haemorrhage among others.10,12–18 Helical CT has become the imaging method of choice in the acute trauma setting, complemented by its ability to acquire high-quality coronal and sagittal reformations in contrast to axial CT images alone. While the sensitivity of left-sided rupture of the
1281
diaphragm has increased with the use of helical CT, the sensitivity of detection of right-sided diaphragm rupture has not increased to the same extent.12,16 The advent of MDCT has provided the capability to perform CT with smaller sections and the ability to produce high-quality, non-axial, two-dimensional multiplanar reformatted (MPR) and volumetric renderings. The purpose of this study was to review retrospectively findings of MDCT in the diagnosis of right-sided hemidiaphragm rupture resulting from blunt trauma. Herein, the authors describe two other signs, which are variants of the collar sign, seen on high-quality coronal and sagittal images. The first sign termed the “hump sign” consists of a portion of the liver herniating through the right diaphragm to form a hump-shaped herniation, clearly visible on coronal and sagittal reformations. The second sign is the “band sign”, a band-like lucency through the mass of the liver where it herniates through the torn diaphragm, again clearly visible on coronal and sagittal reformations. The accuracy of other commonly valued signs was also reviewed.
Materials and methods Permission for this study was obtained from our institutional review board and HIPAA compliance was maintained. A retrospective trauma registry and radiology database review was performed for patients diagnosed with right-sided hemidiaphragm injury caused by blunt trauma who underwent preoperative thoraco-abdominal MDCT at the authors’ Level 1 trauma center, for the period from May 2001 to November 2004. There were 12 such patients found, 11 of whom had surgical confirmation of the injury. The study group consisted of six men and six women, whose ages ranged from 28–86 years (mean 44 years). Mechanisms of injury consisted of motor-vehicle collisions in 11 and one pedestrian struck by a vehicle. All patients underwent a supine anteroposterior (AP) chest radiograph on admission, followed shortly by intravenous and oral contrast-enhanced MDCT (MX 8000, Philips Medical Systems, Best, The Netherlands) using a standardized MDCT protocol. Thoracic and abdominal MDCT examinations were initially performed using a four-detector row system (Philips MX 8000, Philips Medical Systems, Best, The Netherlands) obtaining 4!1.3 mm sections per rotation with a table speed of 10 mm/s. During the last 21 months of the study, abdominal– pelvic CT images were acquired using a 16-detector
1282
O. Rees et al.
row CT (MX 8000 IDT, Philips Medical Systems, Best, The Netherlands) using 16!1.5 detector width and 0.75 s rotation time. For both protocols, intravenous contrast material (300 mg I2/ml, 150 ml) was administered at 3 ml/s using a power injector (Medrad 4, Medrad, Pittsburg, PA, USA). A total volume of 600 ml of 2% sodium diatrizoate (Hypaque sodium, Nycomed, Princeton, NJ, USA) oral contrast material was ingested 30 min before and immediately prior to the examination whenever possible. For both protocols, axial images were reconstructed to 5 mm section thickness for initial review and PACS storage. When diaphragm injury was suspected based on the admission supine chest radiograph and axial CT images, the original data were reconstructed to 2 mm sections with 1 mm overlap. These axial images were used for coronal and sagittal reformations through the diaphragms. In eight patients, a dedicated 16-MDCT study was performed for the hemidiaphragms using 16! 0.5 mm sections and 0.75 s rotation time. For this study, the admission supine AP chest radiographs obtained before MDCT and axial MDCT images with the coronal and sagittal reformations were reviewed by two experienced trauma radiologists (S.E.M., 20 years experience and K.S., 14 years experience) and findings determined by consensus. The chest radiographs were reviewed and the distances between the domes of the apparent left and right diaphragms were recorded (in centimetres). The MDCT studies with MPRs were reviewed for signs of right hemidiaphragm injury. The images were considered diagnostic if any of the following signs were present: Direct visualization of a diaphragm defect, evidence of liver parenchymal herniation into the right hemithorax (hump sign), a band of lucency across the liver at the level of herniation through the diaphragm (band sign), herniated liver with focal constriction at the point where it the liver crosses the torn diaphragm (collar sign), and herniation of other viscera structures (such as bowel) through the right hemidiaphragm (with or without the collar sign), and the dependent viscera sign (liver in direct contact with the
Figure 1 Diaphragm discontinuity and dependent viscera sign. (a) Axial MDCT shows discontinuity in the
right hemidiaphragm (arrow) and thickening of the adjacent diaphragm due to haemorrhage. The liver (L) is in direct contact with the posterior chest wall through the torn diaphragm constituting the dependent viscera sign. (b) Dedicated sagittal diaphragm MPR shows a tear in the anterior right hemidiaphragm (arrow) and herniation of liver into the right hemithorax. There is an effusion (E) above the liver. (c) Coronal MPR shows torn edges of the right hemidiaphragm (arrows) and herniation of the right lobe into the right hemithorax. The gallbladder (G) is in the chest.
Multidetector-row CT of right hemidiaphragmatic rupture caused by blunt trauma
posterior chest wall). Thickening of the right hemidiaphragm or fluid in the adjacent retroperitoneal fat were also noted. Associated thoracic and abdominal injuries were recorded. The apparent height difference between the hemidiaphragms was measured from a supine chest radiograph for the 12 patients in the study and for 50 admission supine chest radiographs from patients admitted to our trauma center who had no radiographic or clinical evidence of diaphragm injury. The range, mean, and variance among the samples were calculated.
Figure 2 The hump sign of liver herniation. (a) Admission chest radiograph shows focal rounded smooth increased density at the medial right lung base (arrow). (b) Axial CT image at the right lung base elevation of the right hemidiaphragm with a subtle lucent line (“band”) across dome of liver (arrow) representing liver compression by the edge of the diaphragm. There are low attenuation lacerations also in the dome of the right lobe.
1283
The medical record of each patient was subsequently reviewed for demographic data, including patient age and sex, mechanism of injury, and operative reports. Surgical reports were correlated with CT findings.
Results Twelve patients with TDR were reviewed. Eleven of the 12 patients had TDR of the right hemidiaphragm confirmed at surgery. The CT diagnosis of rupture of the right hemidiaphragm was based on the evidence of liver herniation through the right diaphragm, seen on either axial or MPR images. Direct visualization of diaphragm discontinuity (torn edge) was present in seven of the 12 patients (58%), all of which were visualized on reformatted images with two seen directly on axial images (Fig. 1). The hump sign was seen in 10 cases (83%), all diagnosed on MPRs (Figs. 2–4). The band sign was present in four of the 12 cases (33%), two diagnosed on axial images and the other two on MPRs (Figs. 2 and 4). The collar sign was present in five of 12 patients (42%), the diagnoses made from axial images in two and by MPRs in the other three (Figs. 5 and 6). The dependent viscera sign was documented in four (33%) cases (Fig. 1). One patient had colon herniating through the torn right hemidiaphragm (Fig. 4), and one patient had fat alone herniating through the defect. The measured distance from the high point of the apparent right hemidiaphragm to the high point of the left hemidiaphragm (Fig. 7) ranged from 0–8.4 cm, with a mean of 5.01 cm in the patients with right hemidiaphragm injury. In the 50 blunt trauma patients without a known diaphragm injury (controls) the range of height differences was K2.8–5.1 cm with a mean of 1.75 cm. The variance of the groups was 4.625 for the patients with right diaphragm injury versus 2.348 for the control group. A t-statistic assuming unequal variance was significant at the pZ0.0001 level. Fig. 7 illustrates the distribution of heights of the apparent right hemidiaphragm in the 12 patients with ruptures versus 50 controls (patients who had sustained blunt trauma but who did not have clinical or imaging findings of diaphragm injury to the time of discharge). The result indicates that an elevation of the apparent peak of the right hemidiaphragm of 4 cm or more above the left on supine radiographs should be regarded as highly suspicious for right hemidiaphragm injury in the setting of blunt torso trauma. Associated abdominal and thoracic injuries were present in all 12 patients and are listed in Table 1.
1284
O. Rees et al.
Table 1 Summary of injuries concurrent with right hemidiaphragm rupture. Associated injury
Number (%)
Liver laceration Splenic laceration (Rt.) Adrenal haematoma Abdominal wall contusion Atelectasis Pneumothorax Haemothorax/effusion Pulmonary contusion Rib fractures Pelvic fractures Lumbar spine fractures Sacral fractures Lumbar hernia Aortic injury
6 2 2 2 9 3 4 4 4 3 4 2 1 1
(50%) (17%) (17%) (17%) (75%) (25%) (33%) (33%) (33%) (25%) (33%) (17%) (8%) (8%)
concurrent medical conditions the diagnosis was established at autopsy.
Discussion
Figure 3 The hump sign on coronal MPR and band sign on axial CT. (a) Axial CT of a 22-year-old man injured in motor vehicle collision shows apparent elevation of the right hemidiaphragm with a cleft (arrow) created by the indentation of the torn edge of diaphragm. (b) Coronal MPR clearly shows the right dome of the liver projecting into the lower right chest cavity and band sign (arrow).
Diaphragm rupture was surgically verified in all but one patient, but was described in the operative note in only five. The length of tear varied, but in all cases the overall length was greater than 10 cm. Two cases had ruptures that measured 10 cm; one case measured 15 cm; in one case the perforation consisted of “three separate ruptures forming a stellate-shaped perforation, with two 5 cm ruptures joining a smaller third rupture to form the perforation”; and in the final case there was a “complete tear of the diaphragm with no remnant either laterally or anteriorly”. In one patient for whom surgery was not performed due to severe
The preoperative diagnosis of diaphragm rupture is difficult not only due to non-specific or absent clinical signs, but also due to the relatively low sensitivity of chest radiographs in detecting the injury.2,4,12 With advances in CT technology and the advent of helical and multidetector CT, with resulting improved detail of coronal and sagittal reformations, the sensitivity and specificity of CT for the detection of diaphragm rupture has increased.14–16 However, the diagnosis of rightsided rupture has not improved to the same magnitude as left-sided rupture, mainly due to the difficulty in visualizing such an injury due to its close proximity to the liver.12,16 The CT signs of diaphragm rupture include discontinuity of the diaphragm, diaphragm thickening, herniation of abdominal organs or fat into the thorax, a waistlike constriction of herniating viscera (collar sign), and the dependent viscera sign.13–18 When confronted with a left-sided rupture the visualization of these signs may be more obvious due to an increased tendency for left-sided abdominal fat and viscera to herniate into the chest cavity, promoted by negative intra-pleural pressure, as well as the proximity of intra-abdominal and retroperitoneal fat to the undersurface of the left hemidiaphragm. However, with right-sided diaphragm tears the hemidiaphragm disruption itself is less likely to be detected directly due to inherent lack of contrast between the diaphragm and the adjacent liver. Second, the most common abdominal organ to herniate through the right hemidiaphragm, if any, is
Multidetector-row CT of right hemidiaphragmatic rupture caused by blunt trauma
1285
Figure 4 Right hemidiaphragm rupture with liver and colon herniation. (a) Admission supine chest radiograph shows marked elevation of the apparent right hemidiaphragm compared to left (horizontal lines). The colon also projects above the level of the left hemidiaphragm. (b) Coronal MPR confirms the marked elevation of the right lobe and hepatic flexure. (c) Dedicated sagittal reformation shows hump of liver created by the herniated right lobe with collar sign at its base (arrows).
the liver itself. Herniation of the liver requires both a large enough vent to allow a portion of the bulky liver to extend above the diaphragm and stretching or disruption of the ligamentous attachments of the liver to the body wall. All but one of the present patients had herniation of a significant part of the liver, typically the right lobe, protruding into the
right chest on the admission chest radiograph, producing the appearance of an apparently elevated hemidiaphragm radiographically or a hump, band, or collar sign by MDCT. Surgical findings, which were available for five patients, documented extensive tears of the hemidiaphragm as previously described.
1286
O. Rees et al.
Figure 6 Collar sign of herniated liver. Coronal MDCT reformation shows the herniation of a portion of the medial right lobe (L) through the adjacent torn diaphragm with constriction at the base of the herniated segment (arrow).
Figure 5 A collar sign of a herniated liver in a 34-yearold man injured in a vehicular collision. (a) Admission chest radiograph shows smooth right hemidiaphragm contour with no elevation relative to the dome of the left hemidiaphragm. (b) Coronal reformation from a dedicated diaphragm MDCT shows herniation of a portion of the right hepatic lobe into the right chest with indentation of the parenchyma at the torn edge of diaphragm (arrow).
Reports describing use of CT for diagnosing diaphragmatic rupture are limited, due to small populations, and have conflicting results.13–18 Killeen et al.16 showed that the sensitivity and specificity for detecting left-sided rupture was 78 and 100%, respectively, while the sensitivity and specificity for detecting right-sided rupture was 50 and 100%, respectively, using single-slice helical CT with coronal and sagittal MPRs. Killeen et al. reported that the use of reformations was an aid in the diagnosis of diaphragm rupture, increasing the sensitivity of right-sided ruptures from 16.7 to 50%.9 Larici et al.15 found that helical CT also performed better for left-sided injuries (100%
sensitivity) in comparison with right-sided injuries (76% sensitivity). However, they reported that the addition of coronal and sagittal reformations added little to the reviewers’ ability to diagnose the injury. MDCT with coronal and sagittal reformations proved quite valuable in the detection of rightsided hemidiaphragm rupture in this study. The present authors found several signs of right hemidiaphragm injury to be present on both coronal and sagittal reformations and either not present or very subtle on corresponding axial images. Previous studies have emphasized diaphragm discontinuity as a sensitive marker of diaphragm rupture. Murray et al.13 found this sign to be present in 73% of all patients with rupture hemidiaphragms, however, their study only included three right hemidiaphragm injuries, two of which had direct visualization of diaphragmatic discontinuity. Bergin et al.14 reported this finding in 80% of their cases and Worthy et al.18 in 71% of cases. In the present study, visualization of diaphragm discontinuity was present in seven (58%) of 12 cases, all noted on nonaxial reformations. Direct diaphragm disruption was less commonly found in the present study compared with other studies that included mainly left-sided ruptures because the present study focused exclusively on right-sided rupture. The most common sign found in the present study was the hump sign, an area of smooth, rounded herniation of liver tissue through the hemidiaphragm, less than the width of the hemidiaphragm, projecting into the thorax. This sign was
Multidetector-row CT of right hemidiaphragmatic rupture caused by blunt trauma
1287
Figure 7 Distribution of distance between the domes of the apparent right hemidiaphragm and left hemidiaphragm. Comparison of 12 patients diagnosed with right hemidiaphragm rupture (left-side group) and 50 controls with no imaging or clinical evidence of diaphragm injury (right-side group). Note that the white line at 4 cm separates all but two patients with the injury from 46 (92%) of 50 without.
present in 10 (83%) patients, and was mostly clearly demonstrated on coronal or sagittal reformations. This sign is often difficult to detect on axial images as it may be misinterpreted as simply a high dome of the hemidiaphragm itself, hence, reformatted images are necessary to fully appreciate the sign. The band sign, a horizontal portion of low density through the liver at the level of the torn diaphragm was viewed in four (33%) cases, two seen on MPRs and two on axial images. Although we are uncertain of the precise aetiology of the band sign, the authors believe it represents under-perfusion of liver parenchyma adjacent to the compressing edge of the torn diaphragm and relative diminished enhancement compared with the adjacent normally perfused liver. The hump and band signs are considered variations of the collar sign. The collar sign, constriction of the herniating viscera at the level of the edges of the torn hemidiaphragm was observed in five of the present 12 cases (42%). Reports concerning the prevalence of the collar sign in right diaphragm rupture are conflicting. Murray et al.13 failed to show the presence of a collar sign in their three cases of right hemidiaphragm rupture. However, Killeen et al.16 showed this sign to be the most sensitive sign of diaphragm rupture in their study, identified
in three (50%) of their six cases of right diaphragm rupture. It should be noted that Murray et al. did not use reformatted images in their case assessment, but Killeen et al. showed that the addition of reformatted images increased sensitivity of this sign from 16.7 to 50%.13,16 The collar sign was present in two cases on axial images in the present study, but the other cases required reformatted images to identify the sign. Hence, the addition of reformatted images increased the sensitivity of this sign from 17 to 42%. Two of the present cases showed signs of visceral herniation. In one case fat was visualized herniating through the diaphragm, whereas the other demonstrated both colon and abdominal fat herniation. The findings were present on both axial views and sagittal reformations. Although the liver is the most common abdominal organ to herniate through the right hemidiaphragm, it is recognized that other visceral organs may less commonly herniate.19 Murray et al.13 reported that one of three patients with right hemidiaphragm rupture had the right kidney and retroperitoneal fat herniating through the injured hemidiaphragm. Bergin et al.14 reported that one of six patients with right hemidiaphragm trauma had small bowel herniating through the right diaphragm defect. Herniation of the viscera
1288
through the ruptured diaphragm may not develop if patients are supported on positive-pressure ventilation as this eradicates negative intra-pleural pressure that normally helps pull abdominal contents into the chest. Right-sided diaphragm injuries from blunt impact are less common than left-sided injuries due to both the protective effects of the liver and the typically greater thickness of the right hemidiaphragm.10,12,19 Thus, it requires a greater force to disrupt the right hemidiaphragm leading to a high frequency of associated injuries.12,19 All of the present patients had associated abdominal or thoracic injuries as listed in Table 1. With the use of MDCT, thinner sections and better quality reformatted images can be produced. This is due to the reformatted images being compiled from very thin, often overlapping, axial sections and that several sections (four or 16) are acquired simultaneously rather than each section being obtained in succession. This reduces the problems posed by gross and respiratory motion during the study allowing smoother reformatted images to be produced. The authors found reformatted coronal and sagittal images to be an invaluable aid in establishing the diagnosis of diaphragm rupture. However, conflicting results in the literature have shown that some studies found reformatted images useful whereas other studies found that they did not improve the diagnostic accuracy at all.15,16 As previous studies used reformatted images produced by single-detector helical CT rather than MDCT, it is probable that the better quality images permitted by MDCT offer greater diagnostic value. An incidental observation made in this study by review of admission chest radiographs was that all but two of 12 patients with right hemidiaphragm rupture had elevation of the “apparent” dome of the right hemidiaphragm more than 4 cm above the top of the left hemidiaphragm. Although some institutions may routinely perform coronal and sagittal MPR images for all major trauma patients having chest or abdominal CT, the finding of an apparently elevated right diaphragm of 4 cm or more above the left mandates exclusion of right diaphragm rupture requiring MPR images orthogonal to the axial plane. Such an elevation of the hemidiaphragm alone cannot be considered diagnostic of right hemidiaphragm injury as it could also occur as an anatomic variant, secondary to right sub-pulmonic effusion, from congenital weakness of the diaphragm (eventration), a hemidiaphragm contusion, or phrenic nerve injury. The main limitations of the present study are the retrospective ascertainment of data and the limited sample size. The latter is a problem common to all
O. Rees et al.
studies investigating the use of CT in right diaphragm disruption due to the relative rarity of this injury. In conclusion, MDCT augmented with sagittal and coronal MPR is a valuable tool for the diagnosis of right hemidiaphragm from blunt trauma. MPRs are essential in demonstrating this injury as liver herniation through the diaphragm is difficult to detect using only axial images. Images produced using MDCT with a detector width of 1.5 mm or less allows more detailed imaging of the diaphragm and increases the chance of diagnosing diaphragm injury due both to the increased probability of obtaining an axial section directly through the site of the torn diaphragm and to the increased quality of coronal and sagittal reformations aiding in visualization of mild to moderate trans-diaphragmatic liver herniation, focal liver constriction of the liver, or the band sign along the edges of the torn diaphragm. A prospective study using these observations should be performed to verify the conclusions of the present study.
Acknowledgements The authors thank Dr Rao Gullipolli, Department of Radiology, University of Maryland School of Medicine, Physics Section.
References 1. Voeller GR, Reisser JR, Fabian TC, Kedsk K, Mangiante EC. Blunt diaphragm injures: A five-year experience. Am Surg 1989;56:28—31. 2. Meyers BF, McCabe CJ. Traumatic diaphragmatic hernia: Occult marker of serious injury. Ann Surg 1993;218:783—90. 3. Lee W, Chen R, Fang J, et al. Rupture of the diaphragm after blunt trauma. Eur J Surg 1994;160:479—83. 4. Gelman R, Mirvis SE, Gens D. Diaphragmatic rupture due to blunt trauma: Sensitivity of plain chest radiographs. AJR Am J Roentgenol 1991;156:51—7. 5. Pipkin NL, Hamit HF. Traumatic perforation of the diaphragm. South Med J 1988;81:1347—50. 6. Flancbaum L, Dauber M, Demas C, Boyarsky AH, Trooskin SZ. Early diagnosis and treatment of blunt diaphragmatic injury. Am Surg 1988;54:195—9. 7. Guth AA, Pachter HL, Kim U. Pitfalls in the diagnosis of blunt diaphragmatic injury. Am J Surg 1995;170:5—9. 8. Rodriguez-Morales G, Rodriguez A, Shatney CH. Acute rupture of the diaphragm in blunt trauma: Analysis of 60 patients. J Trauma 1986;26:438—44. 9. Kearney PA, Rouhana SW, Burney RE. Blunt rupture of the diaphragm: Mechanism, diagnosis, and treatment. Ann Emerg Med 1989;18:105—9. 10. Iochum S, Ludig T, Walter F, et al. Imaging of diaphragmatic injury: A diagnostic challenge. RadioGraphics 2002;22: S113—S6.
Multidetector-row CT of right hemidiaphragmatic rupture caused by blunt trauma
11. Shah R, Sabanathan S, Mearns AJ, Chodhury AK. Traumatic rupture of the diaphragm. Ann Thorac Surg 1995;60:1444—9. 12. Shanmuganathan K, Killeen K, Mirvis SE, White CS. Imaging of diaphragmatic injuries. J Thorac Imaging 2000;15:104—11. 13. Murray JG, Caoili E, Gruden JF, et al. Acute rupture of the diaphragm due to blunt trauma: Diagnostic sensitivity and specificity of CT. AJR Am J Roentgenol 1996;166:1035—9. 14. Bergin D, Ennis R, Keogh C, Fenlon HM, Murray JG. The “dependent viscera” sign in CT diagnosis of blunt traumatic diaphragmatic rupture. AJR Am J Roentgenol 2001;177: 1137—40. 15. Larici AR, Gotway MB, Litt HI, et al. Helical CT with sagittal and coronal reconstruction: Accuracy and detection of diaphragmatic injury. AJR Am J Roentgenol 2002;179:451—7.
1289
16. Killeen K, Mirvis SE, Shanmuganathan K. Helical CT of diaphragmatic rupture caused by blunt trauma. AJR Am J Roentgenol 1999;173:1611—6. 17. Mueller CF, Pendarvis RW. Traumatic injury of the diaphragm: Report of seven cases and extensive literature review. Emerg Radiol 1994;1:118—32. 18. Worthy SA, Kang EY, Hartman TE, et al. Diaphragmatic rupture: CT findings in 11 patients. Radiology 1995;194: 885—8. 19. Mirvis SE. Diagnostic imaging of thoracic trauma. In: Mirvis SE, Shanmuganathan K, editors. Imaging in trauma and critical care. 2nd ed. Philadelphia: WB Saunders; 2003. p. 297—368.
Multidetector-row CT of right hemidiaphragmatic rupture caused by blunt trauma: a review of 12 cases O. Reesa, S.E. Mirvisb,*, K. Shanmuganathanb a
University of Wales College of Medicine, Cardiff, UK; and bDepartment of Radiology and Maryland Shock Trauma Center, University of Maryland School of Medicine, Baltimore, MD, USA
Received 23 June 2004; received in revised form 31 May 2005; accepted 22 June 2005
KEYWORDS Diaphragm (injuries); Diaphragm (CT); Computed tomography; Clinical effectiveness; Trauma; Thorax (injuries)
AIM: To determine the usefulness of multidetector-row CT (MDCT) with multiplanar reformatted (MPR) images in the sagittal and coronal plane in diagnosing acute right hemidiaphragmatic rupture. MATERIALS AND METHODS: Twelve patients were identified who received chest and abdominal MDCT after major blunt trauma diagnosed with right diaphragmatic injury. Sagittal and coronal reformations were performed in all cases. The images were retrospectively reviewed by two experienced radiologists for signs of right diaphragm injury, such as direct diaphragm discontinuity, the “collar sign”, the “dependent viscera sign”, and intra-thoracic location of herniated abdominal contents. RESULTS: Of the 12 cases of right hemidiaphragm rupture, diaphragm discontinuity was seen in seven (58%) cases, the collar sign in five (42%), the dependent viscera sign in four (33%), and transdiaphragmatic herniation of the right colon and fat in another. Two variants of the collar sign were apparent on high-quality sagittal and coronal reformations. The first, termed the “hump sign”, describes a rounded portion of liver herniating through the diaphragm forming a hump-shaped mass, and the second, termed the “band sign,” is a linear lucency across the liver along the torn edges of the hemidiaphragm. The hump sign occurred in 10 (83%) patients and the band sign in four (33%). CONCLUSION: MDCT is very useful in the diagnosis of right hemidiaphragm injury caused by blunt trauma when sagittal and coronal reformatted images are obtained, and should allow more frequent preoperative diagnosis. Q 2005 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
Introduction Traumatic diaphragm rupture (TDR) is a serious injury resulting from blunt abdominal trauma, the majority of which are caused by motor vehicle collisions and vehicles striking pedestrians, and occurs in 0.5–8% of patients undergoing surgical exploration.1–11 Hemidiaphragm rupture more commonly involves the left side and occurs in 56–86% of * Guarantor and correspondent: S.E. Mirvis, Department of Radiology, 22 South Greene Street, Baltimore, MD, USA. Tel.: C1 410 328 8845; fax: C1 410 328 0641. E-mail address: [email protected] (S.E. Mirvis).
cases. Right-sided tears occur in 11–39% of cases, and bilateral and other sites account for 2.4–13.0% of cases.1–11 A higher percentage of right-sided injuries are reported in surgical series, because historically this injury has been far easier to diagnose by direct inspection than by imaging methods.1,8 There are several theories concerning the mechanism of diaphragm rupture. These include avulsion of the attachments of the diaphragm or shearing of the stretched membrane after right or left lateral impact to the chest wall, rib fracture fragments directly penetrating the diaphragm, and a sudden increase in intra-abdominal pressure
0009-9260/$ - see front matter Q 2005 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.crad.2005.06.013
Multidetector-row CT of right hemidiaphragmatic rupture caused by blunt trauma
throughout the abdomen with the relatively weak, unprotected left diaphragm tearing from the force.10,12 Experimental studies also suggest that the right hemidiaphragm is mechanically stronger than the left, and requires a larger force to disrupt it.12,18 The right hemidiaphragm is also protected from abdominal impact by the energy-absorbing liver.12,18 Associated injuries are common in traumatic rupture of the diaphragm due to the relatively large force required to disrupt the diaphragm concurrently damaging adjacent organs. Associated major injuries are reported in 52–100% of patients.1,2,6,8,10–12 The most commonly damaged intra-abdominal organs are the liver, reported as being disrupted in 93% of patients with right-sided injury, and spleen in 24% of those with left-sided injury.12 In all cases of diaphragm tears, splenic injuries occur in 27– 63%.2,6,10,12 Other commonly associated abdominal injuries include pelvic and renal injuries and associated intra-thoracic injuries include haemopneumothorax and multiple rib fractures.1,2,6,8,10–12 The preoperative diagnosis of TDR is often difficult due to a lack of specific clinical signs and the presence of distracting or more obvious serious concurrent injuries. The use of chest radiography for initial screening of TDR has a relatively low sensitivity for diagnosis, reported as diagnostic in 46% of patients with rupture of the left side, and suspicious for TDR in another 18%.4 The use of helical computed tomography (CT) with sagittal and coronal reformations has increased the detection rate of TDR, with sensitivities and specificities for overall detection of right and left ruptures reported between 61–90% and 77–100%, respectively.13–18 However, the sensitivity for diagnosis of rightsided rupture is not as high as for the left side, with a CT sensitivity and specificity of left hemidiaphragm rupture reported as 78–100% and 100%, respectively, and the sensitivity and specificity of right hemidiaphragm rupture reported as 50–83% and 100%, respectively.12,16 The main diagnostic CT signs of rupture of the diaphragm include direct visualization of diaphragm discontinuity, abdominal visceral herniation above the diaphragm, a waist-like constriction of herniating viscera, the so-called “collar sign”, direct contact of the herniated liver with the chest wall without intervening lung, the “dependent viscera” sign, and thickening of the hemidiaphragm from haemorrhage among others.10,12–18 Helical CT has become the imaging method of choice in the acute trauma setting, complemented by its ability to acquire high-quality coronal and sagittal reformations in contrast to axial CT images alone. While the sensitivity of left-sided rupture of the
1281
diaphragm has increased with the use of helical CT, the sensitivity of detection of right-sided diaphragm rupture has not increased to the same extent.12,16 The advent of MDCT has provided the capability to perform CT with smaller sections and the ability to produce high-quality, non-axial, two-dimensional multiplanar reformatted (MPR) and volumetric renderings. The purpose of this study was to review retrospectively findings of MDCT in the diagnosis of right-sided hemidiaphragm rupture resulting from blunt trauma. Herein, the authors describe two other signs, which are variants of the collar sign, seen on high-quality coronal and sagittal images. The first sign termed the “hump sign” consists of a portion of the liver herniating through the right diaphragm to form a hump-shaped herniation, clearly visible on coronal and sagittal reformations. The second sign is the “band sign”, a band-like lucency through the mass of the liver where it herniates through the torn diaphragm, again clearly visible on coronal and sagittal reformations. The accuracy of other commonly valued signs was also reviewed.
Materials and methods Permission for this study was obtained from our institutional review board and HIPAA compliance was maintained. A retrospective trauma registry and radiology database review was performed for patients diagnosed with right-sided hemidiaphragm injury caused by blunt trauma who underwent preoperative thoraco-abdominal MDCT at the authors’ Level 1 trauma center, for the period from May 2001 to November 2004. There were 12 such patients found, 11 of whom had surgical confirmation of the injury. The study group consisted of six men and six women, whose ages ranged from 28–86 years (mean 44 years). Mechanisms of injury consisted of motor-vehicle collisions in 11 and one pedestrian struck by a vehicle. All patients underwent a supine anteroposterior (AP) chest radiograph on admission, followed shortly by intravenous and oral contrast-enhanced MDCT (MX 8000, Philips Medical Systems, Best, The Netherlands) using a standardized MDCT protocol. Thoracic and abdominal MDCT examinations were initially performed using a four-detector row system (Philips MX 8000, Philips Medical Systems, Best, The Netherlands) obtaining 4!1.3 mm sections per rotation with a table speed of 10 mm/s. During the last 21 months of the study, abdominal– pelvic CT images were acquired using a 16-detector
1282
O. Rees et al.
row CT (MX 8000 IDT, Philips Medical Systems, Best, The Netherlands) using 16!1.5 detector width and 0.75 s rotation time. For both protocols, intravenous contrast material (300 mg I2/ml, 150 ml) was administered at 3 ml/s using a power injector (Medrad 4, Medrad, Pittsburg, PA, USA). A total volume of 600 ml of 2% sodium diatrizoate (Hypaque sodium, Nycomed, Princeton, NJ, USA) oral contrast material was ingested 30 min before and immediately prior to the examination whenever possible. For both protocols, axial images were reconstructed to 5 mm section thickness for initial review and PACS storage. When diaphragm injury was suspected based on the admission supine chest radiograph and axial CT images, the original data were reconstructed to 2 mm sections with 1 mm overlap. These axial images were used for coronal and sagittal reformations through the diaphragms. In eight patients, a dedicated 16-MDCT study was performed for the hemidiaphragms using 16! 0.5 mm sections and 0.75 s rotation time. For this study, the admission supine AP chest radiographs obtained before MDCT and axial MDCT images with the coronal and sagittal reformations were reviewed by two experienced trauma radiologists (S.E.M., 20 years experience and K.S., 14 years experience) and findings determined by consensus. The chest radiographs were reviewed and the distances between the domes of the apparent left and right diaphragms were recorded (in centimetres). The MDCT studies with MPRs were reviewed for signs of right hemidiaphragm injury. The images were considered diagnostic if any of the following signs were present: Direct visualization of a diaphragm defect, evidence of liver parenchymal herniation into the right hemithorax (hump sign), a band of lucency across the liver at the level of herniation through the diaphragm (band sign), herniated liver with focal constriction at the point where it the liver crosses the torn diaphragm (collar sign), and herniation of other viscera structures (such as bowel) through the right hemidiaphragm (with or without the collar sign), and the dependent viscera sign (liver in direct contact with the
Figure 1 Diaphragm discontinuity and dependent viscera sign. (a) Axial MDCT shows discontinuity in the
right hemidiaphragm (arrow) and thickening of the adjacent diaphragm due to haemorrhage. The liver (L) is in direct contact with the posterior chest wall through the torn diaphragm constituting the dependent viscera sign. (b) Dedicated sagittal diaphragm MPR shows a tear in the anterior right hemidiaphragm (arrow) and herniation of liver into the right hemithorax. There is an effusion (E) above the liver. (c) Coronal MPR shows torn edges of the right hemidiaphragm (arrows) and herniation of the right lobe into the right hemithorax. The gallbladder (G) is in the chest.
Multidetector-row CT of right hemidiaphragmatic rupture caused by blunt trauma
posterior chest wall). Thickening of the right hemidiaphragm or fluid in the adjacent retroperitoneal fat were also noted. Associated thoracic and abdominal injuries were recorded. The apparent height difference between the hemidiaphragms was measured from a supine chest radiograph for the 12 patients in the study and for 50 admission supine chest radiographs from patients admitted to our trauma center who had no radiographic or clinical evidence of diaphragm injury. The range, mean, and variance among the samples were calculated.
Figure 2 The hump sign of liver herniation. (a) Admission chest radiograph shows focal rounded smooth increased density at the medial right lung base (arrow). (b) Axial CT image at the right lung base elevation of the right hemidiaphragm with a subtle lucent line (“band”) across dome of liver (arrow) representing liver compression by the edge of the diaphragm. There are low attenuation lacerations also in the dome of the right lobe.
1283
The medical record of each patient was subsequently reviewed for demographic data, including patient age and sex, mechanism of injury, and operative reports. Surgical reports were correlated with CT findings.
Results Twelve patients with TDR were reviewed. Eleven of the 12 patients had TDR of the right hemidiaphragm confirmed at surgery. The CT diagnosis of rupture of the right hemidiaphragm was based on the evidence of liver herniation through the right diaphragm, seen on either axial or MPR images. Direct visualization of diaphragm discontinuity (torn edge) was present in seven of the 12 patients (58%), all of which were visualized on reformatted images with two seen directly on axial images (Fig. 1). The hump sign was seen in 10 cases (83%), all diagnosed on MPRs (Figs. 2–4). The band sign was present in four of the 12 cases (33%), two diagnosed on axial images and the other two on MPRs (Figs. 2 and 4). The collar sign was present in five of 12 patients (42%), the diagnoses made from axial images in two and by MPRs in the other three (Figs. 5 and 6). The dependent viscera sign was documented in four (33%) cases (Fig. 1). One patient had colon herniating through the torn right hemidiaphragm (Fig. 4), and one patient had fat alone herniating through the defect. The measured distance from the high point of the apparent right hemidiaphragm to the high point of the left hemidiaphragm (Fig. 7) ranged from 0–8.4 cm, with a mean of 5.01 cm in the patients with right hemidiaphragm injury. In the 50 blunt trauma patients without a known diaphragm injury (controls) the range of height differences was K2.8–5.1 cm with a mean of 1.75 cm. The variance of the groups was 4.625 for the patients with right diaphragm injury versus 2.348 for the control group. A t-statistic assuming unequal variance was significant at the pZ0.0001 level. Fig. 7 illustrates the distribution of heights of the apparent right hemidiaphragm in the 12 patients with ruptures versus 50 controls (patients who had sustained blunt trauma but who did not have clinical or imaging findings of diaphragm injury to the time of discharge). The result indicates that an elevation of the apparent peak of the right hemidiaphragm of 4 cm or more above the left on supine radiographs should be regarded as highly suspicious for right hemidiaphragm injury in the setting of blunt torso trauma. Associated abdominal and thoracic injuries were present in all 12 patients and are listed in Table 1.
1284
O. Rees et al.
Table 1 Summary of injuries concurrent with right hemidiaphragm rupture. Associated injury
Number (%)
Liver laceration Splenic laceration (Rt.) Adrenal haematoma Abdominal wall contusion Atelectasis Pneumothorax Haemothorax/effusion Pulmonary contusion Rib fractures Pelvic fractures Lumbar spine fractures Sacral fractures Lumbar hernia Aortic injury
6 2 2 2 9 3 4 4 4 3 4 2 1 1
(50%) (17%) (17%) (17%) (75%) (25%) (33%) (33%) (33%) (25%) (33%) (17%) (8%) (8%)
concurrent medical conditions the diagnosis was established at autopsy.
Discussion
Figure 3 The hump sign on coronal MPR and band sign on axial CT. (a) Axial CT of a 22-year-old man injured in motor vehicle collision shows apparent elevation of the right hemidiaphragm with a cleft (arrow) created by the indentation of the torn edge of diaphragm. (b) Coronal MPR clearly shows the right dome of the liver projecting into the lower right chest cavity and band sign (arrow).
Diaphragm rupture was surgically verified in all but one patient, but was described in the operative note in only five. The length of tear varied, but in all cases the overall length was greater than 10 cm. Two cases had ruptures that measured 10 cm; one case measured 15 cm; in one case the perforation consisted of “three separate ruptures forming a stellate-shaped perforation, with two 5 cm ruptures joining a smaller third rupture to form the perforation”; and in the final case there was a “complete tear of the diaphragm with no remnant either laterally or anteriorly”. In one patient for whom surgery was not performed due to severe
The preoperative diagnosis of diaphragm rupture is difficult not only due to non-specific or absent clinical signs, but also due to the relatively low sensitivity of chest radiographs in detecting the injury.2,4,12 With advances in CT technology and the advent of helical and multidetector CT, with resulting improved detail of coronal and sagittal reformations, the sensitivity and specificity of CT for the detection of diaphragm rupture has increased.14–16 However, the diagnosis of rightsided rupture has not improved to the same magnitude as left-sided rupture, mainly due to the difficulty in visualizing such an injury due to its close proximity to the liver.12,16 The CT signs of diaphragm rupture include discontinuity of the diaphragm, diaphragm thickening, herniation of abdominal organs or fat into the thorax, a waistlike constriction of herniating viscera (collar sign), and the dependent viscera sign.13–18 When confronted with a left-sided rupture the visualization of these signs may be more obvious due to an increased tendency for left-sided abdominal fat and viscera to herniate into the chest cavity, promoted by negative intra-pleural pressure, as well as the proximity of intra-abdominal and retroperitoneal fat to the undersurface of the left hemidiaphragm. However, with right-sided diaphragm tears the hemidiaphragm disruption itself is less likely to be detected directly due to inherent lack of contrast between the diaphragm and the adjacent liver. Second, the most common abdominal organ to herniate through the right hemidiaphragm, if any, is
Multidetector-row CT of right hemidiaphragmatic rupture caused by blunt trauma
1285
Figure 4 Right hemidiaphragm rupture with liver and colon herniation. (a) Admission supine chest radiograph shows marked elevation of the apparent right hemidiaphragm compared to left (horizontal lines). The colon also projects above the level of the left hemidiaphragm. (b) Coronal MPR confirms the marked elevation of the right lobe and hepatic flexure. (c) Dedicated sagittal reformation shows hump of liver created by the herniated right lobe with collar sign at its base (arrows).
the liver itself. Herniation of the liver requires both a large enough vent to allow a portion of the bulky liver to extend above the diaphragm and stretching or disruption of the ligamentous attachments of the liver to the body wall. All but one of the present patients had herniation of a significant part of the liver, typically the right lobe, protruding into the
right chest on the admission chest radiograph, producing the appearance of an apparently elevated hemidiaphragm radiographically or a hump, band, or collar sign by MDCT. Surgical findings, which were available for five patients, documented extensive tears of the hemidiaphragm as previously described.
1286
O. Rees et al.
Figure 6 Collar sign of herniated liver. Coronal MDCT reformation shows the herniation of a portion of the medial right lobe (L) through the adjacent torn diaphragm with constriction at the base of the herniated segment (arrow).
Figure 5 A collar sign of a herniated liver in a 34-yearold man injured in a vehicular collision. (a) Admission chest radiograph shows smooth right hemidiaphragm contour with no elevation relative to the dome of the left hemidiaphragm. (b) Coronal reformation from a dedicated diaphragm MDCT shows herniation of a portion of the right hepatic lobe into the right chest with indentation of the parenchyma at the torn edge of diaphragm (arrow).
Reports describing use of CT for diagnosing diaphragmatic rupture are limited, due to small populations, and have conflicting results.13–18 Killeen et al.16 showed that the sensitivity and specificity for detecting left-sided rupture was 78 and 100%, respectively, while the sensitivity and specificity for detecting right-sided rupture was 50 and 100%, respectively, using single-slice helical CT with coronal and sagittal MPRs. Killeen et al. reported that the use of reformations was an aid in the diagnosis of diaphragm rupture, increasing the sensitivity of right-sided ruptures from 16.7 to 50%.9 Larici et al.15 found that helical CT also performed better for left-sided injuries (100%
sensitivity) in comparison with right-sided injuries (76% sensitivity). However, they reported that the addition of coronal and sagittal reformations added little to the reviewers’ ability to diagnose the injury. MDCT with coronal and sagittal reformations proved quite valuable in the detection of rightsided hemidiaphragm rupture in this study. The present authors found several signs of right hemidiaphragm injury to be present on both coronal and sagittal reformations and either not present or very subtle on corresponding axial images. Previous studies have emphasized diaphragm discontinuity as a sensitive marker of diaphragm rupture. Murray et al.13 found this sign to be present in 73% of all patients with rupture hemidiaphragms, however, their study only included three right hemidiaphragm injuries, two of which had direct visualization of diaphragmatic discontinuity. Bergin et al.14 reported this finding in 80% of their cases and Worthy et al.18 in 71% of cases. In the present study, visualization of diaphragm discontinuity was present in seven (58%) of 12 cases, all noted on nonaxial reformations. Direct diaphragm disruption was less commonly found in the present study compared with other studies that included mainly left-sided ruptures because the present study focused exclusively on right-sided rupture. The most common sign found in the present study was the hump sign, an area of smooth, rounded herniation of liver tissue through the hemidiaphragm, less than the width of the hemidiaphragm, projecting into the thorax. This sign was
Multidetector-row CT of right hemidiaphragmatic rupture caused by blunt trauma
1287
Figure 7 Distribution of distance between the domes of the apparent right hemidiaphragm and left hemidiaphragm. Comparison of 12 patients diagnosed with right hemidiaphragm rupture (left-side group) and 50 controls with no imaging or clinical evidence of diaphragm injury (right-side group). Note that the white line at 4 cm separates all but two patients with the injury from 46 (92%) of 50 without.
present in 10 (83%) patients, and was mostly clearly demonstrated on coronal or sagittal reformations. This sign is often difficult to detect on axial images as it may be misinterpreted as simply a high dome of the hemidiaphragm itself, hence, reformatted images are necessary to fully appreciate the sign. The band sign, a horizontal portion of low density through the liver at the level of the torn diaphragm was viewed in four (33%) cases, two seen on MPRs and two on axial images. Although we are uncertain of the precise aetiology of the band sign, the authors believe it represents under-perfusion of liver parenchyma adjacent to the compressing edge of the torn diaphragm and relative diminished enhancement compared with the adjacent normally perfused liver. The hump and band signs are considered variations of the collar sign. The collar sign, constriction of the herniating viscera at the level of the edges of the torn hemidiaphragm was observed in five of the present 12 cases (42%). Reports concerning the prevalence of the collar sign in right diaphragm rupture are conflicting. Murray et al.13 failed to show the presence of a collar sign in their three cases of right hemidiaphragm rupture. However, Killeen et al.16 showed this sign to be the most sensitive sign of diaphragm rupture in their study, identified
in three (50%) of their six cases of right diaphragm rupture. It should be noted that Murray et al. did not use reformatted images in their case assessment, but Killeen et al. showed that the addition of reformatted images increased sensitivity of this sign from 16.7 to 50%.13,16 The collar sign was present in two cases on axial images in the present study, but the other cases required reformatted images to identify the sign. Hence, the addition of reformatted images increased the sensitivity of this sign from 17 to 42%. Two of the present cases showed signs of visceral herniation. In one case fat was visualized herniating through the diaphragm, whereas the other demonstrated both colon and abdominal fat herniation. The findings were present on both axial views and sagittal reformations. Although the liver is the most common abdominal organ to herniate through the right hemidiaphragm, it is recognized that other visceral organs may less commonly herniate.19 Murray et al.13 reported that one of three patients with right hemidiaphragm rupture had the right kidney and retroperitoneal fat herniating through the injured hemidiaphragm. Bergin et al.14 reported that one of six patients with right hemidiaphragm trauma had small bowel herniating through the right diaphragm defect. Herniation of the viscera
1288
through the ruptured diaphragm may not develop if patients are supported on positive-pressure ventilation as this eradicates negative intra-pleural pressure that normally helps pull abdominal contents into the chest. Right-sided diaphragm injuries from blunt impact are less common than left-sided injuries due to both the protective effects of the liver and the typically greater thickness of the right hemidiaphragm.10,12,19 Thus, it requires a greater force to disrupt the right hemidiaphragm leading to a high frequency of associated injuries.12,19 All of the present patients had associated abdominal or thoracic injuries as listed in Table 1. With the use of MDCT, thinner sections and better quality reformatted images can be produced. This is due to the reformatted images being compiled from very thin, often overlapping, axial sections and that several sections (four or 16) are acquired simultaneously rather than each section being obtained in succession. This reduces the problems posed by gross and respiratory motion during the study allowing smoother reformatted images to be produced. The authors found reformatted coronal and sagittal images to be an invaluable aid in establishing the diagnosis of diaphragm rupture. However, conflicting results in the literature have shown that some studies found reformatted images useful whereas other studies found that they did not improve the diagnostic accuracy at all.15,16 As previous studies used reformatted images produced by single-detector helical CT rather than MDCT, it is probable that the better quality images permitted by MDCT offer greater diagnostic value. An incidental observation made in this study by review of admission chest radiographs was that all but two of 12 patients with right hemidiaphragm rupture had elevation of the “apparent” dome of the right hemidiaphragm more than 4 cm above the top of the left hemidiaphragm. Although some institutions may routinely perform coronal and sagittal MPR images for all major trauma patients having chest or abdominal CT, the finding of an apparently elevated right diaphragm of 4 cm or more above the left mandates exclusion of right diaphragm rupture requiring MPR images orthogonal to the axial plane. Such an elevation of the hemidiaphragm alone cannot be considered diagnostic of right hemidiaphragm injury as it could also occur as an anatomic variant, secondary to right sub-pulmonic effusion, from congenital weakness of the diaphragm (eventration), a hemidiaphragm contusion, or phrenic nerve injury. The main limitations of the present study are the retrospective ascertainment of data and the limited sample size. The latter is a problem common to all
O. Rees et al.
studies investigating the use of CT in right diaphragm disruption due to the relative rarity of this injury. In conclusion, MDCT augmented with sagittal and coronal MPR is a valuable tool for the diagnosis of right hemidiaphragm from blunt trauma. MPRs are essential in demonstrating this injury as liver herniation through the diaphragm is difficult to detect using only axial images. Images produced using MDCT with a detector width of 1.5 mm or less allows more detailed imaging of the diaphragm and increases the chance of diagnosing diaphragm injury due both to the increased probability of obtaining an axial section directly through the site of the torn diaphragm and to the increased quality of coronal and sagittal reformations aiding in visualization of mild to moderate trans-diaphragmatic liver herniation, focal liver constriction of the liver, or the band sign along the edges of the torn diaphragm. A prospective study using these observations should be performed to verify the conclusions of the present study.
Acknowledgements The authors thank Dr Rao Gullipolli, Department of Radiology, University of Maryland School of Medicine, Physics Section.
References 1. Voeller GR, Reisser JR, Fabian TC, Kedsk K, Mangiante EC. Blunt diaphragm injures: A five-year experience. Am Surg 1989;56:28—31. 2. Meyers BF, McCabe CJ. Traumatic diaphragmatic hernia: Occult marker of serious injury. Ann Surg 1993;218:783—90. 3. Lee W, Chen R, Fang J, et al. Rupture of the diaphragm after blunt trauma. Eur J Surg 1994;160:479—83. 4. Gelman R, Mirvis SE, Gens D. Diaphragmatic rupture due to blunt trauma: Sensitivity of plain chest radiographs. AJR Am J Roentgenol 1991;156:51—7. 5. Pipkin NL, Hamit HF. Traumatic perforation of the diaphragm. South Med J 1988;81:1347—50. 6. Flancbaum L, Dauber M, Demas C, Boyarsky AH, Trooskin SZ. Early diagnosis and treatment of blunt diaphragmatic injury. Am Surg 1988;54:195—9. 7. Guth AA, Pachter HL, Kim U. Pitfalls in the diagnosis of blunt diaphragmatic injury. Am J Surg 1995;170:5—9. 8. Rodriguez-Morales G, Rodriguez A, Shatney CH. Acute rupture of the diaphragm in blunt trauma: Analysis of 60 patients. J Trauma 1986;26:438—44. 9. Kearney PA, Rouhana SW, Burney RE. Blunt rupture of the diaphragm: Mechanism, diagnosis, and treatment. Ann Emerg Med 1989;18:105—9. 10. Iochum S, Ludig T, Walter F, et al. Imaging of diaphragmatic injury: A diagnostic challenge. RadioGraphics 2002;22: S113—S6.
Multidetector-row CT of right hemidiaphragmatic rupture caused by blunt trauma
11. Shah R, Sabanathan S, Mearns AJ, Chodhury AK. Traumatic rupture of the diaphragm. Ann Thorac Surg 1995;60:1444—9. 12. Shanmuganathan K, Killeen K, Mirvis SE, White CS. Imaging of diaphragmatic injuries. J Thorac Imaging 2000;15:104—11. 13. Murray JG, Caoili E, Gruden JF, et al. Acute rupture of the diaphragm due to blunt trauma: Diagnostic sensitivity and specificity of CT. AJR Am J Roentgenol 1996;166:1035—9. 14. Bergin D, Ennis R, Keogh C, Fenlon HM, Murray JG. The “dependent viscera” sign in CT diagnosis of blunt traumatic diaphragmatic rupture. AJR Am J Roentgenol 2001;177: 1137—40. 15. Larici AR, Gotway MB, Litt HI, et al. Helical CT with sagittal and coronal reconstruction: Accuracy and detection of diaphragmatic injury. AJR Am J Roentgenol 2002;179:451—7.
1289
16. Killeen K, Mirvis SE, Shanmuganathan K. Helical CT of diaphragmatic rupture caused by blunt trauma. AJR Am J Roentgenol 1999;173:1611—6. 17. Mueller CF, Pendarvis RW. Traumatic injury of the diaphragm: Report of seven cases and extensive literature review. Emerg Radiol 1994;1:118—32. 18. Worthy SA, Kang EY, Hartman TE, et al. Diaphragmatic rupture: CT findings in 11 patients. Radiology 1995;194: 885—8. 19. Mirvis SE. Diagnostic imaging of thoracic trauma. In: Mirvis SE, Shanmuganathan K, editors. Imaging in trauma and critical care. 2nd ed. Philadelphia: WB Saunders; 2003. p. 297—368.