| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Emergency Radiology |
1 From the Departments of Diagnostic Radiology (P.A.P., S.E.M., K.S., K.L.K.) and Interventional Radiology (D.C.), University of Maryland Medical Center and Shock Trauma Center, 22 S Greene St, Baltimore, MD 21201; and the Department of Radiology, Division of Diagnostic and Interventional Radiology, University Hospital of Geneva, Switzerland (P.A.P.). Received July 15, 1999; revision requested September 21; revision received December 29; accepted January 12, 2000. Address correspondence to S.E.M. (e-mail: smirvis@rad1.ummc.umaryland.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Seventy-two patients with blunt liver injury underwent CT and hepatic angiography. Hepatic injuries were graded with CT-based classification. Scans were assessed for evidence of contrast extravasation and laceration or contusion extending into the hepatic vein(s), inferior vena cava, porta hepatis, or gallbladder fossa. Medical, angiographic, and surgical records were reviewed to determine angiographic findings, surgical indications and findings, and outcomes.
RESULTS: Compared with hepatic angiography, CT was 65% (11 of 17 patients) sensitive and 85% (41 of 48 patients) specific for detection of arterial vascular injury. When CT severity grades 2 and 3 were analyzed, the sensitivity and specificity of CT were 100% (three of three patients) and 94% (34 of 36 patients), respectively (P < .001). Injury involving at least one major hepatic vein was found in 15 (88%) of 17 patients who required liver-related surgery and in 23 (42%) of 55 of the other patients (P < .01).
CONCLUSION: CT-based criteria, including hepatic injury grade, signs of arterial vascular injury, and presence or absence of major hepatic venous involvement assists in selecting patients for hepatic angiography and those at increased risk of ongoing or delayed hepatic bleeding or other posttraumatic complications.
Index terms: Liver, angiography, 761.1242 • Liver, CT, 761.12112 • Liver, hemorrhage, 952.7192 • Liver, injuries, 761.41, 952.412
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
No specific criteria have been clearly established to determine, particularly on the basis of initial imaging examination results, which patients are prone to have failed nonsurgical management or to develop delayed complications from hepatic injury. There has been an evolution in our understanding of the importance of computed tomographic (CT) findings and clinical examination results in patients who have undergone hepatic trauma. The quantity of hemoperitoneum depicted by initial CT was initially considered to be an indicator of hepatic trauma severity (8,9). However, the results of several subsequent studies (6,10–12) indicate that the quantity of hemoperitoneum does not correlate with failed nonsurgical management.
A CT-based grading system has been adapted from the American Association for the Surgery of Trauma classification of blunt hepatic injury (Table 1). However, the direct application of such a CT classification, although reflective of the extent of parenchymal liver damage, cannot reliably predict the need for angiographic assessment of the liver or the probable clinical outcome of attempted nonsurgical management (6,12). Even major hepatic injuries with a severity of up to CT grade 4 typically can be managed without surgery in those patients who maintain hemodynamic stability (6,12–18). Some authors have described wide discrepancies between the CT injury grade and the injury severity determined at surgery (19), with CT generally yielding an underestimation of the extent of injury. However, the advent of spiral CT and improvements in image quality have led to an increasing role of and reliance on CT for evaluating acute traumatic hepatic lesions (6,7,20–25).
|
Intrahepatic vascular injuries have been reported more frequently in association with liver injuries of a higher CT grade than in association with those of a lower CT grade (4). Some authors (4,29) advocate performing mandatory hepatic angiography in all patients with hepatic injuries of CT grade 3 or higher to avoid the risk of missing arterial bleeding at CT. In the present study, our aim was to further determine the value of CT for assisting in decisions regarding the treatment of hemodynamically stable patients with blunt hepatic trauma. The accuracy of CT in depicting hepatic arterial hemorrhage was determined by comparing the CT findings with the results of angiography and surgery.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
All initial abdominal CT scans were obtained within 24 hours after admission—typically in less than 2 hours. The CT scans were obtained by using a Somaton Plus 4 unit (Siemens Medical Systems, Iselin, NJ). Scanning was routinely performed with intravenous contrast enhancement by using a power-injected bolus of 150 mL of 240 mg of iodine per milliliter injected at a rate of 3 mL/sec. A uniphasic injection, with a scanning delay of 60 seconds from the time of initiation of the intravenous injection of contrast material, was used. Whenever possible and on the basis of clinical circumstance, oral contrast material (diatrizoate sodium [1% Hypaque]; Sanofi Winthrop, Morrisville, Pa) was administered: A 450-mL dose was administered 30 minutes before scanning, and an additional 450 mL was administered immediately before scanning. CT was performed from the lung bases to the pelvis with 8-mm contiguous sections, a table speed of 8 mm/sec, and a pitch of 1.
Digital subtraction angiography of the liver was performed by using a digital angiographic system (Toshiba America Medical Systems, Tustin, Calif) with 1,024 x 1,024 resolution. An initial anteroposterior celiac arteriogram was obtained to demonstrate the overall anatomy. Additional anteroposterior and oblique projections of the liver were obtained by using selective catheterization of the common hepatic artery. The indications for hepatic angiography included confirmation of and potential embolization for CT signs of contrast material extravasation (ie, CT blush) in hemodynamically stable patients. Hepatic angiography was performed also to exclude hepatic arterial injury in patients with CT evidence of liver injury without direct CT findings of vascular injury who had unexplained transient hypotension (ie, peak systolic pressure equal to or below 100 mm Hg). Hepatic angiography was performed within 12 hours after CT in 59 patients and within 24 hours after CT in 11 patients; it was delayed in two patients for 4 and 7 days after CT.
The CT images were reviewed and interpreted by means of consensus by three radiologists (P.A.P., K.S., S.E.M.) on the basis of a predetermined protocol without knowledge of the angiographic or surgical findings. This protocol included an initial grading of the lesion by using the CT grades of injury severity for blunt hepatic trauma (Table 1) established by Mirvis et al (13). The CT studies were analyzed specifically for evidence of injury extension into the porta hepatis, major hepatic veins, inferior vena cava, and/or gallbladder fossa. Any of these structures that were in direct contact with the hepatic parenchymal injury were considered to be "involved" by the injury. Concurrent extrahepatic intraabdominal injuries also were recorded.
Extravasation of contrast material–enhanced blood into the hepatic parenchyma was identified as a focal round or oval area with a CT attenuation within 10 HU of the attenuation of an adjacent artery. Focal intrahepatic bleeding was differentiated from an adjacent contrast-enhanced artery by means of consensus on the basis of the size of the focal enhancement in relation to the hepatic vessels at the same distance from the hilum and the appearance of the focal enhancement on sequential transverse images. The foci of contrast material leakage did not follow a contiguous (ie, vascular) course on the subsequent images (Figs 1, 2). Subcapsular or extrahepatic bleeding appeared as an irregular blush of contrast material extending from the liver periphery into a surrounding hematoma.
|
|
|
The medical and surgical records of all the patients, as well as the radiologic reports of the 47 patients who underwent follow-up CT, were examined to determine the outcome of surgical or nonsurgical management and the prevalence and types of liver-related complications that occurred. The surgical report for each patient who underwent surgery was reviewed to determine the indication or indications for surgery, the presence and location of any bleeding site or "oozing," and whether surgical treatment (ie, packing, suturing, and/or resection) was required.
Our study data were analyzed to determine the value of admission CT in predicting the need for hepatic angiography and the potential for early and late complications among all grades of blunt liver injury. To achieve this goal, the following factors were assessed: (a) the association between CT injury grade and injury to specific anatomic sites; (b) the relationship between specific anatomic sites of hepatic injury at CT and angiographic findings, need for surgery, or failed nonsurgical management; (c) the sensitivity, specificity, negative and positive predictive values, and accuracy of CT findings of vascular injury with angiography and surgery as the reference-standard methods; (d) the clinical outcome versus initial treatment (ie, early surgery, angiographic intervention, or observation); and (e) the relationship between delayed hepatic trauma complications that occurred more than 10 days after admission and initial CT findings.
Statistical Analyses
Each CT criterion was compared with the angiographic and surgical results in two by two tables by using statistical software (Stata, College Station, Tex). The 
2 or Fisher exact test was used, when appropriate, to evaluate the univariate association between the tested parameters. A P value of less than .05 was considered to be indicative of a statistically significant difference between two different sample populations.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
|
|
|
|
Hepatic Arterial Contrast Material Extravasation at CT
Hepatic arterial injury with contrast material extravasation was suspected at admission CT in 22 (31%) of the 72 study patients, including six patients with CT grade 3 injury (Fig 1), 15 with grade 4 injury (Fig 2), and one with grade 5 injury (Table 3). Seven of the 72 patients—six with CT grade 4 injury, and one with grade 3 injury—became hemodynamically unstable and underwent hepatic surgery before angiography could be performed. Six of the patients had hepatic hemorrhage at surgery, and an additional patient with a positive CT scan was found to have a huge right retroperitoneal hematoma without active bleeding from the contused left hepatic lobe. All of these patients survived and did not require further treatment for hepatic injury. Seven of 18 hemodynamically stable patients with a CT scan positive for bleeding or pseudoaneurysm—six with CT grade 4 injury and one with grade 3 injury—had a normal hepatic angiogram.
In total, the CT scans of eight (36%) of the 22 patients suspected of having hepatic arterial injury at CT were considered to be false-positive for arterial bleeding (Table 5). All eight studies were rereviewed by the consensus panel with knowledge of the angiographic findings to determine the probable reason for the false-positive result. In two patients, the CT finding probably did represent active bleeding, which retrospectively was found to be extrahepatic in origin and thus did not originate from a branch of the hepatic artery (Fig 4). In another patient, the common hepatic artery was not selectively catheterized for anatomic reasons, and, therefore, the angiographic study was suboptimal because the contrast material injection was limited to the celiac trunk. However, because the angiographic examination was considered to be a reference standard for the present study and because the patient was treated successfully without surgery, that CT study also was considered to be false-positive. Four other false-positive bleeding sites diagnosed by using CT were retrospectively found to represent areas of normal hepatic parenchyma surrounded by contused (ie, low-attenuating) parenchyma (Fig 5). In another patient, the structure incorrectly diagnosed as a pseudoaneurysm probably was an intact hepatic vessel traversing an area of injury with lower attenuation (Fig 6). Careful examination of serial contiguous CT sections was required to establish the true nature of this focally enhancing area.
|
|
|
|
|
|
|
|
Seven (44%) of the 16 initially hemodynamically stable patients with both CT grade 4 hepatic injury and major hepatic venous involvement underwent surgery—five (31%) for liver-related complications—but the majority (n = 9) of these patients were successfully treated without surgery.
Angiographic Results versus Outcomes
In 49 of the 65 patients who were initially determined to be hemodynamically stable, the hepatic angiogram was negative for signs of vascular injury. In three of these patients, the nonsurgical management failed because of ongoing bleeding that was probably of venous origin (Fig 9). In all three cases, there was hepatic venous involvement at CT. However, none of 30 patients with negative angiograms and no hepatic venous involvement at CT demonstrated active hepatic bleeding (Fig 9).
|
Remote Complications and Mortality
Among the 72 patients, 69 survived more than 24 hours; 15 (22%) of these patients had one or more delayed complications—that is, those occurring 10 days or more after admission—which were diagnosed at follow-up CT. Because ongoing or recurrent hepatic bleeding was always diagnosed within 24 hours after admission, it was not considered to be a late complication. One or more delayed complications occurred in seven (47%) of 15 patients who underwent surgery within 10 days after admission—for liver bleeding in six patients and for colon tear in one—and in eight (16%) of 49 surviving patients who did not undergo surgery within 10 days after admission. The complications included biloma in seven patients, intraabdominal infection in six, liver abscess in one, mesenteric abscess in one, intraperitoneal bile leak in four, hepatobiliary fistula in two, and cholecystitis in two. Biloma was the most frequently encountered complication in the group of patients who did not undergo prior surgery (n = 6 [12%]), and infections were most common in the group that did undergo surgery (n = 4 [27%]).
Hepatic venous involvement was more frequently encountered in the 15 patients with delayed complications (n = 12 [80%]) than in the 54 patients without liver-related complications (n = 24 [44%]) (P < .01). In the present study, delayed complications were encountered in the patients with high CT grade injuries (ie, grades 3 and 4). Failed nonsurgical management was more frequent with grade 3 and grade 4 injuries than with grade 2 injuries (P = .339) and more frequent with grade 4 injuries than with grade 3 injuries (P = .486), but these observations were not statistically significant (Table 4). A comparison among the other CT injury grade groups was not possible because of the limited sample size.
In summary, surgery was ultimately required in 15 (23%) of the 65 patients who were treated initially with nonsurgical management (Fig 8). Forty-nine (75%) of these patients survived and left the hospital without undergoing a surgical procedure. Fifty-one (80%) of 64 patients were treated successfully with nonsurgical management (surgery for cholecystitis was not considered a failure of nonsurgical management). Three (4%) deaths were directly or indirectly related to the liver trauma. One patient, who had grade 3 injury, had persistent bleeding after surgery; another patient, who had grade 4 injury, died because of infection of a large segment of necrotic liver after the initially attempted nonsurgical management; and one patient, who had grade 5 injury, had a cardiac arrest during surgery for hepatic hemorrhage.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
In 1989, Mirvis et al (13) proposed a CT liver injury classification system that is partly based on the American Association for the Surgery of Trauma system for grading the severity of hepatic injuries (35) (Table 1). This classification has been shown to be of value in defining the severity of liver damage and, to a certain extent, helping to predict the outcome of initial nonsurgical management of lower grade hepatic injuries (ie, CT grades 1 and 2) (29). However, no specific CT-based criteria have been found to reliably predict which patients who have higher CT grades of hepatic injury and are treated nonsurgically will most likely have complications, particularly delayed or recurrent hepatic hemorrhage.
To assess the value of admission CT in demonstrating hepatic vascular injury and in predicting patient outcome, we compared admission abdominal CT findings with the hepatic angiographic results, surgical findings, and clinical course. Our goal was to determine whether there were any CT signs that might indicate increased likelihood of delayed hepatic injury complications. One major problem in meeting this goal was the determination of a reference standard (ie, angiographic and surgical findings) for the detection of arterial and venous hepatic bleeding to compare with the CT results. Hepatic bleeding could either cease or begin during the interval from CT to angiography or surgery and thereby cause overestimation or underestimation of the accuracy of the CT signs of active bleeding. In this study, which was performed at a major trauma center, the study time between CT and angiography or surgery was minimized because of the close proximity of the CT scanner (24-hour operation), angiography suite, and operating rooms to the patient admission area.
Another potential limitation of this retrospective study was the unavoidable selection of only those patients who underwent hepatic CT and angiography. Thus, the population was biased toward the selection of more severely injured patients because angiography was rarely performed in patients with CT grade 1 and grade 2 injuries. Therefore, the success rate of nonsurgical management of hepatic trauma appeared to have been lower than is typically reported owing to the exclusion of minor injuries from the analysis.
The CT scans obtained in patients with hepatic trauma were specifically assessed for extension of injury (ie, laceration, contusion, hematoma) into the inferior vena cava, hepatic veins, porta hepatis, and gallbladder fossa (Table 3). Injuries involving one or more major hepatic veins were 3.5 times more frequently associated with major hepatic bleeding than were injuries without major hepatic venous involvement. Our study data suggest that the CT finding of major hepatic venous involvement by injury is predictive of hepatic hemorrhage of arterial origin and probably of venous origin as well, although confirmatory hepatic venography was not performed in our study. Study findings indicate that extension of traumatic hepatic lesions into one or more major hepatic veins at CT should be considered an additional marker for severity of injury. This finding assists in determining which patients, within a specific CT injury scale group, are more prone to persistent or recurrent hepatic bleeding and other delayed complications. Conversely, the absence of major hepatic venous involvement suggests a more benign course of traumatic hepatic injury, which supports the selection of nonsurgical management of the hepatic injury if permitted according to the overall clinical picture.
In the current series, liver-related surgical management was 6.5 times more frequently required when the laceration extended into one or more hepatic veins than when it did not (Table 3). No other statistically significant correlation with patient outcome could be established with regard to involvement of the inferior vena cava, porta hepatis, or gallbladder. Major hepatic venous involvement along with the CT grade should not be considered firm criteria for determining whether a hemodynamically stable patient will require surgery. In the present study, nine (56%) of 16 hemodynamically stable patients with both CT grade 4 hepatic injury and major hepatic venous involvement were successfully treated without surgery. Thus, even in this high-risk group, hemodynamically stable patients can be successfully treated nonsurgically.
Another important finding of this study was the apparent correlation between CT grade and occurrence of hepatic bleeding (Table 3), failed nonsurgical management (Table 4), and remote complications. On the basis of the surgical American Association for the Surgery of Trauma classification of hepatic injury, it has been suggested that there are no differences in admission characteristics between patients who have successful and those who have failed nonsurgical management, except in admission systolic blood pressure (21). The results of another study indicate that CT injury grade is of no help in predicting the outcome of nonsurgical management (6). The data from our study do not support these findings. In the present study, CT grade 4 hepatic lacerations were more prone to bleed (P = .008) (Table 3) and lead to remote complications (P = .003) than CT grade 3 lesions. Failed nonsurgical management was more frequent among the grade 3 and grade 4 liver injuries than among the grade 2 injuries and more frequent in grade 4 injuries than in grade 3 injuries (Table 4).
The second goal of the current study was to determine the accuracy of CT in helping to diagnose hepatic vascular injury in the form of free contrast material extravasation by using both angiographic and surgical findings as the reference standards. Hepatic arterial bleeding or pseudoaneurysm was diagnosed by using CT in 22 (31%) of the 72 patients and confirmed in 14 (64%) of 22 of them by using angiography (Table 5). Eight (36%) of all 22 CT scans that were positive for active bleeding were considered to be false-positive for active bleeding or pseudoaneurysm when compared with angiograms. When confronted with a CT finding that is suggestive of vascular injury in a hemodynamically stable patient, the next logical step is to perform hepatic angiography. No hemodynamically stable patient in our series underwent surgery on the basis of only a positive CT scan. Thus, a hepatic CT scan that indicates vascular injury should be confirmed at angiography to mitigate the consequences of a false-positive CT scan.
A CT scan that is false-negative for vascular injury, however, can have more problematic consequences, because it can delay proper intervention until the patient demonstrates a falling hematocrit level or becomes hemodynamically unstable. All the CT studies that were false-negative for active arterial bleeding when compared with the angiograms were cases of CT grade 4 injury and had at least one hepatic vein involved by the hepatic parenchymal injury. This result suggests that to diagnose and control any arterial bleeding, hepatic angiography should be performed in any hemodynamically stable patient with CT grade 4 liver injury when there is associated major hepatic venous involvement, even without direct CT evidence of hepatic vascular injury (Fig 9).
Conversely, hepatic angiographic examination does not appear to be warranted in the absence of active bleeding at CT among patients with CT grade 2 and grade 3 injuries. Patients with major hepatic venous involvement by CT grade 3 or lower injury appear to be at higher risk for substantial venous bleeding (Fig 9) and should therefore be clinically observed. Because angiography is not reliable in depicting hepatic venous bleeding, it is not warranted following CT in such patients. Whether there is need to proceed to hepatic angiography for all CT grade 4 or 5 hepatic injuries, regardless of associated hepatic venous involvement, is still unresolved owing to the relatively small population size and limited statistical basis of the results of the current study. There were only seven initially hemodynamically stable patients with CT grade 2 hepatic injury in the present study, so it was difficult to analyze this small group. These observations require confirmation by means of prospective studies.
Compared with angiography alone, CT was 65% (11 of 17 cases) sensitive for depicting arterial bleeding and had a specificity of 85% (41 of 48). Compared with both angiographic and surgical results, the sensitivity of CT findings for indicating active hepatic bleeding fell to 56% (14 of 25 cases) without a change in specificity (Table 5). This decrease in sensitivity most likely resulted from hepatic venous hemorrhage that was detected at surgery but not at hepatic angiography.
How should we proceed in the assessment of the hemodynamically stable patient with blunt hepatic injury? The results of this study indicate that major hepatic venous involvement along with CT injury grade can help to determine which patients with blunt hepatic trauma may be more prone to ongoing hepatic bleeding or to the development of remote complications. The presence of a vascular lesion of any CT grade or a CT grade 4 or 5 injury associated with major hepatic venous involvement, even without vascular injury, indicates the need for hepatic angiography. The role of hepatic angiography for patients who have CT grade 4 injury without vascular injury or major hepatic venous involvement is not clear from our study data.
In the present study, nonsurgical management was successful in 52 (80%) of 65 patients in whom it was attempted. This result is somewhat poorer than that in other published series (1,5–7,36). However, because in our study we selected only those patients who underwent both abdominal CT and angiography, we most likely selected those patients with higher grades of hepatic trauma and did not include the lower risk patients—that is, those with grade 1 or 2 injury—who, in most cases, did not undergo hepatic angiography. Therefore, the actual percentage of successful nonsurgical management in this study group is probably substantially underestimated owing to the limited inclusion of patients with low CT grade hepatic injury.
In conclusion, our data indicate that CT-based criteria can be used to guide the diagnostic management of blunt hepatic trauma in hemodynamically stable patients. Such criteria, including CT grade of hepatic injury, CT evidence of arterial vascular injury, and presence or absence of hepatic venous involvement within the hepatic injury, can help in the selection of patients who should undergo hepatic angiography and possibly embolization. These criteria appear to be useful in identifying high-risk patients—that is, those prone to persistent or delayed hepatic bleeding or who may develop delayed complications and thus need closer observation and CT follow-up.
If supported by further studies, our observations should help in adapting the current CT-based injury classifications to improve their usefulness in selecting patients for initial nonsurgical management of blunt hepatic injury.
|
FOOTNOTES |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. Lubner, C. Menias, C. Rucker, S. Bhalla, C. M. Peterson, L. Wang, and B. Gratz Blood in the Belly: CT Findings of Hemoperitoneum RadioGraphics, January 1, 2007; 27(1): 109 - 125. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Rieger, A. Mallouhi, T. Tauscher, M. Lutz, and W. R. Jaschke Traumatic Arterial Injuries of the Extremities: Initial Evaluation with MDCT Angiography. Am. J. Roentgenol., March 1, 2006; 186(3): 656 - 664. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 O. Catalano, R. Lobianco, M. M. Raso, and A. Siani Blunt Hepatic Trauma: Evaluation With Contrast-Enhanced Sonography: Sonographic Findings and Clinical Application J. Ultrasound Med., March 1, 2005; 24(3): 299 - 310. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. Yoon, Y. Y. Jeong, J. K. Kim, J. J. Seo, H. S. Lim, S. S. Shin, J. C. Kim, S. W. Jeong, J. G. Park, and H. K. Kang CT in Blunt Liver Trauma RadioGraphics, January 1, 2005; 25(1): 87 - 104. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. A. Poletti, K. Kinkel, B. Vermeulen, F. Irmay, P.-F. Unger, and F. Terrier Blunt Abdominal Trauma: Should US Be Used to Detect Both Free Fluid and Organ Injuries? Radiology, April 1, 2003; 227(1): 95 - 103. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Kawashima, C. M. Sandler, F. M. Corl, O. C. West, E. P. Tamm, E. K. Fishman, and S. M. Goldman Imaging of Renal Trauma: A Comprehensive Review RadioGraphics, May 1, 2001; 21(3): 557 - 574. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||
