�#{149}11,

Review Article

. � � � � : � ‘ �‘ � , . � �

: � #{149}-� ‘I-� � � � 4L �

‘I� r� � , � � �‘:��-)‘�-‘ k� � � � �

� J � �v � �#{149}� 1!� � � 1----

� } � -rl� �

.p � I � � I,

� � � �.. � � � �

-.� � �J11��i� � � -� !) � � �

� I � � � � �.1.’ -‘

1#{149} � � � �IL!!.� �ji� � � � � f

�r � I �

-� r � ; �,-.‘ ‘ � I

. � ;.� � T �..‘ � � #{149} �

r � ; � ; tk L� � f�. � l!�II±�41Fl��r -,�T

Blunt Upper Abdominal Trauma: Evaluation by CT � �Neil T Wolfman � Robert E Bechtold � EricSSch&ling1andJ Wayne Meredith2

CT is the technique of choice for initial examination ot hemo- �. extent of injury to . the head, abdomen, chest, pelvis, anddynamically stable patients after blunt abdominal trauma. it is �r skeleton [2, 5-1 1 ]. �

highly sensitive, specific, and accurate for use in detecting the � � � CT examination of the abdomen has largely replaced otherpresence or absence of injury and defining its extent. Nonoper- : ‘; imaging techniques in the evaluation of hemodynamicallyative management of many posttraurnaticinjunes, particularly in � � stable patients. It is perlormed in patients in whom the ab-the liver, spleen, and kidney, is possible in part because of the - . T . .

diagnostic usefulness of CT. CT can be used effectively to visu- . � domen cannot be evaluated adequately by clinical examina-alize the progression of liver and spleen injuries in those patients tion b��US� of altered mental status (e.g., ethanol or drugchosen for conservative management. CT helps in treatment #{149} � abuse, head injury), in those in whom findings on clinicaldecisions in patients with renal injury by defining the character � � abdominal examination are equivocal, and in those with sig-and extent and distinguishing minor from severe renal trauma. � :� nificant pelvic fractures or history of multiple operations [2, 5,Posttraumatic injuries to the pancreas, bowel, and mesenterycan � -� 6, 8, 9, 1 2]. CT is not performed in hemodynamically unstablebe detected with CT. In these areas, however, signs may be J patients and in patients with obvious signs of peritonitis whosubtle and a significant injury may be missed on an initial require immediate surgery For rapid evaluation of the extentexamination � � �of abdominal injury diagnostic pentoneal lavage (DPL) can be

J � � �- *j� � � I� t�&fOfTfl�d either immediately before or dunng exploratory�,- � I. ��surgery

Trauma is the third leading cause of death in the. United � � DPL is very sensitive for detecting hemorrhage, is quickStates and the leading cause of death in person#{233}less than�T�and simple to perform, and does not require sophisticated40 years old [1 , 2]. It is currently one of America’s costliest � � � equipment. However, the test cannot differentiate inconse-health problems [3, 4]. Two important advances in the treat- � � � quential from significant bleeding [2, 13], resulting in unnec-ment of trauma patients in the past decade have been the � � essary laparotomies in 6-25% of cases [6, 7, 14]. DPL alsocreation and development of the Emergency Medical Service � cannot show the location or extent of injury. Furthermore, itsystem and the widespread use of CT to examine patients � does not show injuries of the retropentoneum, and thus is[2]. The Emergency Medical Service system defines the level T insensitive to trauma to the pancreas, the kidneys, and theof trauma care available at individual hospitals and identifies � � � retroperitoneal portion of the duodenum. Traumatic cannuladedicated trauma centers (levels I and II centers) for desig- . insertion, rents in the pentoneal membrane that allow bloodnated geographic areas. Referring trauma victims to such � from a retropentoneal injury to enter the abdominal cavity,centers for treatment results in reduced morbidity and mor- �. and pelvic fractures can result in false-positive examinationstality rates [2]. CT has become an integral part of the clinical �� � [7, 15].evaluation of traumatized patients because of its high sensi- � � � CT has replaced DPL in many centers for the initial evalu-tivity, specificity, and accuracy in detecting the presence and � � ation of suspected abdominal injury in stable patients [2, 7,

Received August 15, 1991 ; accepted after revision Octob�r 14, 1991 � � � � #{149}� . �

I Department of Radiology, Bowman Gray School of Medicine, Wake Forest University, Medical Center Blvd., Winston-Salem, NC 27157-1088. Address repnntrequests to N. T. Woffman. � � � � .

2 Department of Surgery, Bowman Gray Schod of Medicine Wake Forest University, Winston-Salem, NC 27157 � �

AJR 158 493-501 March 1992 0361 -803x/92/1 583-0493 © Amencan Roentgen Ray Society -‘---p Pu!

� � OflO-..J

494 WOLFMAN ET AL. AJR:158, March 1992

1 1 , 12]. Although many studies [5, 9, 1 1] have confirmed thevalue of CT compared with DPL, these results are somewhatcontroversial. Other reports [2, 16-1 9] show CT to be lessvaluable than DPL, particularly in injuries to the bowel ormesentery, and both CT and DPL may be warranted in suchcases. DPL is best performed after CT to avoid interpretiveerrors because of fluid and free air introduced by the lavageprocedure. Important diagnostic information may still be ob-tamed from CT after DPL by accounting for these factors [18,20].

The use of CT in the examination of patients with bluntabdominal trauma, along with a trend toward nonoperativemanagement of many abdominal injuries, has decreased theneed for exploratory surgery and reduced the frequency ofnontherapeutic laparotomies [2, 6, 7, 9]. In fact, the trendtoward conservative and nonoperative treatment of manyliver, spleen, and kidney injuries is due in part to the ability ofCT not only to define injury but also to exclude significantinjury, thereby avoiding unnecessary surgery. The final deci-sion to operate should be based on CT findings in conjunctionwith the entire clinical picture and judgment of the attendingtrauma surgeon [2, 21]. In this article, we review the CTfindings in patients with blunt upper abdominal trauma andthe impact of these findings on treatment.

Technique

Proper technique is critical for accurate abdominal CTexamination of patients with blunt abdominal trauma. Toopacify bowel, the patient is given approximately 500 ml oforal contrast medium (1 -3% solution of diatnzoate sodium ordiatnzoate meglumine/diatrizoate sodium) orally or via naso-gastric tube approximately 30-45 mm before CT and anadditional 250 ml just before scanning. To avoid imagingartifacts, the patient’s arms are placed above the abdomen ifpossible and monitoring devices, tubes, and wires are posi-tioned out of the scan plane. If a nasogastnc tube is in thestomach, it is partially withdrawn so that the tip is positionedwithin the distal esophagus and repositioned after the exam-ination is completed. A large field of view decreases artifactsfrom structures that cannot be removed from the scan field.Restraints or sedation may be necessary to avoid motionartifacts in patients unable to maintain the proper position.Fast scans (<2 sac) may be performed with many scannersto minimize motion artifacts.

Kelly et al. [22] recommended performing an unenhancedstudy before the contrast-enhanced study to visualize hyper-dense hematomas that may become isodense after IV con-trast administration. In their study, however, all significantvisceral injuries were seen on contrast-enhanced examina-tions. Only minor injuries that did not lead to a change intreatment were seen exclusively on unenhanced studies.Missing an isodense hematoma can be avoided by usingadequate volumes of contrast material and by recognizingthe irregular contour of the clot at its interface with the organin question. Therefore, most authors do not do a preliminaryunenhanced examination [5, 6, 9, 21, 23, 24].

IV contrast material can be given as an initial bolus ofapproximately 50 ml at a rate of 2-3 mI/sec, followed by arapid infusion at 1 mI/sec. Alternatively, a single sustainedbolus of contrast material at a rate of 1-3 mI/sec may beadministered, particularly if dynamic scanning capability isavailable. The dose is altered for children to 2-3 mI/kg bodyweight. A mechanical injector is useful for regulating the flow

of contrast material, and dynamic scanning techniques areused to decrease examination time in this group of patientsin whom expediency can be critical to successful treatment.Scans are usually taken at 1-cm intervals from the dome ofthe diaphragm through the abdomen and at 1 .5- to 2.0-cmintervals through the pelvis. Patients are monitored through-out the examination, and emergent resuscitation equipmentshould be readily available. In addition to soft-tissue windowsettings, lung window settings are obtained for evaluation ofthe lower chest for significant injuries such as pneumothorax,lung parenchymal injuries, or free peritoneal air indicatinghollow viscus injury, and bone window settings are used forbony fractures [5]. It is also useful to view images directly onthe console video monitor, so that window and level settingscan be manipulated to appreciate subtle but potentially sig-nificant findings. After the study, oral contrast material can bewithdrawn from the stomach via a nasogastric tube ifindicated.

CT Findings and Management Decisions in BluntAbdominal Trauma

Many trauma patients have multiorgan injuries. Therefore,a thorough evaluation of all abdominal components is war-ranted in each case.

Hemoperitoneum

Hemoperitoneum is easily seen on CT and may be the onlyor most obvious sign of abdominal injury. Its presence shouldprompt a thorough search for injury to visceral organs. He-moperitoneum is differentiated from other fluid because ofincreased attenuation, averaging 45 H and always greaterthan 30 H if less than 48 hr old [25]. Hemoperitoneum tendsto be identified near the source of bleeding, spreadingthroughout the abdominal cavity and into the pelvis alongpathways common to all abdominal fluid collections, as de-scribed by Meyers [26]. Free collections of intraperitonealblood are most often seen in the Morison pouch, the mostdependent peritoneal recess in the upper abdomen, and arefrequently seen in the perihepatic space and the right paracolicgutter. The pelvis is the most dependent portion of theperitoneal cavity, and large collections of blood may be pres-ent in the pelvis even when little blood is seen in the abdominalrecesses [25].

Federle and Jeffrey [25] used CT to quantify hemoperito-neum into small, moderate, and large collections. They at-tempted to correlate estimates of the amount of hemorrhagewith severity of visceral organ injury and outcome from either

Fig. 3.-Injury of liver and spleen in 60-year-oldwoman after a motor vehicle accident. EnhancedCT scan shows low-density splenic hematomaposteriorly. Spleen also contains multiple lacera-tions (arrow). Pensplenic and perihepatic bloodand liver laceration also are present. Exploratorylaparotomy confirmed CT findings.

AJR:158, March 1992 CT OF BLUNT UPPER ABDOMINAL TRAUMA 495

operative or nonoperative treatment. Kearney et al. [18],however, using Federle and Jeffrey’s classification, did notfind quantification of hemopentoneum useful in predicting theneed for surgery. CT is useful, however, for detecting theresolution of hemoperitoneum [25, 27].

Clotted blood has higher attenuation (>60 H) than doesfree blood [25]. A localized collection of clotted blood, the“sentinel clot,” is an accurate sign of injury to an adjacentorgan [28]. This sign may help in detecting subtle bowel,mesentenc, or splenic lesions.

The presence of hemoperitoneum on a single CT studydoes not indicate that active bleeding is present. Jeffrey et al.[29], using dynamic scanning techniques, recently describedCT signs of active intraabdominal arterial bleeding. Mostfrequently seen is a focal high-density area of 80-1 30 H(higher attenuation than free or clotted blood) that is isodensewith adjacent major arteries (Fig. 1). Seen less frequently, adiffuse high-density area may simulate extravasation of oralcontrast material from perforated bowel, although other signs

Fig. 1.-Renal fracture in 20-year-oid man involved in motor ve-hicle accident. Enhanced CT scanshows kidney has been completelyseparated into two perfused polesby a severe laceration. Hematomais seen between two poles. Focalareas of extremely high attenuationrepresent active bleeding. Patientdied on day of admission.

Fig. 2.-Splenic laceration with subcapsularand penspienic hematoma in 36-year-oldwoman with hypotension after a fall 1 weekbefore admission. Enhanced CT scan showsan irregular, linear, low-density laceration (ar.row) traversing spleen. Lateral contour ofspleen is indented from subcapsular hema-toma. Density of clotted blood (arrowhead) isincreased when compared with nonclotted por-tion of hematoma. Large volume of hemoperi-toneum in right perihepatic space and sur-rounding left lobe of liver also is seen. CTfindings were confirmed at surgery. A spienec-tomy was performed, and hemoperitoneumwas evacuated.

of bowel perforation are not present. In most patients withactive intraabdominal bleeding, CT is rarely performed be-cause of hemodynamic instability. Occasionally, however, apatient with active abdominal hemorrhage may become hemo-dynamically stable and then undergo CT. Detection of hem-orrhage is then of vital importance, so that proper, potentiallylife-saving treatment can be instituted quickly.

Signs of hypovolemic shock on CT include a small aorta, acollapsed inferior vena cava, marked enhancement of thekidneys, and initially diminished density of spleen comparedwith liver after contrast enhancement [30-33]. Taylor andRosenfield [34] described a hypoperfusion complex in childrenthat consists of marked, diffuse dilatation of the intestine byfluid; abnormally intense contrast enhancement of the bowelwall, mesentery, kidneys, and/or pancreas; decreased caliberof the aorta and inferior vena cava; and significant abdominalfluid.

Spleen

The spleen is the most frequently injured organ in patientswith blunt abdominal trauma [5]. CT is extremely sensitiveand specific in determining the presence and extent of splenicinjury in the traumatized patient, and the CT features ofsplenic trauma have been well described [2, 35]. Subcapsularhematomas appear as crescentic fluid collections that flattenor indent the splenic margin (Fig. 2). Initially, hematomas maybe isodense with splenic parenchyma on contrast-enhancedscans, particularly if adequate volumes of contrast materialare not given [20, 35]. Intrasplenic hematomas are seen onCT as rounded low-density areas (Fig. 3). Lacerations appearas linear low-density areas within the spleen (Fig. 2). Multiplelacerations may have the appearance of a fragmented or

Fig. 4.-Shattered spleen in 26-year-old womanafter motor vehicle accident. Enhanced CT scanshows multiple lacerations and fragmentation ofspleen. Large hemoperitoneum also is present,surrounding both spleen and liver. Hemoperito-neum extended into pelvis. Exploratory laparot-omy confirmed CT findings.

496 WOLFMAN ET AL. AJR:158, March 1992

shattered spleen (Fig. 4). Unenhancing portions of the spleenshould suggest injury or thrombosis of the artery to theaffected segment.

Disruption of the splenic capsule results in visible hemo-pentoneum in up to 98% of patients with splenic injury [35].Pensplenic clot also is associated with splenic injury. Bothhemopentoneum and perisplenic clot may be present in casesof splenic injury without evidence of a splenic laceration onCT. Therefore, presence of either should prompt a search forsplenic injury. Occasionally, splenic injury may manifest asheterogeneous parenchyma with a mottled irregular enhance-ment [35]. This finding is almost always associated withpentoneal blood and perisplenic clot and by itself is merelysuggestive of splenic injury. The spleen also may transientlyappear inhomogeneous during the early capillary phase of acontrast-enhanced dynamic CT study.

�s Certain findings on CT may result in a false-positive diag-nosis of splenic injury. Splenic lobulations and congenitalclefts, and a prominent left hepatic lobe extending across themidline to lie immediately adjacent to the spleen, may simulatea splenic laceration [36]. Streak and motion artifacts can alsoresult in incorrect diagnoses. The spleen may seem to enlargeon serial CT scans after blunt abdominal trauma [37]. This isa result of initial splenic contraction from adrenergic stimula-tion at the time of acute injury, which resolves with time andvolume replacement.

Although rare, delayed splenic rupture has been reportedin patients in whom an initial CT scan after injury showed noevidence of splenic abnormality [34, 38-40]. Suboptimal con-trast enhancement techniques may be partially responsiblefor this missed diagnosis [41]. Subtle inhomogeneities of thesplenic parenchyma and minimal thickening of the lateroconal

fascia and the left anterior renal fascia may be the only signsinitially seen in some cases of delayed splenic rupture [40].

King and Shumacker [42] first reported overwhelming sep-sis in five children after splenectomy, two of whom died.

Other reports followed, confirming the increased risk of latesepsis after splenectomy in both children and adults [43].This observation has encouraged nonoperative managementand splenic salvage procedures in appropriate patients [2,43-45]. The role of CT in selecting between conservative andoperative treatment of splenic injuries remains unclear. Mirviset al. [46], using a CT classification of four grades of splenicinjury, found that although patients with severe splenic injury

generally required laparotomy, eight of 23 such patients weretreated successfully without surgery, and four of 1 5 patientswith minor injury, initially treated conservatively, eventuallyrequired celiotomy. Buntain et al. [1 0] had more success inpredicting outcome of splenic injury by using a CT classifica-tion that, in addition to grading splenic trauma, used modifiersfor other intraabdominal and extraabdominal injuries. In theirseries, only patients with minor injuries were selected fornonoperative management, and all others had laparotomy.Resciniti et al. [47] devised a numeric CT scoring system thatgraded severity of splenic injury on a scale of 1 -3 and scored1 point each for the presence of perisplenic fluid, abdominalintraperitoneal fluid, and pelvic intraperitoneal fluid. Question-able observations received 0.5 points. In their series, no

patient with a score lower than 2.5 who was initially managednonoperatively required celiotomy. Umlas and Cronan [48]applied the grading systems of Buntain et al. and Resciniti etal. to determine whether the outcome of nonsurgical manage-ment could be predicted on the basis of CT findings. Theyfound that these systems were not completely reliable. Inparticular, delayed splenic rupture remains a distinct problemin a small percentage of patients in whom the spleen is normal

or shows only a limited injury on the initial CT scan [48, 49].On the other hand, Brick et al. [21] have shown that moderateor severe splenic injury in children or injury associated with amoderate or large amount of hemopentoneum may be treatedsuccessfully without surgery. At this time, although CT isextremely useful in characterizing initial injury and followingthe evolution of splenic trauma, the final decision on whetherto operate or not should be made by the trauma surgeon onthe basis of clinical factors and not entirely on the CT findings.

Liver

The liver is preceded only by the spleen in frequency ofinjury from abdominal trauma. In 22-61 % of patients withhepatic injuries, significant damage is obvious, and becauseof shock or peritonitis such patients have immediate surgerywithout preliminary imaging studies. As with splenic trauma,CT has proved to be highly sensitive, specific, and accuratein defining and characterizing hepatic injury and associatedhemopentoneum in hemodynamically stable patients [5, 1 1,27, 50]. In all reported series, the right hepatic lobe is injuredmore frequently, probably because of its larger size andproximity to the lower ribs [21 , 50, 51].

The CT findings in hepatic injury are similar to those seenin the spleen; they include contusions (the mildest injury),subcapsular hematomas, intraparenchymal hematomas, andsingle or multiple lacerations and fractures through the hepaticparenchyma [7, 27, 50] (Figs. 3 and 5). Lack of enhancementoffractures indicates loss of vascular supply with the potentialfor hepatic necrosis. Periportal tracking (areas of perivascularlow-attenuation surrounding peripheral subsegmental portalvenous branches) also has been described as a sign of hepaticinjury on CT and may be the only finding [51 ] (Fig. 6). AlthoughMacrander et al. [51 ] presume that such tracking is a sign ofblood, Cox et al. [52] ascribe the finding to distension ofpenportal lymphatics because of associated hypotension andsubsequent large fluid volume replacement in traumatizedpatients with no other evidence of abdominal trauma. Hepaticsubcapsular and parenchymal gas seen on CT 2-3 days afterhepatic trauma may be due to hepatic necrosis and may notbe related to infection [53].

The liver has a remarkable ability to heal even after severeinjury. Therefore, nonoperative management of such injuriesin hemodynamically stable patients is now accepted. Unlikethe spleen, the liver does not exhibit delayed rupture. Attempt-ing to define the role of CT in management decisions, Moonand Federle [50] and Meyer et al. [54] reported that limitedhepatic injury without evidence of active bleeding and withlittle or no hemoperitoneum could be managed successfullywithout surgery. More recently, it has been shown that stable

AJR:158, March 1992 CT OF BLUNT UPPER ABDOMINAL TRAUMA 497

Fig. 5.-Liver lacerations in 37-year-old womaninjured in motor vehicle accident. Enhanced CTscan shows multiple irregular, low-density lineardefects within liver, representing lacerations. Lac-erations extend to medial and lateral surfaces ofliver. Patient remained hemodynamicaily stableand was treated nonoperatlveiy.

Fig. 6.-Periportal tracking in 16-year-old boyinjured In motor vehicle accident. Circumferentiallow-attenuation areas surround portal venousbranches. No other signs of injury were seen onthis enhanced CT scan. Patient was treated non-operatively.

patients with severe hepatic injuries and significant hemoper-itoneum also may be treated nonoperatively without signifi-cant sequelae [21 , 27, 55]. Jeffrey [49] states that CT stagingof blunt hepatic injuries has little discriminatory value in pre-dicting outcome of stable patients, as nearly all have anexcellent prognosis. Other abdominal injuries may, however,require laparotomy.

Management decisions in the setting of hepatic injuryshould be based on clinical factors, and CT should be usedto depict the injury and detect healing and resorption ofhemoperitoneum. Peritoneal blood is normally resorbed andtherefore significantly reduced or absent within 1 week onfollow-up CT scans [27]. Otherwise, continued hemorrhageshould be suspected. Subcapsular hematoma usually re-solves within 6-8 weeks [56]. lntraparenchymal hematomasheal much more slowly and may persist for several years, asbile in the hematoma delays clot resorption and adverselyaffects parenchymal healing [56]. Persistent intraparenchymalhematoma appears as a collection with a high attenuation(30-50 H). A water-density posttraumatic cyst or biloma mayresult. Lacerations, conversely, appear to heal more rapidly,and significant healing is seen on serial CT examinations overa 3-week period [27]. Clearly, any patient selected for con-servative therapy requires continued hemodynamic monitor-ing and laboratory assessment; transfusion as needed; andthe availability of nursing, surgical, and imaging facilities ifhemodynamic instability develops [27, 55].

Kidney

Renal injuries after blunt abdominal trauma can be cate-gorized as minor, intermediate, and severe [57-60]. Mostrenal injuries (75-85%) are minor and include contusions,intrarenal hematomas, small subcapsular hematomas, smalllacerations that do not communicate with the collecting sys-tern, and small segmental infarcts. Patients are usually hemo-dynamically stable, have microscopic hematuria, and aretreated conservatively. Intermediate injuries, which accountfor approximately 1 0% of renal trauma, include deep lacera-tions that communicate with the collecting system and resultin urine extravasation. Treatment of such injuries is somewhatcontroversial; some advocate conservative management un-less severe bleeding or clinical deterioration ensues; others

stress early surgery to avoid complications [58-60]. Theremaining 5% of renal injuries are severe, including shatteredkidneys, renal pedicle injuries, and avulsion and laceration ofthe renal pelvis. These catastrophic injuries require immediatesurgery.

Of all imaging techniques, CT most accurately depicts thecharacter and extent of renal injury, best displays penrenalhematomas and extravasation of urine, best distinguishesbetween the categories of renal trauma, and, therefore, ismost useful in case management [57, 58, 61 , 62]. CT alsodetects associated injuries within the peritoneal cavity andretroperitoneum that may influence management. In a stablepatient with a suspected isolated renal injury, excretory urog-raphy will usually suffice as an initial examination [58, 62,63]. Normal findings on excretory urography exclude signifi-cant renal injury. However, evidence of significant renaltrauma on excretory urography often requires further evalu-ation with CT.

The appearance of renal injuries on enhanced CT has beenwell documented [57-60, 62]. Contusion, the mildest renalinjury, results in edema and extravasation of small amountsof blood and urine into the interstitial space [58, 62]. Contu-sion may be subtle and missed on contrast-enhanced CT.When present, it appears as poorly defined areas of de-creased enhancement. Lang et al. [62] described the appear-ance of a small collection of contrast medium in the renalinterstitium on delayed scans as a sign of contusion. Contu-sions usually resolve within 1 week.

Intrarenal hematomas appear as areas of decreased en-hancement that may be poorly defined or well marginated.Subcapsular hematomas are confined by the renal capsule,are often lenticular, and may flatten the renal border. Anapparent low-attenuation region around the surface of thekidney may be caused by respiratory motion during the scanand can result in a false diagnosis of subcapsular hematoma[60]. In such cases, a similar appearance is noted anterior tothe abdominal wall.

Lacerations appear as focal parenchymal injuries with de-creased enhancement. If they involve the collecting system,contrast-laden urine extravasates and is easily seen on CT(Fig. 7). A significant perirenal hematoma is usually presentwith severe lacerations. Because of fascial fusions in theretroperitoneum, hematomas due to or associated with renal

Fig. 7.-Renal laceration extendingInto collecting system in 37-year-oldman Injured In motor vehicle accidentEnhanced CT scan shows contrast cx-travasatlon (arrow). Posterior segmentof kidney is not enhanced, denotingloss of vascular supply. Perirenal he-matoma alsols seen. Nephrectomy wasperformed

Fig. 8.-Left renal pedicle injury in 36-year-oldwoman after motorcycle accident CT scan was ob-tamed after perltoneal lavage and excretory urogra-phy. Left kidney does not enhance because of renalpedicle injury. Perlrenal and anterior pararenal hema-tomas are noted also. Typical appearance of wedge-shaped perfusion defect representing segmental renalinfarct (arrow) is seen in right kidney. Li vertebralburst fracture also is present with surrounding hema-toma. An avulsed left renal artery was found at

498 WOLFMAN ET AL. AJR:158, March 1992

surgery.

Fig. 9-Pancreatic laceration in 29-year-old manwho fell 9 m from a tree. Enhanced CT scan showssite of laceration (arrow) at junction of body and tailof pancreas. Associated peripancreatic fluid andthickening of left anterior pararenal fascia also arenoted. Splenic laceration with a segmental perfusiondefect, seen anteriorly in spleen on this image, wasdue to loss of vascular supply. CT findings were con-firmed at exploratory laparotomy.

tr&ima tend not to cross the midline; the presence of such ahematoma should prompt a search for injury to the aorta, itsbranches, or other midline structures [57]. Lacerations thatcompletely transect the kidney into two separate poles areoften called fractures (Fig. 1). Lacerations usually occur par-allel to the main vascular structures, preserving them, andparenchymal enhancement is observed. The margins of lac-erated fragments may have an inhomogeneous, mottled ap-pearance, possibly the result of vasospasm. The shatteredkidney contains multiple lacerations, some of which may shearacross vascular planes and produce devascularized frag-ments with nonenhancing parenchyma. A section throughthe hilar lip of the kidney may appear as a fracture, butits characteristic posterolateral location should preventconfusion.

A segmental renal infarct may result from injury to anintrarenal or polar arterial branch [60] and appears as awedge-shaped or hemispheric area of nonperfusion with theapex pointing toward the renal hilum (Fig. 8). A thin enhancingnm may be seen [62]. Eventually a deep scar forms in thearea of infarction.

Laceration of the renal pelvis or avulsion of the ureteropelvicjunction causes extravasation of contrast-laden urine. Extrav-asation without evident renal injury should raise suspicion ofsuch injuries of the renal pelvis [60].

In many cases an injury of the renal pedicle resulting inarterial or venous occlusion can be clearly documented onCT because the kidney does not enhance (Fig. 8). A corticalrim of enhancement may be present as a result of collateralblood flow from the capsular arteries. An abrupt cutoff of thecontrast-enhanced renal artery is seen occasionally [64]. Thereliability of CT in detecting injuries of the renal pedicle is indispute. Sclafani et a]. [57] consider CT the method of choiceand confirmatory angiography unnecessary. Lupetin et al.[64], using CT, diagnosed renal artery occlusion in all seven

patients in their series. Lang et al. [62], on the other hand,found CT less reliable in the detection of trauma to the renalartery, as the diagnosis was missed on CT in five of sevenpatients in their series. Injury to the renal vein after blunttrauma is missed more often than injury to the artery; inaddition to lack of enhancement, the kidney may appearenlarged, the rim enhancement may be thicker, and thrombusin the vein may be seen [58].

Pancreas

Pancreatic injuries, including pancreatic duct disruption,fractures, contusions, and traumatic pancreatitis, represent3-1 2% of all abdominal injuries from blunt trauma. They areclinically important, however, because death occurs in 16-20% of cases of pancreatic trauma, and major posttraumaticcomplications (pseudocyst, abscess, hemorrhage, acute re-curling pancreatitis, and fistulae) occur in one of three survi-vors [1 7, 65, 66]. Delay in diagnosis leads to an increase inthe mortality and morbidity rate. The typical clinical findingsof upper abdominal pain, leukocytosis, and increased serumamylase may not be apparent for one or more days afteracute pancreatic trauma. In addition, an increase in the serumamylase level after trauma may be present without pancreaticinjury [67].

Pancreatic injury may be difficult to diagnose on CT [17,65]. Little evidence of pancreatic injury may appear on CTexaminations performed soon after the traumatic event [65].Pancreatic duct disruption is the most significant pancreaticinjury. Although this abnormality cannot be seen directly onCT, associated injury to the pancreas may be detected.Fracture of the pancreas, depicted as a clear separation orlow-density line through the long axis, occurs most commonlyin the neck of the pancreas as a result of compression of theorgan against the spine [67] (Fig. 9). Thickening of the left

;___�% -- ,� � .-V. #{149}�

AJR:158, March 1992 CT OF BLUNT UPPER ABDOMINAL TRAUMA 499

anterior renal fascia, although not specific for pancreaticinjury, is frequently present and should raise the possibility ofpancreatic injury [65]. Other signs include focal or diffuseenlargement of the organ, areas of decreased attenuation,and peripancreatic edema and fluid collections [1 7, 65, 67]. Ifthe possibility of pancreatic trauma persists despite initiallynormal CT findings, a second CT examination in 1 2-24 hrmay be warranted, because signs of injury can be delayed.ERCP, which directly visualizes the pancreatic duct, may alsobe necessary for further evaluation when CT findings areequivocal or CT is technically inadequate [65]. Immediatesurgery is the recommended treatment for pancreaticfracture.

Bowel and Mesentery

Bowel and mesentery injuries are reported to occur in 5%of blunt trauma cases [68]. As these injuries are most com-monly seen in restrained motor vehicle accident victims, theincidence may rise as seat belt usage increases [1 6]. Earlydiagnosis is important. In patients with duodenal perforation,surgery performed within 24 hr of injury has a 5% mortalityrate, whereas delayed diagnosis and treatment leads to amortality rate of 65% [68, 69]. Unfortunately, clinical signsinitially may be subtle or absent.

CT signs of bowel and mesentenc injury include extralu-minal air, extravasation of oral contrast material, pentoneal orretropentoneal fluid, thickened bowel wall, high-density clot(sentinel clot) adjacent to the involved bowel, and focal mes-entenc infiltration [1 6, 68]. Free air in either the peritonealcavity or the retroperitoneum from injury to the retroperitonealduodenum, small bowel, or colon is a relatively specific signof bowel perforation but is seen in only half of cases [68, 69](Fig. 1 OA). The volume of air may be quite small and subtle.Wide window or lung settings will aid in detection. The mostcommon location in which to detect free intraperitoneal air is

the subdiaphragmatic area, anterior to the liver. Extraluminalair also may be present within the leaves of the mesentery orin the retropentoneum, particularly in the anterior pararenalspace. Occasionally, pneumopentoneum results from pneu-momediastinum, pneumothorax, bladder rupture, or previousperitoneal lavage and is not related to bowel trauma [17,70]. Extravasation of oral contrast material from the bowellumen is a specific sign of bowel perforation (Fig. 1 1). Unfor-tunately, its presence is shown on CT in only a minority ofcases. lntraabdominal fluid, although seen in nearly all casesof bowel or mesentenc injury, is not a specific indicator.

Moderate or large amounts of fluid are associated withsignificant bowel or mesenteric trauma. If no associated solidvisceral injury is noted, bowel or mesentenc injury should besuspected [1 6, 68]. Focal mesenteric infiltration is also afrequently seen but nonspecific sign. A localized high-atten-uation mesentenc hematoma or intramural hematoma canhelp localize the site of injury [28, 71] (Fig. 1OB). Thickenedbowel is seen in about 75% of cases with transmural lacera-tion. Therefore, its absence does not exclude bowel injury[68].

Although Rizzo et al. [68] report high sensitivity in detecting

bowel and mesentery injuries, other authors indicate that thisdiagnosis may be difficult with CT [2, 1 7, 33, 72]. This difficultymay be due in part to the subtlety of the findings and the factthat presentation of signs and symptoms may be delayed.The initial CT examination may be done before the clinical

signs are manifest [1 7, 33, 68]. The coordinated use of CTand pentoneal lavage for diagnosis of bowel and mesentericinjuries has been suggested [1 6, 18].

Miscellaneous Abdominal Trauma

Gallbladder injury after blunt abdominal trauma is uncom-mon, occurring in 2-3% of cases [73, 74]. Laceration orperforation, complete avulsion, or intramural contusion can

Fig. 10.-Perforated jejunum in 37-year-old man injured in motor vehicle accidentA, Enhanced CT scan shows free intraperitoneal air due to jejunal perforation anterior to liver.

Wide window settings are used to accentuate display of free air.B, CT scan at level caudal to A shows localized high-attenuation hematoma adjacent to mildly

dilated loop of bowel denoting site of Injury. Free intraperitoneal blood also is seen in both right andleft paracolic gutters. Exploratory laparotomy revealed a perforated jejunum, torn mesentery, andischemlc lleum.

Fig. 11.-Perforated duodenum in 38-year-oldman hit by a train. CT scan shows extravasationof oral contrast material from duodenum at site ofinjury (arrow). Free perltoneal fluid is noted inMorison pouch. Diffuse mesenteric edema also ispresent CTflndingswere confirmed at exploratorylaparotomy.

Fig. 12.-Traumatic rupture of left hemidiaphragm in 20-year-old maninjured in motor vehicle accident CT scan shows that stomach, containingoral contrast medium, has herniated through dlaphragmatic rupture andlies within thoracic cavity. Left hemidiaphragmatic rupture was confirmedat exploratory laparotomy.

500 WOLFMAN ET AL. AJR:158, March 1992

occur. On CT, the gallbladder may contain high-density hem-orrhage with associated peritoneal blood; low-density fluidassociated with bile leakage also may be present [73, 74].Hemobilia is likely if high-attenuation material is seen in thegallbladder, and other causes of increased density, such asstones, contrast material, and milk-of-calcium bile, can beexcluded [75]. Diaphragmatic rupture occurs in 1-2% ofpatients after blunt abdominal trauma and almost alwaysaffects the left side [76, 77]. This injury is difficult to diagnosewith any imaging technique and is often missed on CT [76].Hemiation of abdominal contents through the diaphragm maybe seen occasionally [77] (Fig. 12).

Traumatic adrenal hemorrhage occurs in up to 25% ofpatients after severe trauma [78, 79]. In 85% of cases hem-orrhage is right-sided and in 20% of cases it is bilateral. OnCT, adrenal hemorrhage appears as a hyperdense mass (50-75 H) with streaky infiltration into the penadrenal fat andthickening at the adjacent diaphragmatic crus [78]. Abnor-malities in the subcutaneous fat of the abdominal wall at thesite of trauma also may be seen [79].

REFERENCES

1 . U.S. Bureau of the Census. Statistical abstract of the United States: 1990,1 10th ed. Washington, DC: U.S. Government Printing Office, 1990:

78-81,852. MCCOrt JJ. Caring for the major trauma victim: the role for radiology.

Radiology 1987;163: 1-9

3. MacKenzie EJ, Morris JA Jr, Smith GS, Fahey M. Acute hospital costs of

trauma in the United States: implications for regionalized systems of care.J Trauma 1990;30:1096-1103

4. Harlan LC, Harlan WR, Parsons PE. The economic impact of injuries: a

major source of medical costs. Am J Public Health 1990;80:453-4595. Federle MP. CT of abdominal trauma. In: Federle MP, Brant-Zawadzki M,

eds. Computed tomography in the evaluation of trauma, 2nd ed. Baltimore:Williams & Wilkins, 1986:191-273

6. Peitzman AB, Makaroun MS, Slasky BS, Ritter P. Prospective study of

computed tomography in initial management of blunt abdominal trauma. JTrauma 1986;26:585-592

7. Federle MP, Crass RA, Jeffrey RB, Trunkey DD. Computed tomography

in blunt abdominal trauma. Arch Surg 1982;117:645-650

8. Kaufman RA, Towbin R, Babcock DS, et al. Upper abdominal trauma inchildren: imaging evaluation. AJR 1984;142:449-460

9. Wing VW, Federle MP, Morris JA Jr, Jeffrey RB, Bluth R. The clinicalimpact of CT for blunt abdominal trauma. AJR 1985;1 45:1191-1194

10. Buntain WL, Gould HR, Maull KI. Predictability of splenic salvage by

computed tomography. J Trauma 1988;28:24-341 1 . Goldstein AS, Sclafani SJA, Kupferstein NH, et al. The diagnostic superi-

ority of computerized tomography. J Trauma 1985;25:938-94612. Federle MP, Goldberg HI, Kaiser JA, Moss AA, Jeffrey RB Jr, Mall JC.

Evaluation of abdominal trauma by computed tomography. Radiology1981;138:637-644

13. Gomez GA, Alvarez R, Plasencia G, et al. Diagnostic peritoneal lavage inthe management of blunt abdominal trauma: a reassessment. J Trauma1987;27: 1-5

14. Powell RW, Green JB, Ochsner MG, BarttelbOrt SW, Shackford SR. SiseMJ. Peritoneal lavage in pediatric patients sustaining blunt abdominaltrauma: a reappraisal. J Trauma 1987;27 :6-10

15. Hubbard SG, Bivins BA, Sachatello CR, Griffen WO Jr. Diagnostic errors

with peritoneal lavage in patients with pelvic fractures. Arch Surg

1979;1 14:844-84616. Hayes CW, Conway WF, Walsh JW, Coppage L, Gervin AS. Seat belt

injuries: radiologic findings and clinical correlation. RadioGraphics

1991;1 1:23-2617. Cook DE, Walsh JW, Vick CW, Brewer WH. Upper abdominal trauma:

pitfalls in CT diagnosis. Radiology 1986;159:65-6918. Keamey PA Jr, Vahey 7, Burney RE, Glazer G. Computed tomography

and diagnostic peritoneal lavage in blunt abdominal trauma: their combinedrole. Arch Surg 1989;124:344-347

19. Marx JA, Moore EE, Jorden RC, Eule J Jr. Limitations of computed

tomography in the evaluation of acute abdominal trauma: a prospectivecomparison with diagnostic peritoneal lavage. J Trauma i985;25:933-937

20. Kane NM, Dorfman GS, Cronan JJ. Efficacy of CT followiog peritoneal

lavage in abdominal trauma. J Comput Assist Tomogr 1987;1 1:998-1002

21 . Brick SH, Taylor GA, Potter BM, Eichelberger MR. Hepatic and splenicinjury in children: role of CT in the decision for laparotomy. Radiology

1987;165:643-646

22. Kelly J, Raptopoulos v, Davidoff A, Waite R, Norton P. The value of non-

contrast�enhanced CT in blunt abdominal trauma. AiR 1989;152:41-4623. Mindell HJ. On the value of non-contrast.enhanced CT in blunt abdominal

trauma. MR 1989;152:47-4824. Taylor GA, Guion CJ, Potter BM, Eichelberger MR. CT of blunt abdominal

trauma in children. AJR 1989;153:555-55925. Federle MP, Jeffrey RB Jr. Hemoperitoneum studied by computed tomog-

raphy. Radiology 1983;148:187-192

26. Meyers MA. intraperitoneal spread of infections. In: Meyers MA. Dynamicradiology of the abdomen: normal and pathologic anatomy, 3rd ed. NewYork: Springer-Verlag, 1988:49-89

27. Foley WD, Cates JO, Keliman GM, et al. Treatment ofblunt hepatic injuries:

role of CT. Radiology 1987;164:635-63828. Orwig 0, Federle MP. Localized clotted blood as evidence of visceral

trauma on CT: the sentinel clot sign. AJR 1989;153:747-74929. Jeffrey RB Jr, Cardoza JO, Olcott EW. Oetection of active intraabdominal

arterial hemorrhage: value of dynamic contrast-enhanced CT. AiR1991;156:725-729

30. Shin MS, Borland LL, Ho K-J. Small aorta: CT detection and clinicalsignificance. J Comput Assist Tomogr 1990;14: 102-1 03

31 . Jeffrey RB Jr. Federle MP. The collapsed inferior vena cava: CT evidence

of hypovolemia. AiR 1988:150:431-43232. Sclafani SJA, Becker JA. Radiologic diagnosis of renal trauma. UrolRadiol

1985:7:192-200

33. Berland LL, VanDyke JA. Decreased splenic enhancement on CT in

traumatized hypotensive patients. Radiology 1985:156:469-47134. Taylor CR, Rosenfield AT. Umitations of computed tomography in the

recognition of delayed splenic rupture. J Comput Assist Tomogr1984;8: 1205-1 207

35. Federle MP, Gnffiths B, Minagi H, Jeffrey RB Jr. Splenic trauma: evaluation

with CT. Radiology 1987:162:69-7136. Jeffrey RB, Laing FC, Federle MP, Goodman PC. Computed tomography

of splenic trauma. Radiology 1981:141:729-73237. Goodman LA, Aprahamian C. Changes in splenic size after abdominal

trauma. Radiology 1990:176:629-632

38. Pappas 0, Mirvis SE, Crepps JT. Splenic trauma: false-negative CT diag-nosis in cases of delayed rupture. AiR 1987:149:727-728

39. Fagelman 0, Hertz MA, Ross AS. Delayed development of splenic subcap-

AJR:158, March 1992 CT OF BLUNT UPPER ABDOMINAL TRAUMA 501

sular hematoma: CT evaluation. J ComputA.ssist Tomogr 1985:9:815-81640. Sutton CS, Haaga JR. CT evaluation of limited splenic trauma. J comput

Assist Tomogr 1987;11:167-169

41 . Conrad MR. SpIeniC trauma: false-negative CT diagnosis in cases ofdelayed rupture (letter). AiR 1988;151 :200

42. King H, Shumacker HB Jr. SpleniC studies. I. Susceptibility to infection

after splenectomy performed in infancy. Ann Surg 1952;136:239-24243. Elmore JA, Clark OE, Isler RJ, Homer WA. Selective nonoperative man-

agement of blunt splenic trauma in adults. Arch Surg 1989:124:581-58644. Em SH, Shandling B, Simpson JS, Stephens CA. Nonoperative manage-

mont of traumatized spleen in children: how and why. J Pediatr Surg

1978:13:117-119

45, Zucker K, Browns K, Rossman 0, Hemingway 0, Salk A. Nonoperativemanagement of splenic trauma: conservative or radical treatment? Arch

Surg 1984;1 19:400-40446. Mirvis SE, Whitley NO, Gens OR. Blunt splenic trauma in adults: CT-based

classification and correlation with prognosis and treatment. Radiology

1989;171 :33-3947. Resciniti A, Fink MP, Raptopoulos V. Oavidoff A, Silva WE. Nonoperative

treatment of adult splenic trauma: development of a computed tomographic

scoring system that detects appropriate candidates for expectant man-

agement. J Trauma 1988:128:828-83148. Umlas S-L, Cronan JJ. SpleniC trauma: can CT grading systems enable

prediction of successful nonsurgical treatment? Radiology 1991:178:481-487

49. Jeffrey RB Jr. CT diagnosis of blunt hepatic and splenic injuries: a look tothe future (editorial). Radiology 1989;171 :17-18

50. Moon KL Jr. Federle MP. Computed tomography in hepatic trauma. AiR1983:141:309-314

51 . Macrancor sJ, Lawson TL, Foley WO, Oodds WJ, Erickson SJ, Ouiroz

FA. Penportal tracking in hepatic trauma: CT features. J comput Assist

Tomogr 1989:13:952-95752. Cox JF, Friedman AC, Radecki P0, Lev-Toaft AS, Caroline OF. Penportal

lymphadema in trauma patients. AiR 1990:154:1124-112553. Abramson SJ, Borden WE, Kaufman RA, Ruzal-Shapiro C. Hepatic paren-

chymal and subcapsular gas after hepatic laceration caused by blunt

abdominal trauma. MR 1989:153:1031-103254. Meyer AA, Crass RA, Urn AC Jr, Jeffrey RB, Federle MP, Trunkey 00.

Selective nonoperative management of blunt liver injury using computedtomography. Arch Surg 1985;120:550-554

55. Mirvis SE, Whitley NO, Valnwrsght JR, Gens OR. Blunt hepatic trauma inadults: CT-based classification and correlation with prognosis and treat-mont. Radiology 1989:171:27-32

56. Savolaine ER, Grecos GP, Howard J, White P. Evolution of CT findings inhepatic hematoma. J Comput Assist Tomogr 1985:9:1090-1096

57. Sclafani 5JA, Becker JA, Shaftan GVV, et al. Strategies for the radiologic

management of genitounnary trauma. Urol Radiol 1985;7:231-244

58. Pollack HM, Wein AJ. Imaging of renal trauma. Radiology 1989:172:297-308

59. Foderle MP, Kaiser JA, McAninch JW, Jeffrey RB, Mall JC. The role of

computed tomography in renal trauma. Radiology 1981:141:455-46060. Fanney OR, Casillas J, Murphy BJ. CT in the diagnosis of renal trauma.

RadioGraphics 1990:10:29-40

61 . Sandier CM, Toombs BO. Computed tomographic evaluation of blunt renalinjuries. Radiology 1981;141 :461-466

62. Lang EK, Sullivan J, Frentz G. Renal trauma: radiological studies. Com-parison of urography, computed tomography, angiography, and radio-nuclide studies. Radiology 1985:154:1-6

63. Kisa E, Schenk WG lii. Indications for emergency intravenous pyelography(IVP) in blunt abdominal trauma: a reappraisal. J Trauma 1986:26:

1086-108964. Lupetin AR. Mainwanng BL, Daffner RH. CT diagnosis ofrenal artery injury

caused by blunt abdominal trauma. AiR 1989:153:1065-106865. Jeffrey RB Jr, Federle MP. Crass RA. Computed tomography of pancreatic

trauma. Radiology 1983:147:491-49466. Carr NO. Calms Si, Lees WA, Russell RCG. Late complications of pan-

creatic trauma. Br J Surg 1989;76: 1244-1 24667. Oodds WJ, Taylor AJ, Erickson SJ, Lawson TL. Traumatic fracture of the

pancreas: CT characteristics. J Comput Assist Tomogr 1990:14:375-378

68. Rizzo MJ, Federle MP, Griffiths BG. Bowel and mesenteric injury followingblunt abdominal trauma: evaluation with CT. Radiology 1989:173:143-148

69. Glazer GM, Buy JN, Moss AA, Goldberg Hi, Federle MP. CT detection of

duodenal perforation. AiR 1981:137:333-33670. Bulas 01, Taylor GA, Eichelberger MR. The value of CT in detecting bowel

perforation in children after blunt abdominal trauma. AiR 1989:153:561-564

71 . Plojoux 0, Hauser H, Wettstein P. Computed tomography of intramuralhematoma of the small intestine: a report of 3 cases. Radiology1982:144:559-561

72. Sherck JP, Oakes 00. Intestinal injuries missed by computed tomography.

J Trauma 1990:30:1-773. Jeffrey RB Jr. Fedetle MP, Laing FC, Wing VW. COmputed tomography of

blunt trauma to the gallbladder. J Comput Assist Tomogr 1986:10:

756-75874. Gottesman L, Marks RA, Khoury PT, Moailem AG, Wichem WA Jr.

Oiagnosis of isolated perforation of the gallbladder following blunt trauma

using sonography and CT scan. J Trauma 1984;24:280-28175. Krudy AG, Doppman JL, Bissonette MB, Girton M. Hemobiiia: computed

tomographic diagnosis. Radiology 1983:148:785-78976. Gelman A, Mirvis SE, Gens 0. Diaphragmatic rupture due to blunt trauma:

sensitivity of plaln chest radiographs. AiR 1991;156:51-5777. Heiberg E, Woiverson MK, Hurd RN, Jagannadharao B, Sundaram M. CT

recognition of traumatic rupture of the diaphragm. AJR 1980:135:

369-372

78. Wiims GE, Marchal GJF, Baert AL, et al. CT and ultrasound features ofpost-traumatic adrenal hemorrhage. J Comput Assist Tomogr 1987;1 1:

112-11579. Murphy BJ, Casilias J, Ynzarry JM. Traumatic adrenal hemorrhage: radio-

logic findings. Radiology 1988;169:701-703