Dr. Shanmuganathan and Dr. Mirvis are with the Department of Diagnostic Radiology, University of Maryland Medical Center and Maryland Shock-Trauma Center, Baltimore, MD. Dr. Chen is with the Department of Radiology, National Yang-Ming University and Veterans General Hospital, Taipei, Taiwan.

The availability of fast conventional or spiral scanning, close access to CT from adjacent patient admitting areas, and facilities to monitor critically ill patients in the CT scanning suites, has made CT scan the imaging modality of choice for definitive evaluation of hemodynamically stable patients following blunt trauma. Many studies have shown contrast- enhanced CT (CECT) provides essential information for patient triage and management including accurate localization of the extent and site of hepatic injury, quantification of hemoperitonium, depiction of other associated intra- and extraperitoneal injuries resulting from blunt trauma. ^1-4 CECT can also document injury healing or progression, and demonstrate early and late complications of hepatic trauma. While CECT is the principal imaging modality used at our trauma center to acutely assess the site and extent of hepatic injury, appropriate application of other modalities including sonography, angiography, and nuclear scintigraphy may play a role in diagnosing and managing hepatic injuries.

Studies comparing the outcome of patients with hepatic injury managed without surgery and a control group of patients with surgically managed liver injuries show no difference in the length of hospital stay between the two groups. However, transfusion requirements and intra-abdominal complications were significantly lower in the group treated without surgery. ^5-7 Surgical literature also confirms that up to 67% of celiotomies performed for blunt trauma are unnecessary or nontheraputic, and 86% of the hepatic injuries have stopped bleeding at the time of surgery. ^8,9 Hence, over the last decade trauma surgeons have moved from the traditional approach of early surgical intervention to nonoperative management of blunt liver injury.

CECT not only provides accurate information regarding the grade of liver injury and the amount of hemoperitonium, it also reliably and consistently detects the presence of other solid organ, hollow-viscus, and mesenteric injuries. By providing this constellation of information required to manage hemodynamically stable blunt hepatic injuries without surgery, CECT has played a major role in convincing trauma surgeons to widely accept and promote this method of treatment over the last 5 years. ^6-10 This article will discuss the contrast-enhanced spiral CT (CESCT) findings of blunt hepatic injuries and its role in nonoperative management of liver injuries.

CT scanning technique

Currently, our trauma center uses a spiral CT (Siemens Plus 4, Siemens Medical Systems, Iselin, NJ) to evaluate the abdomen and pelvis of trauma patients. Spiral CT scans are performed with 8-mm collimation and a table speed of 8 mm/sec (pitch of 1). The abdomen is scanned from the lower chest to the iliac crest and the pelvis and then from the iliac crest to the pubic symphysis. A 100 to 120 sec delay between the abdominal and pelvic scans facilitates visualization of contrast excretion in the lower ureters and bladder. Conventional CT can be performed with ¾ 2.0 sec scan and 10-mm slice collimation at 10-mm table increments from the lower chest to the iliac crest and 10-mm collimation at 20-mm table increments from the iliac crest to the symphysis pubis. In our center, intravenous (IV) contrast material (Omnipaque 240, Nycomed, New York, NY), is power-injected (Mark-IV, Medrad, Pittsburgh, PA), using a 150-mL bolus injection and infusion rate of 3 mL/sec for spiral CT. For conventional CT, an initial bolus of 100 mL at a rate of 1.5 to 2 mL/sec is followed by an infusion of an additional 50 mL at a rate of 1.0 mL/sec. Initiation of scanning is delayed 60 sec for spiral CT and 30 to 45 sec for convention CT after the bolus infusion is started. All patients receive oral contrast material consisting of 5 g of Hypaque powder (Nycomed, New York, NY), in 300 mL of water 30 to 45 min before the scan, and if time permits, a second dose is given either orally or via a nasogastric tube when they arrive at the scanning suite.

Use of IV contrast material increases the conspicuity of parenchymal injuries of the liver (figure 1) and helps to demonstrate active bleeding throughout the abdomen and pelvis. ^1,3,11 At our institution, IV contrast material is routinely administered to all patients undergoing CT scans who do not have a known history of major contrast allergies in order to optimize the difference in attenuation between normal enhanced parenchyma of solid organs and injured parenchyma or hematoma. Studies assessing optimal timing of hepatic parenchymal enhancement for hepatic CT scanning have shown that optimal enhancement varies with the volume and rate of infusion of contrast material. ^12-14 Recent advances in CT technology have led to development of very fast conventional or spiral scanners that minimize misregistration from patient motion and decrease volume averaging. Use of such fast devices requires careful coordination of the initiation of contrast opacification of the liver parenchyma and initiation of scanning. Suboptimal enhancement of the liver parenchyma leading to diminished CT sensitivity for hepatic injury detection can result from early or delayed scanning, missing the peak of hepatic parenchymal opacification, markedly aberrant hemodynamics affecting systemic blood flow, or insufficient contrast material or iodine concentration.

Sites and mechanism of liver injury

The liver is the largest solid organ in the abdominal cavity and is enclosed by the lower rib cage anteriorly and laterally. The right lobe of the liver constitutes 80% of its volume, and it is the most frequently injured region from blunt trauma. Of all patients with blunt trauma to the abdomen, some 15% to 20% involve the liver. ¹⁵ Isolated injuries to the liver occur in less than 50% of blunt trauma victims.

The liver is incapable of achieving spontaneous hemostasis because the hepatic veins lie within rigid canals and contract poorly. ^15,16 Therefore, uncontrolled hemorrhage leading to exsanguination can account for 46% to 80% of the mortality in grade IV and V hepatic injuries. ¹⁷

A variety of mechanisms have been proposed to explain hepatic injuries. Injuries occur to the posterior right lobe from compression against the fixed ribs, posterior abdominal wall, or the spine. Contusion of the dome of the right lobe may result from raised right thoracic cavity pressure propagated through the diaphragm. The posterior abdominal wall can act as sites of shear injury at the liver's attachment by the coronary ligaments to the undersurface of the diaphragm, created by motion of the liver relative to the rest of the body during rapid deceleration. Hepatic injuries can also result from transmission of excessively high venous pressure occurring at the time of impact to remote body sites. ¹⁶

CT appearance of liver injuries

Numerous studies have shown CECT can accurately diagnose the four principal types of liver injury shown by CT: hematoma(s), laceration(s), vascular injuries, and active hemorrhage. ^1,3,18,19 Many systems have been proposed to classify liver injuries. At our institution, hepatic injuries are classified using the CT scan classification of Mirvis et al (table 1) ²⁰ that is based on the American Association for the Surgery of Trauma (AAST) surgical grading scale (table 2). ²¹ This helps to compare treatment and outcome among patients with liver injury from different trauma centers.

Hematomas

Single or multiple hematomas may be seen following blunt trauma. Hepatic hematomas may be intraparenchymal and/or subcapsular (figure 2). On unenhanced CT subcapsular hematoma is hyperdense relative to normal liver parenchyma. On CECT subcapsular hematomas are typically seen as a low attenuation lentiform collection of blood between the liver capsule and the enhancing liver parenchyma (figure 2). Subcapsular hematomas often compress the underlying liver parenchyma; this CT finding helps to differentiate subcapsular hematomas from small amounts of blood or fluid in the perihepatic space. Uncomplicated subcapsular hema-tomas typically resolve within 6 to 8 weeks and the attenuation value decreases with age of the injury. ²²

On CECT, acute hematomas appear as irregular high attenuation areas of clotted blood surrounded by lower density non-clotted blood, bile, or an area of parenchymal contusion (figure 3). With healing, a hematoma develops smoother margins, expands in size, and demonstrates gradual decrease in attenuation values until it eventually appears as serous fluid collection.

Laceration

Acute liver lacerations on CECT have sharp or jagged margins and appear as linear or branching low attenuation areas (figure 4). A compressive force may produce multiple parallel lacerations termed "bear-claw" lacerations. Lacerations ¾ 3 cm in depth from the surface can be classified as superficial, and >3 cm deep as deep lacerations. Liver lacerations that extend into the region of one or more of the major hepatic veins (figure 5) were associated with an increased likelihood of a false negative CT for vascular injury, an increased chance of postoperative hepatic hemorrhage, an increase in delayed hepatic-based complications, and an increased chance of failure of nonoperative treatment (PA Polettie, MD, personal communication, September 1999). Unopacified portal or hepatic veins, or dilated biliary ducts can mimic hepatic lacerations with a branching pattern and need careful evalation of serial images to differentiate the various structures.

As liver lacerations heal, the lesion enlarges, its margins become smoother and assume a rounded to oval configuration on follow-up CT. Unlike subcapsular hematomas, liver lacerations persist for several months. These lesions may gradually decrease in size with time or remain as a well-defined hepatic cyst.

Vascular injuries

Fortunately, injuries to the major hepatic veins or retrohepatic vena cava (juxtahepatic venous injuries) are rarely seen following blunt abdominal trauma. Retrohepatic vena caval injuries from blunt trauma are typically associated with a very high mortality (90% to 100%). In most series, this high mortality has been attributed to excessive blood loss and the inevitable coagulopathy resulting from the long delay in recognizing that juxtahepatic venous injuries were present, bleeding, and the difficulty in obtaining exposure to repair these vascular injuries. ^23,24 Retrohepatic vena caval injuries are suspected on CT on the basis of liver lacerations extending into the major hepatic veins or inferior vena cava; or profuse hemorrhage behind the right lobe of the liver, into the lesser sac, or near the diaphragm (figure 5). Prior suspicion of such vascular injuries can assist the surgeon in planning the surgical approach, controlling hemorrhage, and repairing the vascular injuries through vascular isolation techniques requiring a combined thoracoabdominal approach. A combined therapeutic approach for high-grade liver lacerations with injury to the retrohepatic vena cava involves the trauma surgeon and interventional radiologist. Initially, the trauma surgeon attempts to control the massive hemorrhage by temporary perihepatic packing. Recurrent hepatic parenchymal hemorrhage following liver packing has been successfully managed using transcatheter embolization of arterial sources of blood loss, with bleeding from major hepatic veins controlled by intravenous stent placement. ²⁵

Partial hepatic devascularization can result from an injury to the blood vessels in the perihilar region by a deep laceration, thrombosis of the major hepatic veins, or complete avulsion of the dual blood supply of the liver. On CECT these fragments would be seen as wedge-shaped regions extending to the periphery of liver that fail to enhance (figure 6). The ability to scan during the peak of contrast enhancement of the liver parenchyma with spiral CT makes this scanning technique well suited to visualize pseudoaneurysms or contrast extravasation from an injury to the hepatic artery or its major branches. Hepatic vascular injuries may result from hepatic lacerations extending across the major hepatic arterial/venous branches, shearing forces, or excessive stretching. CT can provide a clear indication whether or not to perform hepatic angiography.

Active hemorrhage

On CECT active hemorrhage in the liver is seen as an irregular area of contrast extravasation (figures 5 and 7). The significant difference in the attenuation value of extravasated contrast material (range 85 to 350 H, mean 132 H) and hematoma (range 40 to 70 H, mean 51 H) is helpful in differentiating active bleeding from clotted blood. ²⁶ Studies performed at our institution to determine the value of CECT in detecting active hemorrhage, showed that CT was a valuable imaging modality to show the exact location of active bleeding and help direct appropriate management. ²⁷

CT is useful in evaluating ongoing or recurrent active bleeding in the liver following surgical or angiographic intervention. Extravasated contrast from a prior angiographic study can mimic active hemorrhage by CESCT. Therefore, it is important to perform both pre- and postcontrast CT scans of the liver in patients who have recently had hepatic angiography to distinguish previously extravasated angiographic contrast from ongoing hepatic bleeding. Other parenchymal liver lesions that could mimic active bleeding on CT include calcified liver lesions and contrast enhancement seen in hepatic hemangiomas.

Periportal low density

Periportal low density refers to regions of low attenuation that parallel the portal vein and its branches on CT (figure 8). Often, periportal low density when seen alone in blunt trauma victims is attributed to hemorrhage tracking along the periportal connective tissues and believed to represent an important, if subtle, CT sign of liver injury. ²⁷ Macrander et al ²⁷ reported periportal low density seen on CT in 62% of patients with liver injury. However, studies performed by Patrick et al ²⁸ in children with blunt abdominal trauma and at our institution suggests that periportal low density without parenchymal liver injury occurs as a result of an elevated central venous pressure (CVP). This elevation results from vigorous IV fluid administration prior to obtaining the CT scan or other trauma-related causes of elevated CVP such as tension pneumothorax or pericardial tamponade resulting in distension of the periportal lymphatic vessels. ^28,29 This CT finding has been noted in several nontraumatic clinical conditions including acute transplant rejection, malignant neoplasm of the liver, liver transplantation, cardiac failure, and cardiac tamponade and is usually attributed to dilatation of the intrahepatic lymphatics from obstruction of the normal hepatic lymphatic drainage. Periportal low density seen on CT without evidence of parenchymal liver injury does not represent hepatic injury and does not warrant hospital admission for observation or follow-up CT. Focal periportal low density seen in close proximity to a liver laceration may represent some focal hemorrhage dissecting into the adjacent portal tracts.

Sonography in hepatic trauma

In the United States, sonography is mainly used for the detection of hemoperitoeum as an indicator of intrabdominal injuries. ^30,31 However, up to 34% of abdominal visceral injuries can occur without hemoperitoneum and 17% of these patients require angiographic or surgical intervention. ³² Sonography can be used to visualize parenchymal hepatic injuries directly. The sensitivity to detect hepatic injuries or hemoperitoneum depends to some extent on the quality of imaging equipment, the conditions under which the study is performed, and the level of experience and training of the operator. Hepatic parenchymal injuries are commonly seen on sonography as discrete hyperechoic area. Other sonographic patterns seen include a diffuse hyperechoic liver parenchyma and a discrete hypoechoic focus in the hepatic parenchyma. ³³ Published data report that sonography underestimated the severity of liver injury in 52.8% of 36 patients and entirely missed the lesion in 11% to 36% of the CT or surgically documented high grade (Mirvis grade 3 or greater) hepatic injury. ^33,34 If a hepatic injury is identified by sonography, it could be potentially followed by this method to document resolution or complications.

Angiography in hepatic trauma

Because of the dual blood supply of the liver subselective embolization can be performed without risk. As a primary treatment modality, transhepatic embolization has been used to improve success in nonoperative management of stable patients with high-grade liver injury. ^35,36 Transcatheter embolization has been successfully used to treat active bleeding, pseudoaneurysms, arterio-venous fistulas, and hemobilia.

Schwartz et al ³⁵ reported their 11-year experience of transcatheter embolization for controlling hemorrhage from hepatic injuries and found this technique to be safe and effective in 21 of the 24 patients included in their study. Angiographic embolization can be performed independently or when surgery fails to control bleeding of hepatic origin. At our institution, patients with CT evidence of active hemorrhage in the liver who do not promptly respond to IV fluid resuscitation are referred for urgent surgical or angiographic intervention.

Conclusion

CECT has revolutionized the diagnosis and management of patients with liver trauma. Information provided by CT allows for determination of the extent of liver injury and identification of other nonhepatic abdominal injury. This information, coupled with clinical assessment and interventional angiography, can be used to optimize management of all grades of hepatic injury.

References

1. Federle MP, Goldberg HI, Kaiser JA, et al: Evolution of abdominal trauma by computed tomography. Radiology 138:637-644, 1981.

2. Feliciano DV: Diagnostic modalities in abdominal trauma. Surg Clinic North Am 71:241-256, 1991.

3. Mirvis SE, Shanmuganathan K: Abdominal computed tomography in blunt trauma. Semin Roentgenol 27:150-183, 1992.

4. Patten RM, Spear RP, Vincent LM, et al: Traumatic laceration of the liver limited to the bear area: CT findings in 25 patients. AJR 160:1019-1022, 1993.

5. Croce MA, Fabian TC, Menke PG, et al: Non operative management of blunt hepatic trauma is the treatment of choice for hemodynamically stable patients. Results of prospective trial. Ann Surg 221:747-753, 1995.

6. Sherman HF, Savage BA, Jones LM, et al: Nonoperative management of blunt hepatic injuries at any grade? J Trauma 37:616-621, 1994.

7. Becker CD, Dal I, Baer HU, Vock P: Blunt hepatic trauma in adults: Correlation of CT injury grading with outcome. Radiology 1201:215-220, 1996.

8. Holland MJ, Little JM: Non-operative management of blunt liver injuries. Br J Surg 78:968-972, 1991.

9. Knudson MM, Lim RC Jr, Oakes DD, Jeffrey BR Jr: Nonoperative management of blunt liver injuries in adults: The need for continued surveillance. J Trauma 30:1494-1500, 1990.

10. Patcher HL, Knudson MM, Esrig B, et al: Status of nonoperative management of blunt hepatic injuries in 1995: A multicenter experience with 404 patients. J Trauma 40(1):31-38, 1996.

11. Roberts JL, Dalen K, Bosanko CM, Jafir ZSH: CT in abdominal and pelvic trauma. Radiographics 13:735-752, 1993.

12. Chambers TP, Baron RL, Lush RM: Hepatic CT enhancement. Part II. Alteration in the contrast volume and rate of injection within the same patient. Radiology 193:518-522, 1994.

13. Foley WD, Hoffmann RG, Quiroz FA, et al: Hepatic helical CT: Contrast material injection protocol. Radiology 192:367-371, 1994.

14. Heiken JP, Brink JA, McClennan BL, et al: Dynamic contrast-enhanced CT of the liver: Comparison of contrast medium injection rates and uniphasic and biphasic injection protocols. Radiology 187:327-331, 1993.

15. Williams PL, Warwick R, Dyson M, Bannister LH: Abdominal vicera. In: Grays Anatomy ed 37, pp 1384-1393, 1989.

16. Rouhana SW: Biomechanics of abdominal trauma. In: Nahum AM, Melvin JW (eds): Accidental Injuries: Biomechanics and Prevention, pp 391-401. New York, Springer-Verlag, 1993.

17. Asensio JA, Demetriades D, Chahwan S, et al: Approch to management of complex hepatic injuries. J Trauma 48:66-69, 2000.

18. Mirvis SE, Dunham CM: Abdominal/pelvic trauma. In: Mirvis SE, Young JWR (eds): Imaging in Trauma and Critical Care, pp 148-242. Baltimore, Williams & Wilkins,1992.

19. Gay SB, Sistrom CL: Computed tomographic evaluation of blunt abdominal trauma. Radiol Clin North Am 30:367-388, 1992.

20. Mirvis SE, Whitley NO, Vainwright J, Gens D: Blunt hepatic trauma in adults: CT based injury classification and correlation with prognosis and treatment. Radiology 171:27-32, 1989.

21. Moore EE, Cogbill TH, Jurkovitch J, Shackford SR: Organ injury scale: Spleen and liver (1994 revision). J Trauma 38:333, 1995.

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

23. Beal SL: Fatal hepatic hemorrhage: An unresolved problem in the management of complex liver injuries. J Trauma 30:163-169, 1990.

24. Burch JM, Feliciano DV, Mattox KL: The atriocaval shunt facts and fiction. Ann Surg 207:555-568, 1988.

25. Denton JR, Moore EE, Coldwell DM: Multimodality treatment for grade V hepatic injures: Perihepatic packing, arterial embolization, and venous stenting. J Trauma 42:964-968, 1997.

26. Shanmuganathan K, Mirvis SE, Sover ER: Value of contrast-enhanced CT in detecting active hemorrhage in patients with blunt abdominal or pelvic trauma. AJR 161:65-69, 1993.

27. Macrander SJ, Lawson TL, Foley WD, et al: Periportal tracking in hepatic trauma: CT features. J Comput Assist Tomogr 13:952-957, 1989.

28. Patrick EL, Turner BI, Atkinson GO, Winn KJ: Pediatric blunt abdominal trauma: Periportal tracking at CT. Radiology 183:689-691, 1992.

29. Shanmuganathan K, Mirvis SE, Amerosa M: Periportal low density on CT in patients with blunt trauma: Association with elevated venous pressure. AJR 160:279-283, 1992.

30. Boulanger BR, Brenneman FD, McLellan BA, et al: A prospective study of emergent abdominal sonography after blunt trauma. J Trauma 39:325-330, 1995.

31. Lentz KA, McKenney MG, Nunez DB, et al: Evaluation of blunt abdominal trauma: Role for sonography. J Ultraound Med 15:447-451, 1996.

32. Shanmuganathan K, Mirvis SE, Sherbourn CD, et al: Hemoperitonium as the sole indicator of abdominal visceral injury: A potential limitation of screening abdominal US for trauma. Radiology 212:423-430, 1999.

33. Richards JR, McGahan JP, Pail MJ, et al: Sonographic detection of blunt hepatic trauma: Hemoperitonium and parenchymal injury patterns of injury. J Trauma 47:1092-1097, 1999.

34. Wong-You-Cheong JJ, Shanmuganathan K, Chiu WC, et al: Sonography compared to CT in the detection and grading of liver injuries. Abstract presented at the American Roentgen Ray Society meeting, April 26-May 1, 1998, San Francisco, CA.

35. Schwartz RA, Teitelbaum GP, Katz MD, Penetecost MJ: Effectiveness of transcatheter embolization in the control of hepatic vascular injuries. J Vasc Interven Radiol 4:359- 365, 1993.

36. Hahiwara, Yukioka K, Ohta S, et al: Nonsurgical management of patient with blunt hepatic injury: Efficacy of transcatheter arterial embolization. AJR 169:1151-1156, 1997.