Applied Radiology

Dr. Mirvis is with the Department of Diagnostic Radiology, University of Maryland Medical Center and Maryland Shock-Trauma Center, Baltimore, MD. He is also a member of the Editorial Board of this journal.

Recognition of the spleen's role in immunity­­particularly for phagocytosis of encapsulated bacteria­­and the potential for rare, but often fatal, overwhelming postsplenectomy sepsis has led to greater efforts to conserve the spleen following injury.

The trend towards nonoperative treatment of splenic injuries was first established in the pediatric population. The higher risk of postsplenectomy sepsis coupled with a high rate of success with nonoperative management (NOM) in the pediatric population fostered this approach. In recent years, there has been growing interest in NOM of splenic injury in the adult population.

Pachter et al ¹ described the changing pattern in management of splenic rupture. Over a 6-year period prior to 1990 only 13% of adult patients with splenic injury were managed initially nonoperatively. In the next 6-year period, from 1990 to 1996, this number increased to 54%. The authors noted in this comparison study that splenectomy was associated with an excessive need for blood transfusion and a high rate of postoperative sepsis. In the latter period, the success rate for initial NOM was 98%.

Goan et al ² described using NOM in 24 patients with severe liver and spleen injuries and a moderate to large quantity of hemoperitoneum. Among these patients, 18 had grade 3 or greater injuries on the American Association for the Surgery of Trauma (AAST) Organ Injury Scale (table 1). ³ They report a success rate of 84.6% and noted that the nonoperative group had fewer transfusions, fewer complications, and shorter hospital stays than patients undergoing early surgery. Two of the 24 patients failed NOM for spleen-related causes. They, as well as Pachter et al, ¹ noted that the absolute amount of hemoperitoneum was not a good indicator of the likely outcome of NOM.

Criteria for nonoperative management

The criteria used to select patients for NOM of splenic injury varies among institutions. A universal requirement is that the patient is or becomes hemodynamically stable without requiring on-going aggressive resuscitation. A variety of other clinical factors have been considered in the decision to try NOM including age, presence of abdominal pain, neurologic or mental status, availability of intensive care unit monitoring, transfusion requirements, quantity of hemoperitoneum, and extent of injury as shown on CT scan. Smith et al ⁴ outlined specific clinical criteria for attempting nonoperative treatment including: blood pressure >90 mmHg, pulse <110 after 2 liters of fluid, age >15 and <55 years, CT grade 1 to 3, lack of severe head injury, and a negative abdominal examination. They reported only a 4% failure for NOM using these criteria among patients in the study. Alternatively, Brasel et al ⁵ found that neither age nor neurologic status was a significant factor in predicting the outcome of NOM of splenic injury. This study described a growing incidence of NOM of splenic injury in their institution and an 84% success rate among 116 patients initially managed without surgery.

CT injury grading

Over the last decade, several attempts have been made to show that the CT grade of injury could be used to predict outcome of NOM (figures 1 through 4). A variety of CT grading systems have been proposed and studied, but, in general, these have not shown enough reliability in predicting outcome for a given patient. Mirvis et al, ⁶ in a study of 35 surgically treated patients, found that the failure rate was greater among those with grade 1 or 2 injury than for with grade 3 and 4 injuries as found on preoperative CT. Similarly, Schurr et al ⁷ showed, using logistic regression, that CT injury grade had no predictive value for outcome for a population of 89 patients managed without surgery, 12 of whom ultimately required surgery.

Sutyak et al ⁸ concluded that CT scanning is inaccurate for estimating the extent of splenic injury, often resulting in an underestimation of severity. Further, they found a significant level of disagreement between radiologists in establishing a CT-based injury grade. These factors contribute to the poor correlation between CT appearance and outcome in NOM of splenic injury. Similarly, Shapiro et al ⁹ showed that CT underestimated spleen injury in 36 patients at a statistically significant level, with CT having an average injury score of 3.2 versus a surgical injury score of 3.6 on an injury scale of 5.

CT grading is also hampered by the fact that the CT shows only a "single snapshot in time" of the injured spleen, but does not document progression or healing without serial evaluations. One of the greatest contributions of abdominal CT in the setting of spleen injury is to exclude other intraperitoneal injuries that would otherwise mandate surgical exploration.

CT appearance of arterial IV contrast extravasation

CT scanning with intravenous (IV) contrast enhancement can often demonstrate direct evidence of vascular injury as shown by failure of expected parenchymal enhancement (infarct) or by focal extravascular contrast leak. ^10,11 Faster scan times provided by helical or multirow detector scanning and improved timing of peak arterial enhancement contribute to improving sensitivity for recognition of direct vascular injury. A contrast leak into the splenic parenchyma can occur into the subcapsular, perisplenic, or intraperitoneal space. Typically these appear as oval, rounded, or linear areas of focal high attenuation similar to, or greater in density, than an adjacent artery or the aorta. A contrast leak directly into the parenchyma, sometimes referred to as a contrast "blush," will persist within the parenchyma after arterial contrast wash-out occurs and is usually surrounded by a nonenhancing hematoma from prior bleeding (figures 5 and 6).

A posttraumatic pseudoaneurysm results from an injury to the arterial wall in which the adventitia and perivascular tissues form a wall that focally contains the contrast (blood) leak. In this case, the contrast density should decrease in the splenic parenchyma as arterial enhancement levels diminish. The natural progression of splenic pseudoaneurysm is not clearly known. Though some splenic pseudoaneurysms heal by spontaneous thrombosis without intervention, recent studies have shown that up to 67% of these lesions may ultimately rupture and therefore represent a strong predictor for failure of NOM in splenic injury. ^12,13 In the arterial phase, in which abdominal CT is usually acquired for trauma patients, parenchymal extravasation and pseudoaneurysm have a similar appearance.

Traumatic arteriovenous fistula can result from trauma and also appears as a focal hyperdense lesion similar to parenchymal extravasation or pseudoaneurysm. Ongoing bleeding into the subcapsular, or perisplenic intraperitoneal space typically appears as a linear high-density region with a surrounding hematoma.

CT findings that can mimic splenic vascular lesions include islands of enhanced splenic parenchyma surrounded by low attenuation splenic laceration-contusions, and intact intrasplenic vessels traversing the center or periphery of parenchymal laceration(s) simulating hemorrhage surrounding a focal pseudoaneurysm. Failure to detect splenic vascular lesions by CT can be related to sub-optimal contrast enhancement, particularly in obese patients with decreased tissue contrast resolution or delayed scanning well beyond the peak of splenic-parenchymal enhancement with "wash-out" of contrast occurring in pseudoaneurysms by unenhanced blood.

Understanding these potential diagnostic pitfalls, use of an optimal volume and concentration of IV contrast material, scanning at the peak of visceral contrast enhancement, and objective contrast bolus timing techniques can improve detection of splenic vascular injury. The absolute sensitivity (required rate of bleeding or size of lesion) for detection of arterial bleeding by technically optimized single-slice helical CT is, to the author's knowledge, not known. Hypothetically, multirow detector CT imaging systems with subsecond scanning performed with efficient contrast bolus tracking should improve detection of these lesions.

Predictive value of CT contrast extravasation in blunt splenic injury

Recently Schurr et al, ⁷ Davis et al, ¹² and Sugg et al ¹³ have demonstrated the significance of a splenic "contrast blush" in predicting failure of NOM. In the study by Schurr et al, 12 of 89 patients (13%) initially managed nonoperatively failed this approach. A hyperdense collection was noted within the spleen parenchyma on CT scan in 8 of these 12 (67%). Failure of NOM was highly correlated with the CT contrast blush (P <0.0001). The contrast blushes were demonstrated to be either contained pseudoaneurysms of splenic vessels or foci of active bleeding (figures 5 and 6). Davis et al then showed that angiography of patients with a contrast blush on CT typically demonstrated parenchymal pseudoaneurysms that could be successfully treated by selective embolization. In this study, 26 of 322 patients (8%) managed nonoperatively showed the CT contrast blush in the spleen, with 20 patients having distal embolization of pseudoaneurysms documented by angiography. Davis and colleagues also showed a statistically significant improvement in outcome of NOM when compared to NOM without angiographic evaluation.

In 1991, Sclafani et al ¹⁴ documented the value of proximal splenic artery embolization in patients displaying angiographic abnormalities in improving splenic salvage rate for NOM. Subsequently, he and his coworkers ¹⁵ described the use of proximal splenic artery embolization in 150 patients with blunt splenic injury undergoing NOM. In the later report, 87 of 90 patients were managed successfully with bed rest alone and an additional 60 underwent angiographic study and proximal splenic artery embolization. The overall salvage rate in blunt splenic trauma was 88%, including 22 patients managed surgically. The overall success rate of NOM was 97% in patients managed by either bed rest alone or angiography and embolization of documented vascular injuries. The authors noted that CT grade did not reliably predict outcome, with many patients with CT grades 3 and 4 injury treated successfully with NOM. In addition, they suggested that enhanced abdominal CT could be used as a triage test for splenic angiography.

A recent study at the University of Maryland Shock-Trauma Center verifies that power-injected bolus enhanced spiral CT is a useful study to determine the need for splenic angiography, based primarily on the presence of a CT parenchymal "contrast blush." In this study all patients with evidence of splenic injury on CT underwent urgent splenic arteriography. Patients with angiograms revealing contained parenchymal contrast extravasation or active bleeding had selective embolization and distally in all but one case (figures 5 and 6). There were two failures of embolization, including the single patient who was embolized by proximal coils only. The benefits of distal embolization included preservation of splenic parenchyma with minimal chance of rebleeding due to preserved collateral flow. The advantage of proximal embolization is that it provides a quicker, technically easier procedure with reduction of splenic arterial perfusion pressure with collateral preservation of splenic immune function (RB Kranis MD, personal communication, January 1999). Both proximal and distal embolization can result in splenic infarctions of varying size, but are rarely complicated by infection and/or abscess formation (K Killeen, MD, personal communication, March 2000). The best technique of splenic embolization is yet to be established.

Future studies will attempt to document more precisely the CT findings that mandate angiographic assessment of the spleen for potential embolization. The impact of combined spiral CT and splenic arteriographic evaluation with embolization on splenic preservation remains to be firmly established. It must be emphasized that while NOM of splenic injury has achieved high success rates in recent years, based on use of both clinical and imaging parameters, the need for early, aggressive operative intervention for splenic injuries associated with hemodynamic compromise or other abdominal injuries requiring surgery must remain paramount.

Follow-up CT

The appropriate long-term clinical and imaging follow-up of splenic injury managed nonoperatively has not yet been established. Specifically, the value of obtaining a follow-up CT scan after initial identification of splenic injury by CT is debatable. Patients who develop hemodynamic instability or increasing resuscitation requirements during attempted NOM either undergo repeat CT or proceed directly to celiotomy. Lawson et al ¹⁶ found that routine follow-up CT scanning was not useful in 22 clinically stable patients with splenic injury managed nonoperatively. Similarly, Thaemert et al ¹⁷ assessed the role of follow-up CT for nonoperatively managed splenic injuries in 73 children and adults. Only 1 of 49 stable patients undergoing follow-up CT had management altered due to the CT result, with total CT procedure charges exceeding $54,000. On the other hand, Federle ¹⁸ has pointed out that as many as 8% to 29% of blunt trauma patients with splenic injury have delayed bleeding hours to weeks after injury and follow-up CT might show early progression of injuries. Also, he argues that follow-up CT can be very helpful in demonstrating splenic healing as a prelude to resumption of full physical activity. As noted by Federle, one criterion for return to full physical activity (particularly contact sports) might be demonstration of complete healing by repeat CT at 6 to12 weeks postinjury.

Lynch ¹⁹ described use of serial sonography to follow splenic injury to document healing after initial CT diagnosis. He demonstrated a correlation between injury grade by CT and time to complete healing. At the Shock-Trauma Center at University of Maryland there have been several instances in which repeat follow-up CT in clinically stable patients revealed progression of splenic injury or delayed appearance of pseudoaneurysms requiring splenic angiography or surgical inspection of the spleen (figure 7).

In most cases, if there is no clinical evidence of injury progression and no other injuries that require continued hospitalization, most patients are discharged in 3 to 5 days. Some physicians will obtain a repeat CT scan prior to discharge to exclude any subclinical progression of injury, but this is by no means universal. Also, some physicians may opt to perform a repeat CT study prior to the patient resuming a physically demanding activity to verify complete healing.

Songraphy for splenic injury

Krupnick et al ²⁰ reported a 38% incidence of missed splenic injury and a 53% incidence of injury downgrading in pediatric patients based on sonography (radiologist/technologist performed) as compared with CT for initial diagnosis. This finding argues strongly against the use of sonography for initial the diagnosis of splenic injury. In addition, 7 of 32 splenic injuries (22%) in this series had no associated hemoperitoneum, raising questions about the role of sonography in screening for abdominal injury based on the presence or absence of free intraperitoneal fluid. AR

References

1. Pachter HL, Guth AA, Hofstatter SR, Spencer PC: Changing patterns in the management of splenic trauma: The impact of nonoperative management. Ann Surg. 5:708-719, 1998.

2. Goan YG, Huang MS, Lin JM: Nonoperative management for extensive hepatic and splenic injuries with significant hemoperitoneum in adults. J Trauma 45:360-365, 1998.

3. Moore EE, Cogbill TH, Jurkovich GH, et al: Organ injury scaling: Spleen and liver (1994 revision). J Trauma 38:323-324, 1995.

4. Smith JS, Cooney RN, Mucha P: Nonoperative management of the ruptured spleen: A revalidation of criteria. Surgery 120:745-751, 1996.

5. Brasel KJ, DeLisle CM, Olson CJ, Borgstrom DC: Splenic injury: trends in evaluation and management. J Trauma 44:283-286, 1998.

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

7. Schurr MJ, Fabian TC, Gavant M, et al: Management of blunt splenic trauma: Computed tomographic contrast blush predicts failure of nonoperative management. J Trauma 39:507-513, 1995.

8. Sutyak JP, Chiu WC, D'Amelio LF, et al: Computed tomography is inaccurate in estimating the severity of adult splenic injury. J Trauma 39:514-518, 1995.

9. Shapiro MJ, Krausz C, Durham RM, Mazuski JE: Overuse of splenic scoring and computed tomographic scans. J Trauma 47:651-658, 1999.

10. Shanmuganathan K, Mirvis SE, Reaney SM: CT appearance of contrast extravasations associated with injury sustained from blunt abdominal trauma. Pictorial Essay. Clin Radiol 50:182-187, 1995.

11. 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.

12. Davis KA, Fabian TC, Croce MA, et al: Improved success in nonoperative management of blunt splenic injuries: Embolization of splenic artery pseudoineurysms. J Trauma 44;1008-1015, 1998.

13. Sugg SL, Gerndt SJ, Hamilton BJ, et al: Pseudoaneurysms of the intraparenchymal splenic artery after blunt abdominal trauma: A complication of nonoperative therapy and its management. J Trauma 39:593-595, 1995.

14. Sclafani SJA, Weisberg A, Scalea TM, et al: Blunt splenic injuries: Nonsurgical treatment with CT, arteriography, and transcatheter arterial embolization of the splenic artery. Radiology 181:189-196, 1991.

15. Sclafani SJ, Shaftan GW, Scalea TM, et al: Nonoperative salvage of computed tomography-diagnosed splenic injuries: Utilization of angiography for triage and embolization for hemostasis. J Trauma 39:818-827, 1995.

16. Lawson DL, Jacobson JA, Spizarny DL, Pranikoff T: Splenic trauma: Value of follow-up CT. Radiology 194:97-100, 1995.

17. Thaemert BC, Cogbill TH, Lambert PJ: Nonoperative management of splenic injury: are follow-up computed tomographic scans of any value? J Trauma 43:748-751, 1997.

18. Federle MP: Splenic trauma: Is follow-up CT of value? Radiology 194:23-24, 1995.

19. Lynch JM: CT grade of splenic injury is predictive of time to reach radiographic healing. J Ped Surg 32:1093-1095, 1997. ]

20. Krupnick AS, Teitelbaum DH, Geiger JD, et al: Use of abdominal sonography to assess pediatric splenic trauma. Potential pitfalls. Ann Surg