Computed tomography (CT) has not been used routinely to evaluate penetrating injuries to the thoraco-abdominal region, abdomen proper, or pelvis. This article discusses the authors’ experience with single- and quad-detector helical CT combined with oral, intravenous, and colonic contrast (triple-contrast) to assess hemodynamically stable patients who have penetrating injuries to the torso.
is an Associate Professor and
is a Professor of Radiology, Department of Diagnostic Radiology
and Maryland Shock-Trauma Center, University of Maryland School
of Medicine, Baltimore, MD. Dr. Mirvis is also the
Editor-in-Chief of this journal.
Firearm-related injuries have become a public health problem
with a devastating impact on American society.
There are substantial emotional, physical, and financial costs
associated with firearm injuries, which are common in our society
and often lethal. For every firearm death, it is estimated that 3
to 5 other nonfatal firearm injuries occur.
During the last decade, computed tomography (CT) has been used to
evaluate patients with penetrating injuries to the back and flank,
but it has not been used routinely to evaluate penetrating injuries
to the thoraco-abdominal region, abdomen proper, or pelvis. This
article discusses the authors' experience with helical CT (single-
and quad-detector helical CT) combined with oral, intravenous, and
colonic contrast (triple-contrast) to assess hemodynamically stable
patients who have penetrating injuries to the torso.
Most civilian gunshot wounds result from medium-energy handguns
with a muzzle velocity <1000 ft/sec. The extent of tissue damage
caused by a projectile is more severe if the missile has high
kinetic energy, yaws (tilts) early in its path through tissue, is
of high mass, strikes bone (creating secondary missiles), or
expands, deforms, or fragments.
A permanent cavity results from the missile crushing the tissues it
strikes. Temporary cavitation from stretching results within
several milliseconds, following the passage of the bullet through
the tissues. Severe damage results from formation of the temporary
cavity in fluid-filled organs, such as the bladder or heart, as
well as in organs with dense parenchyma, such as the liver (Figure
1), kidney, and brain.
More elastic tissue, such as lung parenchyma or muscle, resists
severe damage from the temporary tissue cavity. Stab wounds are
low-energy injuries that cause tissue damage only along the wound
Urgent laparotomy is the standard of care in the United States
for patients with gunshot or stab wounds who are hemodynamically
unstable or have overt clinical signs of peritonitis. Prior to the
mid-1960s, mandatory laparotomy was performed routinely in all
patients admitted with penetrating torso injuries.
This approach was based on military experience since World War I.
If this management is applied to all civilian penetrating injuries,
15% to 20% of abdominal gunshot wound victims and 35% to 53% of
abdominal stab wound victims will have an unnecessary
Since the mid-1960s, most trauma surgeons have approached the
decision of whether or not to perform a laparotomy on patients with
penetrating trauma based on the significance of visceral injury or
tissue damage, rather than simply on the presence or absence of
This selective policy used to be applied more readily to patients
with stab wounds and has been accepted more slowly for the
management of hemodynamically stable patients with gunshot wounds.
A selective surgical approach should reduce the number of negative
and nontherapeutic procedures without increasing the number,
morbidity, or mortality of missed or delayed diagnosis of
significant injuries. The optimal method to triage hemodynamically
stable patients for peritoneal violation and significant visceral
injury with equivocal peritoneal signs following penetrating torso
(defined by the area between the inter-nipple line and upper third
of the thigh) injury is controversial. The anatomic location of the
entry wound and the local practice at a given trauma or emergency
center may determine the triage method upon which selective
management will be determined.
Selective surgical management
The two most commonly used methods to select patients with
penetrating torso trauma for initial nonsurgical management include
observation and local wound exploration followed by diagnostic
peritoneal lavage (DPL). Other triage methods include ultrasound,
CT, and diagnostic laparoscopy. Prior studies have shown that
clinical observation with serial physical examination (ideally
performed by the same examiner) may be practiced with excellent
results to reduce the incidence of nontherapeutic laparotomy for
These studies show that 34% to 72% of patients with penetrating
injury to the anterior abdomen or back who are hemodynamically
stable and without overt signs of peritonitis are suitable for
observation alone. Missed injuries that require subsequent
operation occurred in 4% to 6% of patients in the observation-only
group; time of delay in treatment ranged from 3 hours to 5 days.
Compared with mandatory laparotomy, this approach reduced the
nontheraputic laparotomy rate of 53% down to 3% to 8% in some
Assuming that the patient is hemodynamically stable with no or
equivocal clinical signs of peritonitis, local wound exploration is
performed initially in the admissions area under local anesthesia
and sterile conditions to determine whether the peritoneum has been
If the peritoneum has not been violated, the exploration site and
skin are closed and the patient can potentially be discharged,
given an isolated penetrating torso injury. If there is penetration
of the peritoneum or if the wound exploration is equivocal for
peritoneal violation, DPL is performed to determine if there is
intraperitoneal blood. Unfortunately, a positive DPL result may
occur from injury to the peritoneal lining itself or from very
minor and self-limited parenchymal injury, making the result
true-positive for hemoperitoneum, but leading to potentially
In the past decade, triple-contrast CT has become the standard
imaging study to evaluate patients with penetrating injuries to the
back and flank
due to its capacity to optimally reveal injuries to the
retroperitoneal organs. The principal purpose of the CT study is to
divide these injuries into those that are superficial; those with
retroperitoneal penetration without significant visceral injury; or
those with significant retroperitoneal and intraperitoneal
injuries. Isolated retroperitoneal injuries infrequently cause
critical injury to the retroperitoneal viscera. Retroperitoneal
injuries often do not present with overt clinical signs or symptoms
on admission and are poorly detected by DPL alone.
Oral and intravenous contrast material is used routinely for the CT
study of flank and back penetrating injury to optimize detection of
injuries to the intraperitoneal bowel, solid organs, mesentery,
genitourinary system, and retroperitoneal portions of the duodenum
and colon. Most patients with dermal or superficial muscle injury
are discharged within a short time. Patients with minor or low-risk
injuries (small retroperitoneal hematoma or minor renal injury) are
managed with short-term observation. Patients with major visceral
injuries might be managed nonoperatively, but may require surgery
Triple-contrast CT is a popular technique to triage patients with
penetrating flank and back injuries because of its excellent
sensitivity (89% to 100%), negative predictive value (89% to 100%),
and accuracy (92% to 96%) to identify patients likely to require
intervention for injury.
Triage by CT leads to a very low rate of subsequent nontherapeutic
Concerns regarding the accuracy of CT to diagnose hollow viscus
injuries have, in the past, diminished trauma surgeons' enthusiasm
to request CT to triage patients with penetrating injuries to other
anatomic regions in the torso. The bowel is the most commonly
injured abdominal organ resulting from penetrating trauma.
In some studies, surgeons have used DPL or observation as an
adjunct to CT to diagnose potential intraperitoneal injuries.
However, recent retrospective studies by Grossman et al
and Ginzberg et al
have shown that helical CT is accurate in determining the
trajectory of gunshot wounds and the presence of peritoneal
violation in patients with nonflank and back wounds of the torso.
In the retrospective study performed by Ginzberg et al,
triple-contrast CT was used as the initial screening study to
detect peritoneal violation and minimize nontherapeutic laparotomy.
Patients with equivocal helical CT results for peritoneal violation
had cavitary endoscopy to verify CT findings. There were no missed
injuries among 53 patients with negative CT who were observed for
23 hours. For abdominal wounds, CT had a specificity of 54%, a
negative predictive value of 100%, and an overall accuracy of 71%;
for flank wounds, CT had a specificity of 98%, a negative
predictive value of 100%, and an overall accuracy of 98%.
Both retrospective studies concluded that CT was safe and effective
as an initial screening study to detect peritoneal violation in
hemodynamically stable patients with torso gunshot wounds.
Prospective studies were performed at the authors' level-1
trauma center to evaluate triple-contrast helical CT in patients
with penetrating torso trauma.
These studies found that CT is highly accurate in demonstrating
peritoneal violation and in predicting the need for subsequent
In this research, CT studies were positive for peritoneal violation
or a significant retroperitoneal injury in 34% (35 of 104) of
patients and negative in 66% (69 of 104) of patients. Among
patients with a negative CT, 97% (67 of 69) were managed
nonoperatively without missed injuries. CT had 100% (19 of 19)
sensitivity, 96% (69 of 72) specificity, 100% (69 of 69) negative
predictive value, and 97% (101 of 104) accuracy in predicting the
need for laparotomy.
Triple-contrast multidetector spiral CT is obtained from the
inter-nipple line to the symphysis pubis, following administration
of intravenous contrast material. CT was performed using an MX 8000
(Philips Medical Systems, Best, Netherlands) with a collimation of
2.5 mm (2.5 * 4 images per rotation) and a table speed of 10
mm/sec. Intravenous contrast material (300 mg iodine/mL, 150 mL) is
administered at 3 mL/ sec using a power injector (Medrad 4, Medrad,
Pittsburg, PA). A total volume of 600 mL of 2% sodium diatrizoate
(Hypaque sodium, Amersham Health, Princeton, NJ) oral contrast
material is administered 30 minutes before and again immediately
before initiation of CT. An enema of 1 to 1.5 L of 2% sodium
diatrizoate is also administered to opacify the colon on the CT
table before scanning. Delayed images are obtained routinely in the
renal excretory phase to evaluate the renal collecting system and
to verify evidence of active arterial extravasation or
At the authors' institution, any patient with hemodynamic
instability, evidence of peritonitis on clinical examination
(rigidity, rebound, tenderness, or considerable tenderness away
from the wound site), or evisceration must undergo laparotomy.
Other indications for surgery include free intraperitoneal fluid on
focused abdominal sonography for trauma (FAST), rectal bleeding,
hematemesis, or pneumoperitoneum diagnosed by chest or abdominal
Wound tracts and peritoneal violation
CT can demonstrate the extent of a wound tract accurately by
showing air (Figure 2), hemorrhage (Figure 1), bone (Figure 2), or
bullet fragments along the wound.
Prior knowledge of the wound entry site and use of optimum CT
windows and level (window = 550, level = 75) assist in identifying
subtle wound tracts and identification of visceral injury sites.
Low-energy knife wound tracts may be subtle compared with
high-energy gunshot wound tracts that are seen clearly on CT from
the presence of larger amounts of hemorrhage, air, and metal
CT findings of peritoneal violation include the presence of a
wound tract outlined by air, hemorrhage, or bullet fragments due to
the missile or knife entering the peritoneal cavity (trajectory of
missile or knife); intraperitoneal free air or free fluid (Figure
3); bullet fragments; or intraperitoneal organ (Figures 1, 3, and
4), mesenteric, or vascular injury. The most common CT finding
among patients with peritoneal violation is intraperitoneal free
Other CT signs useful to diagnose peritoneal violation include
intraperitoneal visceral injury seen in 60% (21 of 35) of patients
and free intraperitoneal air seen in 43% (15 of 43) in the authors'
These two signs are more often associated with gunshot wounds than
with stab wounds.
At the authors' institution, the CT scan is considered positive
if there is evidence of peritoneal violation; injury to the renal
collecting system, retroperitoneal colon, or duodenum; or major
vascular injury. CT studies are considered negative when there is
no evidence of peritoneal violation (Figure 5), with or without
minor extraperitoneal injuries (minor renal injuries, small
retroperitoneal or pelvic hematomas). Prospective studies have
reported that CT is highly accurate in excluding peritoneal
violation and can be used reliably to select patients for selective
nonsurgical management without a high likelihood of significant
Solid organ injury
Injuries to the liver and spleen are seen more frequently
following penetrating injuries to the thoracoabdominal region
(Figures 1, 3, and 4), upper abdomen, and flank.
The liver is the most commonly seen injured solid organ in patients
with penetrating trauma to the torso.
On contrast-enhanced CT, intraparenchymal hematomas are seen along
the wound tract as low-attenuation areas compared with the normal
enhancing parenchyma of the liver or spleen. Intraparenchymal
hematomas (Figure 1) seen along the wound tract are usually larger
as a result of the higher energy gunshot wound, causing more tissue
damage than a stab wound. Subcapsular hematomas are seen as
low-attenuation collection(s) between the parenchyma and capsule.
Subcapsular hematomas cause direct compression of the underlying
normal parenchyma. Lacerations occur along the wound tract, as a
result of the crushing of the parenchyma. On contrast-enhanced CT,
acute lacerations are seen as linear low-attenuation areas compared
with the normal enhancing parenchyma (Figure 3).
CT has proven very helpful in selecting patients with isolated
liver injury from thoracoabdominal penetrating trauma for
successful nonoperative management in studies reported by Renz and
and the current authors.
In these studies, nonoperative management was successful in all
patients with isolated penetrating liver injury. No patient needed
laparotomy or developed complications requiring intervention. In
the study performed at the authors' trauma center, adjunctive
angiography and embolization were required for 3 patients to treat
active hepatic hemorrhage shown by CT in 1 patient, and clinical
evidence of ongoing hemorrhage in 2 others. Triple-contrast spiral
CT provided valuable information permitting attempted nonsurgical
management, including determination of bullet trajectory,
identification of the extent of liver injury, and, principally,
exclusion of other injuries that would mandate celiotomy.
Bowel and mesenteric injuries
Unlike in cases of blunt trauma, hollow viscus injuries are the
most common injuries seen following all penetrating trauma.
The concerns about the varying sensitivity and specificity of CT in
diagnosing bowel injury have limited the use of CT to evaluate
penetrating torso injury for the past decade. At the authors'
institution, routine administration of oral and rectal contrast
material is used to help to opacify the bowel and increase the
sensitivity of CT to detect small amounts of intermesenteric fluid,
mesenteric contusions, or infiltration of the mesenteric fat that
may provide the only evidence of bowel injury. Also, adequate
distention of the bowel by gastrointestinal contrast material
enhances the ability to demonstrate bowel-wall pathology. Specific
CT findings of bowel or mesenteric injury in patients with
penetrating trauma include extravasation of oral or colonic
contrast (Figure 6), bowel-wall thickening (Figures 3 and 7),
mesenteric bleeding, discontinuity or defect in the bowel wall, or
focal mesenteric hematoma or infiltration. A wound tract extending
up to the wall of a hollow viscus (Figure 2) is also considered a
sign of bowel injury on CT.
In the prospective study performed at the authors' institution,
of the 35 patients in the study with peritoneal violation, 13 (37%)
had bowel or mesenteric injury as shown on CT.
The wound tract extending adjacent to injured bowel was the most
common CT finding seen in 69% (9 of 13) of the patients with bowel
injury. Other CT findings of bowel injury included bowel-wall
thickening in 54% (7 of 13) and oral or rectal contrast
extravasation in 15% (2 of 13). All 7 patients with bowel-wall
thickening had bowel injury that required intervention.
Unlike in cases of blunt trauma, the inability to use the
presence of isolated free intraperitoneal air or fluid as a
diagnostic or suspicious CT sign of bowel injury makes diagnosis of
bowel injury far more challenging in the setting of penetrating
trauma. In penetrating trauma, free intraperitoneal blood may
result from bleeding from an injury to the abdominal wall (the
peritoneal lining itself) or from extraperitoneal bleeding leaking
through a defect caused by a wound tract into the peritoneal
cavity. Free intraperitoneal fluid in the absence of a solid-organ
injury should be considered a CT finding of peritoneal violation
and not a nonspecific finding of bowel injury. The presence of
pneumoperitoneum without evidence of a pneumo-thorax,
pneumomediastinum, or retroperitoneal air decompressing into the
peritoneal cavity is a specific CT finding of bowel injury and
mandates surgery in cases of blunt trauma. However, free air may be
introduced into the intraperitoneal cavity by a bullet or knife
during violation of the peritoneum, thus pneumoperitoneum is a sign
of peritoneal violation and also cannot be used as a specific CT
finding of bowel injury in cases of penetrating injury.
Patients with any penetrating injury to the thoracoabdominal
region with radiographic or CT evidence of a wound trajectory
extending in close proximity to the diaphragm are likely to have a
diaphragmatic injury. Most patients sustaining diaphragm injury
from blunt trauma have long lacerations often permitting immediate
or early herniation of abdominal structures into the chest.
Typically, lacerations of the diaphragm from penetrating injury are
<2 cm long (particularly for knife wounds) and therefore may be
Murray et al
reported a 24% (26 of 110) incidence of occult diaphragm injuries
diagnosed prospectively by laparoscopy in 110 stable patients with
penetrating injury to the thoracoabdominal region but no clinical
indication for celiotomy. In this study, chest radiographs were
normal in 62% (16 of 26) of the patients with diaphragm injury. All
patients with a penetrating injury tract adjacent to the diaphragm
should be considered to have a potential diaphragm injury.
Suspicious CT findings of diaphragm injury in penetrating trauma
include a wound track outlined by air, hemorrhage, or bullet
fragments due to the missile or knife extending adjacent to the
diaphragm; thickening of the diaphragm as a result of hematoma or
edema; free air on the peritoneal aspect of the diaphragm in the
absence of a hollow viscus injury; and a defect in the continuity
of the normal diaphragm or crus with no clear evidence of hematoma
or blood. Diagnostic CT signs of diaphragm injury include the CT
collar sign (constriction of a herniating viscus at the
diaphragmatic rent) (Figure 8), herniation of abdominal content
into the thoracic cavity through a diaphragmatic rent (Figure 8),
or the presence of contiguous organ injury on either side of the
diaphragm in patients with a single gunshot or stab wound (Figure
A review of CT findings in 19 patients with potential diaphragm
injury at the authors' institution found that the most common
diagnostic CT sign of diaphragm injury was contiguous organ injury
on either side of the diaphragm from a single wound seen in 42% (8
of 19) of patients.
Herniation of abdominal fat through a diaphragm defect was seen in
1 patient. CT findings were diagnostic in 47% (9 of 19) of
patients. Other CT findings suggestive of a diaphragm injury seen
in the study include focal thickening of the diaphragm in 37% (7 of
19), and focal discontinuity of the diaphragm in 5% (1 of 19) of
patients. Further prospective studies are needed to determine the
sensitivity, specificity, and accuracy of CT signs to diagnose
diaphragm injury in penetrating trauma.
Initial studies indicate that triple-contrast helical CT is
highly accurate in excluding peritoneal violation in
hemodynamically stable patients with penetrating trauma of the
torso who have no other indication for laparotomy. Patients with
isolated liver injury can be selected using helical CT for
nonoperative management. In particular, further studies are
necessary to determine the accuracy of triple-contrast helical CT
in diagnosing the more challenging penetrating injuries to the
bowel and the diaphragm.