EDUCATION EXHIBIT 1395

RadioGraphics
Bronchial and Nonbron-
chial Systemic Artery
Embolization for Life-
threatening Hemop-
tysis: A Comprehensive
Review1
Woong Yoon, MD ● Jae Kyu Kim, MD ● Yun Hyun Kim, MD
ONLINE-ONLY
CME Tae Woong Chung, MD ● Heoung Keun Kang, MD
See www.rsna
.org/education
/rg_cme.html. Massive hemoptysis is one of the most dreaded of all respiratory emer-
gencies and can have a variety of underlying causes. In 90% of cases,
the source of massive hemoptysis is the bronchial circulation. Diagnos-
LEARNING
OBJECTIVES tic studies for massive hemoptysis include radiography, bronchoscopy,
After reading this
and computed tomography (CT) of the chest. Bronchoscopy and chest
article and taking radiography have been considered the primary methods for the diagno-
the test, the reader
will be able to: sis and localization of hemoptysis. Many researchers currently suggest
䡲 List the most com- that CT should be performed prior to bronchoscopy in all cases of
mon causes of mas-
sive hemoptysis. massive hemoptysis. Bronchial artery embolization (BAE) is a safe and
䡲 Discuss the roles of effective nonsurgical treatment for patients with massive hemoptysis.
CT and bronchos- However, nonbronchial systemic arteries can be a significant source of
copy in the diagnosis
of massive hemopty- massive hemoptysis and a cause of recurrence after successful BAE.
sis. Knowledge of the bronchial artery anatomy, together with an under-
䡲 Describe the tech- standing of the pathophysiologic features of massive hemoptysis, are
niques, embolic ma-
terials, results, and essential for planning and performing BAE in affected patients. In ad-
complications associ- dition, interventional radiologists should be familiar with the tech-
ated with BAE for
massive hemoptysis. niques, results, and possible complications of BAE and with the char-
acteristics of the various embolic agents used in the procedure.
©
RSNA, 2002

Abbreviations: BAE ⫽ bronchial artery embolization, FOB ⫽ fiberoptic bronchoscopy, ICBT ⫽ intercostobronchial trunk

Index terms: Arteries, bronchial, 943.1264, 943.92 ● Arteries, therapeutic embolization, 943.1264 ● Bronchi, anatomy, 671.92 ● Bronchi, interven-
tional procedures ● Lung, CT, 60.1211, ● Lung, hemorrhage ● Pulmonary angiography, 60.124

RadioGraphics 2002; 22:1395–1409 ● Published online 10.1148/rg.226015180

1From the Department of Diagnostic Radiology, Chonnam National University Hospital, Chonnam National University Medical School, 8 Hak-1-
dong, Dong-gu, Gwangju 501-757, South Korea. Recipient of a Certificate of Merit award for an education exhibit at the 2001 RSNA scientific as-
sembly. Received December 19, 2001; revision requested April 23, 2002; final revision received July 26; accepted August 1. Address correspon-
dence to W.Y. (e-mail: radyoon@cnuh.com).
©
RSNA, 2002
 1396 November-December 2002
RadioGraphics Introduction
Life-threatening hemoptysis is one of the most
challenging conditions encountered in critical
care and requires a thorough and timely investiga-
tion. Despite advances in medical and intensive
care unit management, massive hemoptysis re-
mains a serious threat. According to recently pub-
lished data, 28% of chest clinicians had experi-
enced a patient’s death from massive hemoptysis
during a previous 1-year period (1). Conservative
management of massive hemoptysis carries a
mortality rate of 50%–100% (2). The cause of
death is usually asphyxiation, not exsanguination
(3). The reported mortality rates for surgery per-
formed for massive hemoptysis range from 7.1%
to 18.2% (4). However, the mortality rate in-
creases significantly, up to about 40%, when the
surgery is undertaken as an emergency proced-
ure (4).
Bronchial artery embolization (BAE) has be-
come an established procedure in the manage-
ment of massive and recurrent hemoptysis; its use
was first reported in 1973 by Remy et al (5). The
efficacy, safety, and utility of BAE in controlling
massive hemoptysis have been well documented
in many subsequent reports (6 –14). Because of
poor pulmonary reserve and other medical co-
morbid conditions, most patients with massive
hemoptysis are not surgical candidates (2,3).
However, surgery remains the procedure of
choice in the treatment of massive hemoptysis
caused by specific conditions, such as hydatid
cyst, thoracic vascular injury, bronchial adenoma,
and aspergilloma that is resistant to other thera-
pies (15). Even in surgical candidates, BAE is
effective in preparing the patient for elective
rather than high-risk emergency surgery (4).
In this article, we describe the definition and
causes, pathophysiologic features, and diagnosis
of massive hemoptysis and review the angio-
graphic and computed tomographic (CT) anat-
omy of the bronchial circulation. We also discuss
the technique, embolic materials, results, and
complications associated with BAE. In addition,
we describe the role of nonbronchial systemic
artery embolization and present CT findings that
are useful in predicting the existence of nonbron-
chial systemic arterial supply in patients with
massive hemoptysis.
Massive Hemoptysis
Definition and Causes
Massive hemoptysis has been described as the
expectoration of an amount of blood ranging
from 100 mL to more than 1,000 mL over a pe-
RG f Volume 22 ● Number 6
riod of 24 hours, and the most widely used crite-
rion is the production of 300 – 600 mL per day
(2– 4,15). However, depending on the ability of
the patient to maintain a patent airway, a life-
threatening condition may be caused by a rather
small amount of hemorrhage. Thus, a more func-
tional definition of “massive” as an amount suffi-
cient to cause a life-threatening condition should
be used in deciding whether to undertake inter-
ventional management (15,16).
Massive hemoptysis may result from various
causes, and the frequency with which these causes
occur differs greatly between the Western and the
non-Western world. In the non-Western world,
pulmonary tuberculosis, including tuberculosis
bronchiectasis, is the most common underlying
cause of massive hemoptysis. Bronchogenic carci-
noma and chronic inflammatory lung diseases
due to bronchiectasis, cystic fibrosis, or aspergil-
losis are the more prevalent causes of hemoptysis
in Western countries (2,3,15). Other causes in-
clude lung abscess, pneumonia, chronic bronchi-
tis, pulmonary interstitial fibrosis, pneumoconio-
sis, pulmonary artery aneurysm (Rasmussen
aneurysm), congenital cardiac or pulmonary vas-
cular anomalies, aortobronchial fistula, ruptured
aortic aneurysm, and ruptured bronchial artery
aneurysm.
Bronchial artery aneurysm or pseudoaneurysm
is a rare entity, and the term aneurysm is used in-
terchangeably with pseudoaneurysm in the litera-
ture (17–22). The cause of bronchial artery aneu-
rysm is unknown, although bronchiectasis or re-
current bronchopulmonary inflammation, a
mycotic origin, trauma, and Rendu-Osler-Weber
syndrome are often associated with an aneurysm
(21). Bronchial artery aneurysms can be charac-
terized as either mediastinal (juxtaaortic) or in-
trapulmonary. The treatment options for bron-
chial artery aneurysm include transcatheter em-
bolization, placement of a covered stent, and
surgery (21,22).
Pathophysiologic Features
The source of massive hemoptysis is usually the
bronchial circulation (90% of cases) rather than
the pulmonary circulation (5%) (23). In a minor-
ity of cases (5%), massive hemoptysis may origi-
nate with the aorta (eg, aortobronchial fistula,
ruptured aortic aneurysm) or the systemic arterial
supply to the lungs (24 –26). In many acute and
chronic lung diseases, pulmonary circulation is
reduced or occluded at the level of the pulmonary
arterioles because of hypoxic vasoconstriction,
intravascular thrombosis, and vasculitis (27). As a
result, bronchial arteries proliferate and enlarge to
replace the pulmonary circulation. The enlarged
bronchial vessels, which exist in an area of active
or chronic inflammation, may be ruptured due to
 RG f Volume 22 ● Number 6
RadioGraphics erosion by a bacterial agent or due to elevated
regional blood pressure. The arterial blood under
systemic arterial pressure subsequently extrava-
sates into the respiratory tree, resulting in massive
hemoptysis (27,28).
Diagnosis
Diagnostic studies for massive hemoptysis include
radiography, bronchoscopy, and CT of the chest
(15,16,29,30). In patients with massive hemopty-
sis, the diagnostic work-up is usually performed
both to find the cause of the bleeding and to lo-
calize the pulmonary lobes in which bleeding is
suspected. Localization of the bleeding site prior
to BAE is very important if the procedure is to be
performed in a focused and efficient manner.
Conventional radiography is a basic study and
is readily available even under emergency condi-
tions. It may be helpful in diagnosing and localiz-
ing pneumonia, acute or chronic pulmonary tu-
berculosis, bronchogenic cancer, or lung abscess
(29). However, radiographic findings are normal
or nonlocalizing in 17%– 81% of patients with
hemoptysis (16,29 –33). In a retrospective evalua-
tion of 208 patients with hemoptysis, Hirshberg
et al (31) found that radiography was considered
to be diagnostic in only 50% of cases.
Bronchoscopy has long been considered by
chest clinicians to be the primary method for di-
agnosis and localization of hemoptysis (34,35).
Fiberoptic bronchoscopy (FOB) has proved effi-
cacious in evaluating central bronchial lesions. A
vasoactive drug can be infused locally during
bronchoscopy to control bleeding (15).
It has been reported that FOB can help localize
bleeding to one side or the other in 49%–92.9%
of patients with hemoptysis (16,31,33). The role
of FOB in the evaluation of patients who present
with hemoptysis is still controversial, especially in
cases in which chest radiographic findings are
normal or nonlocalizing (29). The diagnostic ac-
curacy of FOB in patients with hemoptysis and
normal chest radiographs is low, ranging from 0%
to 31% (29,31,32,36,37). The overall diagnostic
accuracy of FOB in evaluating patients with he-
moptysis is reported to be 10%– 43% (29 –33). In
a recent article, Hsiao et al (16) documented that
FOB prior to BAE is unnecessary in patients with
hemoptysis of known cause if the site of bleeding
can be determined on conventional radiographs.
FOB has some disadvantages in the diagnosis and
localization of massive, active hemoptysis. It is
difficult to localize the bleeding site with FOB in
patients with massive hemoptysis because of ex-
cessive blood in the bronchi. Moreover, endo-
bronchial therapies are not effective in most cases
Yoon et al 1397
of massive hemoptysis (15). The risks of FOB
include possible airway compromise from seda-
tion, delay in definitive treatment, hypoxemia,
and high cost.
The role of CT in the evaluation of patients
with hemoptysis has been validated (30,32,38).
CT has proved to be of considerable value in di-
agnosing bronchiectasis, bronchogenic carci-
noma, and aspergilloma in patients with massive
hemoptysis (31,39). CT may demonstrate lesions
that may not be visible on conventional radio-
graphs, and contrast material– enhanced CT may
help detect vascular lesions that cause massive
hemoptysis. CT findings can suggest a specific
diagnosis in 50% of patients in whom FOB find-
ings are nondiagnostic and in 39%– 88% of pa-
tients in whom chest radiographs are nondiagnos-
tic (29,30,32,39). CT can also help localize the
site of bleeding in 63%–100% of patients with
hemoptysis, a rate that is higher than that for
FOB (16,33). It has been stated that CT and
FOB are not competitive but complementary
tools for assessing patients with hemoptysis, and
indeed, the combined use of FOB and CT does
yield the best results in evaluating hemoptysis
(33). However, many researchers are currently
suggesting that CT should be performed prior to
bronchoscopy in all patients with hemoptysis
(15,30,32,38). State-of-the-art CT allows rapid
scanning, making timely examination feasible in
critically ill patients. In addition, bronchial and non-
bronchial systemic feeder vessels can be detected at
contrast-enhanced CT. In a prospective study of
40 patients with massive hemoptysis, 27 patients
(67.5%) had a nonbronchial systemic arterial sup-
ply, and CT had an overall accuracy of 84% in
identifying this finding (unpublished data).
Anatomy of the Bronchial Artery
Interventional radiologists who perform BAE
should have a thorough knowledge of bronchial
artery anatomy. The bronchial arteries have vari-
able anatomy in terms of origin, branching pat-
tern, and course (40). The bronchial arteries
originate directly from the descending thoracic
aorta, most commonly between the levels of the
T5 and T6 vertebrae (3). Cauldwell et al (41)
reported four classic bronchial artery branching
patterns: two on the left and one on the right that
presents as an intercostobronchial trunk (ICBT)
(40% of cases); one on the left and one ICBT on
the right (21%); two on the left and two on the
right (one ICBT and one bronchial artery)
(20%); and one on the left and two on the right
 1398 November-December 2002 RG f Volume 22 ● Number 6

RadioGraphics

Figure 1. Diagrams illustrate the types of bronchial arterial supply: Type I, two bronchial arteries on the left and
one on the right that manifests as an ICBT (40.6% of cases); Type II, one on the left and one ICBT on the right
(21.3%); Type III, two on the left and two on the right (one ICBT and one bronchial artery) (20.6%); and Type IV,
one on the left and two on the right (one ICBT and one bronchial artery) (9.7%).

Figure 2. Right intercostobronchial artery. On a se-
lective right ICBT angiogram, an intercostal branch
(solid arrows) and the right bronchial artery (open ar-
rows) are seen to arise from a common trunk.
Figure 3. Common bronchial trunk. On a selective
bronchial angiogram, the right intercostobronchial ar-
tery (black arrow) and a pathologic left bronchial artery
(white arrow) are seen to arise from a common trunk
and supply hypervascular lesions in the left lower lobe.

(one ICBT and one bronchial artery) (9.7%) (Fig

1). The right ICBT is the most consistently seen
vessel at angiography (80% of individuals) (Fig
2). The right ICBT usually arises from the right
posterolateral aspect of the thoracic aorta and the
normal right and left bronchial arteries from the
anterolateral aspect of the aorta. Right and left
bronchial arteries that arise from the aorta as a
common trunk are not uncommon at angiogra-
phy (Fig 3). The true prevalence of a common
bronchial artery trunk is unknown.
The bronchial arteries supply the trachea, ex-
tra- and intrapulmonary airways, bronchovascular
bundles, nerves, supporting structures, regional
lymph nodes, visceral pleura, and esophagus as
well as the vasa vasorum of the aorta, pulmonary
artery, and pulmonary vein (27).
Bronchial arteries that originate outside the
area between the T5 and T6 vertebrae at the level
of the major bronchi are considered to be anoma-
lous (41,42). The reported prevalence of bron-
chial arteries with an anomalous origin ranges
from 8.3% to 35% (42,43). These aberrant bron-
chial arteries may originate from the aortic arch,
internal mammary artery (Fig 4), thyrocervical
trunk (Fig 5), subclavian artery, costocervical
trunk, brachiocephalic artery, pericardiaco-
phrenic artery, inferior phrenic artery, or abdomi-
nal aorta. Aberrant bronchial arteries can be dis-
tinguished anatomically and angiographically
from nonbronchial systemic collateral vessels in
that they extend along the course of the major
bronchi. In contrast, nonbronchial systemic col-
lateral vessels enter the pulmonary parenchyma
 RG f Volume 22 ● Number 6 Yoon et al 1399

RadioGraphics

Figure 5. Aberrant bronchial artery. (a) Thoracic aortogram shows a hypertrophic right bronchial artery (arrow)
and intercostal arteries that supply hypervascular lesions in both upper lobes. There is no evidence of hypervascular-
ity in the right middle lobe or the lower lobes despite the visualization of significant parenchymal lesions at chest radi-
ography. (b) Image obtained after selective injection of an aberrant bronchial artery that arises from the left thyrocer-
vical trunk shows hypertrophic arteries that supply hypervascular lesions in the right lower and left upper lobes. Un-
like systemic collateral vessels, the aberrant bronchial arteries extend along the course of the primary bronchi.

through the adherent pleura or via the pulmonary

Figure 4. Aberrant bronchial artery in a 46-year-old
man who presented with massive hemoptysis. Thoracic
aortography and a catheter search demonstrated only hy-
pertrophic intercostal arteries with no identification of the
right bronchial artery. Despite the embolization of mul-
tiple pathologic intercostal arteries, massive hemoptysis
recurred after 1 day, and the patient underwent repeat
angiography. Selective right internal mammary angio-
gram shows an aberrant right bronchial artery (arrows)
that originates from the proximal portion of the right in-
ternal mammary artery. This aberrant bronchial artery
was embolized, and the massive hemoptysis did not recur
during a follow-up period of 1 year.
ligament, and their course is not parallel to that
of the bronchi (42). The majority of aberrant
bronchial arteries originate from the aortic arch
(41,42). The prevalence of bronchial arteries with
origins outside the aorta is unknown. Interven-
tional radiologists should be aware of the possible
presence of aberrant bronchial arteries, especially
when a significant bronchial arterial supply to ar-
eas of abnormal pulmonary parenchyma is not
demonstrated during a catheter search or at de-
scending thoracic aortography (44). In addition,
bronchial arteries of anomalous origin should be
suspected and investigated angiographically in
patients who present with recurrent hemoptysis
despite successful embolization and in those in
whom the source of bleeding has not been de-
tected (42).
Two kinds of spinal arteries may be seen at
bronchial and intercostal angiography during
BAE. Dorsal and ventral radicular arteries (Fig 6)
are small vessels that arise from segmental spinal
arteries and supply the dorsal and ventral roots.
An average of eight anterior medullary arteries
reinforce the anterior spinal artery, which is the
major independent source of spinal cord perfu-
sion. The artery of Adamkiewicz, or greater ante-
rior medullary artery, reinforces the circulation of
 1400 November-December 2002 RG f Volume 22 ● Number 6

RadioGraphics

Figure 6. Radicular arteries. Left intercostal angio- Figure 7. Anterior medullary artery. Selective left sev-
gram shows two radicular arteries (arrows) that arise enth intercostal angiogram shows an anterior medullary
from the proximal portion of intercostal arteries. artery with the typical hairpin configuration (arrows).

Figures 8, 9. Bronchial artery. (8) Contrast-enhanced CT scan shows a pathologic right bronchial artery (arrow)
that originates from the anteromedial aspect of the thoracic aorta and a hypertrophic left bronchial artery in the aor-
topulmonary window (arrowheads). (9) Contrast-enhanced CT scan shows a pathologic left bronchial artery (arrow)
that originates from the anterior wall of the descending thoracic aorta.

lumbar enlargement of the spinal cord. This uni-
lateral vessel has been observed to arise between
T9 and T12 in 75% of cases (45). Anterior med-
ullary arteries have a characteristic “hairpin” con-
figuration at angiography (Fig 7). Radicular arter-
ies are often visualized during BAE. In our experi-
ence, unintentional embolization of radicular
arteries does not cause clinical problems like spi-
nal cord ischemia. It has been stated that the
presence of radicular arteries is not considered to
be a contraindication for BAE (8,9,43). Anterior
medullary arteries are rarely observed, but em-
bolization and repeat angiography should be
avoided because spinal cord ischemia may occur
with embolization. It has been suggested that spi-
nal arteries will arise from the intercostal branch
of the right ICBT in 5%–10% of cases, but it is
generally believed that the true prevalence is con-
siderably lower (3).
In adults, normal bronchial arteries measure
less than 1.5 mm in diameter at their origin and
0.5 mm at their point of entry into a bronchopul-
monary segment (27). A bronchial artery larger
than 2 mm at CT is most likely abnormal (46).
Hypertrophic bronchial arteries are easily visual-
ized as enhancing nodular or tubular structures
within the mediastinum and around the central
airway on contrast-enhanced CT scans (47) (Figs
8, 9). The primary locations of enlarged bronchial
arteries at CT are the retroesophageal area, retro-
 RG f Volume 22 ● Number 6 Yoon et al 1401

RadioGraphics

Figure 10. Value of preliminary thoracic aortography. (a) Descending thoracic aortogram demonstrates two hyper-
trophic bronchial arteries (solid arrows) and one intervening intercostal artery (open arrow) that supply a hypervas-
cular lesion in the right upper lobe. (b) On a selective upper bronchial angiogram, the bronchial artery that supplies
the large hypervascular lesion is seen to have an anomalous origin from the inferior aspect of the aortic arch. Marked
bronchopulmonary shunting is also noted. (c) Selective lower bronchial angiogram shows a hypertrophic artery that
supplies the hypervascular lesion, along with marked bronchopulmonary shunting. (d) Selective intercostal angio-
gram shows a hypertrophic artery that supplies the hypervascular lesion, along with bronchopulmonary shunting.
The two bronchial arteries and the intervening intercostal artery were selectively embolized with polyvinyl alcohol
particles. Postembolization angiograms (not shown) demonstrated occlusion of each vessel, with no opacification of
the hypervascular lesion and no pulmonary arterial shunting.

tracheal area, retrobronchial area, posterior wall
of the main bronchus, and aortopulmonary win-
dow. Mediastinal lymph nodes, the azygos vein,
and an enhancing esophageal wall can mimic the
bronchial arteries at CT (48).
Bronchial Artery Embolization
Technique
Prior to BAE, the number and origin sites of
bronchial arteries from the aorta should be care-
fully evaluated to determine the optimal angio-
graphic approach. This can be accomplished with
a preliminary descending thoracic aortogram. Ab-
normal bronchial arteries are visualized on an ini-
tial thoracic aortogram in the majority of affected
patients (Fig 10). A descending thoracic aorto-
gram is also useful in the detection of nonbron-
chial systemic arteries that supply parenchymal
 1402 November-December 2002 RG f Volume 22 ● Number 6

RadioGraphics

Figure 11. Value of preliminary thoracic aortography. (a) Descending thoracic aortogram shows a tortuous
branch of the right inferior phrenic artery (arrows) that supplies a hypervascular lesion in the right lower lobe.
(b) Selective right inferior phrenic angiogram shows a nonbronchial systemic artery that supplies the hypervas-
cular lesion (arrows) in the right lower lobe. Filling of the pulmonary artery is also seen.

Figure 12. Superselective embolization with a microcatheter. (a) Selective intercostal angiogram shows hy-
pervascular areas and a small amount of pulmonary arterial shunting (open arrow) at the periphery of the left
lower lung field. Note the small radicular artery (solid arrow) that arises from the proximal portion of the inter-
costal artery. (b) Postembolization angiogram shows a microcatheter (arrows) that has been advanced into the
intercostal artery beyond the origin of the radicular artery.

lesions (Fig 11) (49). Although cobra-type curved types of catheters (eg, Simmons-1, headhunter,
catheters are most commonly used for catheter- Yashiro-type) should be prepared for optimal se-
ization of the bronchial artery, several different lection of bronchial arteries. The usefulness of a
microcatheter for selective BAE has been empha-
 RG f Volume 22 ● Number 6 Yoon et al 1403

RadioGraphics

Figures 13, 14. (13) Systemic artery–pulmonary artery shunt. Right bronchial angiogram demonstrates an en-
larged bronchial artery that supplies a hypervascular lesion in the right upper lobe. There is massive bronchopulmo-
nary shunting with visualization of the branches of the right upper lobar pulmonary arteries (arrows). (14) Systemic
artery–pulmonary vein shunt. Right ICBT angiogram shows a hypervascular staining lesion in the right lower lung as
well as shunting with branches of the right inferior pulmonary vein (arrows).

artery and safe positioning in the bronchial circu-

Figure 15. Contrast material extravasation. Right
bronchial angiogram shows multifocal hypervascular
lesions in the right lung and extravasation of contrast
material into the right lower lobe bronchus (arrow).
sized in many recent articles (2,3,50,51). This
superselective catheterization permits stabiliza-
tion of the catheter position within the bronchial
lation beyond the origin of spinal cord branches,
which prevents severe complications (Fig 12).
After catheterization of the bronchial artery,
bronchial angiography is performed with manual
injection of contrast medium.
Angiographic findings in massive hemoptysis
include hypertrophic and tortuous bronchial ar-
teries, neovascularity, hypervascularity (Figs 3, 5,
10), shunting into the pulmonary artery or vein
(Figs 13, 14), extravasation of contrast medium
(Fig 15), and bronchial artery aneurysm (Fig 16).
Although extravasation of contrast medium is
considered a specific sign of bronchial bleeding,
this finding is rarely seen, and its reported preva-
lence ranges from 3.6% to 10.7% (12,16). Thus,
the determination of which arteries are to be em-
bolized should be based on a combination of CT,
bronchoscopic, and angiographic findings with
clinical correlation. All angiograms, including
intercostal arteriograms, must be carefully scruti-
nized for opacification of spinal arteries to avoid
inadvertent embolization.
 1404 November-December 2002
RadioGraphics
Figure 16. Bronchial artery aneurysm. (a) Selective left bronchial angiogram demonstrates a large, hypervascular
lesion in the left upper lobe with an aneurysm (arrow). (b) On an angiogram obtained after embolization with polyvi-
nyl alcohol particles, neither the hypervascular lesion nor the aneurysm is visualized.
Embolic Materials
A variety of embolic materials are used for BAE.
Absorbable gelatin sponge is widely used because
it is inexpensive, easy to handle, and has a con-
trollable embolic size. However, disadvantages of
absorbable gelatin sponge are its resolvability and
lack of radiopacity. Its use may lead to recanaliza-
tion of the embolized artery and may sometimes
be responsible for recurrent bleeding (50). Polyvi-
nyl alcohol particles are nonabsorbable embolic
materials, and particles 350 –500 ␮m in diameter
are the most frequently used worldwide (51).
Their use may prevent the early recurrence of he-
moptysis due to recanalization of the embolized
artery, as might be anticipated with absorbable
gelatin sponge.
It is essential to avoid the use of embolic mate-
rials that can pass through the bronchopulmonary
anastomosis. Experimental study has demon-
strated a bronchopulmonary anastomosis of 325
␮m in the human lung (52). Pulmonary infarc-
tion via bronchial artery–pulmonary artery shunts
or systemic artery embolization via bronchial ar-
tery–pulmonary vein shunts may occur when em-
bolic agents less than 325 ␮m in diameter are
used. In addition, it is important to avoid using
embolic agents that produce distal occlusion to
such an extent that normal peripheral branches
that supply the bronchi, esophagus, or vasa vaso-
rum of the pulmonary artery or aorta become oc-
RG f Volume 22 ● Number 6
cluded, possibly leading to disastrous complica-
tions (eg, bronchial, esophageal, pulmonary arte-
rial, or aortic wall necrosis) (3). To avoid the
complications indicated earlier, we recommend
the use of polyvinyl alcohol particles with a diam-
eter of 350 –500 ␮m for BAE.
Bronchopulmonary fistula is a common finding
at bronchial angiography in patients with massive
hemoptysis. Regardless of whether a bronchopul-
monary fistula is demonstrated at bronchial an-
giography, we perform BAE with the same tech-
nique and the same size of embolic materials
(polyvinyl alcohol particles over 350 ␮m in diam-
eter or absorbable gelatin sponge), and we have
not encountered any clinical problems related to
pulmonary infarction or systemic embolization.
Thus, we recommend the use of the same strategy
even when a bronchopulmonary fistula is visual-
ized at bronchial angiography.
Liquid embolic agents (eg, isobutyl-2 cyanoac-
rylate, absolute ethanol) are not currently used
because of the high risk of severe complications
such as tissue necrosis. Stainless steel platinum
coils are generally not used for BAE because they
tend to occlude more proximal vessels and may
preclude repeat embolization if hemoptysis re-
curs. However, they may be used to occlude a
pulmonary artery aneurysm and may occasionally
be used in the internal mammary artery to pre-
vent embolization of a normal vascular territory
and development of collateral vessels (Fig 17)
(3,53).
 RG f Volume 22 ● Number 6
RadioGraphics
Figure 17. Coil embolization. (a) Selective right internal mammary angiogram shows small branches that supply a
hypervascular staining lesion in the right upper lobe, as well as a pseudoaneurysm (arrow). (b) On a postembolization
angiogram obtained after deposition of two coils (arrows) at both the proximal and distal portions of the origin site of
the small branches and use of additional polyvinyl alcohol particles, neither the hypervascular lesion nor the pseudo-
aneurysm is visualized.
Results
Previous studies have shown that BAE is very ef-
fective in controlling acute massive hemoptysis.
The initial nonrecurrence rates for BAE have
been reported to be 73%–98%, with a mean fol-
low-up period ranging from 1 day to 1 month (7–
14,54). Immediate success rates have increased
recently because of the introduction of superse-
lective embolization and the refinement of em-
bolic agents and techniques. However, the long-
term success rate of BAE in hemoptysis is unfa-
vorable. Long-term recurrence rates have been
reported to be 10%–52%, with a mean follow-up
period ranging from 1 to 46 months (2,3,55).
However, the long-term success rate can be im-
proved with repeat BAE. Hemoptysis may recur
after successful BAE if the disease process is not
controlled with drug therapy or surgery because
embolization does not address the underlying dis-
ease but rather treats the symptom. In this sense,
BAE is a palliative procedure that prepares the
patient for elective surgery for localized disease or
continued antimicrobial therapy (51).
Recurrent bleeding may be caused by recan-
alization of embolized vessels, incomplete em-
bolization, revascularization by the collateral cir-
culation, inadequate treatment of the underlying
disease, progression of basic lung disease, or non-
bronchial systemic arterial supply (2,3,51,55).
Recurrence rate may also be influenced by the
cause of the hemoptysis. Recurrent bleeding is
more common in patients with chronic tuberculo-
sis, aspergilloma, or neoplasm. In one study of
Yoon et al 1405
103 patients who underwent BAE, 16 patients
(15.5%) required repeat embolization; all 16 had
hemoptysis due to chronic tuberculosis (56). It is
believed that this was due mainly to hypertrophy
of the collateral nonbronchial systemic arteries. In
a series by Katoh et al (55), 75% of patients with
aspergilloma experienced recurrence of hemopty-
sis after undergoing initial embolization. Haya-
kawa et al (11) reported that BAE failed within 1
month in 42% of patients with neoplasm. A neo-
plasm receives its blood supply from multiple
feeder vessels other than the bronchial artery and
invades the vascular structure aggressively.
Complications
Several complications of BAE have been reported
in the literature. Chest pain is the most common
complication, with a reported prevalence of 24%–
91% (12,43,57). Chest pain is likely related to an
ischemic phenomenon caused by embolization
and is usually transient. In addition, dysphagia
due to embolization of esophageal branches may
be encountered, with a reported prevalence of
0.7%–18.2% (12,57). Dysphagia also regresses
spontaneously. Subintimal dissection of the aorta
or the bronchial artery during BAE is the other
minor complication, with a reported prevalence of
1%– 6.3% (8,9,11,13,14). There are usually no
symptoms or problems related to the subintimal
dissection.
 1406 November-December 2002 RG f Volume 22 ● Number 6

RadioGraphics

Figure 18. Nonbronchial systemic artery. Right sub-
clavian angiogram shows an enlarged superior thoracic
artery, lateral thoracic artery, and subscapular artery
and numerous small branches from the subclavian and
axillary arteries that supply a hypervascular lesion in
the right upper lobe, along with systemic artery–pulmo-
nary artery shunting (arrows).
Figure 19. Nonbronchial systemic artery. Selective
left thyrocervical angiogram shows tortuous branches
that supply a hypervascular lesion in the left upper
lung, with a pseudoaneurysm (solid arrow) and sys-
temic artery–pulmonary artery shunting (open arrow).

The most disastrous complication of BAE is

spinal cord ischemia due to the inadvertent occlu-
sion of spinal arteries. The prevalence of spinal
cord ischemia after BAE is reported to be 1.4%–
6.5% (12,13,50,58). As discussed earlier, the vi-
sualization of radicular branches on bronchial or
intercostal angiograms is not an absolute contra-
indication for BAE. However, when the anterior
medullary artery (artery of Adamkiewicz) is visu-
alized at angiography, embolization should not be
performed.
Other rare complications that have been re-
ported in the literature include aortic and bron-
chial necrosis, bronchoesophageal fistula, non–
target organ embolization (eg, ischemic colitis),
pulmonary infarction, referred pain to the ipsilat-
eral forehead and orbit, and transient cortical Figure 20. Nonbronchial systemic artery. Selective
blindness (59 – 64). It is hypothesized that cortical right lateral thoracic angiogram shows a hypervascular
blindness develops because of embolism to the lesion with systemic artery–pulmonary artery shunting
occipital cortex, either via a bronchial artery–pul- (arrows) in the periphery of the right upper lobe.
monary vein shunt or via collateral vessels be-
tween the bronchial and vertebral arteries (64). tion of the bronchial artery. Many investigators
have documented that a concerted search for
Nonbronchial nonbronchial systemic arterial supply should be
Systemic Arterial Supply made (3,10,53,56,58). In the presence of pleural
Nonbronchial systemic arteries can be a signifi- thickening, nonbronchial systemic feeder vessels
cant source of massive hemoptysis, especially in that originate from various arteries (eg, intercostal
patients with pleural involvement caused by an artery [Figs 10, 12], branches of the subclavian
underlying disease. Missing the nonbronchial sys- and axillary arteries [Figs 18 –20], internal mam-
temic arteries at initial angiography may result in mary artery [Fig 21], inferior phrenic artery [Fig
early recurrent bleeding after successful emboliza- 11]) (3,10,47,53,55) may develop along the pleu-
ral surface and become enlarged as a result of the
inflammatory process. Pleural thickening that is
 RG f Volume 22 ● Number 6 Yoon et al 1407

RadioGraphics

Figure 21. Prediction of nonbronchial systemic arterial supply. (a) Contrast-enhanced CT scan shows irregular
pleural thickening at the left apical thorax with tortuous vascular structures within a hypertrophic extrapleural layer of
fat (arrows). (b) On a left subclavian angiogram, a tortuous, enlarged internal mammary artery (solid arrows) and its
branches are seen to supply a hypervascular lesion in left upper lobe. A hypertrophic lateral thoracic artery (open ar-
rows) also supplies the hypervascular lesion.

Figure 22. Prediction of nonbronchial systemic arterial supply. (a) Contrast-enhanced CT scan demonstrates dif-
fuse pleural thickening at the upper thorax (solid arrows) and tortuous, enhancing vascular structures within a hyper-
trophic extrapleural layer of fat (open arrows). Hypertrophic bronchial arteries are also seen in the aortopulmonary
window. (b) Selective intercostal angiogram shows tortuous hypertrophic vessels with systemic artery–pulmonary
artery shunting (arrows).

noted at chest radiography negatively influences
the long-term success rate of BAE (65). CT may
help predict the presence of nonbronchial sys-
temic collateral vessels as a source of bleeding in
patients with massive hemoptysis. In our experi-
ence, pleural thickening of more than 3 mm and
tortuous enhancing vascular structures within
hypertrophic extrapleural fat seen at contrast-
enhanced CT are signs of nonbronchial systemic
arterial supply in patients with massive hemopty-
sis (Figs 21, 22) (unpublished data). Use of CT
 1408 November-December 2002
RadioGraphics to predict the presence of nonbronchial systemic
vessels that supply a parenchymal lesion is impor-
tant prior to BAE because it helps in localizing
the site of bleeding and in selecting systemic ves-
sels for the interventional approach.
Conclusions
Massive hemoptysis constitutes a significant and
often life-threatening respiratory emergency.
Bronchial and nonbronchial systemic artery
embolization is a safe and effective nonsurgical
treatment for patients with massive hemoptysis.
Knowledge of bronchial artery anatomy, together
with an understanding of the pathophysiologic
features of massive hemoptysis, are essential for
performing BAE. The pulmonary artery can be
the source of massive hemoptysis in a minority of
cases, and pulmonary arterial disease should be
considered in cases of recurrent hemoptysis. CT
is useful in diagnosing the disease that causes
massive hemoptysis, localizing the bleeding site,
and selecting vessels to be embolized. In addition,
CT may be helpful in predicting the presence of
nonbronchial systemic collateral vessels, which
can be a significant source of recurrent hemopty-
sis after successful BAE.
References
1. Haponik EF, Fein A, Chin R. Managing life-threaten-
ing hemoptysis: has anything really changed? Chest
2000; 118:1431–1435.
2. Najarian KE, Morris CS. Arterial embolization in the
chest. J Thorac Imaging 1998; 13:93–104.
3. Marshall TJ, Jackson JE. Vascular intervention in the
thorax: bronchial artery embolization for hemoptysis.
Eur Radiol 1997; 7:1221–1227.
4. Fernando HC, Stein M, Benfield JR, Link DP. Role
of bronchial artery embolization in the management
of hemoptysis. Arch Surg 1998; 133:862– 866.
5. Remy J, Voisin C, Ribet M, et al. Treatment, by em-
bolization, of severe or repeated hemoptysis associ-
ated with systemic hypervascularization. Nouv Presse
Med 1973; 2:2060 –2068. [French]
6. Wholey MH, Chamorro HA, Rao G, Ford WB,
Miller WH. Bronchial artery embolization for massive
hemoptysis. JAMA 1976; 236:2501–2504.
7. Remy J, Arnaud A, Fardou H, Giraud R, Voisin C.
Treatment of hemoptysis by embolization of bron-
chial arteries. Radiology 1977; 122:33–37.
8. Uflacker R, Kaemmerer A, Neves C, Picon PD. Man-
agement of massive hemoptysis by bronchial artery
embolization. Radiology 1983; 146:627– 634.
9. Uflacker R, Kaemmerer A, Picon PD, et al. Bronchial
artery embolization in the management of hemopty-
sis: technical aspects and long-term results. Radiology
1985; 157:637– 644.
10. Keller FS, Rosch J, Loflin TG, Nath PH, McElvein
RB. Nonbronchial systemic collateral arteries: signifi-
cance in percutaneous embolotherapy for hemoptysis.
Radiology 1987; 164:687– 692.
11. Hayakawa K, Tanaka F, Torizuka T, et al. Bronchial
artery embolization for hemoptysis: immediate and
long-term results. Cardiovasc Intervent Radiol 1992;
15:154 –159.
RG f Volume 22 ● Number 6
12. Ramakantan R, Bandekar VG, Gandhi MS, Aulakh
BG, Deshmukh HL. Massive hemoptysis due to pul-
monary tuberculosis: control with bronchial artery
embolization. Radiology 1996; 200:691– 694.
13. Mal H, Rullon I, Mellot F, et al. Immediate and long-
term results of bronchial artery embolization for life-
threatening hemoptysis. Chest 1999; 115:996 –1001.
14. Kato A, Kudo S, Matsumoto K, et al. Bronchial ar-
tery embolization for hemoptysis due to benign dis-
eases: immediate and long-term results. Cardiovasc
Intervent Radiol 2000; 23:351–357.
15. Jean-Baptiste E. Clinical assessment and manage-
ment of massive hemoptysis. Crit Care Med 2000;
28:1642–1647.
16. Hsiao EI, Kirsch CM, Kagawa FT, Wehner JH,
Jensen WA, Baxter RB. Utility of fiberoptic bron-
choscopy before bronchial artery embolization for
massive hemoptysis. AJR Am J Roentgenol 2001;
177:861– 867.
17. Servois V, Denys A, Silbert A. Mycotic aneurysm of
the bronchial artery: a rare cause of hemoptysis (let-
ter). AJR Am J Roentgenol 1992; 159:428.
18. Cearlock JR, Fontaine AB, Urbaneja A, Spigos DG.
Endovascular treatment of a posttraumatic bronchial
artery pseudoaneurysm. J Vasc Intervent Radiol
1995; 6:495– 496.
19. Yanagihara K, Ueno Y, Kobayashi T, Isobe J, Itoh
M. Bronchial artery aneurysm. Ann Thorac Surg
1999; 67:854 – 855.
20. Shimokawa S, Ishizake N, Watanabe S. Ruptured
bronchial artery aneurysm. Ann Thorac Surg 2000;
69:1641–1642.
21. Pugnale M, Portier F, Lamarre A, et al. Hemomedi-
astinum caused by rupture of a bronchial artery aneu-
rysm: successful treatment by embolization with
N-butyl-2-cyanoacrylate. J Vasc Intervent Radiol
2001; 12:1351–1352.
22. Tringali S, Tiffet O, Berger JL, Cuilleret J. Bronchial
artery aneurysm disguised as a leiomyoma of the
esophagus. Ann Thorac Surg 2002; 73:632– 633.
23. Remy J, Remy-Jardin M, Voisin C. Endovascular
management of bronchial bleeding. In: Butler J, ed.
The bronchial circulation. New York, NY: Dekker,
1992; 667–723.
24. MacIntosh EL, Parrott JCW, Unrhu HW. Fistulas
between the aorta and tracheobronchial tree. Ann
Thorac Surg 1991; 51:515–519.
25. Hakanson E, Konstantinov IE, Fransson SG. Man-
agement of life-threatening haemoptysis. Br J Anaesth
2001; 88:291–295.
26. Pearse EO, Bryan AJ. Massive hemoptysis 27 years
after surgery for coarctation of the aorta. J R Soc Med
2001; 94:640 – 641.
27. Deffenbach ME, Charan NB, Lakshminarayan S,
Butler J. The bronchial circulation: small, but a vital
attribute to the lung. Am Rev Respir Dis 1987; 135:
463– 481.
28. Liebow AA, Hales MR, Lindskog GE. Enlargement
of the bronchial arteries, and their anastomosis with
the pulmonary arteries in bronchiectasis. Am J Pathol
1949; 25:211–231.
29. Naidich DP, Funt S, Ettenger NA, Arranda C. He-
moptysis: CT— bronchoscopic correlations in 58
cases. Radiology 1990; 177:357–362.
30. McGuinness G, Beacher JR, Harkin TJ, Garay SM,
Rom WN, Naidich DP. Hemoptysis: prospective
high-resolution CT/bronchoscopic correlation. Chest
1994; 105:1155–1162.
31. Hirshberg B, Biran I, Glazer M, Kramer MR. Hemop-
tysis: etiology, evaluation, and outcome in a tertiary re-
ferral hospital. Chest 1997; 112:440– 444.
32. Tak S, Ahluwalia G, Sharma SK, Mukhopadhya S,
Pande JN. Haemoptysis in patients with a normal
 RG f Volume 22 ● Number 6
RadioGraphics chest radiograph: bronchoscopy-CT correlation. Aus-
tralas Radiol 1999; 43:451– 455.
33. Abal AT, Nair PC, Cherian J. Haemoptysis: aetiology,
evaluation, and outcome—a prospective study in a
third-world country. Respir Med 2001; 95:548–552.
34. Cahill BC, Ingabar DH. Massive hemoptysis: assess-
ment and management. Clin Chest Med 1994; 15:
147–167.
35. Dweik RA. Stoller JK. Role of bronchoscopy in mas-
sive hemoptysis. Clin Chest Med 1999; 20:89 –105.
36. Peters J, McClung HC, Teague RB. Evaluation of
hemoptysis in patients with a normal chest roent-
genogram. West J Med 1984; 141:624 – 626.
37. Poe RH, Israel RH, Marin MG, et al. Utility of fiber-
optic bronchoscopy in patients with hemoptysis and a
nonlocalizing chest roentgenogram. Chest 1988; 92:
70 –75.
38. Naidich DP, Harkin TJ. Airways and lung: correla-
tion of CT with fiberoptic bronchoscopy. Radiology
1995; 197:1–12.
39. Millar AB, Boothroyd AE, Edwards D, et al. The role
of computed tomography (CT) in the investigation of
unexplained hemoptysis. Respir Med 1992; 86:39 –
44.
40. Lippert H, Pabst R. Bronchial arteries. In: Lippert H,
Pabst R, eds. Arterial variations in man. Munich,
Germany: Bergmann Verlag, 1985; 18 –19.
41. Cauldwell EW, Siekert RG, Lininger RE, Anson BJ.
The bronchial arteries: an anatomic study of 105 hu-
man cadavers. Surg Gynecol Obstet 1948; 86:395–
412.
42. Sancho C, Escalante E, Dominguez J, et al. Emboliza-
tion of bronchial arteries of anomalous origin. Cardio-
vasc Intervent Radiol 1998; 21:300–304.
43. Cohen AM, Doershuk CF, Stern RC. Bronchial ar-
tery embolization to control hemoptysis in cystic fi-
brosis. Radiology 1990; 175:401– 405.
44. McPherson S, Routh WD, Nath H, Keller FS.
Anomalous origin of bronchial arteries: potential pit-
fall of embolotherapy for hemoptysis. J Vasc Intervent
Radiol 1990; 1:86 – 88.
45. Rosenthal D. Spinal cord ischemia after abdominal
aortic operation: is it preventable? J Vasc Surg 1999;
30:391–399.
46. Furuse M, Saito K, Kunieda E, et al. Bronchial arter-
ies: CT demonstration with arteriographic correla-
tion. Radiology 1987; 162:393–398.
47. Do KH, Goo JM, Im JG, et al. Systemic arterial sup-
ply to the lungs in adults: spiral CT findings. Radio-
Graphics 2001; 21:387– 402.
48. Song JW, Im JG, Shim YS, Park JH, Yeon KM, Han
MC. Hypertrophied bronchial artery at thin-section
CT in patients with bronchiectasis: correlation with
CT angiographic findings. Radiology 1998; 208:187–
191.
49. Phillips S, Ruttley MST. Bronchial artery emboliza-
tion: the importance of preliminary thoracic aortogra-
phy. Clin Radiol 2000; 55:317–319.
This article meets the criteria for 1.0 credit hour in category 1 of the AMA Physician’s Recognition Award. To obtain
credit, see www.rsna.org/education/rg_cme.html.
Yoon et al 1409
50. Tanaka N, Yamakado K, Murashima S, et al. Super-
selective bronchial artery embolization for hemoptysis
with a coaxial microcatheter system. J Vasc Intervent
Radiol 1997; 8:65–70.
51. White RI Jr. Bronchial artery embolotherapy for con-
trol of acute hemoptysis: analysis of outcome. Chest
1999; 115:912–915.
52. Pump K. Distribution of bronchial arteries in human
lung. Chest 1972; 62:447– 451.
53. Hashimoto M, Heianna J, Okane K, Izumi J, Watarai
J. Internal mammary artery embolization for hemop-
tysis. Acta Radiol 1999; 40:187–190.
54. Cremashi P, Nascimbene C, Vitulo P, et al. Thera-
peutic embolization of the bronchial artery: a success-
ful treatment in 209 cases of relapse hemoptysis. An-
giology 1993; 44:295–299.
55. Katoh O, Kishikawa T, Yamada H, Matsumoto S,
Kudo S. Recurrent bleeding after arterial emboliza-
tion in patients with hemoptysis. Chest 1990; 97:
541–546.
56. Goh PYT, Lin M, Teo N, Wong DES. Embolization
for hemoptysis: a six-year review. Cardiovasc Inter-
vent Radiol 2002; 25:17–25.
57. Tonkin ILD, Hanissian AS, Boulden TF, et al. Bron-
chial arteriography and embolotherapy for hemopty-
sis in patients with cystic fibrosis. Cardiovasc Inter-
vent Radiol 1991; 14:241–246.
58. Wong ML, Szkup P, Hopley MJ. Percutaneous em-
bolotherapy for life-threatening hemoptysis. Chest
2002; 121:95–102.
59. Girard P, Baldeyrou P, Lemoine G, Grunewald D.
Left mainstem bronchial artery stenosis complicating
bronchial artery embolization. Chest 1990; 97:1246 –
1248.
60. Munk PL, Morris C, Nelems B. Left main bronchial-
esophageal fistula: a complication of bronchial artery
embolization. Cardiovasc Intervent Radiol 1990; 13:
95–97.
61. Lemoigne F, Rampal P, Petersen R. Fatal ischemic
colitis after bronchial embolization. Nouv Presse Med
1983; 12:2056 –2057. [French]
62. Remy-Jardin M, Wattinne L, Remy J. Transcatheter
occlusion of pulmonary arterial circulation and collat-
eral supply: failures, incidents, and complications.
Radiology 1991; 180:699 –705.
63. Ramankantan R, Ketkar M, Maddali K, Deshmukh
H. Referred pain to the ipsilateral forehead and orbit:
an unusual phenomenon during bronchial artery em-
bolization. Cardiovasc Intervent Radiol 1999; 22:
275–277.
64. Liu SF, Lee TY, Wong SL, Lai YF, Lin AS. Tran-
sient cortical blindness: a complication of bronchial
artery embolization. Respir Med 1998; 92:983–986.
65. Tamura S, Kodama T, Otsuka N, et al. Embolo-
therapy for persistent hemoptysis: the significance of
pleural thickening. Cardiovasc Intervent Radiol 1993;
16:85– 88.