A review of medical records from 1921 to 1996 identified 299 patients
who had a primary aneurysmal bone cyst with no preexisting condition.
Of these, forty-four (15%) had a lesion in the sacrum,
ilium, ischium, or pubis. Two patients with an iliac lesion, one
patient with a sacral lesion, and one with a pubic lesion were excluded
from the study because of inadequate documentation or because they
had been lost to follow-up. Of the forty patients, thirty-five were
treated primarily and five were referred for treatment after a local
recurrence.
Plain radiographs of the pelvis were made for all patients before
diagnosis and were evaluated by a musculoskeletal radiologist. All
available preoperative and postoperative imaging studies were reviewed.
Data about the site and extent of the lesion and the presence of
cortical thinning, erosion, or breakthrough in twenty-nine patients
were obtained from the examination of these imaging studies. No
preoperative radiographic images were available for eleven patients,
and the overall dimensions of the lesions in these patients were
determined from the radiology reports and the surgeons’ descriptions.
The method of surgical removal was recorded, and the surgeons’ operative
findings as well as the pathologists’ gross descriptions
were reviewed. The pathological specimens were obtained by incisional
biopsy of the lesion in all patients. Previously described histological
criteria were used for diagnosis5,9,17,18.
The histological sections were available and reviewed for each patient.
For the most recent follow-up evaluation, nineteen patients were
examined by a physician at the authors’ institution and eleven,
by a local physician; a written questionnaire was used to obtain
information from the other ten patients. The mean duration of follow-up
for the forty patients was thirteen years (range, three to fifty-three
years).
Of the forty patients, twenty-two were male and eighteen were
female. At diagnosis, the patients’ ages ranged from two to
forty-four years (mean, 15.5 years). The majority of patients (twenty-six)
were in the second decade of life (see Appendix).
Location and Size
The lesion involved the sacrum in twelve patients, the pubis
in seven, the ischium and the posterior column of the acetabulum
in seven, the pubis and the anterior column of the acetabulum in
six, the ilium in five, and the ilium and the superior dome of the
acetabulum in three. Sixteen of the forty lesions were adjacent
to the acetabulum, and four were adjacent to the sacroiliac joint.
Nine of the forty lesions were found to be intra-articular at the
time of the operation: seven involved the hip joint and two, the
sacroiliac joint.
The lesions were classified further according to location with the
use of a modification of the system proposed by Enneking and Dunham19: seventeen involved region I (iliosacral);
sixteen, region II (acetabular); and seven, region III (ischiopubic)
(Fig. 1).
The average dimension of the lesions at the widest point was 6 cm
(range, 2.5 to 14 cm).
The lesions were staged according to the system of the Musculoskeletal
Tumor Society20-23. Stage 2 refers
to active lesions and stage 3, to aggressive, benign lesions. With
the use of these criteria, twenty-eight patients had a stage-3 lesion
and twelve had a stage-2 lesion.
Clinical Presentation
Pain was the main symptom in thirty-eight patients (95%). The
location of the pain was the thigh in fifteen patients, the groin
in eight, the low back in seven, the buttocks in four, and the leg
in four. Additional signs and symptoms included a limp in twenty
patients (50%), decreased range of motion of the hip joint
in sixteen (40%), a localized mass in three (7.5%),
and a palpable presacral mass on rectal examination in one (2.5%).
Neurological signs and symptoms were present in eight of the twelve
patients with a sacral lesion: two had urinary retention; three,
leg paresthesias; two, leg weakness; and one, paresthesias and loss
of anal tone.
The average duration of symptoms in the patients with a primary
lesion was seven months (range, one to eighteen months) before diagnosis.
In eighteen (51%) of these patients, the pain was associated
with unrelated previous injuries. In some patients, the clinical
signs and symptoms were initially thought to be the result of nonspecific
low-back pain, disc herniation, coccygodynia, urinary tract infection,
muscle or tendon injury, ischial bursitis, stress fracture, or myositis.
The most frequently misinterpreted symptoms included radiating pain
in the lower extremity that involved the medial or anterior aspect
of the thigh. Other causes of missed or delayed diagnoses included
the making of knee, rather than pelvic, radiographs for the diagnosis
of medial knee pain; inadequate physical examination (not including
rectal palpation); and the use of gonad shields that obscured an
aneurysmal bone cyst of the sacrum.
Histological Findings
After a histological review, nine of the forty lesions were reclassified
because of a previously incorrect diagnosis. Three lesions were
initially misdiagnosed as a giant-cell tumor; three, as a simple
cyst; two, as osteitis fibrosa cystica; and one, as a low-grade
chondrosarcoma. The histological features were typical of a primary
aneurysmal bone cyst in all forty cases. The lesions consisted of
spaces separated by septa, and the septa almost always contained
benign-appearing giant cells. The number of giant cells ranged from
sparse to abundant, and no giant cells were found in only one case. The
septum was composed of loosely arranged spindle cells in addition
to the giant cells. Bone formation with a typical reactive pattern
was found in all but eight cases. The amount of bone formation was
also quite variable, ranging from small foci with osteoid production
to extensive areas of trabecular-appearing osteoid. Foci of calcification,
typically found in aneurysmal bone cysts at other sites, were found
in twenty-two of the forty cases. The calcification occurred either
in the septa or in areas that appeared more solid and appeared either as
foci of calcification in chondroid-like areas or as fine lines of
calcification simulating trabeculae of bone. One lesion had unusual
sarcoid-like granulomas in the wall.
Radiographic Findings
Preoperative images of the primary lesions in twenty-nine patients
were available for review. All lesions appeared lytic on the radiographs,
and trabeculation was present in seven. Expansion of the bone was
a predominant feature and was present in all but two of the cases.
The overlying cortex was normal in two cases; bowed and thinned
but intact in eleven; and interrupted in sixteen, nine of which
had a definite partial rim over the lesion. A rim was detected most
easily on cross-sectional imaging (Fig. 2-B). The radiographic margin between
the lesion and the adjacent bone was sharply defined in fifteen
cases and indistinct in seven other cases. Pathological fracture
was present in five cases.
Computed tomography and T2-weighted magnetic resonance imaging
provided optimal evaluation of the internal contents and the extent
of the lesion. Internal septation was seen in four cases, and multiple
fluid levels were seen in three.
Treatment
Primary Lesions
All thirty-five patients with a primary lesion had surgical treatment;
fourteen had intralesional curettage, and twenty-one had excision-curettage
(resection of the majority of the lesion). Intralesional curettage
was performed for small lesions of £5 cm and included complete
exposure by removal of cortical bone at the site of the lesion and
meticulous removal of all abnormal tissue with use of a curet, a
rongeur, a motorized burr, and suction22.
Excision-curettage was performed for lesions >5 cm and
for lesions with a large soft-tissue mass. For lesions adjacent
to the articular cartilage of the acetabulum or the sacroiliac joint, excision-curettage
was performed so that the articular cartilage was preserved. In
these cases, the remainder of the lesion was removed by intralesional
curettage.
The choice of surgical approach was based on the location and
extent of the lesion. Size and intra-articular involvement were
important considerations in the overall management of these lesions.
Nine of the eleven primary sacral lesions were reached through a
posterior approach; one, through an anterior approach; and one,
through a combined anterior and posterior approach. An ilioinguinal
approach was used for lesions of the ilium and pubis. A Kocher-Langenbeck
posterior approach, ilioinguinal approach, or extended iliofemoral approach
was used for acetabular lesions, and a gluteal or Kocher-Langenbeck
posterior approach was used for ischial lesions.
Surgical findings revealed cystic lesions with a thin cortical rim.
At times, the wall of the cyst was formed by periosteum alone. Septa
divided the large fluid-filled cavities. Nonpulsatile bleeding was
common after the cavities were opened. The amount of blood loss
ranged from 200 to 4200 mL (mean, 1185 mL).
Adjunctive chemical cautery with phenol and acid alcohol was
used in six patients. Chemical cautery was not used for patients
with a sacral lesion.
After removal of the lesion, bone-grafting to fill the cavity was
performed in twenty-one patients. Arthrodesis of the sacroiliac
joint was performed in two patients. Hip arthrodesis was performed
in one patient with a lesion of the acetabulum associated with a
pathological fracture and medial migration of the femoral head.
Before 1945, adjuvant radiation therapy was used in two patients
with an aggressive sacral lesion. Preoperative angiography and selective
arterial embolization were performed successfully without any complications
for one sacral lesion and four lesions of the innominate bone (Figs. 2-A, 2-B, 2-C, 3-A, 3-B, and 3-C). Polyvinyl-alcohol
particles measuring 150 to 250 mm in diameter (Contour; Interventional
Therapeutics, San Francisco, California) and 3-mm embolization coils
were used.
Three of the eleven sacral lesions involved the sacroiliac joint. Sacroiliac
joint arthrodesis was performed in one patient who had an intra-articular
expansion of the lesion. The mean blood loss in the eleven patients
was 1470 mL (range, 500 to 4200 mL). Encasement of the caudal sac,
sacral nerve roots, or sciatic nerve by the lesion was noted in
six of the eleven patients. In all but three, the neural elements
were dissected free and preserved. The remaining three patients
had unilateral involvement of the sacral segments, and the lower
sacral nerve roots were resected with the lesion to achieve a clear
margin. No recurrence was noted within a mean follow-up interval
of 18.5 years (range, three to fifty-three years) after the initial operation.
Four patients had a lesion of the ilium, without any involvement
of the acetabulum. No recurrence was noted in these patients within
a mean follow-up interval of 17.8 years (five, six, fifteen, and
forty-five years). At the latest follow-up evaluation, all four
patients were disease-free; two were asymptomatic, one had mild
pain and a limp, and one had mild sacroiliac pain.
Thirteen patients had extension of the lesion into the acetabular
region. Ten of these lesions were treated with excision-curettage
and three, with intralesional curettage. Bone-grafting was used
in ten patients. The mean estimated blood loss was 1220 mL (range,
200 to 4000 mL).
The lesion was located in the ilium and extended into the superior
dome of the acetabulum in three patients, two of whom had a recurrence
within the mean follow-up interval of 9.3 years (six, ten, and twelve
years). At the time of the latest follow-up, all three patients
were disease-free and asymptomatic.
The lesion was located in the pubis and the anterior column of the
acetabulum in five patients. Excision-curettage and bone-grafting
were performed in one patient, and excision-curettage, bone-grafting,
and phenolization were performed in the other four patients. At
the latest follow-up evaluation, all five patients were disease-free:
four had no pain, and one had a flail hip and moderate discomfort
after a local recurrence and a failed hip arthrodesis.
The lesion was in the ischium and the posterior column of the acetabulum
in five patients. Two of these patients were treated with curettage
and three, with excision-curettage. Bone-grafting was used in four
patients. Recurrence was noted in one of these five patients. At
the latest follow-up evaluation, all seven patients were disease-free
and asymptomatic.
Seven patients had a lesion in the ischiopubic region. All of the
lesions involved the pubis, with no extension into the anterior
column of the acetabulum. Three of the seven patients were treated
with excision-curettage and bone-grafting and four, with intralesional
curettage. Only one of these seven patients had a recurrence of
the lesion. At the latest follow-up evaluation, all five patients
were disease-free and asymptomatic.
Recurrent Lesions
Ten patients had a recurrent aneurysmal bone cyst of the pelvis:
five of the thirty-five patients treated initially for a primary
lesion and the five patients who were referred with a recurrent
lesion. In the five patients initially treated for a primary lesion,
local recurrences were noted within eighteen months after the initial
operation. Four of these five patients underwent surgical treatment
with curettage and bone-grafting or with excision-curettage and
bone-grafting. Only one of the four patients had a second recurrence,
which required repeat curettage and bone-grafting. Another patient
with a local recurrence after excision-curettage of the lesion and
hip arthrodesis was treated successfully with external beam irradiation.
Of the five patients who were referred for treatment of a recurrent
lesion, one had a subsequent recurrence, which was treated successfully
with repeat curettage and bone-grafting.
Complications
No intraoperative vascular or major neural complications occurred.
A tear occurred in the dura mater in two patients and was repaired
successfully. In three patients with a sacral lesion, it was necessary
to cut the lower (third, fourth, and fifth) sacral nerve roots because
of involvement by the lesion. In a fourth patient the lesion encased
the sciatic nerve, producing neuropathy.
Two deep infections occurred. One infection was in a patient in
whom a sacral lesion was resected through a posterior approach.
Chronic osteomyelitis of the sacrum with sinus drainage developed
and persisted. At the time of follow-up, the patient had no functional
restriction and refused further treatment. The other infection occurred
in a patient with a recurrent lesion of the ilium. It was treated
successfully with débridement, intravenously administered
antibiotics, and antibiotic-impregnated methylmethacrylate.
Mechanical small-bowel obstruction developed in another patient
after resection of a sacral lesion through an anterior approach.
This was treated successfully with laparotomy, lysis of the adhesions,
and closure of the retroperitoneum with an omental flap.
Clinical Results
At the latest follow-up evaluation, all forty patients were disease-free
and twenty-eight (70%) were asymptomatic. After treatment
for a sacral lesion, one patient had chronic osteomyelitis of the
sacrum with sinus drainage but without pain. Another patient had
dysesthesias of the posterior aspect of the leg and thigh due to
neuropathy of the sciatic nerve, and three patients had mild low-back
pain associated with activity and prolonged sitting. After excision
of a lesion of the ilium, one patient had mild pain in the left
leg and a limp, with no other functional limitations, and another
patient had degenerative changes of the sacroiliac joint and mild
low-back pain. A patient who had a flail hip after excision of an
acetabular lesion and a failed arthrodesis reported moderate discomfort and
moderate limitation of activity. Leg-length discrepancy was noted
in three patients.
Aneurysmal bone cysts of the pelvis and sacrum are a challenging
therapeutic problem. In the present series, acetabular lesions were
associated with such complications as pathological fracture, protrusio
acetabuli with medial migration of the femoral head into the pelvis,
and hip subluxation and dislocation. Expanded forms of sacral aneurysmal
bone cysts extended to more than one segment, encasing the cauda equina
and sacral nerve roots.
Pain in the low back, groin, and thigh as well as a limp and decreased
hip joint motion were the main presenting symptoms in this series.
Sacral lesions were often associated with radicular symptoms. Persistent
radicular symptoms in young patients require further investigation
because primary bone tumors may simulate lumbar disc syndrome24.
During the early course of the condition, lytic areas may not be
recognized on plain radiographs. Although technetium bone-scanning
is a sensitive method for the detection of lesions about the pelvis25, aneurysmal bone cysts are visualized
most effectively with computed tomography or magnetic resonance
imaging. The definition of the lesion, the demonstration of its
extent, and the identification of its nature are better accomplished
with weighted magnetic resonance imaging than with computed tomography26-28. Multiple fluid levels within
a multiloculated lesion, best seen on T2-weighted magnetic resonance
images, support the diagnosis of aneurysmal bone cyst, although
they are not specific for this entity.
The method of treatment of aneurysmal bone cysts of the pelvis
and sacrum must be individualized and depends on the location, extent,
and aggressiveness of the lesion. Complete intralesional excision,
including exteriorization of the lesion and removal of the overlying
cortex of the entire lesion so that the cavity can be seen clearly,
must be performed in a systematic, thorough manner. The current
technique of curettage uses power burrs to extend the excision past
the reactive zone that surrounds the lesion. Lesions of £5
cm that exhibit minimal destruction or expansion of cortical bone
and do not threaten the integrity of the acetabulum or the sacroiliac
joint are best treated with intralesional curettage, with or without bone-grafting.
Lesions of >5 cm that exhibit large areas of destruction
or major expansion of cortical bone and threaten the integrity of
the acetabulum or the sacroiliac joint require a more aggressive
approach with use of the excision-curettage technique. If the stability
of the hip is endangered because of the resection and curettage,
autogenous tricortical iliac crest bone grafts should be used to
restore structural integrity. Large bone defects may require reconstruction
with structural allografts, as was performed in one patient in the
present series who had a destructive lesion of the anterior column
of the acetabulum.
In the present series, adjunctive therapy involved chemical cauterization
and radiation. Chemical cauterization with phenol was used in six
patients who had a relatively large primary lesion in order to kill
any cells of the lesion on the surface of the curetted cavity29,30. It was not used for sacral lesions
to avoid nerve root damage. In the present series, all recurrences
were in patients who had not had chemical cauterization after curettage,
although, because of the small sample size, no significant difference was
found in the prevalence of recurrence between patients who had received
supplemental phenol and those who had not (p > 0.05, two-tailed
Fisher exact test). Cryotherapy was not used in the present series;
however, it has been recommended that this modality be used adjuvantly
with surgical treatment to provide local control and to achieve,
with bone-grafting, consolidation of the lesions31,32.
Adjuvant radiation was used relatively early in this series; currently,
however, it is not recommended for aneurysmal bone cysts, to minimize
the risk of postradiation sarcoma31,33,34.
Radiation is used only when no surgical options are available.
Substantial intraoperative bleeding was common in the present
series, and it was observed most often with curettage of lesions
of either the sacrum or the acetabulum. Selective arterial embolization
may diminish this complication. Preoperative angiography and selective
embolization therapy are now performed for lesions with substantial
soft-tissue expansion and for lesions larger than 5 to 8 cm to minimize
intraoperative hemorrhage35-39.
It can be considered a primary treatment for lesions whose size
or site makes other types of treatment difficult or hazardous36,38-40.
Intra-articular aneurysmal bone cysts of the acetabulum were managed
successfully in most patients, and hip arthrodesis was performed
only in one patient with a primary lesion. Excision of a pelvic
aneurysmal bone cyst may be difficult when the lesion extends around
iliac vessels, the lumbar plexus, or the sciatic nerve. In the present
series, no major vascular or neural damage occurred. In all cases,
neural elements were dissected free and preserved. Only in three
cases were the lower sacral nerve roots sacrificed, unilaterally,
during the resection of a large lesion of the sacrum. However, patients
with a sacral lesion were more likely to present with neurological
symptoms and they had a higher rate of neurological complications.
This finding is consistent with those observed by others11.
In the present series, all recurrent lesions developed within eighteen
months after the initial intralesional curettage of the primary
lesion. Other authors have noted that recurrent lesions developed
within several months to two years after the primary surgical treatment1,9,18,41,42. Overall, the management
of the pelvic and sacral lesions was successful, with only five
recurrences (14%) among thirty-five primary lesions. Capanna
et al.43, in their review of twenty-three
aneurysmal bone cysts of the pelvis, noted a recurrence rate of
13%. Incomplete excision is associated with a relatively
high recurrence rate1,13, but
complete excision of the lesion is a highly successful method of
treatment for these lesions9,44.
Malignant transformation of an aneurysmal bone cyst is extremely
rare45.
Recurrences were treated predominantly with excision-curettage
and bone-grafting. Only one referred case of a recurrent lesion
of the acetabulum, incorrectly diagnosed as low-grade chondrosarcoma,
was managed with wide local excision. Excision-curettage and bone-grafting
was successful for the treatment of the recurrent lesion in all
but two patients. The second recurrence in these two patients was
treated successfully with subsequent excision-curettage and bone-grafting, without
any further recurrence. Recurrence of aneurysmal bone cysts can
be avoided by use of extensive imaging; by preoperative planning
with use of information from cross-sectional imaging, preoperative
arteriography, and selective arterial embolization; and by aggressive
surgical treatment.