Between January 1975 and July 1996, sixty-four patients with pelvic
chondrosarcoma were treated surgically at our institution (Table I). Patients
who were first seen with a recurrent lesion after initial surgical
resection elsewhere were excluded. No patient had regional or distant
metastasis. All operative procedures were performed by an orthopaedic
oncologist.
There were forty-one male and twenty-three female patients, with
a mean age at diagnosis of forty-seven years (range, fifteen to
eighty-eight years). The patients were followed for a minimum of
three years or until death, with the living patients followed for
a median of 140 months (range, thirty-nine to 295 months). No patient
was lost to follow-up. Approval from the Institutional Review Board
at our institution was obtained before the study was begun. Medical
records, including notes on the operative procedure, clinical follow-up
examinations, and pathologic findings, were reviewed in detail.
Preoperative staging studies generally included routine radiography,
computed tomography scanning, magnetic resonance imaging, and nuclear
scanning. The grade and size of the tumor, type of surgical resection,
surgical margin, development of local recurrence or metastasis,
and patient outcome were evaluated. At the time of follow-up, patients
either were seen in the clinic or responded to a telephone questionnaire
about their status.
Criteria for Distinguishing Between Benign
Cartilage Tumors and Chondrosarcomas
A preliminary diagnosis of chondrosarcoma was frequently based
on the clinical and radiographic findings. A definitive diagnosis
was then established by correlation of the clinical and radiographic
findings, the gross pathologic findings, and the histopathologic
findings. In addition to the characteristic histopathologic hallmarks
of malignancy (described below), an important pattern distinguishing
chondrosarcoma from enchondroma or osteochondroma is permeation,
which is best recognized when the tumor is seen to fill the marrow
spaces entrapping preexisting osseous trabeculae. Secondary chondrosarcomas
that developed in patients with Ollier disease, a single osteochondroma,
or multiple osteochondromas were clearly malignant lesions according
to the review of the musculoskeletal pathologist. The diagnosis
of chondrosarcoma arising in an osteochondroma is based on the radiographic, gross,
and histologic features. Radiographically, the lesion has a thick,
irregular cartilage cap and lack of uniform mineralization. Grossly,
the cartilage cap is abundant (often >2 cm in thickness)
with myxoid change in the matrix (seen as cystic spaces), and the
surface is irregular and bosselated. Histologically, small foci
of cartilage are separated from the main mass and permeate through
the surrounding tissue. The usual organized columnar arrangement
of chondrocytes is lost.
No benign cartilage tumors, particularly large osteochondromas,
were included in the present series.
Criteria for Grading of the Degree of Malignancy
in Chondrosarcomas
Grading of primary chondrosarcoma is difficult and varies among
institutions. The histopathologic analysis is based primarily on
morphologic and cytologic criteria, including cellularity, nuclear
size (including anisocytosis), and intensity of nuclear staining.
The three-tiered grading system described by Lichtenstein and Jaffe
was used18. Cytologically, increased
cellularity and cytological atypia are the most important features,
and these characteristics are used to determine the grade of the
chondrosarcoma. With the current grading system used at our institution,
the majority of pelvic chondrosarcomas are classified as low-grade
(grade-1 or 2), with few high-grade (grade-3 or 4) tumors. In the present
series, thirty-three patients had a grade-1 lesion; twenty-three
patients, a grade-2 lesion; one patient, a grade-3 lesion; and seven
patients, a grade-4 lesion (dedifferentiated chondrosarcoma). The
last diagnosis is made when a high-grade spindle-cell sarcoma is
juxtaposed with a low-grade chondrosarcoma.
Another histologic differentiation was made between chondroblastic
osteosarcoma and high-grade chondrosarcoma. To be called a chondroblastic
osteosarcoma, a lesion has to have sheets of spindle cells surrounding
chondroid lobules and lace-like osteoid has to be present between
the malignant spindle cells.
Tumor size was defined as the maximum diameter of the tumor at
the time of pathologic analysis. The mean tumor size was 10 cm (range,
2 to 24 cm). The staging system described by Enneking et al.19 was used to determine which patients
had stage-IA disease (low-grade and intracompartmental), stage-IB
disease (low-grade and extracompartmental), or stage-IIB disease
(high-grade and extracompartmental). Forty-nine patients had a primary
pelvic chondrosarcoma, and fifteen had a lesion that developed from
a previous benign cartilage lesion (Table I). Tumor location was assigned according
to the Musculoskeletal Tumor Society classification10,20. Fifteen patients had a type-P1
(iliac) lesion, with five of the lesions extending into the sacrum;
thirty-nine had a type-P2 (acetabular) lesion; and ten had a type-P3
(ischiopubic) lesion.
Adjuvant treatment (chemotherapy) was given to only two patients,
and they both had a dedifferentiated lesion.
The decision to perform either a hemipelvectomy or a limb-salvage
procedure was made by the orthopaedic oncologist on the basis of
whether clear margins and functional restoration could be obtained.
As a general rule, if two of three structures (the lumbosacral plexus,
femoral neurovascular bundle, and hip joint) could be maintained,
the surgeons believed that a reasonably functional extremity could
be salvaged. In this series, a limb-salvage procedure was performed
on fifty-one patients, whereas thirteen patients required a hemipelvectomy.
Eleven of the thirteen patients requiring amputation had a periacetabular
lesion. When the preoperative imaging studies showed involvement
of vital neurovascular structures and an adequate margin could not
be achieved by limb-sparing resection, an external hemipelvectomy
was performed. Considering the indication for limb-salvage procedures
in this series, we thought that the expected recurrence rate should
be comparable with that following amputation and should produce
a superior functional result. A multidisciplinary team participated
in the surgical removal of many of the pelvic tumors. Neurosurgeons
often assisted if the tumor involved the lumbosacral plexus. Urologists
performed preoperative cystoscopy and inserted urethral stents if
the tumor was close to the genitourinary structures. When necessary,
vascular, general, and plastic surgeons also assisted in the tumor
resection or soft-tissue reconstruction.
The resection of the tumor in the patients amenable to limb salvage
was performed with a modification of the system reported by Enneking
and Dunham10 (Fig. 1). Type-I resection
includes only the ilium; type-IA, the ilium and the gluteal muscles;
and type-I/S, the ilium and a portion of the sacrum. Type
II is a periacetabular resection, and type-IIA includes the hip
joint. Type III is a resection of all or a portion of the ischium
and pubis. Multiple combinations of these resections were used,
depending on the tumor location and the extent of osseous invasion
(Table II).
Of the fifteen patients who had an iliac lesion, eight had removal
of all or a portion of the ilium (type-I excision) or of both the
ilium and the surrounding gluteal muscles (type-IA excision). When
the plane of dissection extended across the sacroiliac joint to
include a portion of the sacrum or spinal column, the excision was
classified as a type I/S or IA/S (seven patients).
In four of the seven patients who had a type-I/S or type-IA/S
excision, the procedure included lumbar or sacral laminotomy and
sacrifice of nerve roots.
A variety of resections were performed for the thirty-nine periacetabular
lesions. Four patients underwent resection of a portion of the acetabulum
without removal of any portion of the surrounding ilium or ischium
(a type-II excision). Seven patients required removal of all or
a portion of the acetabulum as well as all or a portion of the pubis
and ischium (type-II/III excision). Eight patients required
an extracapsular excision of the proximal portion of the femur in
addition to all or a portion of the acetabulum and all or a portion
of the ischium and pubis (type-IIA/III excision). Five
patients required resection of all or a portion of the acetabulum
in addition to all or a portion of the ilium and sacrum (type-I/II
or type-I/S/II excision). Four patients required
excision of the entire hemipelvis extending into the sacrum. Eleven
patients required a hemipelvectomy.
Of the ten patients with an ischiopubic lesion, eight had a type-III
excision of all or a portion of the pubis and ischium and two had
a hemipelvectomy.
The primary oncologic goal at the time of surgery was to achieve
uncontaminated, clear margins, with recognition of the difficulty
of obtaining wide margins in the pelvis according to the standard
definition. If the intrapelvic extension of the tumor was resected
with the overlying iliacus muscle, the surgical margin was considered
wide. If the tumor abutted the surrounding genitourinary structures
or bowel and the peritoneum or bladder could be clearly delineated
and separated easily from the tumor, the resection was considered
marginal. If there was involvement of the viscera or portions thereof,
a partial resection of the involved structures (that is, a partial cystectomy)
was performed. Similarly, for tumors in the posterior part of the
pelvis, an extended limb-sparing resection or an amputation through
the sacrum or vertebrae was often necessary to achieve an adequate
margin. In this particular series, surgical margins were classified
as wide in forty patients and marginal in thirteen. The wide or
marginal excision was considered to be contaminated if a focal area
of tumor was noted on frozen section. The excision was then widened,
at the same operative setting, to yield negative margins. Five patients
had a contaminated wide excision, and three patients had a contaminated
marginal excision. Three patients had an intralesional excision.
An intralesional margin was defined as a resection violating the
substance of the tumor, with positive margins noted on final histologic
evaluation.
Reconstructive procedures were individualized (Table II) on the basis
of such variables as patient age, functional demands, and extent
of the tumor and according to surgeon preference. The primary goal
of each procedure was to completely excise the tumor with wide margins;
the secondary goal was to maintain a stable, functional limb. If
the acetabulum was maintained, minimal reconstruction was necessary. However,
if a periacetabular resection was completed, maintenance of a functional
limb often required reconstruction. General reconstructive options
for patients undergoing limb salvage included arthrodesis, creation
of a pseudarthrosis, allograft reconstruction with or without total
hip arthroplasty, and insertion of a saddle prosthesis21-33.
The most common type of reconstruction was an arthrodesis, which
was done in sixteen patients. Five had a sacroiliac arthrodesis
(Figs. 2-A, 2-B, and 2-C), and eleven had
an iliofemoral arthrodesis (Figs. 3-A, 3-B, and 33-C). Seven patients had an attempted
creation of a pseudarthrosis, which was sacrofemoral in one, iliofemoral
in one, and ischiofemoral in five. A saddle prosthesis was used
in two patients (Figs. 4-A, 4-B, and 4-C). A segmental acetabular allograft
alone was used in one patient, and an acetabular allograft combined
with a total hip replacement was used in three patients. The patients
who had an ischiopubic resection did not require reconstruction.
Of the five patients who had an attempted sacroiliac arthrodesis,
four had solid healing and one required removal of the allograft
because of infection. Of the eleven patients who had an attempted
iliofemoral arthrodesis, eight had union, two had a nonunion that
did not require surgical intervention and were left with a pseudarthrosis,
and one required an amputation because of infection. The only patient
with a large segmental acetabular allograft alone had good healing.
Of the three patients who had a large acetabular allograft and a
total hip replacement, one had solid union, one had a nonunion requiring
removal of the allograft and placement of a saddle prosthesis, and
one required hardware removal and amputation because of local recurrence.
At the time of the final follow-up, patient function was assessed
with a lower-extremity functional evaluation developed by the International
Society of Limb Salvage and modified by the Musculoskeletal Tumor
Society, as described by Enneking et al.34.
Five points each were assigned on the basis of pain, disability,
emotional acceptance, support, walking, and gait. The points were
totaled (with a maximum possible total score of 30 points), and
percentages were assigned. Finally, patients were followed clinically
and radiographically on a regular basis postoperatively. The timing
of local recurrence or distant metastasis was recorded, and the
recommended treatment was noted for all patients in whom a recurrence
developed.
Summary statistics were presented as the mean (and the standard
deviation) for continuous variables and as frequencies and percentages
for categorical data. Rates of overall survival, local recurrence,
metastasis, and disease-free survival were estimated with use of
Kaplan-Meier survival analysis. Risk factors for these end points
were compared with use of the log-rank test. The effects of continuous
variables and time-dependent variables were assessed with use of
Cox proportional-hazards models. The threshold of significance of
all statistical tests was set at a = 0.05. All analysis
was done with the SAS software program (version 6.12; SAS Institute,
Cary, North Carolina) on a Sun Ultra 2 computer (Sun Microsystems,
Mountain View, California).
Oncologic Evaluation
At the time of the final follow-up (maximum duration, nearly twenty-five
years), forty-five (70%) of the original sixty-four patients
were alive. All were contacted by telephone and were asked to complete
a questionnaire. Of the living patients, twenty-eight returned for
updated clinical evaluations. The remainder of the patients were
followed by a local physician, and the majority sent follow-up radiographs
for review.
Survival
Forty-four patients (69%) were alive without evidence
of disease at a mean of 152 months (range, thirty-nine to 295 months).
Thirteen patients (20%) died of the disease at a mean of
thirty-five months (range, four to 139 months). Six patients died
of other causes but had no evidence of chondrosarcoma at the time
of death. These patients had been followed for a mean of seventy-five
months (range, eighteen to 190 months). One patient remained alive
but had slowly progressive pulmonary metastases at 110 months. The
overall survival rate was 94% at one year, 82% at
five years, 80% at ten years, and 77% at fifteen
years. Survivorship curves with 95% confidence intervals,
with death due to disease as the end point, are shown in Figure 5. Univariate
analysis revealed that high grade (p < 0.001), high stage
(p < 0.001), and a primary lesion (p = 0.029)
significantly affected the overall survival rate. Multivariate analysis also
revealed that high grade was significantly related (p = 0.0001)
to the overall survival rate.
The overall survival rates were calculated according to tumor grade.
The ten-year survival rate was 97% for patients with a grade-1
lesion, 75% for those with a grade-2 lesion, and 14% for
those with a dedifferentiated chondrosarcoma (Fig. 6). The one patient
with a grade-1 lesion who died had a local recurrence, classified
as a grade-3 lesion, and subsequent systemic metastasis. The disease-free
survival rate was 80% at one year, 69% at five
years, 67% at ten years, and 63% at fifteen years.
Univariate analysis revealed that high grade (p < 0.001),
high stage (p < 0.001), a primary lesion (p = 0.019), and
an age of more than forty-five years (p = 0.028) significantly
affected disease-free survival. Multivariate analysis revealed that
high grade (p < 0.0001) and less than a wide margin (p = 0.05)
were significantly related to disease-free survival.
As mentioned above, patients with a secondary chondrosarcoma
had a significantly better overall rate of survival compared with
those with a primary chondrosarcoma. Of the fifteen patients with
Ollier disease or single or multiple osteochondromas, thirteen were
alive and disease-free and two had died of other causes at the time
of the final follow-up. When the thirteen patients with a pelvic
chondrosarcoma secondary to an osteochondroma were excluded, the
overall survival rate was slightly lower: with death due to disease
as the end point, the survival rate was 92% at one year,
78% at five years, 75% at ten years, and 71% at
fifteen years.
Local Recurrence
Twelve patients (19%) had a local recurrence at a mean
of twenty-four months (range, seven to fifty-six months). Ten of them
had additional surgical intervention, and seven were alive and disease-free
at the time of the final follow-up. Two of the patients who had
a local recurrence had concurrent distant metastasis. Less than
a wide margin at the time of the initial operation correlated with
an increased chance of local recurrence (p = 0.014) (Fig. 7). In addition,
high grade correlated with an increased chance of local recurrence
(p = 0.001). Local recurrence was associated with an increased
risk of death (p = 0.0001). The cumulative probability
of local recurrence at ten years was 8% for forty patients
with a wide surgical margin at the initial operation, 38% for
thirteen patients with a marginal margin, and 43% for eleven
patients with contamination during the operation. Three patients
in the series had an intralesional curettage as the primary surgical
treatment. One of them, who had a grade-2 lesion, had systemic metastasis
twenty-one months after surgery and died three months later. The
second patient, who had a grade-1 chondrosarcoma, had a local recurrenceæa grade-3
tumor as previously mentioned. This patient had an internal hemipelvectomy
after the first recurrence. There was a second recurrence, and the
patient required an external hemipelvectomy. Systemic metastasis
subsequently developed, and the patient later died. The third patient
remained free of disease with no evidence of local recurrence nearly twelve
years after the initial operation. Age, gender, and tumor size were
not found to be correlated with local recurrence.
Metastasis
Distant metastases developed in eleven patients (17%), including
the two with concomitant local recurrence, at a mean of twenty-one
months (range, two to seventy-nine months). Metastasis developed
after a grade-1 chondrosarcoma had recurred as a grade-3 tumor in
one patient; it developed from a grade-2 lesion in four patients,
and it developed from a dedifferentiated chondrosarcoma in six patients.
Ten of the eleven patients died of the disease at a mean of twelve months
(range, four to eighty-one months) after the diagnosis of the metastasis.
One patient was alive but had multiple unresectable pulmonary metastases
thirty-one months after the initial diagnosis of systemic spread.
High tumor grade (p < 0.001) and stage (p < 0.0001)
were highly correlated with distant metastasis. Distant metastasis
increased the overall risk of death (p = 0.0001). Tumor
size, location, and margins were not correlated with development
of distant metastasis.
Function
All living patients were contacted to evaluate their current functional
status, which was graded as a percentage of a possible total score
of 30 points. The patients who lived more than twenty years after
the initial operation often lost function secondary to comorbid
conditions. All patients who had a limb-salvage procedure were able
to walk at the time of the final follow-up, and they had substantially
improved function compared with those who had hemipelvectomy (77% and 43%,
respectively) (Table III). However, all patients contacted
were satisfied with the results and their current ability to perform
activities of daily living. Patients who had had a type-I or type-III
resection tended to have higher functional scores than did those
in whom the acetabulum or sacrum had been involved. This finding
was also observed in a previous series of patients with a pelvic
tumor and subsequent reconstruction20.
Patients who had not required a reconstruction had higher scores
compared with those who had had a reconstructive procedure, a finding
that was related to the extent of the initial lesion. Of the patients
who had had a reconstruction, those with a stable extremity had
higher scores than those with a pseudarthrosis.
Twenty-four (53%) of the forty-five living patients
required no assistive devices for walking; five required a cane;
nine, crutches; five, a walker; and two, a wheelchair. At the time
of the final follow-up, twenty-two (49%) of the forty-five patients
had no pain, fourteen patients had mild pain, two had moderate-to-severe
pain, and seven had no pain at the surgical site but complained
of varying degrees of phantom pain.
Complications
Considerable morbidity and mortality after extensive pelvic resection
and reconstruction have been reported in the literature20,35,36. In the present series of
sixty-four patients, there were no intraoperative or perioperative
deaths. Forty-three patients (67%) had no important complications
(Table IV).
Twenty-one patients had a total of twenty-nine complications. Of
these twenty-one patients, fifteen (71%) had had a limb-salvage
procedure and six (29%) had had a hemipelvectomy. Of the
fifty-one patients who initially had a limb-salvage procedure, fifteen
(29%) had a complication. Six of the thirteen patients
who initially required an amputation had postoperative complications.
Chondrosarcoma of the pelvis poses a difficult treatment problem.
Chemotherapy and radiation therapy are not reliably effective, in
either the neoadjuvant or adjuvant setting, for the treatment of
classic chondrosarcoma. Therefore, the adequacy of surgical resection
determines the outcome5,14,16.
In general, the literature suggests that patients with pelvic chondrosarcoma
do worse than patients with chondrosarcoma in an extremity1. Presumably, this tendency results
from the anatomic complexity of the pelvis and the close proximity
of the tumor to surrounding neurovascular and visceral structures.
Recent studies on pelvic chondrosarcoma have documented a five-year
survival rate of approximately 65% and a ten-year survival
rate of approximately 54%8,11,16.
These findings are consistent with the results of other series on
chondrosarcoma of the entire skeleton, which have demonstrated five-year
survival rates of 67% to 79% and ten-year survival
rates of 50% to 66%1,5,14.
The five-year survival rate of 82% and the ten-year survival rate
of 80% in the present series compare favorably with those in
the previous studies.
Sheth et al.16, in a series
of sixty-seven patients with pelvic chondrosarcoma, reported a ten-year
survival rate of 52%; however, thirteen patients in their
series had a dedifferentiated chondrosarcoma. The fact that they
included a higher number of patients with a dedifferentiated lesion
may partially account for their lower overall survival rate. In
addition, they used different inclusion criteria and included eight
patients with a recurrent lesion, who clearly had a worse prognosis.
It is of note that twenty-eight (43%) of the sixty-five
patients in their series who could be evaluated had positive margins
at the time of surgery, which likely contributed to the higher rate (28%)
of local recurrence.
The grade of the tumor proved to be extremely significant in terms
of overall and disease-free survival in the current series. A marked
difference was demonstrated among the ten-year survival rates associated
with grade-1 lesions (97%), grade-2 lesions (75%),
and dedifferentiated lesions (14%). These findings are
similar to those reported by Sheth et al.16.
Grade has been shown to be prognostically significant (p < 0.001)
for chondrosarcomas in all skeletal sites5,14 as
well as for those in the pelvis37,38.
It is difficult to compare clinical series of patients with chondrosarcoma
because the grading varies among institutions. Although most pathologists
agree with the three-tiered system of grading cartilage lesions,
there seems to be differences in the percentages of grade-1, 2,
and 3 tumors among large centers. The percentage of grade-1 chondrosarcomas
has ranged from 26% to 61%; that of grade-2 chondrosarcomas,
from 34% to 61%; and that of grade-3 chondrosarcomas,
from 9% to 28%1,5,14,37.
Mayo Clinic series tend to have a higher proportion of grade-1 and
grade-2 tumors1,5, whereas those
from the Rizzoli Institute and M.D. Anderson Cancer Center have
a higher proportion of grade-3 tumors14,16,39.
While it is possible that referral patterns are different for patients
with pelvic chondrosarcoma, it is unlikely that patients with higher-grade
tumors are referred to centers other than ours. It has been the
practice of the musculoskeletal pathologists at our institution
to classify the majority of chondrosarcomas as grade 1 or 2. As
evidenced by a large series of 344 chondrosarcomas throughout the
musculoskeletal system reviewed at the Mayo Clinic1,5, the percentage of high-grade tumors
has remained consistent over the eighty years spanning these studies.
As mentioned, it would be unlikely that all patients with high-grade
tumors were being referred elsewhere. The mean tumor size in the current
series was 10 cm. In general, these are extremely large tumors with
intrapelvic or extrapelvic extension and are representative of pelvic
chondrosarcoma. We believe that it is not the tumors but the grading
systems in the various clinical reports that are different. It is
most important for the orthopaedic oncologist to understand the
particular grading scheme at his or her own institution. However,
we recommend a wide surgical resection of all pelvic chondrosarcomas,
regardless of their specific grade. Grade-1 tumors are likely to
recur locally and need to be treated aggressively.
The local recurrence rate of 19% (twelve patients) in
the current series compares favorably with those in other studies
of pelvic chondrosarcoma, which have ranged from 28% to
45%8,11,16. We believe
that aggressive surgical resection with an attempt to obtain adequate
surgical margins in all cases has led to this decrease in the local
recurrence rate. Although there has been a trend toward the removal
of low-grade chondrosarcomas of the extremities by curettage, we
believe that this approach should be avoided in the pelvis because
of the consequences of recurrence in this location. Of the twelve
patients who did have a local recurrence in the current series,
seven were disease-free after additional aggressive surgery.
Systemic metastasis developed in eleven patients (17%)
in the current series. This rate compares favorably with those in other
reports on pelvic chondrosarcoma, in which the prevalence of distant
metastasis has ranged from 19% to 36%8,11,16, as well as with those in general
series on chondrosarcoma of the entire skeleton, in which the prevalence
of systemic disease has ranged from 6% to 37%1,5,14,37. Metastases are more likely
to develop in patients with a higher-grade lesion and are almost
uniformly present in patients with dedifferentiated chondrosarcoma40. All but one of our patients with
metastases died of the disease, at an average of one year after
the diagnosis of the metastasis. However, one patient with a dedifferentiated
chondrosarcoma was alive and disease-free at 152 months after the operation.
In addition, a patient with a grade-3 chondrosarcoma died of other
causes with no evidence of the chondrosarcoma 190 months after the
operation. The majority of patients in this series did not receive
adjuvant chemotherapy. A continued search for better chemotherapeutic
agents for patients with dedifferentiated chondrosarcoma might extend overall
survival in this subset. In view of the fact that, in one of our
patients, a grade-1 chondrosarcoma recurred at a grade-3 level with
subsequent development of systemic metastasis and death, we again
stress the importance of complete resection of the initial lesion.
If the tumor recurs, it may do so at a higher grade, which puts
the patient at increased risk for systemic disease and death. Local
recurrence and systemic metastasis were independent predictors of
poor overall survival in the present series.
Patients who had a primary chondrosarcoma of the pelvis had worse
rates of overall and disease-free survival than those in whom a
secondary chondrosarcoma had developed from a previously benign
cartilaginous lesion. When the thirteen patients (20%)
in whom a secondary chondrosarcoma developed from an osteochondroma
were analyzed separately, overall survival improved approximately
5% at all time-points. The percentage of patients with
a secondary chondrosarcoma in the present series was higher than
that in other series. Sheth et al.16 reported
that 10% of their patients had a secondary tumor. In the
current series, the same grading system was used for all patients
regardless of whether they had a primary or secondary chondrosarcoma.
The patients with a chondrosarcoma arising from a previous osteochondroma
had clear evidence of a malignancy on the basis of standard histologic
criteria. Every tumor in this series was reviewed in detail by a
highly trained musculoskeletal pathologist. Of the thirteen patients with
a secondary chondrosarcoma from an osteochondroma, eleven were considered
to have a grade-1 tumor and two had a grade-2 tumor. It is likely
that patients with single or multiple osteochondromas were followed
more closely over an extended period of time; therefore, transformation
into a malignant lesion was discovered earlier, when it was easier
to remove it completely.
The indications for hemipelvectomy have not changed over the
course of this study. If adequate margins cannot be obtained with
a limb-salvage procedure, or if two of three structures (the femoral
neurovascular bundle, lumbosacral plexus, or hip joint) are involved
by tumor, a hemipelvectomy is performed. The improved imaging modalities,
better surgical techniques, and clearer understanding of the margins required
for complete tumor resection that have been developed over the past
twenty-five years have allowed a higher percentage of patients to
safely undergo limb-salvage procedures.
The functional outcome analysis in this series revealed that patients
who had had limb salvage had an improved functional score compared
with those who had had a hemipelvectomy. In our series, hemipelvectomy
was indicated only if inadequate margins would be obtained with
a limb-salvage procedure. After attaining adequate resection margins,
limb-salvage procedures could be completed in 80% of our patients.
In the study by Sheth et al.16,
external hemipelvectomy was performed in thirty-five (52%) of
sixty-seven patients. Patients who had had partial sacral or periacetabular
resection had a worse functional outcome than did those who had
had a type-I or type-III resection. This finding may be due to the
fact that resection of the sacroiliac joint or periacetabular region
creates deficiencies that are difficult to reconstruct into a stable
extremity. In the majority of patients in the present series, the
desired reconstructive result was achieved. Our reconstructive choice
is based on multiple factors, including the experience of the surgeon,
the age and physical condition of the patient, and the available
bone stock after tumor resection. In general, patients had better
function when pelvic stability was restored, a finding that is similar
to those of O’Connor and Sim12,20.
Older patients with lower activity demands who had a P2 lesion were
often treated with planned iliofemoral or ischiofemoral pseudarthrosis.
Fewer complications are associated with this type of surgery, although
the extremity is not as stable as that in patients who have a successful
iliofemoral arthrodesis.
Although pelvic chondrosarcoma is a difficult treatment problem,
the present series revealed that the majority of patients survive
for a long time after aggressive surgical resection. We continue
to see patients with recurrent pelvic chondrosarcoma, which suggests
that the initial surgical resection of these tumors is often inadequate41. It is extremely important to obtain
an adequate margin during resection of these pelvic tumors, regardless
of their grade. This may require amputation or resection of portions
of the pelvic viscera and neurovascular structures. A continued search
for better reconstructive options should improve the functional
outcome of these patients. Improved chemotherapeutic options for
high-grade and dedifferentiated chondrosarcoma may improve overall
patient survival.