Orthopaedics & Traumatology: Surgery & Research (2009) 95, 393—401
Pelvic chondrosarcomas: Surgical treatment options
X. Deloin∗, V. Dumaine, D. Biau, M. Karoubi,
A. Babinet, B. Tomeno, P. Anract
Orthopaedic and Traumatology Department A, Cochin Hospital, 27, rue du Faubourg-Saint-Jacques, 75014 Paris, France
Accepted: 12 May 2009
Introduction: Chondrosarcoma (CS) is a primary malignant bone tumor with cartilaginous dif-
ferentiation. The only available treatment is carcinological surgical resection since the usual
adjuvant treatments are ineffective. The pelvic location creates specific technical difficulties
both for exeresis and reconstruction. Our objective was to evaluate the carcinological and
functional outcomes of inter-ilioabdominal amputation and conservative surgery.
Materials and methods: We retrospectively studied 59 cases of pelvis chondrosarcoma managed
in our department between 1968 and 2003. Demographic, anatomopathological, surgical and
survival data were analyzed. Survival was estimated by the Kaplan-Meier curves and the cumu-
lative incidence method. Multivariate analysis was used to identify all possible independent
Results: There were 33 men and 26 women, with an average age of 48 years. The average
follow-up duration was 94 months. Eleven patients had a grade 1 chondrosarcoma, 36 a grade 2
chondrosarcoma, five were grade 3, and seven were dedifferentiated chondrosarcoma. Eleven
patients underwent an inter-ilioabdominal disarticulation, and 48 had a more conservative
surgery. Resection margins proved healthy in 46 patients (78%). Eighteen patients (31%) had
a local recurrence, and 12 (20%) had metastases. At last follow-up, 30 patients (51%) were
still alive without any sign of recurrence. Twenty-three patients (39%) died from the disease.
Multivariate analysis showed that margin invasion was associated with a definitely increased
local recurrence rate. A high tumoral grade was correlated with a greater risk of metastases
occurrence. These two last factors (margin status and tumor grade) as well as acetabulum
involvement were correlated with a reduced survival rate.
Function was better among patients treated by conservative surgery, and among them, even
better when the peri-acetabular area remained intact.
Our study confirmed that resection margins quality is a major prognostic factor both for local
control and for survival. On the other hand, local recurrence is an adverse survival prognosis
factor and is itself correlated with resection margins quality.
E-mail address: firstname.lastname@example.org (X. Deloin).
1877-0568/$ – see front matter © 2009 Elsevier Masson SAS. All rights reserved.
394 X. Deloin et al.
Peri-acetabular chondrosarcoma location (in zone 2) appears to be a poor oncological prog-
nosis factor since, in this location, obtaining healthy margins appears particularly difficult.
Compared to resection, inter-ilioabdominal amputation did not prove its superiority concern-
ing resection margins quality or survival. However, resection guaranteed a better functional
Conclusion.— Chondrosarcoma of the pelvic girdle remains of worse prognosis than periph-
eral bones chondrosarcoma since the critical prognosis factor is the resection margins quality.
This location, and especially the peri-acetabular zone, poses difficult specific technical prob-
lems when conservative surgery is selected. Various imaging techniques should help better
envision tumor resection extent. Inter-ilioabdominal amputation should only be resorted to
in non-metastatic patients, when the tumor does not seem to be removable with sufficient
healthy margins guarantee, or when local conditions make it impossible to hope for a good
Level of Evidence.— Level IV; therapeutic retrospective study.
© 2009 Elsevier Masson SAS. All rights reserved.
Chondrosarcoma (CS) is a malignant skeletal tumor with
cartilaginous differentiation. In terms of incidence, it repre-
sents the second most frequent bone tumor in adults , and
is preferentially located in the pelvis in 22 to 39% of cases
[2—7]. CSs of the pelvic girdle remain asymptomatic in the
long-term and may thus be large at the time of diagnosis.
From a therapeutic viewpoint, CS is individualized
from other primary skeletal tumors as it is radioresistant
and chemoresistant: its only treatment is carcinological
surgical exeresis. However, depth in the pelvis, tumor
size and its connections with local articular, nervous,
vascular and visceral structures very often make carcinolog-
ical exeresis difficult and require complex reconstructions
The objective of this work is to evaluate the carcinolog-
ical and functional results of different surgical treatments
of CS of the pelvic girdle.
Materials and methods
We report data from a retrospective, monocentric study in
a series of 59 patients with CS of the pelvis and treated
surgically by the Orthopaedics Service of Cochin Hospital
between 1968 and 2003. Patients treated initially in another
establishment were not included.
In all cases, the diagnosis of CS was confirmed histolog-
ically by a referred pathologist specialized in the diagnosis
of tumors of the locomotor apparatus. The following clin-
ical data were analyzed: sex, age, initial clinical picture,
duration of symptoms at the time of diagnosis. Tumor loca-
tion was studied on radiographs as well as by scanner and
magnetic resonance imaging when they were available. The
topographical characteristics of the tumor were expressed
according to the Muskuloskeletal Tumor Society classifica-
tion [10,13]. Macroscopic data, i.e., tumor size defined by
its largest diameter, and microscopic findings, i.e. grade
according to O’Neal and Ackerman , were collected.
Resection margins were analyzed according to the classifi-
cation of Enneking et al. . Type of surgical intervention,
local recurrence, metastases, and patient outcome were
studied. Function was evaluated by walking, and pain graded
according to the 30-point score of the Musculoskeletal Tumor
Follow-up was calculated from the intervention data at
Cochin Hospital until the day of death or the last follow-
up. Survival was estimated by Kaplan-Meier curves and the
cumulated incidence method in the presence of events com-
peting with the event of interest.
Cox’s proportional model was used to estimate the effect
(hazard ratio or relationship of instant risks) of different
variables on the events considered. Multivariate analysis
served to identify a set of variables independently prognos-
tic of the events. The analyses were performed with the R
2.5.1 statistical analysis program; all tests were bilateral
and significant from a threshold of 0.05.
Thirty-three men and 26 women, with an average age of 48
years at the time of diagnosis (23 to 78 years) were managed
for CS of the pelvic girdle. Average follow-up was 94 months
(3—311 months). The patients were followed for a minimum
of four years or until their death.
Fifty-two patients had a biopsy before exeresis. In seven
cases, the histological diagnosis was confirmed by the study
of exeresis pieces without preliminary biopsy.
Forty-seven patients (80%) had a primary CS, and 12
(20%) had a secondary CS with a preexistent benign tumor
Grade 2 CSs were most frequent (n=36). This series had
seven dedifferentiated CS.
Fourteen patients (24%) had a zone 1 (iliac) tumor, includ-
ing four with sacral invasion; 42 patients (71%) had a zone
Pelvic chondrosarcomas: Surgical treatment options395
Epidemiological and histological characteristics of
Tumor grade 
2 (acetabular) lesion, and three patients (5%) had zone 3
The average duration of symptoms before the diagnosis was
14.5 months. Pain was the most frequent symptom (n=45).
Eight patients consulted because of palpable tumefaction.
The other patients consulted because of abnormal radiolog-
One patient had a pulmonary metastasis at the time of
Eleven inter-ilioabdominal disarticulations and 48 conser-
vative interventions were performed. In three cases of
dedifferentiated CS, chemotherapy was associated with
Zone 1 CS (n=14)
Eleven patients underwent partial or total resection of the
ilium. In two cases, exeresis included the sacrum. An ampu-
tation was undertaken because of major invasion of the
Zone 2 CS (n=42)
There were five isolated resections of the acetabulum.
Twenty patients had partial or total resection of the
acetabulum associated with partial or total resection of the
pubis or ischium.
Four partial or total resections of the acetabulum were
undertaken in association with all or part of the ilium.
Three patients had resection of the acetabulum, ilium
and ischium and/or the pubis.
Ten inter-ilioabdominal disarticulations were performed.
Zone 3CS (n=3)
All cases were treated by exeresis of all or part of the
ischium and the pubis.
They were considered to be healthy in 46 cases (78%),
marginal in two cases (3%) and contaminated in 11 cases
In the amputation group (n=11), the margins were
healthy in seven cases (63%), and contaminated in four cases
In the conservative treatment group (n=48), the margins
were healthy 39 times (81%), exeresis was marginal twice
(5%), and the margins were contaminated seven times (14%).
After resection of the iliac wing(n=13), reconstruction was
carried out in three cases, by autografts (n=2) or allografts
After resection of all or part of the acetabulum (n=32),
reconstruction was conducted according to the following
• total hip prostheses, n=10,with six Puget interventions
• saddle prostheses, n=8;
• arthrodesis, n=7, with three femoro-iliac arthrodesis,
three femoro-obturator arthrodesis and one femorocoty-
• bone reconstructions of the cotyle, n=4;
• neocotyles at the expense of the iliac wing, n=2.
In one case, no reconstruction was undertaken: resection
involved a non-weight-bearing zone of the posterior cotyle
No reconstructions were carried out after isolated resec-
tion of the pubis (n=3).
Reconstruction according to the Puget technique.
396X. Deloin et al.
Thirty patients (51%) were alive without signs of recurrence
at last follow-up, with a survival average of 123 months (48
to 272 months).
Twenty-three patients (39%) died from the disease. Their
average survival was 69 months (3 to 312 months).
recurrence at the time of their death, within an average of
59 months (6 to 119 months).
Thus, the global survival rate was 90% (CI: 82—98%) at
one year, 66% (CI: 55—80%) at five years, 52% (CI: 39—70%)
at 10 years and 45% (CI: 32—64%) at 15 years (Fig. 2).
Multivariate analysis searched for risk factors of death.
This analysis showed that invasive resection margins
(p=0.00001, RR=7.28) (Fig. 3), high histological grade (3 or
dedifferentiated) (p=0.0005, RR=4.57) (Fig. 4), and tumors
invading the acetabulum (p=0.049, RR=3.24) were statisti-
cally pejorative factors for survival.
Eighteen patients (31%) had a local recurrence within an
average period of 37 months (5 to 230 months), associated
in four cases (22%) with one or more metastases at the time
of diagnosis (Table 2).
Among the 14 patients with a local recurrence without
metastasis at the time of diagnosis, three (21%) had sec-
ondary metastases within an average of 27 months after the
first recurrence (15 to 43 months).
Local recurrence was treated surgically in 12 cases: inter-
ilioabdominal disarticulation from diagnosis in one case,
intralesional exeresis surgery in 11 cases, multiple surgeries
ilioabdominal amputation afterwards. A patient with local
recurrence and synchronous adrenal metastasis received
chemotherapy (dedifferentiated CS).
studied event was patient death in relation to the disease.
Kaplan-Meier analysis of global patient survival. The
patient survival as a function of resection margin quality.
Comparison according to the Cox model of global
A patient with local recurrence and synchronous pul-
monary metastasis was treated by chemotherapy alone.
Five patients received no treatment.
All patients who had a local recurrence died within an
average time period of 38 months (0—240 months) after the
diagnosis of recurrence. Reoperated patients lived an aver-
age of 55 months (1—240 months), non-reoperated patients
lived an average of 3.5 months (0—6 months), but among
them, three had metastases at the time of local recurrence
patient survival as a function of histological tumor grade. Low
grade: grades 1 and 2; high grade: grade 3 and dedifferentiated
Comparison according to the Cox model of global
Pelvic chondrosarcomas: Surgical treatment options397
Characteristics of local recurrences.
Initial CS grade
Time period after initial intervention
Delayed diagnosis — death
Multivariate analysis objectively established that the risk
of local recurrence was strongly correlated with invasion of
the resection margins (p=0.001). The probability of local
recurrence at 10 years was 15% for patients with healthy
resection margins, and was 54% for patients whose margins
were contaminated (Fig. 5).
Characteristics of metastases.
Without local recurrence
With local recurrence
Initial CS grade
Delayed diagnosis — death
On the other hand, high histological grade (3 and ded-
ifferentiated) was not a risk factor for local recurrence
(p=0.3), just like tumor location (p=0.28) (Fig. 6).
Twelve patients (20%) developed metastases, associated
with local recurrence in seven cases (58%), synchronous in
four, or secondary in three (Table 3).
The average time of occurrence was 41 months (3 to
189 months). Metastases were most frequently localized in
the lungs (n=11), associated in two cases with liver metas-
tases. We observed a case of adrenal metastasis during the
evolution of dedifferentiated CS.
function of resection margin quality. a: event probability local recurrence metastases.
Analysis according to the cumulative incidence method of local recurrence risk (Fig. 5a) and metastasis (Fig. 5b) as a