CLINICAL ORTHOPAEDICS AND RELATED RESEARCH

Number 397, pp. 12–18

© 2002 Lippincott Williams & Wilkins, Inc.

PELVIC SURGERY

Combined Posterior Pelvis and Lumbar

Spine Resection for Sarcoma
Bruno Fuchs, MD; Michael J. Yaszemski, MD, PhD;
and Franklin H. Sim, MD

The oncologic outcome in the treatment of pelvic patients of L4 combined with L5 was done to
sarcomas is not comparable with that in the ex- achieve an adequate margin. In four patients,
tremity. Particular problems arise when the tu- the resection was followed by reconstruction
mor involves the posterior pelvis and the sacrum with a vascularized fibula or autograft. At a
or even the lower lumbar spine. Because of the mean followup of 56 months (range, 3–220
difficulty in achieving local control and the ma- months), 12 of 18 patients were alive and with-
jor functional deficits after extensive surgery, out disease. Six patients died; four of these pa-
some authors suggest conservative, nonsurgical tients had metastasis to the lung at presentation.
treatment for these patients. The purpose of the The two remaining patients died of complica-
current study was to analyze the oncologic out- tions of adjuvant treatment. The authors con-
come of patients who were treated for a pelvic clude from this small series of patients that an
sarcoma necessitating resection of the ilium, aggressive surgical resection for localized, non-
part of the sacrum and part of the lower spine, metastatic, high-grade sarcomas of the poste-
which is defined as extended hemipelvectomy. rior pelvis may be justified to provide local con-
Between 1979 and 1999, 11 males and seven fe- trol and improve survival.
males with a mean age of 34.5 years (range,
14–67 years) had an extended hemipelvectomy
for a sarcoma of the posterior pelvis. The tu- The treatment of patients with a sarcoma in the
mors included seven osteogenic sarcomas, six
pelvis represents one of the most challenging
chondrosarcomas, and five fibrosarcomas: 13
were classified as high-grade lesions. The mean
problems for the orthopaedic oncologist. The
size of the tumor was 11
9
6 cm. In 11 pa- patient is confronted with one of the most mu-
tients, a hemivertebrectomy of L5 and in seven tilating surgical procedures, particularly when
a hemipelvectomy is combined with amputa-
tion of its associated lower limb. Very often,
pelvic tumors tend to be large in diameter at
From the Mayo Clinic, Department of Orthopedics, diagnosis,23 involving, or at least are in close
Rochester, MN.
proximity to, the bladder, rectum, and neu-
Reprint requests to F.H. Sim, MD, Chair, Division of Or-
thopedic Oncology, Department of Orthopedics, Mayo rovascular structures. For these reasons, the
Clinic, 200 First Street SW, Rochester, MN 55905. indication for such a procedure requires care-

12
Number 397
April, 2002 Posterior Pelvis and Lumbar Spine Resection 13

ful evaluation and thorough discussion with 14–67 years). There were seven osteogenic sarco-
the patient. mas, six chondrosarcomas, and five fibrosarcomas.
There are only a few reports in the literature In 13 patients, the lesion was defined as high-grade,
dealing with the oncologic outcome of pa- and in five patients, the lesion was defined as low-
grade. The average size of the tumor was 16
9

tients having a hemipelvectomy for sarco-
6 cm. Four patients presented with metastatic le-
mas.5,6,8,11,17,19,25 Analyzing the impact of the sions to the lung, and one patient had a metachro-
anatomic localization on the patient’s out- nous lesion (both in the pelvis). Eight of these 18
come, O’Connor and Sim15 reported on a local patients had an external hemipelvectomy (local re-
recurrence rate of 27% in 23 iliosacral lesions. section in combination with the amputation of the
Furthermore, patients with high-grade pelvic associated limb) and 10 patients had an internal
lesions had a significantly worse prognosis hemipelvectomy (local resection with preservation
than patients with low-grade pelvic lesions.18 of the associated limb). In seven patients, the prox-
The surgical margin is an additional factor imal resection included the hemivertebrae of L4–L5
considered to be of importance and influenc- and in 11 patients, the proximal resection included
ing the prognosis in patients with pelvic sar- the hemivertebra of L5. The margins obtained were
adequate in 17 patients and inadequate in one pa-
comas.11 All these factors have led some au-
tient. In this anatomic site with close proximity to
thors12 to consider nonoperative treatment in neurovascular structures, in most cases, it is not pos-
patients with large, high-grade sarcomas com- sible to obtain a traditional wide margin with a cuff
bined with a significant medial extension of of normal tissue. Therefore, in this location, the
tumor to the sacroiliac joint or sacrum, given standard system to classify margins is difficult to
the poor prognosis and high morbidity in these
patients. Based on the senior author’s (FHS)
earlier experience of a high rate of local recur-
rences mostly attributable to inadequate mar-
gins in such a patient group,15 which is sup-
ported by other reports,11,18 the authors think
that the recurrence rate may be decreased by
more aggressive resection to obtain adequate
margins, justifying a major surgery in selected
patients.
Therefore, the purpose of the current study
was to retrospectively analyze the oncologic
outcome of a selected group of patients with
such adverse prognostic factors, having a pos-
terior pelvic and partial lower lumbar spine re-
section for sarcoma.

MATERIALS AND METHODS

Between December 1979 and July 1999, 18 pa-
tients had an extended hemipelvectomy for a sar-
coma involving the sacroiliac joint.
An extended hemipelvectomy is defined as the re-
section of part of the lumbar spine (at least interver-
tebral disc between L5 and S1); part of the sacrum;
and part of (at least) the ilium, either combined with Fig 1. An anteroposterior radiograph of the pel-
external or internal hemipelvectomy (Fig 1). vis obtained postoperative shows the resection
At the time of the operation, the average age of of the right hemipelvis and the hemivertebrae L4
the 11 males and seven females was 35 years (range, and L5.
 Clinical Orthopaedics
14 Fuchs et al. and Related Research

apply.4 The authors suggest using the definition of cotomy for lung nodules, and a fourth patient
an adequate margin (including wide and marginal had the diagnosis of a metachronous multifocal
resection) and an inadequate margin (including in- osteogenic sarcoma (calcaneus, pelvis, man-
tralesional resection) at this particular anatomic site. dible). He now is 2 years after surgery, cur-
Only in four patients was a reconstruction attempted,
rently having samarium treatment.1
either by a vascularized fibula (in three patients) or
a bone-graft from the anterior iliac crest (in one pa-
Three of 18 patients (15%) had a local recur-
tient). Only one patient with an external extended rence. In one patient, the tumor recurred 4 years
hemipelvectomy in this series had a primary recon- after an inadequate (intralesional) resection of
struction with stabilization. an osteogenic sarcoma and in a second patient,
the tumor recurred 10 years after the resection
RESULTS of a Grade 2 chondrosarcoma. A third patient
had an extended external hemipelvectomy in
At an average followup of 56 months (range, November 1998, 13 months after an intrale-
3–220 months), these 18 patients were re- sional internal hemipelvectomy for osteogenic
viewed retrospectively and data were obtained sarcoma at an outside institution. At the 1-year
from office records, hospital records, and from followup, the patient worked full-time as a high
the orthopaedic oncology files. At the time of school teacher and ambulated with a prosthesis.
the data assessment for this report, 12 patients However, he required resection of a recurrent
were alive and six patients died (Table 1). The mass in the ipsilateral flank (Fig 2).
overall 5-year survival rate was 72% (13 of 18 Wound healing complications were com-
patients). Six patients died after a mean of 20 mon after these extensive resections, often re-
months after surgery. Four of these six pa- quiring multiple debridements and muscle
tients had lung metastasis diagnosed at the ini- flaps. In addition, there was a lower quadrant
tial presentation. Moreover, one patient died hernia, a compartment syndrome of the hand,
of a cardiomyopathy subsequent to extensive and a resection of an ischemic terminal ileum.
chemotherapy and the other patient died after Comparing the oncologic outcome of pa-
severe ischemia of the terminal ileum caused tients having an internal extended hemipelvec-
by radiotherapy. The mean duration until death tomy with patients having an external extended
(20 months) is in contrast to the followup of hemipelvectomy, no major difference was de-
those patients who still were alive at the time tected between these two patient groups ex-
of this assessment (73 months). However, four cept for the longer followup of the patients
of the 12 patients who currently are considered with the limb salvage procedure. Both patients
disease-free are at high-risk for having pro- who died of unrelated causes had an external
gression of the disease. One patient is 6 months hemipelvectomy (Table 2).
after the resection of his third local recurrence.
A second patient had a craniotomy for a brain DISCUSSION
metastasis of a histologically proven osteo-
genic sarcoma 7 months at the time of data col- Patients who present with a primary pelvic
lection. A third patient is 3 years after a thora- sarcoma present a complex treatment problem

TABLE 1. Patient Status at Followup

Patient Status Number of Patients Followup (months)

Dead with disease 4 21 (range, 6–62)

Dead without disease 2 18 (3 and 34)
Alive with disease 0 —
Alive without disease 12 73 (range, 3–220)
Number 397
April, 2002 Posterior Pelvis and Lumbar Spine Resection 15

dictor of survival in this patient group. The pa-

tients analyzed in the current report represent
a selected group with the adverse prognostic
factors in view of the anatomically difficult lo-
calization, and the majority of tumors being
high-grade. The current study shows that an
adequate surgical margin can be obtained even
when the tumor is localized in the posterior
pelvis, and extends into the sacrum or along
the lumbar spine. Obtaining an adequate re-
section margin has improved the outcome
with an overall survival of 72%. However, one
Fig 2. A computed tomography scan of a 33- has to take into account that four patients had
year-old patient shows a paravertebral recur- treatment for progression of disease and cur-
rence in the flank which developed proximal to rently are considered disease-free. When con-
the prior resected tumor bed 1 year after the pa-
tient had an extended hemipelvectomy. sidering the failure rate in the current series, it
is noteworthy that four of six patients who
died initially presented with lung metastasis.
with an increased risk for failure, and compro- This suggests that this extensive surgery may
mised survival. Although refined techniques be best reserved for patients with localized
for pelvic surgery have evolved during the disease. Careful planning and a multidiscipli-
past decades,2,7,9,15,20,22 allowing partial pelvic nary surgical approach may provide an ade-
resection with limb sparing surgery, the indi- quate margin at this anatomic site, which has
cation for resection of a tumor localized to the been a significant factor for the overall sur-
sacroiliac joint or sacrum still is controver- vival.11 Local recurrence at this location re-
sial.12 Shin et al18 reported that patients with mains problematic, with three local failures in
high-grade sarcomas in the pelvis continue to this series. This may be explained by an inad-
have a bad prognosis. Fifteen of the 17 pa- equate resection in two patients, but one tumor
tients (88%) survived with low-grade tumors recurred despite obtaining adequate margins.
compared with six of 24 patients (25%) with Because the tumor in this particular patient re-
high-grade tumors. Kawai et al,11 analyzing curred proximally to the prior tumor bed, it is
102 patients with localized pelvic sarcomas assumed that this represents a tumor cell seed-
who had surgical excision, reported that an in- ing that occurred during the prior surgeries.
adequate surgical margin was a negative pre- Therefore, doing an extended hemipelvec-

TABLE 2. Comparison of Patients Having Internal Versus

External Extended Hemipelvectomy
Hemipelvectomy External Internal

Number 8 10
Dead (entire series) 3 3
Alive (entire series) 5 7
Alive without disease 4 7
Grade of tumor
High-grade (Grades III and IV) 6 7
Low-grade (Grades I and II) 2 3
Size of tumor (cm) 12/8/7 11/9/6
Mean followup (months) 35 67
 Clinical Orthopaedics
16 Fuchs et al. and Related Research

tomy in a patient who had prior surgeries at the low threshold to place a myocutaneous rectus
same site may represent an increased risk. flap16,22 or even a free muscle transfer from the
There still is an ongoing debate whether to lower leg in the case of an external hemipelvec-
do an external or internal hemipelvectomy for tomy24 to avoid these complications.
pelvic tumors. Kawai et al11 found an external Achieving a good oncologic outcome in
hemipelvectomy to be a negative predictor of limb salvage procedures of the pelvis remains
survival. However, the patients in their study challenging. Moreover, the issue of recon-
who had an external hemipelvectomy also struction after resection at this anatomic site
tended to have the largest and most advanced remains unclear. Skeletal reconstruction is not
tumors. Therefore, they recommend an external necessary after resection of the anterior pelvis
hemipelvectomy particularly when there is because pelvic stability is maintained. Re-
sacral involvement. The results of the current construction after periacetabular resection is
study do not additionally assist in making the extremely challenging with the options of il-
decision of when to do an external or internal iofemoral arthrodeses or pseudarthrosis, mas-
hemipelvectomy. Comparing these two patient sive allograft or autoclaved autograft with hip
groups (Table 2), no difference was detected arthroplasty, and pelvic or saddle prosthesis.
between survival, disease status, margins ob- For patients with resection of the ilium and
tained during surgery, size and grade of the tu- loss of pelvis stability, iliosacral arthrodeses
mor, and recurrence rate. Interestingly, patients provide good function.14 However, particular
having an internal extended hemipelvectomy problems arise regarding reconstruction in pa-
had the longer followup, which cannot be ex- tients with extended hemipelvectomy as pre-
plained by the aforementioned parameters. The sented in the current study. Although some au-
current authors think that patients with these tu- thors are not enthusiastic to reconstruct defects
mors should be treated by a multidisciplinary after sacroiliac resections because the func-
approach with involvement of colorectal, uro- tional results are not predictable,13 others de-
logic, neurosurgical, and plastic surgical teams. scribe using vascularized and nonvascularized
If attempts at limb salvage would sacrifice fibular iliosacral arthrodeses.3,21 In the current
neural structures producing limb anesthesia, series, four patients having an internal ex-
paralysis, or excessive shortening, or if the pre- tended hemipelvectomy had reconstruction
operative investigations showed that adequate with a fibular graft. One patient had a second
margins could not be envisaged, then external postirradiation osteosarcoma develop in the
hemipelvectomy is indicated. A vascular re- opposite left sacroiliac region 2.5 years after
construction after tumor resection in the pelvis sacroiliac joint resection on the right side. Af-
was not necessary in this series. ter the extended internal hemipelvectomy on
In the current series, there was a high number the left side, the patient had reconstruction
of revision surgeries because of wound compli- with a fibular graft. Although this reconstruc-
cations. The greatest risk factors for having a tion failed with graft fractures, the patient was
wound problem develop is the inability to pre- able to walk with crutches and with an ankle-
serve the internal iliac vessels and the necessity foot-orthosis (Fig 3). The issue of spinal insta-
to do a sacral osteotomy.12 Even though some bility after extended hemipelvectomy must be
think10 that preservation of the internal iliac ves- considered. Reconstructive options include
sels has no influence on posterior flap necrosis, fixation with pedicle screws to the remaining
it has to be considered that a sacral osteotomy sacrum and ilium combined with rods.21 Al-
decreases the blood supply to a posterior flap by though only one patient in the current series
detaching the sacral attachment of the gluteus having an external extended hemipelvectomy
maximus and increases the potential dead space had primary reconstruction, this may be nec-
under the flap, compromising wound healing. At essary as a later secondary procedure. Progres-
the authors’ institution, there currently is a very sive scoliosis could be documented as shown
Number 397
April, 2002 Posterior Pelvis and Lumbar Spine Resection 17

A B

Fig 3A–C. (A) A 22-year-old patient had a right in-

ternal hemipelvectomy for a radiation-induced os-
teosarcoma, 2.5 years before having a second os-
teosarcoma develop at the left sacroiliac joint. (B)
He had a left extended internal hemipelvectomy
and reconstruction with a vascularized fibula. (C)
Twenty months after the last surgery, the recon-
struction failed, but the patient was able to walk
C
with two crutches and an ankle-foot orthosis.

in Figure 4. Two years after the extended ex-

ternal hemipelvectomy, this patient had a 38
scoliosis develop, and complained of mechan-
ical pain with increased activity. This patient
may require reconstruction in the future. The
question remains as to what is the best type of
reconstruction in this setting. Based on the re-
sults of the current study, these issues may
need to be confronted as the survival in this
patient group increases. Additional investiga-
tion into development of biomechanically sound
stabilization procedures is warranted.
It is concluded from the current study that
the resection of high-grade sarcomas located
in the posterior pelvis is justified because local
control can be achieved in combination with a
multimodality treatment. However, this type
of surgery is not indicated in patients with
metastatic disease at onset or with other se- Fig 4. A radiograph of a 14-year-old patient
verely compromising medical factors at initial shows a progressive scoliosis which developed 2
presentation. years after an external extended hemipelvectomy.
 Clinical Orthopaedics
18 Fuchs et al. and Related Research

Acknowledgment Hindquarter amputation for pelvic tumors. Clin Or-

thop 350:187–194, 1998.
Dr. Bruno Fuchs was supported by a scholarship from 13. Missenard G, Dubousset J, Genin J: Résection large
the Schweizerische Gesellschaft für Orthopädie. de la sacro-iliac: Technique, reconstruction, résultats
anatomiques et fonctionnels. Rev Chir Orthop Repara-
trice Appar Mot 77:14–24, 1991.
References 14. O’Connor MI: Malignant pelvic tumors: Limb-
1. Aas M, Moe L, Gamlem H, et al: Internal ra- sparing resection and reconstruction. Semin Surg
dionuclide therapy of primary osteosarcoma in Oncol 13:49–54, 1997.
dogs, using 153Sm-ethylene-diamino-tetramethylene- 15. O’Connor MI, Sim FH: Salvage of the limb in the
phosphonate (EDTMP). Clin Cancer Res 5S (10 treatment of malignant pelvic tumors. J Bone Joint
Suppl): 3148S–3152S, 1999. Surg 71A:481–494, 1989.
2. Apffelstaedt JP, Driscoll DL, Spellmann JE: Com- 16. Ross DA, Lohmann RF, Kroll S, et al: Soft tissue re-
plications and outcome of external hemipelvectomy construction following hemipelvectomy. Am J Surg
in the management of pelvic tumors. Ann Surg On- 176:25–29, 1998.
col 3:304–309, 1996. 17. Sheth DS, Yasko AW, Johnson ME, et al: Chon-
3. Cannon SR, Tillman RM, Grimer RJ, et al: Treat- drosarcoma of the pelvis. Cancer 78:745–750, 1996.
ment of Primary Bone Tumors of the Ilium by 18. Shin KH, Rougraff BT, Simon MA: Oncologic out-
Local Resection and Fibular Strut Graft (Non- comes of primary bone sarcomas of the pelvis. Clin
Vascularized). In Campanacci M, Capanna R (eds). Orthop 304:207–217, 1994.
Proceedings of the Eighth International Symposium 19. Spanier SS, Shuster JJ, VanderGriend RA: The ef-
on Limb Salvage. Florence, Italy 77, 1995. fect of local extent of the tumor on prognosis in os-
4. Enneking WF: Musculoskeletal Tumor Surgery. teosarcoma. J Bone Joint Surg 72A:643–653,
New York, Churchill Livingstone 515–518, 1983. 1990.
5. Fahey M, Spanier SS, Vander Griend RA: Osteosar- 20. Stephenson RB, Kaufer HP, Hankin FM: Partial pelvic
coma of the pelvis. J Bone Joint Surg 74A:321–330, resection as an alternative to hindquarter amputation
1992. for skeletal neoplasms. Clin Orthop 242:201–211,
6. Frassica FJ, Frassica DA, Pritchard DJ, et al: Ewing sar- 1989.
coma of the pelvis. J Bone Joint Surg 75A:1457–1465, 21. Takeda N, Kaneda K, Abumi K, et al: Reconstruc-
1993. tion With Vascularized Fibula Graft and Pedicle
7. Gradinger R, Rechl H, Hippe E: Pelvic osteosar- Screw Spinal Fixation for Malignant Pelvic Tumors.
coma. Clin Orthop 270:149–158, 1991. In Campanacci M, Capanna R (eds). Proceedings of
8. Grimer RJ, Carter SR, Tillman RM, et al: Osteosar- the Eighth International Symposium on Limb Sal-
coma of the pelvis. J Bone Joint Surg 81B:796–802, vage. Florence, Italy 76, 1995.
1999. 22. Weber KL, Nelson H, Gunderson LL, et al: Sacro-
9. Harrington KD: The use of hemipelvic allografts or pelvic resection for recurrent anorectal cancer. Clin
autoclaved grafts for reconstruction after wide resec- Orthop 372:231–240, 2000.
tions of malignant tumors of the pelvis. J Bone Joint 23. Wurtz D, Peabody TD, Simon MA: Delay in the di-
Surg 74A:331–341, 1992. agnosis and treatment of primary bone sarcoma of
10. Karakousis CP, Enrich L, Driscoll D: Variants of the pelvis. J Bone Joint Surg 81A:317–325, 1999.
hemipelvectomy and their complications. Am J Surg 24. Yamamoto Y, Minakawa H, Takeda N: Pelvic re-
158:404–408, 1989. construction with a free fillet lower leg flap. Plast Re-
11. Kawai A, Healey JH, Boland PJ, et al: Prognostic constr Surg 99:1439–1441, 1997.
factors for patients with sarcomas of the pelvic 25. Yang RS, Eckardt JJ, Eilber FR, et al: Surgical indi-
bones. Cancer 82:851–859, 1998. cations for Ewing’s sarcoma of the pelvis. Cancer
12. Masterson EL, Davis AM, Wunder JS, Bell RS: 76:1388–1397, 1995.