External and internal hemipelvectomy for sarcomas of the pelvic girdle: consequences of limb-salvage treatment

European Journal of Surgical Oncology 1997; 23:540-546

External and internal hemipelvectomy for sarcomas of the pelvic girdle: consequences of limb-salvage treatment S. John Ham*,1", Heimen Schraffordt Koops*, Ren~ P. H. Vetht, Jim R. van Hornt, Willem H. Eisma:l: and Harald J. Hoekstra* Departments of*Surgical Oncology, t Orthopedic Surgery and ~Rehabilitation, Groningen University Hospital, Groningen, the Netherlands

The outcome of different limb-saving treatment modalities for pelvic girdle sarcoma is controversial. The oncological and functional results after 11 external and 10 internal hemipelvectomies and the consequences of limb-salvage treatment v~ere studied in 21 consecutive patients with primary bone (19 patients) or soft tissue sarcoma (two patients) of the pelvic girdle. Following external hemipelvectomy, 10 patients (91%) died after a median follow-up of 1.6 years (range: 0.3-7.1). Isolated local recurrences occurred in three patients (27%), with concomitant distant failure in one (9%), while isolated distant failure occurred in six patients (55%). The rate of flap necrosis and wound infection following external hemipelvectomy were both 25%. Following internal hemipelvectomy, nine patients (90%) were alive without evidence of disease after a median follow-up of 6.6 years (range: 2.3-16.0). Concomitant local and distant failures were found in one patient (10%). Reconstruction-related complications necessitated revisional procedures in five of seven patients (72%), leading to external hemipelvectomy in one. Patients with a locally advanced pelvic girdle sarcoma who are unable to undergo an internal hemipelvectomy have a worse prognosis than patients who undergo an internal hemipelvectomy. An internal hemipelvectomy is not attended by an increased risk of local failure, but is by long-term local complications, requiring extensive surgical procedures.

Key words: pelvic sarcoma; external hemipelvectomy; partial internal hemipelvectomy; limb salvage; saddle prosthesis; autoclaved autogeneic allograft.

Introduction Disarticulations at the level of the pelvic girdle for soft tissue and bone sarcomas can be regarded as one of the most mutilating surgical procedures in cancer treatment. Due to the introduction of multimodality therapy consisting of surgery, chemo- and/or radiation therapy the number of patients indicated for major amputations has declined from 40% in the past to 5-10%) 3 A curative hemipelvectomy is now usually performed for carefully selected patients with extensive growth of soft tissue or bone sarcomas of the upper thigh, hip or pelvis, and for tumours not amenable to hip disarticulation or limb-saving procedures. 4-7 Palliative hemipelvectomy is only occasionally performed for metastatic disease.4'8-~° The introduction of the internal hemipelvectomy, with endoprosthetic or allograft replacement when indicated, allowed the functional preservation of a lower limb following entire or partial removal of the hemipelvis.9.H32 In this study the 25-year experience of bone and sarcoma

Correspondence to: H. J. Hoekstra, Division of Surgical Ontology, Department of Surgery, Groningen University Hospital, P.O. Box 30.001, 9700 RB Groningen, the Netherlands. E-mail: h.j .hoekst [email protected]. 0748-7983/971060540+ 07 $12.00/0

surgery of pelvic girdle malignancies treated by external or internal hemipelvectomy at the Groningen University Hospital is analysed. In particular the progress in extremity sarcoma surgery during the past two decades, and the consequences for the future treatment of these kind of malignancies are discussed.

Materials and methods During the period 1970-95, 21 consecutive patients, 14 males (67%), seven females (33%), median age 43 years (range: 15-70), were treated with curative intent for sarcoma of the pelvic girdle. There were 19 bone sarcomas (90%), and two soft tissue sarcomas (10%). Information about the primary tumour was obtained from plain radiographs of pelvis, hip and chest, bone scintigraphy, angiography, computerized tomography (CT) and/or magnetic resonance imaging (MRI). CT of the lungs and bone scintigraphy were performed to rule out distant metastases. The definitive histological diagnosis was known at the time of treatment. Neoadjuvant chemotherapy according to different protocols was received by three osteosarcoma patients, one Ewing sarcoma patient, and one patient with a malignant © 1997W.B. SaundersCompany Limited

54 !

Hemipeh,ecton O, for sarcomas o f the peh,ic gh'dle

fibrous histiocytoma (MFH) of bone. Another patient (Pt. 18) with an osteosarcoma of the proximal femur and periacetabular region had an isolated metastatic lesion in the lung at initial presentation, and was first treated with neoadjuvant chemotherapy followed by metastasectomy; histological examination after resection revealed no viable tumour, and this patient was therefore treated with curative intent. Patient characteristics like tumour histology, grading, site, therapy, complications and survival are listed in Table I. The type of surgery (external or internal hemipelvectomy) was generally based on tumour size and/or local involvement of adjacent tissues and structures. External hemipelvectomy was performed in 11 patients (52%), median age 49 years (range: 15-70), in seven patients as primary treatment, in two patients (Pts 5, 11) following local recurrence, and in two patients (Pts 3, 10) following treatment of a pathological fracture. In one patient (Pt. 12) external hemipelvectomy was performed due to failure of limb-salvage procedures. During surgery, the patient was ideally placed in the lateral position. ~'~3'~ The standard procedure for external hemipelvectomy involved disarticulation through the symphysis pubis and sacroiliac joint. The hemipelvectomy procedure was qualified as radical when the bone was divided proximally through the sacroiliac joint or sacral ala, or conservative when a portion of the iliac bone was preserved. H°'35 The advantage of the conservative hemipelvectomy is that it may be used to help stabilize a prosthesis post-operatively, u°'33'~'36'~7 Posterior flap hemipelvectomy was performed in 10 patients (83%), and anterior flap hemipelvectomy in two patients (17%) (Pts 9, I I ). Surgical margins were classified according to the criteria reported by Enneking et al. 3~ Wide resection after hemipelvectomy performed for resection of the primary tumour was achieved in six of the 11 patients (55%), marginal resection in three patients (27%) (Pts I, 8, 9), and intralesional resection in two patients (18%) (Pts 5, 6). Internal hemipelvectomy signified the removal of the entire hemipelvis or a portion of it, with preservation of the lower extremity, j'4'~°'39 In this series, partial internal hemipelvectomy was performed in 10 patients (48%), median age 29 years (range: 16--43). Three major types of resection were distinguishable according to Enneking. H'4° Wide resection after partial internal hemipelvectomy was achieved in seven patients (70%), marginal resection in one patient (10%) (Pt. 17), and intralesional resection in two patients (20%) (Pts 14, 15). Following partial internal hemipelvectomy, endoprosthetic reconstruction was performed in five patients (50%). A saddle prosthesis (Type MKII, Waldemar-Link, Hamburg, Germany) was used as primary endoprosthetic reconstruction in three patients, and after failure of other endoprosthetic reconstructions in two. Autogeneous replacement of the resected segment after autoclavation was performed in two patients (20%). In this procedure, the part of the pelvic bone to be replaced was autoclaved? 5 After this, the patient's hemipelvic segment was restored to its original bed. Post-operatively, partial weight-bearing was delayed for 2 to 3 months, and unrestricted full weight-bearing was not allowed until the pelvic osteotomy site showed evidence of bone union. ~5 In

three patients (30%) no reconstructive procedures were performed following tumour resection. Patients were regularly followed for functional evaluation, long-term complications, and the detection of local recurrence and/or distant failures by radiographs of the chest and bone scans. Functional evaluation for patients treated with partial internal hemipelvectomy was performed according to the most recent system of the Musculoskeletal Tissue Society (MSTS)fl Overall and disease-free survival rates following external and internal hemipelvectomy were calculated according to the method of Kaplan and Meier. 42 Differences between these two groups were calculated using the log-rank test. Calculations were performed using the Graphpad Prism software for Windows 95. 43

Results

After external hemipelvectomy (Pts I-I1) 10 patients (91%) died of their disease after a median follow-up of 1.6 years (range: 0.3-7.1). One patient (Pt. 11) is still alive without disease 3 years after initial diagnosis of a grade I chondrosarcoma, and I.I years after undergoing hemipelvectomy. In four of the 11 patients (36%) local recurrence developed following external hemipelvectomy after a median follow-up of I1 months (range: 2-24); in two patients of six following wide resection, and in two of three following marginal resection. Distant metastases occurred in seven patients (64%) after a median follow-up of 8 months (range: 5-37). In one patient (Pt. 12) conservative external hemipelvectomy was performed 14.3 years after internal hemipelvectomy and endoprosthetic reconstruction for grade IIA chondrosarcoma, due to chronic infection. In total, therefore, 12 external hemipelvectomy procedures were performed in this series. Flap necrosis occurred in three of 12 patients (25%) following external hemipelvectomy; in two of these three patients (Pts I 1, 12) endoprosthetic reconstruction had been performed earlier, and in one patient (Pt. 1) no gluteus maximus muscle was preserved attached to the posterior flap. A wound infection was documented in three patients (25%), and a urinary tract infection occurred in three patients (25%). Considerable phantom limb pain requiring prolonged treatment was documented in five patients (42%). After internal hemipelvectomy for primary sarcoma nine patients (90%) were alive without recurrent disease after a median follow-up of 6.6 years (range: 2.3-16.0). Local recurrence and distant spread to the lungs was found in only one patient (10%) 9 months after initial diagnosis. Fracture of the saddle prosthesis necessitated total endoprosthetic revision in one patient (Pt. 16) 5.4 years after the initial reconstruction. The saddle prosthesis is now in place without complications in another patient (Pt. 18) 6.3 years after implantation. One patient (Pt. 19) died due to disseminated disease 1 year after implantation of a saddle prosthesis without endoprosthesis-related complications. In one patient (Pt. 12) internal hemipelvectomy and resection of the caput femoris was followed by implantation of a pelvic endoprosthesis and a cemented total hip

S.J. Ham

542

et al.

o =I.LI

"" 0

0

0

0

0

0

0

0

0

0

"~

""'

"-

"-

"~'

~"

"--

6_-

~ Z"

Z~Z~'

Z"

Z ''

Z"

0

Z~

c~ ~o

~

11

~

~

~

~

~,

z

~

~

~

z

z.

z.

.

.

.

.z

.z

z

z

o

z

z

~,

o

~

z

8 --s o

z~

o

o

o

o

z

z

o

o

o

o

o

o

Z

Z

Z

Z

Z

Z

o

=

.-=_ ~=~

.~.~,

.~ . . . .

uo~

--',.,i>~

."

o.--

~

.~

~

~-~

o

.~. . o. . o.o

~

o='==~

~, , = ~ ~ .... ~

.=

~

-.~ E o ~ ,=

6

~=

r~=. o=o =

~.

~

._--,

~

~

=

. . . .

==

z=.

=;

.=-

~.

--'N

6.

,-r

=

=~.: :

=0,, = . . . . .

~

==

~-iZZ

' 7 . . ~u

~ ..u~

-- ~;

,~

~

"~

o,

o

.~

. . . .

•. ._~

~=.~._o . . .. o.~ . o , , ~E.~

o

o

Z

Z

-- ~=

~ ==

"~ ~

.~

~ ~

~ o

o o ,,~.o o . . . .

_.-

m.

.. . . .

8u

. ' , . ; ' E~ o=~. = ~r o

r,~

e.~..~.=

=~

o ~"

~ Z

o 7-

g

~ o o

zz~

~

z

z

z

o

o

.=

.~

'0

U

o ~Z

Z

Z

Z

Z

>,

o

o

o

o

~

"~ ~. P9. "~

~

r:o o'~

.~ . _ - , o>._ ,., ~o : ,.,

,., > - o - F.

o =

-.=.

"~ >,

~._~~_~ ,,~

-o=

- ~ = ~

:-3

~

~'~ ~

°~

.==

e~

M Z

._~

~:

~

=-

= '~

~

= -~,

.-

~o uo~,..,.,:= :

.~ ~

~ o"~ o-8

~ -Su

z

0 ~.~

,~,Z

~

~

~=

=

~z

:~

:'5

"~-=

oo

zz

•i=

~

Z

o

~

=

.~

Z

mZ

o

z

E

z

o

-o

,=

~

.

=

_~.-

~

~

,~

~

Eu~<

o

o

o

u

o

z

z

m

:=

~

.~

o ~

o Z

o Z

o z

o o Z ~

o z

~

4

f

•-

~=

;:

o -o ,.: "o

o

,= o

.o~

,= ,= =' " = = ~

o =.~_

~

u o'0.-~

o o ==

o--~

=,

o"0~.

o

,=~[~.-,-.

K

o ~ . ~ _ _ . ~.

o z

~

o Z

o o Z Z

o z

~ cJ

~ ~

.-

~,.~,

~

E- ~

"8 :5 0,.,.,

o z

o

= ~

-

.~.~o

E

_.q

k..

~: ._" ~

o-o

~._o ~ . ~ ~_.=

o

~.o ._ ,0

~ ._- 6

~.~

=~E

~

EE

~

~

E

E

E

E

i m

Hemipeh,ectomy for sarcomas of the pelvic girdle prosthesis, 25 and in another patient (Pt. 17) internal hemipelvectomy was followed by transposition of the femur, implantation of a snap-fit acetabular prosthesis and a cemented femoral prosthesis) 4 In both patients endoprosthetic revision using a saddle prosthesis was performed due to femoral loosening and chronic infection, respectively. However, in both patients the saddle prosthesis had to be removed 2 years and 1.6 years later, respectively, because of chronic infection. External hemipelvectomy was performed in one patient, and implantation of a new saddle prosthesis in the other patient. After autocl~:ved autogeneic allograft replacement following partial pelvic resection for grade il chondrosarcoma (Pts 20, 21) loosening of the femoral prosthesis occurred in one patient, and protrusion of the acetabulum in the other, necessitating revisional procedures in both. No local recurrence was found 4.3 years after 4.2 years after the surgical procedure, respectively. However, in both patients weight-bearing continued to be partial because of pseudarthrosis at the distal junction between the graft and the ischiopubic bone in one patient (Pt. 20) and pain experienced at full weight-bearing in the other (Pt. 21). The mean functional evaluation scere was 17.6 points (range: 14-27) for the five patients alive after partial internal hemipelvectomy followed by a reconstructive procedure. For the three patients alive after partial internal hemipelvectomy without reconstruction, the functional evaluation score ranged from 23 to 30 points (Table I). Although the two patient groups of external and internal hemipelvectomy are not comparable with respect to tumour type or stage, the overall and disease-free survival rates for the group of patients undergoing internal hemipelvectomy were significantly better than for the-group of external hemipelvectomy patients, P<.0002 and P<.0001, respectively (Fig. I).

Discussion

Bone or soft tissue sarcomas of the pelvic region are often large tumours at initial presentation, which may involve important structures, and are frequently not resectable for cure. In the past the only curative radical surgical option for these tumours was hemipelvectomyY '~ Improvement of surgical techniques and development of implants have made limb-salvage procedures feasible with partial pelvic resections and reconstructions in selected cases, especially for benign tumours and low-grade malignancies. ".39'45'46 However, external and internal hemipelvectomies for highgrade sarcomas have an increased risk for local recurrence, as the surgical margins are minimal. Tumours arising near the sacroiliac region are often non-resectable for cure, having minimal resection margins, and are therefore at high risk of local failure) 2-~4'47-49In selected cas~s positive margins of resection may occur unless the procedure is extended to the sacral ala or to the contralateral side of the symphysis, respectively, lntraoperative radiotherapy may be a valuable adjunct for local control in the treatment of pelvic girdle sarcomas. 5o.s~ External rather than internal hemipelvectomy is indicated following tumour spill during biopsy,47incomplete resection,

543

(a)

lOOI 80 60 "~ 40 CO

20

0

2

4

6

(b)

8 I0 12 14 Follow-up (years)

16

18

20

100 ,1 .L,Ik

.IL.I.

8O

~ 60 ,o

0

2

I

I

4

6

I

t

I

I

8 10 12 14 Follow-up (years)

I

I

16

18

20

Fig. 1. (a) Overall survival rates and (b) disease-free survival rates ( I , series 1: external hemipelvectomy; A, series 2: internal hemipelvectomy).

local haemorrhage and tumour-contaminated haematoma because of the high risk of local recurrence. "'45 Infiltration of the iliac or femoral vessels, and the necessity to resect one of the two major nerves to the leg has been regarded as a major indication for external hemipelvectomy by some authors, s't°'52whereas others occasionally replaced the vessels or accepted resection of one nerve." Furthermore, external hemipelvectomy might be the treatment of choice after the failure of limb-salvage procedures, because of local recurrence or infection, as was the case in two of our patients. Stump coverage after conversion of failed limbsalvage procedure to hemipelvectomy may be jeopardized in these cases, when an adequate myocutaneous flap cannot be created because of tumour spill, areas of infection or previous incision. As reported before, the basic condition for the viability of a posterior flap is the preservation of a portion of the gluteus maximus muscle attached to the flap, which will prevent ischaemie necrosis in most patients, w°'33'36'53 in contrast to the rate of flap necrosis after removal of the gluteus maximus muscle, which was reported to be 26-80%. 4" Even after division of the common iliac vessels instead of the external iliac vessels, blood supply to this posterior flap is adequate, because ofc'ross-circulation from the opposite internal iliac artery and from lumbar branches? '4'53 After anterior flap hemipelvectomy,

544

S. ,I. H a m et al.

preservation of the quadriceps muscle provides a cushion of viable tissue on the sacrum on which a prosthesis may comfortably rest without traumatizing the overlying skin. 36'37The superficial femoral artery and preferably the femoral nerve have to be preserved to sustain the myocutaneous flap. I'='t0'36'37 However, necrosis after anterior flap hemipelvectomy has been reported to be as high as 20-37%. 4'~° If the construction of a posterior or an anterior flap of adequate length is not possible, a Marlex mesh can be used to approximate the fascia lata to the abdominal fascia, followed by skin grafting 4 weeks later when there is granulation of tissue through the mesh? Following external hemipelvectomy complication rates of 53-60% have been described.4'u Wound infection was found in 25% of external hemipelvectomy patients in this series. The overall rate of wound infection after hemipelvectomy reported in the literature is 11-35%, usually in the form of cellulitis, treatable with antibiotics.4't°'3"~'uWound infection may predispose the patient to the development of posthemipelvectomy hernia. 55 Genito-urinary complications were common after hemipelvectomy. In the series of Prewitt et al. ~ the majority of complications were of genito-urinary origin (42%); 21% of males required trans-urethral prostatic resection to manage urine retention, which was probably due to denervation of one side of the bladder, leading to decreased detrusor contraction. Phantom limb pain is experienced by virtually all patients after external hemipelvectomy in the early post-operative period. Besides analgesics, transcutaneous nerve stimulation or anticonvulsive drugs work well in some patients. 36'~7 The estimated 5-year survival after external hemipelvectomy performed with curative intent in two large series for a heterogeneous patient population was 21% and 43%, respectively.4'~° Local recurrence was found in 37%, which is consistent with our findings.4 Prewitt et al. ~ reported no local recurrence in 53 patients with high-grade soft tissue sarcoma treated with external hemipelvectomy with curative intent. Distant metastases occurred in 70% of their patients. Five-year actuarial disease-free and overall survival rates were 30% and 38%, respectively. In this series, only one of the I1 patients (9%) treated with external hemipelvectomy performed for primary pelvic sarcoma was alive !.1 years after hemipelvectomy. After partial internal hemipelvectomy for resection of sarcoma, nine patients (90%) were alive after a median follow-up of 6.6 years. The incidence of local recurrence was 36% for external hemipelvectomy and 10% for internal hemipelvectomy. The differences in survival and local recurrences may be explained by the size and extent of the primary tumour. Huth et al. 5" also reported higher local recurrence rates and lower overall survival rates for external hemipelvectomy patients than in those undergoing internal hemipelvectomy)2 However, results between these two procedures are difficult to compare as the extent of the tumour often dictates the operative procedure. Motion in the hip area is minimal following internal hemipelvectomy without reconstruction, unless the hip joint is retained. ~'45'52 Leg-length discrepancy occurs after this procedure and may vary from 3 to 13cm. 45"52 When a reconstruction is performed, the type depends on the type and extent of resection. According to Enneking, three major

Table 2. Three major types of resection of the pelvis according to

Enneking 7),pc I: Iliac resection or Type IA: including the gluteal muscles Type II: Periaeetabular resection or Type IIA: periaeetabular

including the hip joint Type 111: lsehiopubic resection (Type ll~ Lateral mass of os sacrum)

types of resection can be distinguished in the pelvis (Table 2). ".40 In general, Type I resection causes only modest and temporary disability and endoprosthetic or allograft replacement is usually unnecessary. ~°'~'47 After type IA resection significant abduction instability of the hip may cccur due to loss of the adductor mechanism, which can be reduced by iliopsoas transfer." After Type I1 resection, reconstruction is unnecessary when only one third of the acetabulum has to be resected. After resection of the entire acetabulum, mostly including the femoral part of the hip joint (type IIA), or after combined type II(A) and III resection, several techniques have been developed for reconstruction of the hip and hemipelvis. After Type Ill resection function is normal, except for weakness of the hip adductors, and no reconstructive procedure is needed." Reconstructive techniques include the use of endoprosthetic replacements,9..,~ .,6..,~..,9.3~.3_,.s6-ss pelvic allografts,.,o ,_,5,- osteoarticular allografts59 and replantation of the resected hemipelvis after it has been autoclaved, t5'23 Other procedures include the ischiofemoral or ileofemoral (pseud)arthrodesis, 16.17.19.22.47.52.6oa partial transposition of the ipsilateral femur with fusion to the sacrum and the pubic/ ischial bone after a combined Type 1 and II(A) resection; ~4 and a sacrofemoral fusion using a (vascularized) fibular graft after a combined type I (IA), I1 (IIA) and 111 resection. ~4 An endoprosthesis assures a stable reconstruction without limb-length differences, allows early weight-bearing without pain and offers an acceptable functional replacement of the hip joint. 9When an endoprosthesis is used for reconstruction, a pelvic prosthesis including the hip joint can be used. 25"3t'6~ Complication rates after these procedures are high, and are mainly caused by dislocation, loosening and infection, t6 An alternative procedure is to use a saddle prosthesis, in which fewer complications were expected. 9'~4'29 However, Windhager et al. 3~ recently reported functional results of the saddle prosthesis after resection of periacetabular sarcomas to be inferior to those of custommade pelvic prosthesis? ~ Furthermore, the complication rate in their series was higher after reconstruction using a saddle prosthesis than a custom-made prosthesis? ~ Aboulafia et al? ° also reported a high complication rate (53%) in patients with a saddle prosthesis following tumour resection, with a mean follow-up of 33 months (range: 15-62). 30 Endoprosthesis-related complications in their series were dissociation (12%) and dislocation (12%) of the prosthesis. Reconstruction using an allograft replacement after periacetabular resection with or without a hip prosthesis may result in excellent function by restoring hip motion and preserving hip length. However, high rates of infection and failures have been reported for this procedure, t6'62Total hip replacement with minimal reaming of the allografted

545

Hemipelvectomy for sarcomas o f the pelvic girdle

acetabulum or a bipolar prosthesis may be used to replace the hip joint. According to Harrington) 5 there is no evidence for gradual eroding of the bipolar prosthesis through the avascular part of the allograft. This author found the most bothersome complication in his series to be the incidence of late stress fractures in 21% of patients 5 to 8 years after periacetabular resection, including the hip joint, with either a massive allograft reconstruction or replacement of the resected bone after it had been autoclaved. No evidence of different behaviour of autoclaved massive grafts or allografts of the same size was found. However, it has also been suggested that revascularization of autoclaved long-bone grafts occurs more slowly than it does in similar allografts, but that fatigue fractures of such long bone grafts occur less frequently. 23 In one of our two patients (Pt. 20) treated with an autoclaved autogeneic graft, no evidence of revascularization of the distal part of the graft was found after more than 4 years and partial weight-bearing continues. Osteoarticular allograft replacement obviates the need for total hip replacement or proximal femoral replacement, and the risks of dislocation and loosening of the hip prosthesis should be avoided, t5 However, the use of such osteoarticular grafts always lead to progressive narrowing of the joint space and osteoarthrosis, which, in some patients, resembles the rapid deterioration of neuropathy arthroplasty. 59 Other surgical options following periacetabular tumour resection are the ischiofemoral and the ileofemoral arthrodesis. Drawbacks of the ischiofemoral arthrodesis are the impossibility of complete weight-bearing without pain and poor functional results after failure of the arthrodesis, whereas the disadvantages of the ileofemoral arthrodesis are impaired gait and shortening of the limb. -'-''47 Furthermore, failure of arthrodesis has beeri reported to be 50% or more. tLI6'19'47'52

Conclusion Bone and soft tissue sarcoma of the pelvic region still represent an enormous problem. Wide surgical margins are often difficult to obtain. When an internal hemipelvectomy provides no adequate surgical margin, the only alternative is a hemipelvectomy. A hemipeivectomy is only feasible when, based on MRI, microscopically free surgical margins may be expected. During the last two decades, improvement in the limb-saving treatment of pelvic girdle sarcomas was achieved. An internal hemipelvectomy is an excellent alternative in well-selected cases. Different types of reconstruction after total or partial internal pelvic resection are possible, but complication rates are high in all these procedures. Patients have to be informed about the consequences of internal hemipelvectomies with reconstructions. Based on our results and the experience obtained during the last two decades we conclude that adequate surgical resection is the mainstay of succcessful treatment with respect to local tumour control and survival. The use of adjuvant radiation therapy or neoadjuvant chemotherapy may increase the success of surgery. The use of intraoperative radiation therapy in selected cases may be of great value to improve local tumour control, s°'5~

References I. Karakousis CE The technique of major amputations for malignant tumors. J Sarg Oncol 1983; 23: 43-55. 2. Li6nard D, Lejeune FJ. Upper and lower limb disarticulation in soft tissue and bone sarcomas. Worm J Surg 1988; 12: 340--4. 3. Eggerrnont AMM, Sehraffordt Koops, Li6nard D, et ak Isolated limb perfusion with high-dose tumor necrosis factora in combination with interferon-y and melphalan for nonresectable extremity soft tissue sarcomas: a multicenter trial. J CIhl Oncol 1996; 14: 2563-5. 4. Apffelstaedt JP, Driscoll DL, Spellman JE, Velez AF, Gibbs JF, Karakousis CE Complications and outcome of external hemipelveetomy in the management of pelvic tumors. Ann Sarg Oncol 1996; 3: 304-9. 5. Shiu KH, Rougraff BT, Simon MA. Oncologic outcome of primary bone sarcomas of the pelvis. Clin Orthop 1994; 304: 207-17. 6. Douglass HO, Razack M, Holyoke D. Hemipelvectomy. Arch Sarg 1975; 110: 82-5. 7. Miller TR. 100 cases of hemipelvectomy: a personal experience. Sarg Clin North Am 1974; 54: 905-13. 8. Malawer MM, Buch RG, Thompson WE, Sugarbaker PH. Major amputations done with palliative intent in the treatment of local bony complications associated with advanced cancer. J Surg Oncol 1991; 47: 121-30. 9. Lei B van der, Hoekstra HJ, Veth RPH, Ham SJ, Oldhoff J, Schraffordt Koops H. The use of the saddle prosthesis for reconstruction of the hip joint after tumor resection, of the pelvis. J Surg Oncol 1992; 50: 216-9. 10. Karakousis CE Emrich LJ, Driseoll DL. Variants of hemipelvectomy and their complications. Am J Surg 1989; 158: 404-8. II. Enneking WF, Dunham WK. Resection and reconstruction for primary neoplasms involvingthe innominate bone. J Bone Joint Surg (Br) 1978; 60-A: 747-51. 12. Aboulafia AJ, Malawer MM. Surgical management of pelvic and extremity osteosarcoma. Cancer 1993; 7t (Suppl. 10): 3358-66. 13. Gradinger R, Rechl H, Hipp E. Pelvic osteosareoma. Resection, reconstruction, local control, and survival. Clin Orthop 1991; 270: 149-58. 14. Veth RPH, Schraffordt Koops H, Nielsen HK, Oldhoff J, Verkerke G J, Postma A. A critique of techniques for reconstruction after internal hemipelvectomy for osteosarcoma. Cancer Treat Res 1993; 62: 221-9. 15. Harrington KD. The useofhemipelvicallografts or autoclaved grafts for reconstruction after wide resections of malignant tumors of the pelvis. J Bone Joint Surg (Am) 1992; 74-A: 331-41. 16. Campanacei M, Capanna R. Closing remarks, in: Enneking WF (ed.) Limb Salvage in Musculoskeletal Oncology. New York: Churchill Livingstone, 1987: 187-91. 17. Capanna R, Guernelli N, Ruggieri P, et al. Periacetabular pelvic resections. In: Enneking WF (ed.) Limb Salvage in Musculoskeletal Oncology. New York: Churchill Livingstone, 1987: 14t-6. 18. Eilber FR, Eckardt J J, Grant TG. Resection ofmalignant bone tumors of the pelvis: evaluation of local recurrence, survival, and function. In: Enneking WF (ed.) Limb Salvage in Musculoskeleton Oncology. New York: Churchill Livingstone, 1987: 136-41. 19. Enneking WF, Menendez LR. Functional evaluation ofvarious reconstructions after periacetabular resection of iliac lesions. In: Enneking WF (ed.) Limb Salvage hi Musculoskeletal Oncology. New York: Churchill Livingstone, 1987:117-35. 20. Mankin HJ, Doppelt S, Tomford W. Clinical experience with allograft implantation. The first ten years. Clin Orthop 1983; 174: 69-86. 21. Mnaymneh W, Malanin T, Mnaymneh LG, Robinson D. Pelvic allograft. A case report with a follow-up evaluation of 5.5 years. Clin Orthop 1990; 255: 128-32. 22. Tomeno B, Lanquepin A, Gerber C. Local resection with limb salvage for the treatment of periacetabular bone tumors:

546

23.

24.

25.

26. 27.

28. 29.

30.

31.

32. 33.

34. 35. 36. 37.

38. 39. 40.

S . J . Ham et al. functional results in nine cases. In: Enneking WF (ed.) Limb Salvage h~ Musculoskeletal Oncology. New York: Churchill Livingstone, 1987: 147-56. Harrington KD, Johnston JO, Kaufer HN, Luck JV Jr., Moore SM. Limb salvage and prosthetic joint reconstruction for lowgrade and selected high-grade sarcomas of bone after wide resection and replacement by autoclaved autogeneic grafts. Clin Orthop 1986; 211: 180-214. Mutschler W, Burri C, Kiefer H. Functional results after pelvic resection with endoprosthetic replacement. In: Enneking WF (ed.) Lhnb Salvage in Musculoskeletal Oncology. New York: Churchill Livingstone, 1987: 156-66. Nielsen HKL, Veth RPH, Oldhoff J, Schraffordt Koops H, Scales ST. Resection of a peri-acetabular chondrosarcoma and reconstruction of the pelvis. A case report. J Bone Joint Surg (Br) 1985; 67-B: 413-5. Johnson JTH. Reconstruction of the pelvic ring following tumor resection. J Bone Joblt Surg (Am) 1978; 60-A: 747-51. Healey JT, Lane JM, Marcove R, Duane K, Otis JC. Limbsparing excision and reconstruction for periacetabular malignant tumors. Read at the International Symposium on Lhnb Salvage in Musculoskeletal Oncology, Kyoto, Japan, Oct. 1987. blieder E. Die Sattelprothese. In: ENDO-klinik (eds). Primiir und Revisions AIIoarthroplastik Haft und Kniegelenk. Berlin: Springer-Verlag, 1987:257-7 I. Nieder E, Keller A. The saddle prosthesis Mark I1, EndoModell. In: Yamamuro T (ed.) New Developments for Lhnb Salvage in Musculoskeletal Tumors. Tokyo: Springer-Verlag, 1989: 481-90. Aboulafia AJ, Buch R, Mathews J, Li W, Malawer MM. Reconstruction using the saddle prosthesis following excision of primary and metastatic periacetabular tumors. Clin Orthop 1995; 314: 203-13. Windhager R, Karner J, Kutschera HP, Polterauer P, SalzerKuntschik M, Kotz R. Limb salvage in periacetabular sarcomas. Review of 21 consecutive cases. Clin Orthop 1996; 331: 265-76• Uchida A, Myoui A, Araki N, Yoshikawa H, Ueda T, Aoki Y. Prosthetic reconstruction for periacetabular tumors. Clin Orthop 1996; 326: 238-45. Sugarbaker PH, Chretien PB. Posterior flap hemipelvectomy. In: Sugarbaker PH, Malawer MM (eds). Musculoskeletal Surgery for Cancer. Principles and Tecbniques. New York: Thieme Medical Publishers, 1992: 121-37. Chretien PA, Sugarbaker PH. Surgical technique of hemipelvectomy in the lateral position. Surgery 1981 ; 90: 900-9. Ariel IM, Shah JP. The conservative hemipelvectomy. Surg Gynecol Obstet 1977; 144: 407-13. Sugarbaker PH, Chretien PA. Hemipelvectomy for buttock tumors utilizing an anterior myocutaneous flap of quadriceps femoris muscle. Ann Surg 1983; 197: 106-5. Sugarbaker PH, Chretien PB. Anterior flap hemipelvectomy. In: Sugarbaker PH, Malawer MM (eds). Musculoskeletal Sirgery for Cancer. Principles and Techniques. New York: Thieme Medical Publishers, 1992: 138-49. Enneking WF, Spanier SS, Goodman MA. The surgical staging of musculoskeletal sarcoma. J Bone Johlt Surg (Am) 1980; 62A: 1027-30. Eilber FR, Grant TT, Sakai D, et aL Internal hcmipelveetomy: excision of the hemipelvis and limb preservation. Cancer 1979; 43: 806-9. Enneking WF. Musculoskeletal Tumor Surgery. The Pelvis. New York: Churchill Livingstone, 1983: 483-90.

41. Enneking WF, Dunham W, Gebhardt MC, Malawer M, Pritchard DJ. A system for the functional evaluation of reconstructive procedures after surgical treatment of tumors of the musculoskeletal system. Clin Orthop 1993; 286: 241-6. 42. Kaplan E, Meier P. Non-parametric estimation from incomplete observations. J Am Stat Assoc 1958; 8: 423-46. 43. GraphPad Prism'% version 2.0. San Diego, CA: GraphPad Software, 1994--5. 44. Higinbotham NL, Marcove RC, Casson P. Hemipelvectomy: a clinical study of 100 cases with five year follow-up on 60 patients. Surgery 1966; 59: 706-8. 45. Karakousis CP. Internal hemipelvectomy. In: Sugarbaker PH, Malawer MM (eds). Musculoskeletal Surgery for Cancer. Prhwiples and Teclmiques. New York: Thieme Medical Publishers, 1992: 150-63. 46. Enneking WF. Local recurrence of malignant lesions of the hip and pelvis. J Bone Joh~t Surg (Am) 1966; 48-A: 991-1007. 47. O'Connor MI, Sire FH. Salvage of the limb in the treatment of malignant pelvic tumors. J Bone Johlt Surg (Am) 1989; 71A: 481-94. 48. Stephenson RB, Kaufer H, Hankin FM. Hindquarter amputation for skeletal neoplasms. Clin Orthop 1991; 270: 149-58.

49. Ham SJ, Dobbelaar, Schraffordt Koops H, Kroon H, Hoekstra HJ. Osteosarcoma of the pelvis. (Submitted). 50. Hoekstra H J, Sindelar WF, Kinsella TJ. Surgery with intraoperative radiotherapy for sarcomas of the pelvic girdle: A pilot experience, hit J Radiat Oncol Biol Pbys 1988; 15: 1013-6. 51. Hoekstra HJ, Sindelar WF, Szabo BG, Kinsella TJ. Hemipelvectomy and intraoperative radiotherapy for bone and soft tissue sarcomas of the pelvic girdle. Radiother Oncol 1995; 37:160-3. 52. Huth JF, Eckardt J J, Pignatti G, Eilber FR. Resection of malignant bone tumors of the pelvic girdle without extremity amputation. Arch Surg 1988; 123: 1121-3. 53. Karakousis CP, Vezeridis MP. Variants of hemipelvectomy. Am J Surg 1983; 145: 273-7. 54. Prewitt TW, Alexander HR, Sindelar WF. Hemipelvectomy for soft tissue sarcoma: Clinical results in fifth-three patients. Surg Oncol 1995; 4: 261-9. 55. Kraybill WG, Sandiford SB, Johnson FE. Posthemipelvectomy hernia. J Surg Oncol 1992; 51: 38-41. 56. Burri C, Claes L, Gerngross H, Mathys R. Total internal hemipelvectomy. Arch Orthop Traumat Surg 1979; 94: 219-000. 57. Steinbrink K, Engelbrecht E, Fenelon GCC. The total femoral prosthesis. A preliminary report. J Bone Johlt Surg (Br) 1982; 64--B: 305-12. 58. Karpf PM, Mang W. Das Retikulumzellsarkom des Beckens. Forthsclo" Med 1978; 31: 1559-0000. 59. Thompson RC Jr., Manival C. Neuropathy arthropathy as a possible case of failure of a whole joint allograft. A case report. Clhl Orthop 1988; 234: 124-8. 60. Steel HH. Partial or complete resection of the hemipelvis. An alternative to hindquarter amputation for pariacetabular chondrosarcoma of the pelvis. J Bone Joint Surg (Am) 1978; 60-A: 719-30. 61. D~nham WK. Acetabular resections for sarcoma. In: Enneking WF (ed.) Limb Salvage hl Musculoskeletal Oncology. New York: Churchill Livingstone, 1987: 170-84. 62. Rosenberg AG, Mankin HJ. Complications ofallograft surgery. In: Epps CH (ed.) Complications #1 Orthopedic Surgery, vol. 2. • Philadelphia, PA: J.B. Lippincott, 1986: 1385-417.

Accepted for publication 7 October 1997