From 1992 through 1995, the senior author (D. S. S.) performed
six resections of parosteal osteosarcomas located on the posterior
cortex of the distal part of the femur through dual medial and lateral incisions.
Three patients were male and three were female. The average age
of the patients was thirty-two years (range, twenty-two to fifty-five
years).
Before the operation, all patients were evaluated with a complete
physical examination, plain radiographs, a whole-body technetium
bone scan, and computed tomography scans or magnetic resonance images
of the lesion. Radiographs of the chest were made for all patients,
and computed tomography of the chest was performed for four of the
six patients. All patients had the biopsy at the time of the resection.
Operative Technique
The operation can be carried out with the patient under general
anesthesia or spinal anesthesia. The patient is positioned supine.
A well padded pneumatic tourniquet is placed on the proximal-most portion
of the limb. The leg is elevated for two to three minutes, and the
tourniquet is inflated. The knee is flexed to 75 degrees, and the
foot is held to the operating table. A longitudinal incision is made
on the medial aspect of the thigh, extending from the midportion
of the thigh to approximately six centimeters distal to the knee.
The fascia is opened, and the interval between the vastus medialis
and the sartorius is identified (Fig. 2). The vastus medialis is retracted
anteriorly, and the gracilis and sartorius are retracted posteriorly. The
femoral vein and artery are identified in the popliteal space and
are retracted posteriorly. Small vascular branches to the distal
part of the femur and the tumor are ligated and divided. The sciatic nerve
is often not visualized as it is posterior to the plane of dissection.
Distally, the medial head of the gastrocnemius is identified and
cut one centimeter distal to its origin. The posterior part of the capsule
of the knee joint is visualized. The medial aspect of the distal
part of the femur is exposed, usually with some adductor magnus
insertion left on the bone posterior to the site of the medial longitudinal
osteotomy (Fig. 3).
A similar dissection is performed on the lateral side of the
distal part of the femur. A longitudinal incision is made, extending
from the midportion of the thigh to the Gerdy tubercle. The tensor
fascia is opened in line with the incision. The vastus lateralis
muscle is split longitudinally, with its osseous attachment to the
femur left undisturbed along the incision to expose the lateral
aspect of the distal part of the femur. The long head of the biceps
is identified and is dissected free from the underlying short head
(Fig. 4).
The posterior portion of the vastus lateralis and the short head
of the biceps femoris are left attached to the femur and remain
with the operative specimen. The popliteal space is identified through
the dissection between the long and short heads of the biceps femoris
muscle. The lateral head of the gastrocnemius is identified and
is cut one centimeter distal to its origin (Fig. 5). The posterior
part of the capsule of the knee joint is visualized.
Arthrotomies are made medially and laterally within the knee.
The posterior part of the capsule is cut transversely just distal
to its attachment on the femur. On the basis of the preoperative
studies used to determine the anatomical extent of the tumor and
with use of cautery, the lateral and medial osteotomy sites are
demarcated along the cortex of the femoral shaft. The medial and
lateral longitudinal osteotomies and the proximal transverse osteotomy
are made with an oscillating saw. The distal transverse osteotomy
is the most difficult osteotomy. The tumor usually extends into
the intercondylar notch, and the distal osteotomy must be distal
to the tumor. This requires removal of a portion of the origin of
the posterior cruciate ligament and a portion of both femoral condyles.
The osteotomies of the medial and lateral condyles are started with
an oscillating saw, but the central portion of the osteotomy is
made with an osteotome. The resected tumor is now removed from the
limb. The surgeon and pathologist examine the specimen. Frozen-section
analysis is done if there is a close or questionable margin. Additional
bone should be taken if the margin is inadequate.
A fresh-frozen distal femoral allograft is fashioned to match
the femoral defect. It is initially secured by lag-screw fixation
and then is reinforced with a plate and screws (Fig. 6-A). The incisions
are closed over large suction drains. The wounds are dressed, and
the limb is placed in a posterior splint with the knee held in 15 to
30 degrees of flexion.
The posterior splint is changed to a hinged brace once the patient
is comfortable. The patient is then allowed to walk with toe-touch
weight-bearing using crutches. The patient is allowed to move the knee
as soon as he or she is comfortable and is encouraged to increase
the range of motion as much as possible within the first ten days.
Once there is evidence of healing between the allograft and the host
bone, progressive weight-bearing is allowed and strengthening of
the muscles is started. The patient is usually fully weight-bearing
without external aids by four to six months after the operation.
Evaluation
After the resection, each specimen was examined to determine
the histological grade, the extent of the lesion, and the operative
margin. The same musculoskeletal pathologist examined all specimens.
The observations described below are based on these examinations.
All tumors were classified as surface lesions located in the
popliteal paraosseous space. In three specimens, the tumor was confined
to the cortex of the femur. Three specimens had focal microscopic spread
into the medullary canal. One specimen had invasion into the overlying
muscles (the biceps femoris, vastus lateralis, vastus medialis,
and medial head of the gastrocnemius). The patient with this tumor
received postoperative radiation to the site.
The pathologist determined the operative margin after a thorough
examination of the gross specimen and the histological sections.
The definition of margins was that suggested by Enneking et al.5. An intralesional margin was a resection
with tumor at the margin. A marginal margin was a resection through
the reactive tissues surrounding the tumor. A wide margin was a
resection in which the tumor was completely surrounded by normal,
nonreactive tissue.
The outcome of the reconstruction was scored as excellent, good,
fair, or failure with the grading system described by Mankin et
al.8. The result was considered
excellent when the patient had no evidence of disease, had a normally functioning
limb, and was able to return to his or her preoperative functional
activities. The result was considered good when the patient had
no evidence of disease or pain, had decreased function of the involved
limb but needed neither a brace nor an ambulatory support, and was
able to return to most activities but not active sports. The result
was considered fair when the patient had no evidence of tumor recurrence
but had a substantial functional deficit that required a brace or
an ambulatory support and that precluded the patient from returning
to many of his or her preoperative activities. The outcome was considered
to be a failure when the patient required a resection of the allograft
reconstruction or an amputation of the limb because of recurrence
of the tumor, fracture of the graft or the reconstructed limb, or
infection or when the patient had died because of the tumor.
The patients were seen at three-month intervals for the first
two years, six-month intervals for the next three years, and then
yearly for the remainder of their lives. At each visit, the patient
was examined and radiographs of the involved extremity and of the
chest were made.
The patients were observed throughout the treatment and follow-up
periods by the senior authors (D. S. S. and M. C. G.). At the time
of final follow-up, the patient was evaluated with regard to subjective
symptoms of pain and weakness. Objectively, the limb was judged
for joint mobility, clinical and radiographic signs of union, and
tumor recurrence.
Most reports have defined parosteal osteosarcoma, an uncommon
lesion, as a slow-growing, well differentiated malignant bone tumor
with indolent growth resulting in late invasion of the underlying cortex
and limited involvement of the medullary canal10.
Many procedures have been described for the treatment of parosteal
osteosarcoma. Inadequate resection can result in local recurrence
and metastasis2. In their series
of thirty-nine patients, Okada et al. found that incomplete resection
was associated with an increased risk of local recurrence9. Enneking et al. concluded that a
radical procedure is not necessary6.
The literature suggests that resection with a wide operative margin
is the treatment of choice6,11,13.
As emphasized by Campanacci et al.2,
we believe that it is important to define what constitutes a wide
operative margin. However, unlike some investigators2,10, we believe that focal medullary
involvement (stage IB) does not necessitate removal of an entire segment
of bone to achieve a wide operative margin. Our specimens were resected
with tumor-free margins anteriorly; however, each anterior margin was
within the medullary canal. Enneking et al. raised the question
of whether a patient who has underlying medullary involvement (stage
IB) has a greater risk of recurrence6.
However, univariate analysis of patients with parosteal osteosarcoma
treated at the Mayo Clinic demonstrated no association between local
recurrence and medullary involvement9.
Like the Mayo Clinic study and the study by Campanacci et al.2, our study did not demonstrate an
association between medullary invasion and recurrence. In our series,
in which the follow-up was longer than two years, intramedullary
resection with tumor-free margins did not have an adverse effect
on the outcome. Analysis of intraoperative frozen sections of the
medullary margin of the excised parosteal osteosarcoma can be important
to ensure a tumor-free anterior margin. Thus, although medullary
involvement (stage IB) may be associated with an increased risk
of differentiation, local recurrence, and metastasis3,10,12,13, a tumor-free margin remains
the critical factor determining overall prognosis.
The best operative resection of a tumor is one that removes all
tissue that must be removed while sparing all disease-free tissue.
In the past, oncological surgeons tended to remove more tissue than
necessary; thus, amputations for sarcomas were routine. More recently,
limb salvage has become accepted and has been proven to be safe
for the majority of sarcoma resections. The accuracy of computed
tomography scans and magnetic resonance images has improved greatly
and has allowed surgeons to determine the extent of a tumor precisely.
Thus, during the resection of a tumor, less normal tissue can be
removed without compromising the wide operative margin. We believe that
this explains why the increase in the use of limb salvage has not
resulted in an increase in the prevalence of local recurrence. We
suggest that, in some situations, even less radical resections are safe
and that the added function that they afford the patient is substantial.
In conclusion, we have described an operative technique that
facilitates a more conservative resection of a small-to-moderately
sized parosteal osteosarcoma of the posterior aspect of the distal part
of the femur. The two incisions allow the surgeon to clearly visualize
the sites of the osteotomies and to accurately define the margins
of the resection. The neurovascular bundle is easily dissected from
the posterior aspect of the tumor and is retracted out of harm's
way. Most of the distal femoral articular cartilage is salvaged,
the joint remains stable, and function after the resection is closer
to normal compared with that afforded by previously described techniques6,7,9,11,13. To date, there have been
no local recurrences in our series. The resection is technically
demanding and must be done with extreme care, but when it is done
well it results in excellent function and a low risk of local recurrence.
Note: The authors thank Jaye Schlesinger for providing illustrations
and Dr. Andrew Rosenberg for his pathological expertise.