We searched the files on all dedifferentiated peripheral chondrosarcomas
that had been seen at the Rizzoli Institute between 1970 and 2002. We reviewed
all cases in which a high-grade sarcoma had been diagnosed at the same
location as a preexisting osteochondroma. The presence of a preexisting
osteochondroma had to be demonstrated either histologically or
radiographically. Any extent of a high-grade sarcomatous component qualified
the tumor as dedifferentiated. The high-grade sarcomatous component and the
lowgrade cartilaginous component can present either synchronously or
metachronously. In the first case, both components are appreciated
side-by-side in the same lesion. In the second case, the high-grade sarcoma is
seen at the exact location where an exostosis or secondary peripheral
chondrosarcoma was previously excised and the cartilaginous component can be
completely absent from the dedifferentiated lesion.
Of the twenty cases that were identified, two cases were excluded. In one
case the lesion was reclassified as a dedifferentiated central chondrosarcoma
(arising in an enchondroma), and in the other case important data (patient
records and radiographs) were not available. Of the remaining eighteen cases,
fifteen had been treated at our institution and three had been seen in
consultation.
Clinical and follow-up data were obtained from the medical records and from
the patients themselves if possible. The stage of disease was determined
according to the classification system described by Enneking et
al.8. In that
system, low-grade lesions are classified as stage I (with A indicating an
intracompartmental lesion and B indicating an extracompartmental lesion) and
high-grade lesions are classified as stage II (with A indicating an
intracompartmental lesion and B indicating an extracompartmental lesion).
Metastatic disease is classified as stage III (with A indicating a low-grade
lesion and B indicating a high-grade lesion).
Plain radiographs were available for all eighteen patients, computed
tomography scans were available for twelve, and magnetic resonance imaging
scans were available for two. Radiographs were evaluated for evidence of a
preexisting osteochondroma, the presence of aggressive destructive changes,
and a bimorphic appearance, characteristic of
dedifferentiation4.
A bimorphic pattern was defined as being present if a lesion had features
suggesting a chondroid tumor adjacent to a markedly different area, such as a
dominant region of increased opacity within the mineralized tumor, a large
unmineralized soft-tissue mass, a different pattern of mineralization, or a
combination of these findings.
Histologic slides were available for review in all eighteen cases. The
slides were evaluated with regard to several histologic features, including
the histologic grade of the cartilaginous component, the histologic subtype
and grade of the dedifferentiated component, and the percentage of the lesion
that was represented by the dedifferentiated portion. The histologic grade was
determined according to the system of Broders et
al.9. The percentage
of the lesion that was represented by the dedifferentiated component was
estimated by two pathologists (P.B. and F.B.), who reviewed all of the
available slides of the resected specimens. For lesions that had been treated
at our institution prior to 1995, there were five to ten slides available per
lesion. For those that had been treated more recently (beginning in 1995), a
complete mapping of each resected specimen had been performed and thirty to
120 slides were available per lesion, depending on the size of the lesion.
Statistical Analysis
Overall survival analysis was performed with use of the KaplanMeier and Cox
regression methods. When two groups with different treatment strategies were
confronted, discrete variables were compared with use of the chi-square test,
and continuous variables were compared with use of the two-sample t test or
the Mann-Whitney test, if necessary and appropriate. Patients with preexisting
metastatic disease were excluded from the analysis of treatment effects. All
statistical tests were two-sided, and p values of <0.05 were considered
significant.
Clinical Features
The patients included twelve men and six women with an average age of
forty-six years (range, twenty-two to seventy-four years). Eight lesions arose
in patients with multiple hereditary exostoses, and the remaining ten lesions
occurred in patients with solitary osteochondromas. The most frequently
reported symptoms were pain (fifteen patients) and a swelling or growing mass
(eleven patients). One patient with a pelvic lesion presented with urinary
incontinence and limited range of motion of the right hip. The average
duration of symptoms was eighteen months (range, two to forty-eight months).
The most common locations were the pelvis (six lesions) and the femur (five
lesions, including two in the proximal part of the bone and three in the
distal part of the bone). The other lesions were located in the proximal part
of the tibia (three), the humerus (two), the scapula (one), and the sternum
(one). The distributions of the lesions in terms of age, gender, and anatomic
location are presented in Figure
1.
Four patients had an Enneking stage-IIIB lesion (indicating metastatic
disease) at the time of presentation. Eleven patients had a stage-IIB lesion.
For the three patients who were seen in consultation, the stage of the lesion
could not be ascertained.
Radiographic Features
In sixteen cases, there was radiographic evidence of a preexisting
osteochondroma. In the remaining two cases, the patient had multiple
hereditary exostoses and the lesion occurred at exactly the same location
where an exostosis had been previously diagnosed. In all eighteen cases,
radiographs showed a large lesion on the surface of the bone with malignant
characteristics such as indistinct margins, inhomogeneous mineralization, and
a large soft-tissue mass. Only eight lesions demonstrated a bimorphic pattern,
characteristic of dedifferentiation (Fig.
2). In six of these cases the dedifferentiation was seen on plain
radiographs, but in two cases computed tomography and magnetic resonance
imaging revealed a bimorphic pattern that could not be detected on plain
radiographs. Ten lesions appeared as a conventional secondary chondrosarcoma
arising in an exostosis.
Pathological Features
On the average, the lesions measured 12 × 10 × 10 cm (range, 5
× 4 × 4 cm to 33 × 25 × 20 cm).
The cartilaginous component of the dedifferentiated lesions was classified
as grade-1 chondrosarcoma in thirteen cases and grade-2 chondrosarcoma in two
cases. Fifteen lesions showed the characteristic bimorphic pattern of a
dedifferentiated chondrosarcoma, with a high-grade sarcomatous component
juxtaposed to a low-grade cartilaginous lesion
(Fig. 3). In three patients,
the neoplasm consisted entirely of a high-grade sarcoma and no cartilaginous
component was appreciated (metachronous dedifferentiation). Two of these
patients had multiple hereditary exostoses with a previously diagnosed
exostosis at the exact same location, and one had a history of excision of a
solitary osteochondroma at the same location seven years earlier.
The dedifferentiated component was considered to be an osteosarcoma in nine
cases (including six cases in which it was osteoblastic and three in which it
was fibroblastic), a malignant fibrous histiocytoma in eight, and a
fibrosarcoma in one. In all cases, the dedifferentiated component showed
aspects of a high-grade (Broders grade-4) sarcoma. The dedifferentiated
component represented an average of 59% (range, 20% to 100%) of the entire
tumor.
Treatment and Outcome
No treatment or outcome information was available for the three patients
who had been seen in consultation. The fifteen patients who had been managed
at our institution had a median duration of follow-up of fourteen months
(average, thirty-two months; range, three to 111 months). The treatment and
outcome information for the patients who had been managed at our institution
is shown in Table I.
Of the four patients who had metastatic disease at the time of
presentation, one was managed with nonoperative palliative treatment. The
other three patients underwent surgery: two had an en bloc resection, and one
had an above-theknee amputation. None of these patients underwent adjuvant
chemotherapy, and all died within one year (range, five to twelve months)
after the time of diagnosis.
Of the eleven patients with a nonmetastatic dedifferentiated chondrosarcoma
associated with an osteochondroma who had been managed at our institution,
seven had limb-sparing surgery (with six patients having an en bloc resection
and one having local excision) and four underwent an amputation (with three
patients having an above-the-knee amputation and one having a hindquarter
amputation). In one of the patients who had a limb-sparing resection, a
marginal margin was detected. In all other patients, wide margins were
achieved.
Before April 1994, no adjuvant chemotherapy was used at our institution for
patients who had a dedifferentiated chondrosarcoma arising in an
osteochondroma. After April 1994, all patients with this diagnosis were
evaluated by an oncologist for the possibility of adjuvant chemotherapy
treatment. Whenever the patients met all of the necessary health criteria,
adjuvant chemotherapeutic treatment was proposed. Six of the patients in the
present study, including the patient who had had a marginal resection, three
patients who had had a wide resection, and two who had had a wide amputation,
underwent postoperative chemotherapy. The specifics of the chemotherapy
regimen were determined by the medical oncologist consultant in each case. The
choice of drugs and doses was based on the general health status of the
patients and their tolerance to side effects. Chemotherapy regimens consisted
of a combination of Adriamycin (doxorubicin), cisplatin, methotrexate, and
ifosfamide in five patients. In one of these patients the administration of
methotrexate was interrupted after the first cycle because of side effects.
One patient underwent a single-drug regimen with Adriamycin.
Five of the six patients who underwent postoperative chemotherapy were free
of disease at the time of the most recent follow-up, at an average of
sixty-five months (range twenty-two to 111 months) after the time of
diagnosis. The other patient died of disease at thirteen months after the time
of diagnosis. All five patients who had only operative treatment of
nonmetastatic disease died of metastases at an average of twenty months
(range, three to fifty-one months) after the time of diagnosis.
Of the fifteen patients who were managed at our institution, 60% survived
for more than one year after the time of diagnosis. The two and five-year
survival rates were 47% and 29%, respectively, and the median survival time
was fourteen months (Fig. 4).
When the overall survival rates for the patients with stage-IIB lesions were
compared, the patients who had been managed with surgery and adjuvant
chemotherapy had a better overall survival rate than did those who had been
managed with surgery alone (p = 0.03) (Fig.
5). The patients who had been managed with surgery and
chemotherapy had had a shorter duration of symptoms (average, fourteen
compared with thirty-one months) and a smaller percentage of dedifferentiation
in the lesions (52% compared with 76%) than those who had been managed with
surgery alone. However, with the small number of patients studied, we could
find no significant difference between these two groups in terms of the
duration of symptoms, the degree of dedifferentiation, the age of the patient,
the size of the lesion, the histological grade or subtype of the
dedifferentiated component, or the location of the lesion.
Dedifferentiation in cartilaginous tumors is a well-known event. Several
reports have been published on dedifferentiation in
chondrosarcoma3-7,10.
However, only a few cases of dedifferentiated chondrosarcoma at the site of a
preexisting exostosis have been described in detail in the English-language
literature11-14.
To our knowledge, the present study of eighteen cases describes the largest
reported series of dedifferentiated chondrosarcoma arising in
osteochondroma.
Between 1970 and 2002, 273 cases of chondrosarcomas in preexisting
exostoses were treated at our institution. Of these, fifteen were identified
as dedifferentiated chondrosarcomas, which accounts for a dedifferentiation
rate of approximately 5.5%. This percentage is lower than that for central
chondrosarcomas, which we found to be
15.5%15.
The distribution of gender and localization of the lesion resembled that of
conventional secondary chondrosarcomas arising in exostoses. Our patients with
dedifferentiated chondrosarcomas were slightly older than those with
conventional chondrosarcomas in preexisting exostoses but were younger than
those with dedifferentiated central chondrosarcomas (forty-six compared with
fifty-nine years). Almost half of the patients had multiple hereditary
exostoses, whereas in dedifferentiated central chondrosarcomas fewer than 5%
of the cases were linked to multiple cartilaginous lesions as in Ollier
disease and Maffucci
syndrome1,2,15.
The prognosis for patients with a dedifferentiated chondrosarcoma arising
in an exostosis is poor. Four of our eighteen patients had metastatic disease
at the time of presentation. All four patients died within one year. The
patients with localized disease had considerably better outcome if they had
adjuvant chemotherapy after a wide surgical resection rather than surgical
treatment only. In the literature, the benefit of adjuvant chemotherapy in the
treatment of dedifferentiated chondrosarcomas has never been proven. Frassica
et al.4 found no
difference in outcome between patients managed with or without adjuvant
chemotherapy, and this finding was recently confirmed by Dickey et
al.10 in the
extension of that study. Mitchell et
al.7 reported better
outcomes for patients with dedifferentiated chondrosarcomas that were treated
with surgical excision in combination with chemotherapy. However, in that
retrospective study, a selection bias could not be excluded and the group
numbers were small. We recently reviewed the outcomes for 123 patients with
dedifferentiated central chondrosarcomas who were seen at our
institution15. In
that series, we found no difference in overall survival between patients
managed with surgery alone and those managed with a combination of surgery and
chemotherapy. That retrospective study included only twenty-five patients who
underwent surgery and chemotherapy, and the regimens varied widely. However,
in all patients who underwent preoperative chemotherapy, the tumor response
had been poor or very poor, as has been reported by
others7,10.
Thus, it seems that dedifferentiated central chondrosarcomas are relatively
resistant to chemotherapy. The good outcome for the chemotherapy and surgery
group in the current series is therefore surprising and difficult to explain.
The patients who had dedifferentiated peripheral chondrosarcoma were ten years
younger than those who had dedifferentiated central chondrosarcoma. Therefore,
they probably tolerated better the high doses of aggressive chemotherapeutic
agents, which could have made the chemotherapy more effective. Furthermore,
the chemotherapy group in the current study included relatively recent cases.
All of the patients who survived after surgery and adjuvant chemotherapy were
managed after 1994. They may have had some benefit from modern diagnostic
imaging techniques as well as from modern chemotherapeutic agents and
protocols. Finally, in contrast to dedifferentiated central chondrosarcomas,
dedifferentiated peripheral chondrosarcomas occur relatively frequently in
patients with a history of multiple cartilaginous lesions. Four of the six
patients in the present study who were managed with surgery and chemotherapy
had multiple hereditary exostoses and were aware of the risks of malignant
transformation of their lesions. They may have presented at our institution in
an early stage of dedifferentiation, with a greater possibility of a positive
outcome.
It is difficult to draw conclusions from the present retrospective review
of a relatively small number of patients. Another limitation of the present
study is that there were no patients who were managed with neoadjuvant
chemotherapy, and therefore no assessment of the tumor response could be
performed. However, the present study shows that in patients who have a
dedifferentiated chondrosarcoma arising in an osteochondroma, a relatively
good survival rate can be achieved when wide surgical treatment is combined
with adjuvant chemotherapy. Therefore, it seems rational to continue the use
of chemotherapy whenever possible in these patients. ?