Dedifferentiated chondrosarcoma of the pelvis: clinical outcomes and current treatment

Chondrosarcoma (CS) is a rare malignant bone tumour composed of cartilage matrix-producing cells. It is the second most common primary bone sarcoma with an incidence of 1 in 200,000/year. It may arise in the medullary cavity of bone (central CS) or secondary to a malignant transformation of a benign cartilage tumour [1]. The most important factor for guiding management and prognosis is determining the histological grade of the tumour. There is a high degree of inter-observer variability when determining histological grade [2]. Dedifferentiated CS is defined as one area of a lower grade cartilage tumour that lies directly adjacent to an area of high-grade non-cartilaginous sarcoma [3]. Only 10% of all CS dedifferentiate, which is fortunate as these high-grade tumours are associated with a 5-year survival of around 10–24% [1, 4, 5].

Chondrosarcomas, in general, are resistant to chemotherapy and conventional radiotherapy. Occasionally, short-term local control can be achieved but has no proven benefit on distant spread or overall survival [6‐8]. Consequently, surgery remains the mainstay of treatment for dedifferentiated CS. Dedifferentiated CS arising from the pelvis is known to be a negative prognostic factor, further lowering the survival rate [9, 10]. For this disease, there is limited evidence defining the presenting disease stage, accuracy of pre-operative diagnosis and treatment factors influencing outcome [9].

Current treatment for pelvic dedifferentiated CS consists of either palliative, limb-salvage through pelvic resection with or without reconstruction, or limb-sacrifice with hindquarter amputation [5, 6]. Tumour excision with wide margins should provide the best prognosis although this is often difficult due to the proximity of vital structures. This association and the probability of achieving wide margins with limb-salvage has yet to be described in pelvic dedifferentiated CS.

We report our experience with dedifferentiated CS of the pelvis including diagnosis, survival and surgical outcomes. The questions we attempted to answer in this study were; (1) what is the prognosis of patients with dedifferentiated CS of the pelvis? (2) Do wide margins or type of surgical intervention influence outcome in patients for attempted curative resection? (3) Does the use of adjuvant therapy affect outcome?

A retrospective review was conducted of a prospectively maintained database to identify patients with a histological diagnosis of dedifferentiated CS managed at a single tertiary musculoskeletal oncology centre. Minimum follow-up was 12 months or until death. 116 patients were diagnosed with dedifferentiated CS between 1995 and 2016.

31 patients were identified with pelvic dedifferentiated CS. They had a mean age of 55.6 years (range 33 to 76 years). There were 19 males (61.3%) and 12 females (38.7%). The ilium (P1) was involved in 51.6% (16/31) of tumours. 48.4% of tumours (15/31) involved the periacetabular (P2) region either in isolation or in combination with other pelvic regions (Fig. 1).

Local tumour staging comprised of plain radiography and magnetic resonance imaging (MRI). Distant staging comprised of chest computed tomography (CT) and bone scintigraphy. All patients were managed by a specialist sarcoma multidisciplinary team (MDT). Histological diagnosis was based on biopsy material and reviewed by a specialist sarcoma pathologist prior to surgical treatment. In 11/18 patients (61.1%) undergoing surgical resection the histological diagnosis of a dedifferentiated chondrosarcoma was made on the biopsy material pre-operatively. In the remaining seven patients (38.9%) pre-operative biopsy failed to accurately grade the tumour when compared to the post resection histology. Of these seven patients, three had a biopsy that suggested a low-grade CS, two a high-grade CS and in two, a spindle cell sarcoma with no cartilaginous elements (Table 1).

Wide margins, the aim of surgery, were defined as greater than 4 mm on final pathological specimen analysis [11]. Indications for hindquarter amputation (HQA) were involvement of two of the following three structures: sciatic nerve, hip joint, external iliac vessels or where it was deemed the safest way to achieve wide margins (Figs. 2a, b, 3a–c). Chemotherapy and radiotherapy was guided by oncologists on an individualised basis to attempt systemic or local short-term tumour control.

Patients were divided into palliative or surgical treatment groups. 13 patients (41.9%) were in the palliative group. Patients were not offered surgery due to either advanced metastatic disease (in 10 patients), the tumour was deemed unresectable (entire hemipelvis involvement, including the sacrum and sciatic notch), too medically unfit for surgery or died during neoadjuvant chemotherapy (one patient each) (Table 2). Patients in the surgical group were analysed according to the surgical procedure and surgical margins.

18 patients (58.1%) underwent attempted curative treatment by surgery. 13 patients underwent limb-salvage surgery (LSS) and five HQA. None of the patients in the surgical group had metastases at the time of diagnosis. For those treated by LSS, six (46.2%) had internal hemipelvectomies without reconstruction, three (23.1%) had a stemmed acetabular prosthesis (“ice-cream” cone), three (23.1%) had custom implants and one (7.7%) had combined internal hemipelvectomy and a proximal femoral replacement.

One patient died from a post-operative cardiac arrest and another suffered a periprosthetic infection which was successfully treated through debridement, implant retention and antibiotics.

Statistical analysis was performed using R and deducer statistical software packages, and considered statistically significant at p < 0.05. Analysis of local recurrence-free survival (LRFS), metastasis-free survival (MFS), disease-free (DFS) and overall survival (OS) was performed using the Kaplan–Meier survival method with 95% confidence intervals (CI). To assess the effect of different factors on survival outcomes, the Cox proportional hazards (PH) model was used.

In all chondrosarcomas, survival correlates with histological grade, and the dedifferentiated CS subtype is known to confer the worst prognosis [1, 9, 10, 12]. In accordance with other studies, dedifferentiated CS has a peak incidence in the 6th decade of life, a male predilection, and a male-to-female ratio of 3:2 observed in our series [12‐14].

Dedifferentiated CS has been associated with a poor prognosis [10]. The best chance of cure was linked to excision of the tumour with clear margins [10]. Achieving clear margins is much more difficult in the pelvis, even with HQA. The role of chemotherapy and radiotherapy is largely palliative [7, 8, 15‐18]. The presenting clinical scenario and treatment of pelvic dedifferentiated CS has not been specifically examined previously, however, data extrapolated from previous studies clearly reveals pelvic disease confers a dreadful prognosis (Table 4) [5, 9, 10, 12, 19, 20].

In the present study concerning dedifferentiated CS of the pelvis, 42% of patients presented with such advanced disease that surgery was not a viable treatment option. Unsurprisingly, patients able to undergo surgery with curative intent had greater overall survival than the palliative group. The only long-term survivors with no evidence of disease (16.1% of the cohort) underwent attempted curative surgical resection with wide or marginal margins.

Patients who underwent primary HQA had higher rates of disease-free survival than those who underwent LSS, 60% vs 15.4%, respectively. This likely correlated to the higher rates of wide margins achieved with HQA. In patients who had LSS, 66.7% had local recurrence and in the those who did not develop local recurrence, two are long term survivors and the other two died at 3 and 4 months. Other studies of predominantly non-dedifferentiated pelvic CS have revealed satisfactory margins can be achieved with limb-salvage [6, 12, 13]. However, one study identified a significantly higher chance of obtaining a clear margin with amputation [21]. Our data suggests that wide margins are more achievable with HQA, which was translated into improved local control and survival. The primary operation to consider in those with dedifferentiated CS should be a HQA with wide margins to offer patients the best chance of cure. In this series, none of the three patients who had HQA with wide margins died, compared to only three survivors in the 15 patients who had any other combination of surgery and margin type.

Obtaining a wide margin has been reported to be a positive prognostic indicator for non-dedifferentiated CS [11, 14, 22]. However, it remains unclear whether achieving a wide margin around dedifferentiated CS impacts local recurrence [12‐14, 23]. In our series, there was an 80% chance of achieving disease cure when wide margins, greater than 4 mm, were obtained. Of the five patients who were completely disease-free at last follow-up, four had a wide margin and the other a marginal margin. There were no long-term survivors in those with intralesional tumour resection.

Another issue with pelvic dedifferentiated CS, reflecting the tumour heterogeneity and large volume, is that 7 of the 18 curative patients did not have the correct diagnosis on initial biopsy. Three of these patients were thought to have low-grade CS and LSS was carried out with very close margins. Although one died postoperatively the other two both developed recurrent disease at 2 months. This highlights the importance of obtaining wide margins in any pelvic CS. Four of the five patients who were disease-free at last follow-up had accurate pre-operative biopsies. Saifuddin et al. have suggested taking biopsies from areas of reduced signal intensity on T2-weighted MRI may result in higher detection rates of dedifferentiated CS [24].

For isolated local recurrence of dedifferentiated CS treated with amputation, no patients achieved a cure or long-term survival (mean survival = 9.7 months). LR and metastases occurred synchronously in 90%. A study reporting surgical outcomes for recurrent CS of the pelvis also showed no overall survival advantage when treating recurrent high-grade CS tumours [19]. Therefore, the priority should be initial curative resection to reduce the chance of recurrence as rapid disease progression, with subsequent death can be anticipated.

In our series, one of the notable findings was that all recurrent disease, whether local or systemic arose within 24 months. This is likely a reflection of the extremely aggressive nature of this tumour. A high chance of cure can be expected if a patient lacks any evidence of disease after 24 months follow-up. A short time to recurrent disease was also evident in previous studies (Table 4).

We found no benefit for chemotherapy or radiotherapy used in a palliative setting, reaffirming the belief that chemotherapy and radiotherapy are ineffective at controlling dedifferentiated CS [7].

The limitations to this series are that the data is based on a retrospective analysis from a single institution, with inherent selection bias both in those treated surgically and in the type of surgical procedure. This study was not able to analyse whether having wide margins from the dedifferentiated component of the tumour or if having a small proportion of dedifferentiation within the tumour was associated with increased survival as this was not routinely reported by pathologists, however, these would be important areas to research in the future. Additionally, the sample size was small, reflecting the rarity of this disease, therefore whether amputation is truly superior at achieving radical section margins cannot be proven, and a larger multi-institutional review would help clarify this.

Dedifferentiated CS of the pelvis confers a poor prognosis. 12-month survival is 15.4% in those treated palliatively and 55.6% when treated with curative intent. It is a particularly aggressive disease, presenting at an advanced, inoperable state in almost half of patients. The factors that influenced disease cure were achieving a wide surgical margin (greater than 4 mm), which was more common with HQA. Margins less than 4 mm have a very high risk of local recurrence and death. HQA for local recurrence did not result in disease control and 90% of disease recurrence occurred synchronously. Obtaining the correct pre-operative diagnosis is also an important factor. We recommend early surgery with the widest possible margins to optimise chances for long-term survival.