Mesenchymal Chondrosarcoma: Clinicopathologic Study of 20 Cases Open Access

The consultation records of a senior bone pathologist (H.D.D.) spanning 45 years (1964–2009) were reviewed, yielding 20 patients from local, national, and international institutions who carried the diagnosis of mesenchymal chondrosarcoma. Information reviewed included clinical presentation, radiographs, and microscopic sections of the primary tumor stained with hematoxylin-eosin. When available, additional clinical and diagnostic results were reviewed, including office and operative reports, immunohistochemistry staining patterns, and imaging examinations, such as computed tomography, magnetic resonance imaging, and positron emission tomography. Further follow-up information and clinical data, including treatment details, were obtained by direct communication with the consulting pathologists, orthopedic surgeons, medical oncologists, and/or family members. For patients without follow-up, the social security death registry was used to determine whether the patient was deceased.

The review of current literature was performed by a PubMed query for articles with “Mesenchymal Chondrosarcoma” in the title or abstract from 1959 to the present. Articles selected for more detailed review were those in the English language, published during the past 10 years discussing molecular biology, cytogenetic, diagnostic, and therapeutic features of mesenchymal chondrosarcoma.

Eighteen tumors (90%) originated in bone and 2 tumors (10%) were of extraskeletal origin. Of the skeletal tumors, locations included craniofacial bones (50%, n  =  9), ribs and chest wall (22%, n  =  4), sacrum and spinal elements (17%, n  =  3), and the lower extremity (11%, n  =  2). Extraskeletal tumors were located about the scapula (50%, n  =  1) and the lower extremity (50%, n  =  1). This case series identified 11 females (55%) and 9 males (45%) with mesenchymal chondrosarcoma, with ages ranging from 7 to 37 years (mean age, 22.1 years; median age, 23.5 y) (Table 2). There was a difference in the mean age between male and female patients, with the tendency of male patients to be younger at presentation (mean age, males, 19.8 years; mean age, females, 24.0 years). Patients were most commonly in their second or third decade of life (Figure 1). Clinical presentation was available for 16 patients (80%), with most patients' primary symptomatic complaint being pain (50%, n  =  8) and/or a mass (44%, n  =  7). Neurologic symptoms from compression of neural structures were present in 3 cases (19%). Two cases were asymptomatic. The duration of symptoms before diagnosis ranged from 2 weeks to 1 year.

Plain radiographs typically consisted of a calcified, eccentric, osteolytic lesion, demonstrating extraosseous extension (Figures 2, through 4, A). Eighteen out of 20 cases (90%) originated within bone. In contrast to skeletal lesions, tumors originating within soft tissue appeared as tumors close to, but distinct from, the adjacent bone. Two cases (10%) in the present study were initially misdiagnosed as aneurysmal bone cysts based on radiographs, underscoring the benign, albeit locally aggressive, appearance. Computed tomography imaging commonly demonstrated a well-defined, calcified mass in bone with local destruction and soft tissue extension (Figures 5 through 8, A, and 9). Characteristic chondroid matrix calcification was inconsistently present. Magnetic resonance imaging showed variable T2-weighted signal characteristics, ranging from mild to high intensity, and a heterogeneous appearance on T1 fat-suppressed, postgadolinium sequences (Figures 4, B and C, 8, B, and 10). Positron emission tomography scan results were equivocal, demonstrating metabolically active masses in 4 patients (20%), but no abnormal uptake in 2 patients (10%). The maximum standard uptake values in one patient with widespread metastatic disease ranged from 4 to 6.5. The interpreting physician noted suspicion for primary malignancy or metastatic disease in reports of these studies.

Gross appearance was typically grey or tan and poorly circumscribed, with focal hemorrhagic and necrotic areas (Figure 11).

Microscopic examination demonstrated biphasic histology in all cases (n  =  20; 100%). Findings included sheets of undifferentiated, round or spindled mesenchymal cells, interspersed with islands of hyaline cartilage (Figure 12). Round- and spindle-shaped variants of the mesenchymal component were observed with similar frequency (Figure 13). Sixteen tumors (80%) demonstrated predominantly spindle cells, rather than round cells (Figures 14 and 15), and 11 (55%) demonstrated a hemangiopericytomatous pattern (Figures 16 through 19, A and B). Histologic features also included foci of osteoid production, ossification, or calcification (Figures 20 through 22). Overall, a wide spectrum of histologic variability was observed, ranging from tumors with predominately round or spindled cells to tumors represented mainly by hyaline cartilage islands.

Twelve samples (60%) had immunohistochemistry stains available for review. Vimentin positivity of the mesenchymal component was confirmed in 7 of 7 cases (100%) in which the stain was performed. S100 positivity in the cartilaginous component (Figure 23, A and B) was confirmed in 9 of 11 cases (82%), and CD99 positivity in the noncartilaginous component was noted in 4 of 6 cases (67%).

Details regarding treatment were available for 17 patients (85%), and outcome was available for 15 patients (75%). With the exception of 1 patient (5%) who underwent an intralesional procedure of the primary bone tumor, records indicate that wide or radical resections were performed for all patients. Despite the intralesional procedure initially performed for a working diagnosis of an aneurysmal bone cyst of the tibia of patient 1, the patient required multiple, additional procedures after diagnosis with mesenchymal chondrosarcoma. These procedures included an ipsilateral amputation and, eventually, a contralateral above-knee amputation for metastatic disease before the patient died, 12.6 years after initial presentation. At least 3 patients underwent additional operations for control of local disease; patient 2 required resection of the 10th rib for recurrent disease and died 2.2 years after presentation, patient 9 underwent wide excision for positive margins on initial excision and was lost to follow-up, and patient 12 underwent radical craniofacial resection after initial tumor resection and died 6 years after presentation.

During the first 28 years of consultation in this study, 90% of patients (9 of 10) were treated with surgery and no chemotherapy. Among these first 10 patients, 67% (n  =  6) died 2.2 to 21.8 years after presentation, 11% (n  =  1) were alive with disease 12.5 years after presentation, and 33% (n  =  3) were lost to follow-up. Thereafter, a trend favored the incorporation of adjuvant treatment, with 86% of patients treated with at least surgery and chemotherapy. Among those patients, 57% (n  =  4) were alive with no evidence of disease 1.8 to 4 years after presentation, 29% (n  =  2) were alive with disease 5.5 to 12 years after presentation, and 14% (n  =  1) had died 3.5 years after presentation. Chemotherapy regimens, when noted, were highly variable. Patient 15 was treated with an osteosarcoma protocol and died after 3.5 years. After developing widespread metastatic disease about 9 years after initial diagnosis and management, patient 14 was treated with 3 cycles of adriamycin and ifosfamide. Upon progression of disease, the treatment plan was 3 cycles of ifosfamide and etoposide at the most recent follow-up. Patient 20 was managed with adjuvant adriamycin and ifosfamide for 5 cycles and was alive with no evidence of disease 1.8 years after presentation. Neoadjuvant chemotherapy was administered in 1 case. Patient 17 was treated with the Ewing sarcoma (ES) protocol AEWS0031 (alternating courses of vincristine, doxorubicin, and cyclophosphamide with ifosfamide and etoposide) before resection of the tumor. In this setting of neoadjuvant chemotherapy, the only difference noted on histology of the specimen was abundant bone in excess of that usually seen on histology of mesenchymal chondrosarcoma. This patient was alive, with no evidence of recurrent mesenchymal chondrosarcoma, at follow-up, 4 years after presentation; however, he did develop acute myeloid leukemia, clinically interpreted as a secondary malignancy, and was treated with a bone marrow transplant. Overall, 5 patients (25%) underwent radiation therapy in addition to surgery and chemotherapy, except one case that was treated with surgery and radiation therapy alone (patient 12). Among these patients, 40% (n  =  2) were dead 6 to 12.6 years after presentation, 20% (n  =  1) were alive with metastatic disease 5.5 years after presentation, and 40% (n  =  2) were alive with no evidence of disease 2.6 to 4 years after presentation. Overall, 75% (n  =  3) of the 4 patients who were alive without disease received at least surgery and chemotherapy, with follow-up ranging from 1.8 to 4 years after initial presentation.

Despite being a recognized entity for more than 50 years,1 mesenchymal chondrosarcoma continues to present substantial diagnostic, prognostic, and management challenges, in large part because of its rarity. A review of case series within the literature and key features from these studies are summarized in Table 3.

Radiographic appearance is generally not specific and is, at times, variable. Lesions are typically osteolytic and locally destructive, with soft tissue extension and finely stippled calcification.2 Our findings are in keeping with that report. There has been a limited attempt to improve diagnostic accuracy. Font et al12 correlated computed tomography appearance with histopathology of orbital mesenchymal chondrosarcoma, noting that hemangiopericytomatous areas are enhancing and chondrosarcoma-like areas appear nonenhancing upon gadolinium administration.

The histology of mesenchymal chondrosarcoma is typically characterized by a unique, biphasic pattern, composed of mesenchymal and chondrocytic components.6,27 The background of undifferentiated mesenchymal, small, round or spindled cells is interrupted by highly differentiated islands of hyaline cartilage. The primitive, round cell component often contains a hemangiopericytomatous vascular pattern, and the cartilage component may include sites of osteoid production, ossification, or calcification. The diagnosis is more readily apparent if both components are present; however, the absence of cartilage in the biopsy sample can cause this tumor to be confused with other small, round, blue-cell tumors. The histologic differential diagnosis for mesenchymal chondrosarcoma includes ES; lymphoma; neuroblastoma; desmoplastic, small, round-cell tumors; and small cell osteosarcoma.

Typical immunohistochemistry findings include positivity of the mesenchymal portion for vimentin, Leu7, and CD99 and positivity of the cartilaginous regions for S100 protein.6 Once again, absence of either component can lead to a skewed immunohistochemical picture. Utility of immunohistochemistry is limited by the sample provided. CD99, although immunoreactive in all 11 mesenchymal chondrosarcoma tumors in one study,28 is also a marker for ES and for primitive neuroectodermal tumors, further clouding the distinction between these tumors.29 

In an attempt to improve diagnostic specificity, Wehrli et al30 compared immunohistochemistry profiles of 12 mesenchymal chondrosarcoma samples with several other small, round, blue-cell tumors and demonstrated expression of Sox9 in both cartilaginous and mesenchymal regions in 21 of 22 mesenchymal chondrosarcoma tumors (95.5%) evaluated. Sox9, a master regulator for cartilage development, was not expressed in any of the 73 other small, round, blue-cell tumors evaluated. In further support of that study, Fanburg-Smith et al31 noted Sox9 expression in 21 of 22 cases of mesenchymal chondrosarcoma (95.5%). In addition, collagen type II was shown to be localized to the extraskeletal matrix of the small cell regions in mesenchymal chondrosarcoma samples and was shown to be a specific and sensitive marker for mesenchymal chondrosarcoma relative to other small, round, blue-cell tumors.32 In a recent study,33 the marker FLI-1 was shown to be useful for excluding mesenchymal chondrosarcoma and for identifying ES when only small, round blue cells are available for study; 6 of 8 ES samples (75%) stained for FLI-1, and 0 of 10 mesenchymal chondrosarcoma samples (0%) stained for FLI-1. Sox9, collagen type II, and FLI-1 may prove to be diagnostic aids, even in mesenchymal chondrosarcoma samples lacking cartilage islands. In another report, the matrix metalloproteinase, MT1-MMP, was found in 29% of mesenchymal chondrosarcoma samples localized to the mesenchymal component.34 This was compared with ES samples, where none exhibited the marker. The authors concluded that, in a subset of patients, these markers may be useful.

Cytogenetic studies demonstrate extreme heterogeneity (see Table 4) and some overlap among small, round, blue-cell tumors.10,13,35–41 Findings in mesenchymal chondrosarcoma that are related to ES include the reciprocal translocation (11;22)(q24;q12)40 and trisomy 8.13,40,42 In one case of mesenchymal chondrosarcoma with tumor cells approaching tetraploidy, unidentifiable, extra material was present on chromosome 22 with a breakpoint at 22q13,35 whereas the ES breakpoint is at 22q12.43 An identical 13;21 Robertsonian translocation was discovered in 2 other cases of mesenchymal chondrosarcoma,37 but this abnormality is not characteristic of any specific tumor. These findings raise the possibility that a small subset of patients with ES may have been misdiagnosed in the event that sample error failed to identify cartilaginous portions of mesenchymal chondrosarcoma.

Prognosis is extremely variable with published 10-year overall survival rates ranging from 21%21 to 67%23 (Table 1). Furthermore, some patients live for long periods with metastatic disease, whereas other patients die shortly after diagnosis. This is further complicated by variations in adjuvant treatment protocols. A number of hypotheses regarding prognosis of mesenchymal chondrosarcoma exist. Hemangiopericytomatous features have been associated with decreased survival.2,21 In the present study, at follow-up, of patients demonstrating predominantly hemangiopericytomatous features, 4 (57%) died, 2 (29%) were alive with disease, 1 (14%) was alive with no disease, and the disease status of the remaining patients was unknown. Improved survival has been reported in patients who receive chemotherapy,21 a finding reflected in the current study in which 71% (5 of 7) of patients who received chemotherapy were alive at follow-up. Other proposed theories include a better chemotherapy response in patients with the small cell variant,2 a negative correlation of proliferation rate with prognosis,44 and that boney origin of tumor is associated with better survival.21 

Management of mesenchymal chondrosarcoma is inadequately studied because of the rare and variable nature of the disease, but treatment patterns have evolved and are identified throughout the literature. The current management consensus underscores the importance of both local and systemic control.21 Surgery is the mainstay of local treatment with studies showing better survival in patients who underwent wide surgical resection.21 All patients in the present study with known management underwent surgery for local control, all but 1 having had wide or radical excision. Even though chemotherapy is somewhat more controversial, most patients now receive adjuvant chemotherapy. In this series, 86% of patients (6 of 7) diagnosed in the past 15 years received chemotherapy. One study21 showed dramatic results supporting the use of chemotherapy: disease-free survival in patients between 5 and 10 years after surgical remission of disease was 76% with chemotherapy and 17% without. Specific drug combinations, however, have not been studied or consistently used. Current National Comprehensive Cancer Network (NCCN) guidelines recommend mesenchymal chondrosarcoma be treated with vincristine, doxorubicin, and cyclophosphamide alternating with ifosfamide and etoposide, a treatment regimen essentially following an ES protocol. In the present study, among the patients who received chemotherapy, 1 regimen was that used for ES, 1 patient received a regimen intended for osteosarcoma, and 4 other patients were given combinations of adriamycin, ifosfamide, and vincristine. This is similar to treatment described in other recent case series.21,23 The efficacy of radiation therapy is even less clearly defined, and its role in local control remains anecdotal.

Future treatment may develop as improved understanding of tumorigenesis evolves. Park et al45 examined p53 expression in 33 mesenchymal chondrosarcoma specimens. Most showed overexpression, and few showed no expression, but no mutations were detected in exons 5 through 9, which are the usual mutation locations in other malignancies. Despite the wild-type sequence, p53 protein was detected abnormally in the cytoplasm, rather than its functional intranuclear location. The authors45 conclude that an epigenetic, functional, deactivation mechanism, such as a cellular or viral protein, may explain the malignant transformation because sequence error cannot explain the aberrant p53 protein localization. Another study46 examining mesenchymal chondrosarcoma tumorigenesis compared gene expression profiles of the mesenchymal component to the cartilaginous component by immunohistochemistry. The investigation was based on the assumption that the mesenchymal component is dedifferentiated and highly malignant, whereas the cartilaginous component is well-differentiated and more indolent. Malignant mesenchymal cells showed increased expression of the platelet-derived growth factor receptor α (PDGFR-α) cell proliferation pathway, suggesting that therapies targeting this pathway may be useful in treating mesenchymal chondrosarcoma. There are several PDGFR inhibitors that could be applied to cases of mesenchymal chondrosarcoma, such as dasatinib, imatinib, sorafenib, and sunitinib.47 Bevacizumab is also known to bind to the tyrosine kinases of PDGFR.47 To our knowledge, no attempt has been documented in the literature to treat mesenchymal chondrosarcoma in humans with this class of drugs, but a theranostic treatment model (a method in which treatment is customized for an individual's disease) using one patient's mesenchymal chondrosarcoma lung metastatic tissue xenografted to mice determined that bevacizumab with docetaxel was more effective than the other tested drugs.48 Fujisawa et al49 performed in vitro and in vivo analyses in nude mice showing dose-dependent inhibitory effect of tumor growth factor β-1 (a cartilage-inducing factor and antiproliferative, depending on the environment) on mesenchymal chondrosarcoma cells in vitro and on the development of pulmonary mesenchymal chondrosarcoma metastases in vivo. The relevance in clinical settings of these findings remains uncertain to date.

The present study is not without limitations. Because of the exceedingly rare incidence of mesenchymal chondrosarcoma, the few patients discussed here preclude statistically significant conclusions. In addition, the 45-year time span reflects changing patterns of systemic medications and protocols, precluding any useful comparison. Follow-up data were also elusive, especially regarding patients referred more than 2 decades ago. Conversely, this article is the largest known case series in the past 10 years of this entity to be reviewed by a single musculoskeletal pathologist. Our case series represents important demographic, clinical, radiographic, and histologic features of mesenchymal chondrosarcoma, which contribute to the overall understanding of the disease. Given the rarity of this tumor and the diagnostic pitfalls discussed, clinicians may want to consider confirming the diagnosis with other markers and obtaining larger biopsy specimens to limit sample error. Future cooperative studies may help standardize treatment and improve our understanding of this rare malignancy.