A phase 2 trial of imatinib mesylate in patients with recurrent nonresectable chondrosarcomas expressing platelet-derived growth factor receptor-α or -β

An Italian Sarcoma Group study




Chondrosarcoma (CS) is a rare and heterogeneous sarcoma in which, after failure of surgery and radiotherapy, chemotherapy plays only a marginal role. Different molecular pathways have been shown to be activated in CS; in particular, both isoforms of platelet-derived growth factor receptor (PDGFR) are expressed and phosphorylated. These observations prompted investigation of the activity of imatinib mesylate (IM) in patients with advanced CS in a phase 2 trial.


Between January 2007 and June 2009, patients with metastatic, nonresectable CS were treated with 400 mg of IM administered twice daily until disease progression or unacceptable toxicity. Two criteria determined patient trial eligibility: ≥1 prior line of chemotherapy and immunohistochemical expression of either PDGFR-α or PDGFR-β. The primary objective of the trial was objective response. As secondary objectives, the authors selected progression-free survival (PFS) at 4 months, overall survival, and clinical benefit (EUDRACT number 2006-006446-33).


Twenty-six patients were enrolled and all demonstrated PDGFR positivity and phosphorylation. No objective response was demonstrated. The 4-month PFS rate was 31% (95% confidence interval [95% CI], 16%-53%). The median overall survival was 11 months (95% CI, 6 months-15 months). Neither long-lasting freedom from disease progression nor clinical benefit was observed. The IM dose was temporarily reduced in 60%15 of the patients because of toxicity.


IM was found to be relatively well-tolerated, but failed to demonstrate meaningful clinical activity in terms of both objective response and freedom from disease progression. Advanced CS remains an incurable disease, and effective targeted therapies are still awaited. Cancer 2011. © 2010 American Cancer Society.

Chondrosarcomas (CS) are a rare and heterogeneous family of mesenchymal tumors characterized by the production of cartilaginous tissue.1 As of 2010, the group has been further characterized into different histological subtypes based on several features: 1) morphology (white cell, mesenchymal, and grading); 2) chromosomal translocations (myxoid); 3) simultaneous presence of various mesenchymal lineages of differentiation (dedifferentiated form); and 4) site of origin within/outside the bone (osseous and extraosseous). Surgery is the mainstay of treatment of CS, with the aim of wide surgical margins.2 In selected cases, radiotherapy may achieve encouraging results when adequate doses are delivered by conformational techniques or proton beam/carbon ion radiotherapy.3, 4 In general, chemotherapy has a marginal role, if any, except in the case of mesenchymal chondrosarcoma.5 Indeed, international guidelines either fail to suggest or point out the minimal effectiveness of medical therapy. Therefore, nonresectable primary presentation (eg, spine localization) and the development of distant metastases are almost invariably fatal events.6

However, in the last decade, the molecular pathways involved in CS biology and progression have been extensively studied and have revealed new information for therapeutic approaches. Pivotal studies have shown tyrosine kinase receptors are involved in tumor aggressiveness. Both the alpha (α) and beta (β) isoforms of platelet-derived growth factor receptor (PDGFR) were found to be correlated with prognosis.7, 8 PDGFR-α, PDGFR-β, and their cognates were identified in CS tissue samples, which confirmed pathway activation.9 Imatinib mesylate (IM) is a small molecule, acting as a highly selective inhibitor of some protein tyrosine kinases, including the chimeric BCR-ABL fusion protein found in certain leukemias such as chronic myeloid leukemia; PDGFR-α and PDGFRβ; and KIT, the receptor for the stem cell factor. Targeted therapies and IM have been among the greatest breakthroughs of the last decade and gastrointestinal stromal tumors (GIST) have provided a paradigm on how to effectively develop targeted therapies.10 IM was later demonstrated to be active in other mesenchymal tumors (dermatofibrosarcoma protuberans and chordoma) in addition to GIST,10-12 because they too involve activation of the PDGF/PDGFR pathway. These findings substantiated that selective inhibition of PDGFR may also halt disease progression in pathologies formerly without effective medical therapy, and prompted us to prospectively explore the activity of IM in patients with advanced CS after standard therapy failure.



The current study was a nonrandomized, multicenter, open-label phase 2 trial in patients with histologically proven CS with nonresectable or metastatic disease. In all patients, the histological diagnosis was centrally reviewed and PDGFR expression was confirmed by immunohistochemical staining. Immunostaining was performed on 4-um thick tissue sections after antigen retrieval by immersion in 0.01 M of citrate buffer (pH 6.0) for 40 minutes. Incubation with the primary PDGFR-α polyclonal antibody (Lab Vision/NeoMarkers, Lab Vision Corporation, Fremont, Calif) and/or PDGFR-β rabbit polyclonal antibody (Thermo Scientific, Fremont, Calif) were performed for 60 minutes at dilutions of 1:150 and 1:25, respectively. Detection was made using the Envision system (Dakocytomation Inc, Carpinteria, Calif), a polymer-based, biotin-free detection system, for 30 minutes followed by the chromogen 3,3-diaminobenzidine for 10 minutes. The presence of membranous and/or cytoplasmic immunostaining was determined semiquantitatively (ie, <30%, 30%-60%, and >60% positive tumor cells were graded as 1, 2 or 3, respectively). All patients were aged ≥18 years and had received at least 1 line of chemotherapy that was completed >4 weeks before trial entry. In addition, any radiotherapy had to be completed >1 month before trial entry. Patients were required to have a life expectancy of ≥3 months and an Eastern Cooperative Oncology Group performance status <3.

Treatment Schedule and Evaluation

All patients received IM at a dose of 400 mg twice a day. According to toxicity, the dose could be reduced or temporarily suspended until recovery from the observed adverse event/s. Therapy was resumed at the maximum tolerated dose according to the physician's decision.

Before the initiation of treatment, all patients were staged using computed tomography scans of the chest and abdomen, magnetic resonance imaging (if limb lesions were involved), full blood count and serum chemistry, electrocardiogram, and physical examination. All tests were repeated after 2 months and at 3-month intervals thereafter until disease progression or unexpected toxicity occurred. Toxicity was assessed according to the National Cancer Institute Common Terminology Criteria for Adverse Events (version 3.0).13 Treatment was continued until disease progression or unacceptable toxicity.

Outcome Evaluation

Response was assessed according to the Response Evaluation Criteria in Solid Tumors (RECIST) (version 1.0).14 According to these criteria, complete response (CR) and partial response (PR) required confirmation within 4 weeks after a response was first demonstrated whereas stable disease (SD) required confirmation within ≥8 weeks. Nondimensional patterns of tumor response were thoroughly examined in each patient by checking for changes in density and/or contrast medium uptake in target as well as nontarget lesions.15

Progression-free survival (PFS) was computed from the first date of study treatment to the date of documented disease progression according to RECIST or to the date of death from any cause, whichever came first. Survival was computed from the first date of treatment to the date of death. Clinical benefit was assessed by means of the Pain Analgesic Score (PAS). Pain was evaluated by a score obtained using the Brief Pain Inventory Score (BPI) form. Analgesic medication use was recorded according to the PAS.16

Statistical Analysis

We designed this exploratory trial to detect clinically significant activity of IM in an “orphan” disease for which, to the best of our knowledge, no trials had been conducted in 20 years. Bearing in mind this framework, any predefined level of activity for a null hypothesis, as well as for a hypothesis of interest, would have been completely arbitrary and, as such, a mere statistical trick. Therefore, we selected as a primary endpoint objective response and, as a secondary endpoint, and above all, PFS, under the hypothesis that any shrinkage or prolonged stability should be regarded as a treatment success. To mitigate the potential that CS heterogeneity might jeopardize the observed results, we deliberately chose a relatively large sample (25 patients) to detect any activity in terms of objective response. Such an approach would maximize the detection of differential activity within the main subentities that comprise the CS family of tumors. At the same time, we turned to secondary endpoints (PFS, survival, and BPI) to identify any other drug effect on the disease. Notwithstanding the previously mentioned limitations to setting precise statistical parameters, we calculated the sample size according to a 2-stage design: we assumed an expected response rate of 10% from a second-line chemotherapy and an activity of IM of 30%. Therefore, if no responses had been detected in the first 25 patients with type I and type II error rates set at 0.05 and 0.10, respectively, then the probability of having an active drug would have had to have been <10%.17 All patients who received at least 1 dose of IM were included in an intention-to-treat analysis and the results of treatment were expressed as a response rate with a 95% confidence interval (95% CI). PFS and overall survival were estimated according to the Kaplan-Meier method.


Patient Characteristics

Between January 2007 and June 2009, 26 patients diagnosed with metastatic and/or nonresectable CS were enrolled at 6 centers. All patients had received at least 1 prior line of chemotherapy and their median age was 61 years (range, 23-81 years). We included 16 patients diagnosed with conventional CS and 10 patients diagnosed with bone CS with prominent myxoid features (CSm). All 26 enrolled patients demonstrated immunohistochemical PDGFR positivity and phosphorylation, which was cytoplasmic in all cases, with occasional concomitant membrane labeling noted in 3 (12%) cases. Patient characteristics are described in Table 1.

Table 1. Patient Characteristics
  • ECOG indicates Eastern Cooperative Oncology Group.

  • a

    Chondrosarcoma of bone with prominent myxoid features.

  • b

    There were patients with disease progressing at both sites.

Median age at study entry (range), y61 (23-81) 
ECOG performance status
Location of primary tumor
Chondrosarcoma histotype
First-line chemotherapy
Site of disease progressionb

Study Treatment

Of the 26 patients screened to enter into the study, all had IM initiated at a dose of 400 mg twice a day and were included in the analysis for toxicity and response. No patients had to permanently discontinue treatment because of side effects. The median dose administered of imatinib was 83% (range, 74%-91%) of that expected.

Response to IM

We did not observe any CRs or PRs by either dimensional criteria (RECIST) or nondimensional criteria. Among the 26 patients, there were 8 (31%) with SD and 18 (69%) with progressive disease. The median PFS was 3 months (95% CI, 2.2 months-7.7 months) (Fig. 1). The PFS rate at 4 months was 31% (95% CI, 16%-53%). The median overall survival was 11 months (95% CI, 6 months-15 months) (Fig. 2) and 10 (39%) patients remained alive at 1 year of follow-up. Objective response, PFS, and overall survival did not differ between patients with conventional CS and those with CSm (Table 2). Among those patients who achieved SD, we did not observe any long-lasting freedom from disease progression. We were unable to demonstrate any trend in favor of patients with membranous PDFGR-positive staining. Finally, BPI was assessable in 24 of the 26 patients. We could not demonstrate any significant clinical benefit in patients, even among those who achieved SD (defined as a mean basal BPI of 4.2 compared with a mean best response BPI of 4.8; P = .18).

Figure 1.

Kaplan-Meier plot for progression-free survival is shown.

Figure 2.

Kaplan-Meier plot for overall survival is shown.

Table 2. Outcome Variables
  • RECIST indicates Response Evaluation Criteria in Solid Tumors; PDGFR, platelet-derived growth factor receptor; BPI, Brief Pain Inventory.

  • a

    Skeletal chondrosarcomas with myxoid features.

  • b

    Progression-free survival difference between membranous and cytoplasmic positivity.

  • c

    Difference between basal and best response BPI.

Response (RECIST)
 Complete response00
 Partial response00
 Stable disease831
 Progressive disease1869
Median progression-free survival, mo3 
Progression-free survival at 4 mo831
Median overall survival, mo11 
Conventional chondrosarcoma progression-free survival, mo2.6 
Myxoida chondrosarcoma progression-free survival, mo3.6 
PDGFR membranous vs cytoplasmic positivity, difference and p valueb0.9.23
BPI scale, difference and p valuec0.6.18


In general, IM was relatively well tolerated by CS patients as a second-line therapy at a dose of 800 mg per day. Adverse events were observed in 18 (69%) patients, and dose reductions were required in 15 (58%) patients. We did not observe any drug-related deaths. Edema was the most common grade 3 to 4 toxicity, and was noted in 3 (12%) patients, followed by skin rash in 2 (8%) patients, leukopenia in 2 (8%) patients, and muscle cramps in 1 (4%) patient. As expected, the observed side effects were reversible either with dose reductions or short drug interruptions.


Although patients with CS have benefited from improvements in both surgical and radiotherapy techniques, as of 2010, systemic therapies still failed to control the disease in the advanced phase. Indeed, chemotherapy has demonstrated some activity in specific CS subtypes only, among them mesenchymal and dedifferentiated CS.18, 19 Recent insights into CS biology8, 9 have shed some light on the mechanisms behind the proliferative advantage of this tumor and generated the hypothesis that a potential CS therapy may lay in targeting the PDGFR pathway. In fact, PDGFR has already been successfully targeted by IM in other mesenchymal tumors.10-12, 20 Moreover, in the current study, we prospectively demonstrated that in all treated patients, PDGFR was phosphorylated, suggesting its activation, which strengthened the initial hypothesis. However, this multicenter clinical trial failed to demonstrate clinically significant activity of IM in patients with CS. Because drug activity may display different patterns of response, we prospectively collected different endpoints beyond standard objective response as a clinical benefit by means of BPI, a validated tool. This not withstanding, we were unable to demonstrate IM activity using any of the selected endpoints.

It is well known that the expression of a target is a necessary, but not sufficient, condition to elicit a response to targeted therapies.21 The failure to demonstrate an improvement with this small inhibitor might be because of several reasons. As Lagonigro et al9 indicated, no PDGFR mutations were shown in their series. However, these data may only partially explain the inactivity of the drug because even if chordoma does not display PDGFR mutations, IM has been shown to be active.22 More likely, other molecular mechanisms are responsible for IM failure. The best available model for the molecular mechanisms driving response or resistance to targeted therapies, namely activation of downstream effectors23 or alternative growth factor transduced activation,24 remains that demonstrated for GIST.

Moreover, in GIST, the likelihood of developing resistance to tyrosine kinase receptor (TKR) inhibitors depends not only on specific KIT and PDGFR-α mutations but also on the accumulation of chromosomal abnormalities.25 Recent data26 have demonstrated that CS harbor genomic unbalances in 90% of cases with some recurrent amplifications and homozygous deletions. These observations raise the suspicion of a sarcoma that is too complex to be controlled by shutting off only 1 of its driving mechanisms.

Despite the lack of activity shown by IM, the results of the current study indicate 2 potential lines of therapy for future study. The first centers on whether the combination of IM and a chemotherapeutic agent with low to moderate activity, such as doxorubicin, might enhance the effectiveness of both drugs, as observed when patients with GIST fail IM therapy.27 In the absence of other treatment strategies, this combination might deserve further attention. The second hypothesis, recently suggested by Schrage et al,28,;29 proposes that the multikinase inhibitor dasatinib and the cyclooxygenase (COX)-2 inhibitor celecoxib might be appropriate because of activation of relevant CS pathways. Finally, other approaches that will need to be considered in the future rely on the blockade of other pathways known to be important in chondrogenesis, such hedgehog signaling, which is in a negative feedback loop with parathyroid hormone-related protein.30

In summary, as expected, IM was found to be relatively well tolerated but was inactive in terms of both freedom from disease progression and tumor shrinkage. Advanced CS still remains an incurable disease for which targeted therapies must be pursued to overcome its chemoresistance and improve the stagnant dismal prognosis of this tumor.


Sponsored by the Italian Sarcoma Group through an unrestricted grant by Novartis Inc. Drs. Grignani, Aglietta, Comandone, Frustaci, and Ferraresi received reimbursement for meeting participation from Novartis. Drs. Stacchiotti and Casali received funds for clinical studies and research activities in which their institution is involved, acted in a consultant/advisory role, and received travel coverage for medical meetings from Novartis.