Fax: (646) 422-2076
A multicenter Phase II study of bortezomib in recurrent or metastatic sarcomas†
Article first published online: 28 FEB 2005
Copyright © 2005 American Cancer Society
Volume 103, Issue 7, pages 1431–1438, 1 April 2005
How to Cite
Maki, R. G., Kraft, A. S., Scheu, K., Yamada, J., Wadler, S., Antonescu, C. R., Wright, J. J. and Schwartz, G. K. (2005), A multicenter Phase II study of bortezomib in recurrent or metastatic sarcomas. Cancer, 103: 1431–1438. doi: 10.1002/cncr.20968
Presented, in part, at the 39th Annual Meeting of the American Society of Clinical Oncology, Chicago, Illinois, May 31–June 3, 2003 (abstract 3291).
- Issue published online: 18 MAR 2005
- Article first published online: 28 FEB 2005
- Manuscript Accepted: 29 NOV 2004
- Manuscript Received: 17 NOV 2004
- National Cancer Institute (CTEP protocol 1757). Grant Numbers: P01-CA47179, N01-CM17105
- Robert and Deborah Bloch Sarcoma Research Fund
- Phase II;
Based on evidence of activity in preclinical and Phase I studies, the authors undertook a study of bortezomib, a reversible proteasome inhibitor, for patients with metastatic sarcomas.
Two arms were opened, each using a Simon two-stage design. Arm A included patients with osteogenic sarcoma, Ewing sarcoma, and rhabdomyosarcoma. Arm B accrued patients with other types of soft tissue sarcomas. Patients were not allowed to have received previous chemotherapy for metastatic disease. The initial dose of bortezomib was a 1.5 mg/m2 intravenous push twice weekly followed by a rest week. The dose was escalated to 1.7 mg/m2 if patients tolerated Cycle 1 well. The dose escalation was eliminated due to toxicity observed in the first six patients.
Painful neuropathy, myalgias, and asthenia were the most significant observed toxicities. The most frequent toxicities included fatigue, diarrhea, constipation, and nausea. Pharmacodynamic data from 18 patients with complete data collection did not show consistent differences between patients with or without Grade 2 or Grade 3 neuropathy (toxicity graded according the National Cancer Institute Common Toxicity Criteria). Arm A had low accrual and was closed. One confirmed partial response among 21 evaluable patients was observed on Arm B in a patient with leiomyosarcoma. Due to the inactivity of this agent, the study was closed after the first stage of accrual.
Bortezomib has minimal activity in soft tissue sarcoma as a single agent. If studied further in sarcomas, bortezomib should be investigated in combination with agents with demonstrated preclinical synergy. Cancer 2005. © 2005 American Cancer Society.
More than 50 varieties of sarcomas are recognized, and each has a unique pattern of metastasis, disease recurrence, and sensitivity to chemotherapy.1 Although adjuvant chemotherapy is mandatory for the management of osteogenic sarcoma, Ewing sarcoma, and rhabdomyosarcoma, the use of adjuvant chemotherapy for extremity soft tissue sarcomas remains controversial.2 Approximately one-half of patients with soft tissue sarcoma will experience disease recurrence and die of metastatic disease. Metastatic or recurrent sarcomas of bone and soft tissue remain difficult to treat. Response rates to standard chemotherapy agents are low, and responses are not durable.3 New therapies are needed to determine activity in the metastatic setting, which can then be moved to the adjuvant setting, much as is being done with imatinib for patients with gastrointestinal stromal tumors (GIST).
Bortezomib (PS-341, LDP-341) is a selective, potent, and reversible inhibitor of the proteasome, and has been approved for use in the U.S. and Europe for recurrent or refractory multiple myeloma.4 Bortezomib also has significant activity in the treatment of mantle cell lymphoma and other forms of non-Hodgkin lymphoma.5 It is believed that disruption of the many physiologic processes dependent on proteasome function is the reason for the activity of bortezomib in these malignancies, such as regulation of cell cycle proteins, p53, and nuclear factor-kappa beta (NF-κβ) signaling pathways.6
MATERIALS AND METHODS
We conducted the current multicenter study using a protocol reviewed and approved by the Cancer Therapeutics Evaluation Program of the National Cancer Institute (NCI). The protocol was also reviewed and approved by the Memorial Sloan-Kettering institutional review board and by the institutional review board at the cooperating centers. Written informed consent was obtained from all patients. Patients were enrolled on one of two arms of the study based on histology. Arm A included tumors typically found in pediatric patients more commonly than in adults, namely, Ewing sarcoma, rhabdomyosarcoma, and osteogenic sarcoma. Arm B included soft tissue sarcomas not included in Arm A (i.e., histologies typically found in adult populations).
Patients were eligible if they met the following criteria: 1) sarcoma diagnosis histologically or cytologically confirmed at the treating institution; 2) measurable disease defined by response evaluation criteria in solid tumors (RECIST); 3) radiologically documented evidence of disease progression within the 3-month period before study entry; 4) 1 previous line of adjuvant chemotherapy (Arm A), or no previous chemotherapy except adjuvant therapy > 1 year before study entry (Arm B); 5) recovery from toxicity of previous therapy; 6) age ≥ 18 years; 7) Eastern Cooperative Oncology Group performance status of 0 or 1; and 8) normal organ function, as indicated by an absolute neutrophil count of ≥ 1500/μL, platelet count ≥ 100,000/μL, a total bilirubin level ≤ 1.5 mg/dL, aspartate aminotransferase and alanine aminotransferase level ≤ 2.5 times the institutional upper limit of normal, and serum creatinine level ≤ 1.5 mg/dL.
Subjects were excluded if they 1) had received chemotherapy, radiotherapy, or immunotherapy within 4 weeks (6 weeks for nitrosoureas or mitomycin C) before entering the study; 2) were actively receiving other anticancer therapy; 3) had a history of another neoplastic disease within the past 5 years, excluding basal cell carcinoma of the skin or cervical carcinoma in situ, that was treated adequately; 4) had a history of known brain or active brain metastases; 5) were pregnant or nursing; 6) refused to use adequate contraception; 7) had significant psychiatric or social situations likely to limit compliance with the study; or 8) had an uncontrolled intercurrent illness (including human immunodeficiency virus infection) or history of significant atherosclerotic disease. There was no restriction on previous neuropathy because patients were not to have been previously treated for metastatic disease.
Twenty-five patients were enrolled between November 2001 and November 2003, 4 on Arm A (1 was ineligible) and 21 were enrolled on Arm B. Patients were enrolled at 3 institutions: Memorial Sloan-Kettering Cancer Center (n = 22), University of Colorado Health Sciences Center (n = 2), and New York Presbyterian Hospital/Cornell Medical Center (n = 1).
Patients were assigned to treatment Arm A or B, then bortezomib was administered at an intravenous dose of 1.5 mg/m2 per dose by intravenous push twice weekly for 2 weeks, followed by 1 week of rest. A cycle of therapy consisted of 4 doses of bortezomib over 21 days. Dose escalation to 1.7 mg/m2 per dose was permitted initially beginning in Cycle 2 for patients not experiencing > Grade 2 toxicity during Cycle 1. This dose escalation step was discontinued when 2 of the first 6 patients developed Grade 3–4 neuropathy, myalgias, or fatigue during treatment. All treatment was given in an outpatient setting. Antiemetic drugs were not routinely given at the start of therapy, because there was no significant nausea in Phase I studies. Patients who developed nausea or emesis with bortezomib therapy received prochlorperazine 10 mg orally before each dose of bortezomib, or 8 mg of ondansetron for refractory nausea or emesis.
Patients received a minimum of two cycles of therapy unless unacceptable toxicity or disease progression occurred. Patients were evaluated weekly for toxicity graded according to the NCI-Common Toxicity Criteria (CTC) Version 2.0. In addition, a specific worksheet regarding neurologic toxicity was completed at each patient visit. With the exception of alopecia, a patient experiencing Grade 3 or 4 toxicity attributable to therapy had subsequent therapy delayed until resolution of toxicity to < Grade 2. Treatment with bortezomib was then resumed at 1 dose level lower than the previous dose (dose level −1 was 1.3 mg/m2, dose level −2 was 1.0 mg/m2). If treatment was delayed > 21 days or required a dose reduction below dose level −2, the patient was removed from study.
Evaluation of Response and Toxicity
Disease reevaluation was performed using the same modality used for baseline measurements every two cycles for the first six cycles of therapy, then every three cycles thereafter. RECIST was used as the basis for disease reevaluation.9 Toxicity was assessed using the NCI-CTC, Version 2.0.
Each arm (Arm A and Arm B) was designed as an individual Phase II two-stage study, using the Simon-optimized design.10 Each arm was designed so that if bortezomib has a response rate of < 5%, the drug would not be considered worthy of further study in that group. For each group, the type I error was 0.05 and the type II error was 0.10. In the first stage, 21 eligible patients were to be enrolled. If there was 0 or 1 response in the first 21 eligible patients, the drug was not to be examined further because of the lack of utility. If there were ≥ 2 responses in the first 21 eligible patients, accrual on each group was to be increased to 41 eligible patients. The probability of failing to reject a treatment with a ≤ 5% response rate was 0.05. If the drug had a true response rate of 5%, the chance of stopping the study after 21 eligible patients was 72%. The probability of rejecting a drug with a response rate of ≥ 20% was 0.10. Bortezomib was considered promising for further development if ≥ 5 responses in 41 eligible patients were noted. After accruing 21 patients to Arm B, the study was closed after the first stage for lack of activity in Arm B and poor accrual in Arm A (n = 4, including 1 ineligible registrant).
Proteasome Inhibition Assay
Whole blood specimens were collected from all patients immediately before the first dose of therapy, 1 hour after, and 24 hours after the first dose of therapy, as well as before, 1 hour after, and 24 hours after the fourth dose of bortezomib therapy (the last dose of the first cycle). Chymotryptic/tryptic activity (ChT:T) and specific chymotryptic activity (SpA) levels were measured in assays as described by Lightcap et al.11 Samples were collected, immediately put on wet ice, and stored at −70 °C. Sample identification was entered into a password-protected database. Samples were batched and sent to Millennium Pharmaceuticals, Inc. (Cambridge, MA) for analysis.
Patient characteristics are indicated in Table 1. Because only four patients were enrolled in Arm A (pediatric sarcomas), the patient cohorts are considered together. A total of 25 patients were eligible and registered for therapy between November 2001 and November 2003. All four patients on Arm A and one on Arm B had received previous chemotherapy (doxorubicin/ifosfamide in three patients, cisplatin/doxorubicin in one patient, and MAID [doxorubicin, ifosfamide, dacarbazine, mesna] in one patient). There were no apparent sequelae of previous chemotherapy, which had been given > 1 year before enrollment in the study save for 1 patient (6 months before enrollment). One patient withdrew consent before treatment. A second patient was found ineligible for Arm A (pediatric histology) after starting therapy, on the basis of previous therapy for metastatic disease as well as a histologic diagnosis that made the patient ineligible for that arm of the study. This patient was included in the analysis of toxicity, but not for the analysis of response. Two patients were removed from the study for toxicity before completing one cycle of therapy without evidence of progression of disease. Thus, 24 patients received therapy, 23 were evaluable for toxicity, and 21 patients were evaluable for response between the 2 arms of the study.
|Median age (range) (yrs)||56 (22–79)|
|Median KPS||90 (90–100)|
|No. of patients who received previous adjuvant therapy > 1 yr before enrollment||3|
|Patients with previous therapy for metastatic disease||1a|
|Locally advanced disease only||1|
|Metastatic disease only||11|
|Head and neck|
|MFH/high-grade sarcoma NOS||6b|
|Alveolar soft parts sarcoma||2|
|Follicular dendritic cell sarcoma||1|
|Extraskeletal osteogenic sarcoma||1|
|Gastrointestinal stromal tumor||1|
The distribution of patients by age, tumor type, and primary site is similar to a cross-section of patients seen in our clinical practice and was not skewed toward one or another subtype of sarcoma or site of origin. Median performance status at the start of treatment was excellent (the Karnofsky performance score was 90%).
Given the significant toxicity observed in three of the first six patients, the dose escalation step of the protocol was eliminated. Of the 23 patients who received bortezomib therapy and were eligible for toxicity evaluation, 2 had toxicity (Grade 3 fatigue and Grade 2 neurotoxicity) that led to their withdrawal of consent before the end of Cycle 1 of therapy without evidence of disease progression. The patient rapidly developing peripheral neuropathy had neither diabetes nor other comorbid illnesses that might predict its development.
Of the remaining 21 patients eligible for response receiving ≥ 1 cycle of therapy, all but 1 were able to tolerate ≥ 2 cycles of therapy (8 doses). Three patients progressed clinically before the first restaging scans. The median number of treatment cycles was 2 (range, 1–7). Dose reductions for toxicity were necessary for 3 patients in Cycle 1 and for 6 of 18 patients during Cycle 2. The overall response rate was 5% (95% confidence interval [95% CI], 0–10%), with 1 confirmed partial response (PR) in a patient with a retroperitoneal leiomyosarcoma metastatic to the soft tissue, liver, and spleen. This patient received therapy for only 1 further dose of bortezomib after the PR, and was removed at her request for Grade 2 neurotoxicity. She had a RECIST-confirmed PR in a subsequent computed tomographic scan. The responding patient had a leiomyosarcoma of the retroperitoneum with relatively low volume disease to the lungs, paraspinal muscle, bone, and spleen. The PR was largely due to complete resolution of a mass in the spleen whereas other sites responded modestly or remained stable. A biopsy of the splenic mass was not performed to confirm her diagnosis of metastatic sarcoma, and the mass did not progress when she later demonstrated disease progression in the liver, raising the possibility that it was not metastatic leiomyosarcoma. No other patients with leiomyosarcoma had prolonged disease-free progression (i.e., none received treatment for more than four cycles of bortezomib chemotherapy).
At the time of last follow-up (September 2004), 19 patients were alive and 6 had died. The median survival was 10.1 months (range, 0.7–30.2 months). Eight of 21 evaluable patients had a best clinical result of stable disease (SD) and 13 patients had disease progression after a median of 1.4 months (range, 0.9–4.2 months). Seven of the patients who developed disease progression while receiving bortezomib had SD or a PR after receiving doxorubicin or ifosfamide-based chemotherapy subsequently. Because of the lone response in the first 21 patients enrolled on the adult sarcoma histology arm, the study was closed for lack of efficacy on Arm B as well as the lack of accrual on Arm A despite opening the study at an additional 2 centers. The 23 patients evaluable for toxicity received a total of 56 cycles of therapy, 7 cycles of 1.7 mg/m2 per dose, 34 cycles of 1.5 mg/m2 per dose, 7 cycles of 1.3 mg/m2 per dose, and 8 cycles of 1.0 mg/m2 per dose, for a net dose intensity of 1.4 mg/m2 per dose.
CTC Version 2.0 toxicity is indicated in Table 2. The most common side effects were fatigue, diarrhea, constipation, nausea, and neuropathy. All but 1 patient assessable for toxicity (22 of 23) experienced ≥ Grade 1 fatigue (12 with Grade 1, 8 with Grade 2, and 1 each with Grade 3 and Grade 4), with 1 patient removed from study for Grade 4 fatigue after 1 cycle of therapy.
|Event||Gradeb 1||Grade 2||Grade 3||Grade 4||Any grade|
|Infection without neutropenia||1||1||2|
|Deep venous thrombosis||1||1|
Grade 4 toxicity was infrequent, and included fatigue (1 patient) and myalgias (1 patient). Grade 3 toxicities included diarrhea (n = 2) and neuropathic pain (n = 2) as well as single episodes of fatigue, constipation, nausea/emesis, abdominal pain, infection, dyspnea, deep venous thrombosis, and tumor pain.
Myalgias occurred in 7 patients (i.e., Grade 1 in 3 patients, Grade 2 in 3 patients, and Grade 4 in 1 patient) that required discontinuation of therapy. The patients with lesser myalgias typically had proximal muscle aches that lasted 3–7 days after therapy with resolution by the start of the subsequent cycle of therapy.
Neuropathy was largely sensory in nature, and was painful, exacerbated by cold, and frequently required narcotics given with other adjuvants such as amitriptyline or gabapentin to maximize control of the discomfort. Eleven patients had some evidence of neuropathy (i.e., Grade 1 in 6 patients, Grade 2 in 3 patients, and Grade 3 in 2 patients). The neuropathy resolved at least in part over time. The patients with Grade 2 or Grade 3 neuropathy were still using narcotics > 8, > 12, 15, > 16, and > 30 months after stopping bortezomib therapy. The patients with neuropathy requiring narcotics had received a total of five, one, two, three, and four cycles of bortezomib therapy, respectively.
Whole blood specimens were examined to evaluate inhibition of the 20S subunit of the proteasome before and after the first and last dose of bortezomib during Cycle 1, to examine the degree of inhibition at the start and end of the cycle, and to examine any relation between 20S proteasome inhibition and response or toxicity (Fig. 1). Three samples were taken at baseline, and blood samples were taken 1 and 24 hours after the first dose of bortezomib in Cycle 1, as well as just before, 1 hour after, and 24 hours after the fourth (last) dose of bortezomib in Cycle 1.
In all, 18 patients had ≥ 3 samples analyzed for 20S proteasome inhibition and are included in the analysis. Only 7 patients gave samples 24 hours after the last dose of therapy for the cycle was received. The remainder of the patients declined to return to the clinic for the follow-up test. Table 3 and Figure 1 show the results of SpA and ChT:T for the whole blood samples before and after bortezomib therapy. The peak inhibition of 20S proteasome was approximately 60% by either the SpA or ChT:T assays, consistent with the pharmacodynamics of this compound observed in previous Phase I and Phase II studies.
|Characteristics||Before therapy||1 hr after 1st dose||24 hrs after 1st dose||Before 4th dose||1 hr after 4th dose||24 hrs after 4th dose|
|SpA: % inhibition (mean ± SD)||0 ± 0||60 ± 9||22 ± 13a||0 ± 1||61 ± 9||30 ± 13b|
|ChT/T ratio (mean ± SD)||−1 ± 1||61 ± 8||32 ± 14a||0 ± 1||58 ± 11||28 ± 16b|
As has been observed previously in other Phase I and Phase II studies with bortezomib, we observed a statistically significant decrease in 20S proteasome inhibition 1–24 hours after each dose of bortezomib by SpA or ChT:T assay (P < 0.001 after the first dose and P < 0.04 after the fourth dose, Wilcoxon signed rank test). In contrast, there was not a statistically significant difference between SpA or ChT:T comparing the 24-hour time points after the first and last doses of bortezomib (P = 0.08 for SpA, P = 0.94 for ChT:T, Wilcoxon). Thus, in the seven patients examined, there did not appear to be factors affecting bortezomib inhibition of the proteasome over the course of the first cycle of treatment, such as a change in bortezomib clearance.
We analyzed the patients with or without Grade 2 or Grade 3 neuropathy to determine if their degree of 20S proteasome inhibition correlated with the development of neuropathy. Data were available to compare 20S proteasome inhibition in 4 of the 5 patients with neuropathy with 14 patients who did not develop Grade 2 or Grade 3 neuropathy. There was an inverse trend between neurotoxicity and 20S proteasome inhibition 1 hour after the first dose of bortezomib (SpA 52% [neuropathy] vs. 61% [no neuropathy], 2-tailed t test, P = 0.08). The difference in 20S proteasome inhibition by SpA was significant in the 24-hour time point after the first treatment (34% [neuropathy] vs. 18%, [no toxicity], P = 0.003). However, the association was not observed in the proteasome inhibition data 1 hour after the fourth dose of the cycle of therapy (65% [neuropathy] vs. 60% [no neuropathy], P = 0.33).
20S proteasome inhibition as determined by ChT:T activity also yielded inconsistent results with respect to an association with neuropathy. The inhibition 1 hour after the first dose of therapy was 52% (neuropathy) versus 63% (no toxicity), P = 0.04, in contrast to the data from 1 hour after the fourth dose (67% [neuropathy] vs. 55% [no toxicity], P = 0.008). At 24 hours after the first dose of therapy, there was no statistically significant difference between patients who developed neuropathy versus those who did not (38% [neuropathy] vs. 29% [no toxicity], P = 0.16). The responding patient did not have a complete data set to assess the relation of her response to 20S proteasome inhibition.
Systemic treatment of recurrent or metastatic sarcoma remains inadequate. Doxorubicin and ifosfamide may each yield a 20% response rate by RECIST, but even this may be an overestimate when radiographic review is centralized.12 The excitement of targeted therapy such as imatinib for all sarcomas is tempered by the lack of responses seen in most sarcoma subtypes, except GIST and dermatofibrosarcoma protuberans.13–15 As for other novel agents for sarcomas, significant responses have been observed in patients receiving ecteinascidin 743 (ET-743), but most of these responses have been limited to 2 of the more than 50 subtypes of sarcoma, namely, liposarcoma and leiomyosarcoma.16
Only 1 response to bortezomib among 21 evaluable patients occurred in the current study. Toxicity was significant, and led to the elimination of a dose escalation built into the study after the first cycle of therapy. The doses of bortezomib used in our study were higher than the dose approved for use in multiple myeloma. Five patients developed Grade 2 or Grade 3 painful neuropathy and 1 developed myalgias that were difficult to manage. These patients required both narcotics and other adjuvants to minimize the discomfort, and some patients continue to have symptoms and continue to receive medication > 1 year after stopping bortezomib therapy.
In the current study, there appeared to be no direct relation between the maximum dose of bortezomib delivered and the development of neuropathy, because 3 of 18 patients without peripheral neuropathy received ≥ 1 dose at the escalated 1.7 mg/m2 dose, whereas 1 of 5 patients with Grade 2 or Grade 3 neuropathy received a dose escalation for ≥ 1 dose. Only 5 of 23 evaluable patients had received previous adjuvant chemotherapy (and only 1 with cisplatin-containing chemotherapy, without sequelae). Thus, peripheral neuropathy did not appear to be related to previous exposure to neurotoxic chemotherapy. Because most patients progressed quickly while receiving this regimen, it was difficult to confirm this finding. The neurotoxicity was consistent with the data from previous Phase I and Phase II studies, in which it ranged from 0% to 47%.
In a carefully performed Phase I study involving mostly patients with prostate carcinoma, there was no clear association between serum bortezomib level and the development of toxicity.8 Rather, the degree of proteasome inhibition predicted the development of side effects, specifically diarrhea, constipation, vomiting, hypotension, and fatigue. It will be particularly important to address the pharmacodynamics and toxicity of bortezomib in patients who receive this drug for longer courses of therapy, albeit at lower doses, such as responding patients with myeloma or non-Hodgkin lymphoma.17, 18
The pharmacodynamics of bortezomib are similar to those in other Phase I and Phase II studies that examined 20S proteasome in blood mononuclear cells and in tissue specimens.17–19 Approximately 60% inhibition of 20S proteasome activity was observed 1 hour after the first and last doses of bortezomib in Cycle 1, and 20S proteasome activity returned to normal before the fourth dose of the cycle was given. There do not appear to be cumulative effects of the 4 doses given over 2 weeks, because the level of proteasome inhibition was the same 1 hour and 24 hours after the first and fourth doses of bortezomib. In the Phase I study of Papandreou et al.,8 proteasome inhibition was saturated, regardless of a bortezomib dose in the 70–75% range. therefore, 20S proteasome inhibition may not be the ideal surrogate serum drug level. However, in this same study, the degree of 20S proteasome inhibition correlated to the development of side effects.
If bortezomib is to be developed in the future for patients with sarcomas, it should be used in combination with other agents, or with a better understanding of a particular subtype of patients who might respond to bortezomib monotherapy or combinations, such as patients with leiomyosarcoma, based on the responding patient from the current study. Doxorubicin has been observed to increase the sensitivity of a myeloma cell line to subsequent bortezomib exposure.20 In addition, there appears to be synergy between bortezomib and gemcitabine in a bladder carcinoma model.21 Such studies point to possible new combinations that may merit investigation in patients with sarcoma. The side effects in this and other Phase I and Phase II studies with bortezomib also prompt consideration of new schedules of administration for patients with sarcomas and other solid tumors.
The authors are grateful to the patients who enrolled in the current study and to their families and friends who supported them. They also thank the Bioassay Laboratory at Millennium Pharmaceuticals, Inc., for their assistance in performing the 20S proteasome assays. They are grateful to Dr. Anthony Elias at the University of Colorado Health Sciences Center for management of the protocol after the departure of Dr. Kraft from that institution.
- 1Soft tissue sarcoma. In: DeVitaVTJr., HellmanS, RosenbergSA, editors. Cancer: principles and practice of oncology. 6th edition. Philadelphia: Lippincott Williams & Wilkins, 2001: 1841–1891., , .
- 9New guidelines to evaluate the response to treatment in solid tumors. European Organization for Research and Treatment of Cancer, National Cancer Institute of the United States, National Cancer Institute of Canada. J Natl Cancer Inst. 2000; 92: 205–216., , , et al.
- 13Imatinib mesylate (STI-571 Glivec, Gleevec) is an active agent for gastrointestinal stromal tumours, but does not yield responses in other soft-tissue sarcomas that are unselected for a molecular target. Results from an EORTC Soft Tissue and Bone Sarcoma Group phase II study. Eur J Cancer. 2003; 39: 2006–2011., , , et al.