Document 6447024
Transcription
Document 6447024
Carboplatin/Ifosfamide Window Therapy for Osteosarcoma: Results of the St Jude Children’s Research Hospital OS-91 Trial By William H. Meyer, Charles B. Pratt, Catherine A. Poquette, Bhaskar N. Rao, David M. Parham, Neyssa M. Marina, Alberto S. Pappo, Hazem H. Mahmoud, Jesse J. Jenkins, James Harper, Michael Neel, and Barry D. Fletcher Purpose: To determine the activity of carboplatin/ifosfamide in patients with previously untreated osteosarcoma and to estimate patient outcomes after a multiagent chemotherapy protocol that eliminated cisplatin. Patients and Methods: Sixty-nine patients with newly diagnosed, previously untreated osteosarcoma received three cycles of carboplatin (560 mg/m2 ⴛ 1) and ifosfamide (2.65 g/m2/d ⴛ 3). Assessment of response was evaluated after two (week 6) and three (week 9) chemotherapy cycles. At week 9, histologic response was assessed. Adjuvant therapy comprised two additional carboplatin/ifosfamide cycles, doxorubicin, and high-dose methotrexate. Patients were stratified at enrollment: stratum A, resectable primary tumor without metastases; stratum B, unresectable primary tumor; and stratum C, metastatic disease at diagnosis. Week 6 response was compared with that of a historic group that received only ifosfamide during the initial window evaluation. Results: The clinical and radiographic response rate to three cycles of carboplatin/ifosfamide was 67.7% (95% confidence interval, 55.0% to 78.8%). Compared with the historic population who received only ifosfamide, the combination of carboplatin and ifosfamide reduced the progressive disease rate at week 6 (31.9% v 9%, P ⴝ .003). For patients in stratum A, the 3-year event-free survival and survival were 72.3% ⴞ 6.7% and 76.4% ⴞ 6.4%, respectively. Patients who received carboplatin-based therapy had less long-term renal toxicity and ototoxicity. Conclusion: This pilot trial suggests that carboplatin/ifosfamide combination chemotherapy has substantial antitumor activity. In the context of a multiagent chemotherapy protocol comprising high-dose methotrexate and doxorubicin, we found that the addition of carboplatin/ifosfamide resulted in patient outcomes comparable to trials using cisplatin-based therapy with less long-term toxicity. J Clin Oncol 19:171-182. © 2001 by American Society of Clinical Oncology. STEOSARCOMA IS the most common primary malignant bone tumor in children and adolescents. At least two thirds of patients with resectable primary tumors and without metastatic disease at diagnosis will be cured with resection of the primary tumor and effective multiagent chemotherapy.1-5 Patients with unresectable primary tumors and those with metastatic disease present at diagnosis have a much less favorable outcome.6,7 The improved cure rate for osteosarcoma is primarily a result of effective use of multiagent chemotherapy.1 In single-agent trials, the greatest proportion of antitumor responses has been noted with the use of doxorubicin,8,9 cisplatin,10-12 high-dose methotrexate,13,14 and ifosfamide.15-17 Although none of these agents has consistently caused tumor regressions in more than approximately 30% of patients with measurable tumors, when used in adjuvant combination therapy trials, cure rates of 60% to 70% are routinely reported. Toxicity for these multiagent trials is substantial, and for some patients, such toxicity is permanent. In particular, cisplatin causes irreversible renal impairment18 and ototoxicity.19-21 These side effects occur in a substantial proportion of survivors of osteosarcoma.22,23 In addition, even with the use of newer, more effective antiemetic agents, cisplatin continues to be one of the most emetogenic chemotherapy agents. Carboplatin was synthesized in hopes of identifying an effective platin analog with less short- and long-term toxicity than cisplatin. Carboplatin causes more early bone marrow suppression than cisplatin but much less permanent renal toxicity and ototoxicity.24,25 The relative efficacy of these platin analogs seems to be tumor-specific. In some clinical settings, carboplatin is as effective as cisplatin. Randomized trials that compare cisplatin and carboplatin in O From the Departments of Hematology/Oncology, Biostatistics and Epidemiology, Surgery, and Pathology and Laboratory Medicine, St Jude Children’s Research Hospital, and University of Tennessee Memphis, Memphis, TN; Department of Pathology, University of Arkansas for Medical Sciences, Little Rock, AR; and University of Nebraska Medical Center, Omaha, NE. Submitted June 14, 1999; accepted July 28, 2000. Supported by grant no. PO1 CA-23099 and Cancer Center Core Grant CA-21765 from the National Cancer Institute, National Institutes of Health, Bethesda, MD, and by the American Lebanese Syrian Associated Charities, Memphis, TN. Address reprint requests to William H. Meyer, MD, Hematology/ Oncology Section, Department of Pediatrics, University of Oklahoma Health Sciences Center, PO Box 26901, Oklahoma City, OK; email [email protected]. © 2001 by American Society of Clinical Oncology. 0732-183X/01/1901-171 Journal of Clinical Oncology, Vol 19, No 1 (January 1), 2001: pp 171-182 Downloaded from jco.ascopubs.org on August 22, 2014. For personal use only. No other uses without permission. Copyright © 2001 American Society of Clinical Oncology. All rights reserved. 171 172 MEYER ET AL ovarian cancer and lung cancer have shown similar response rates and survival outcomes.26 However, cisplatin has superior efficacy in other adult tumors, including germ cell tumors,27,28 bladder cancer,29 and head and neck cancers30 (reviewed in26). Carboplatin has not been as adequately tested in many childhood and adolescent malignancies, including osteosarcoma. In preclinical osteosarcoma models, carboplatin has antitumor activity. In human osteosarcoma cell lines, colony formation was completely abolished by both carboplatin and cisplatin, with 10-fold greater exposures for carboplatin, which is consistent with much greater dosing clinically for this agent.31 Dogs that received adjuvant therapy with carboplatin after amputation for osteosarcoma have prolongation of survival, similar to a cohort that received adjuvant cisplatin.32 In a human osteosarcoma xenograft model, carboplatin, 20 mg/kg intraperitoneally, reduces tumor growth.33 A phase I trial of carboplatin that included 10 patients with recurrent osteosarcoma reported one complete response.34 The sole single-agent relapse phase II trial of carboplatin conducted in children failed to show antitumor activity in patients with recurrent osteosarcoma.35 Of the 12 assessable patients in this study, 11 had received prior cisplatin therapy; tumors in these patients possibly had developed acquired resistance to platinating agents. Such acquired cross-resistance after exposure to cisplatin has been reported in an osteosarcoma cell line.36 More recently, Petrilli et al37 evaluated responses to intra-arterial carboplatin therapy in 33 consecutive patients with osteosarcoma, reporting clinical and radiologic responses in 81% and 73% of patients, respectively. In several tumors that commonly occur in children and adolescents, including neuroblastoma, rhabdomyosarcoma, and osteosarcoma, agents that unequivocally are active against tumors previously not exposed to anticancer agents will show less activity in classic phase II trials. For this reason, pediatric investigators have used an investigative up-front window approach,38-42 testing the new agent or combination of agents before exposure to known effective drugs. In osteosarcoma, this strategy has been useful in demonstrating antitumor activity for single-agent ifosfamide43 and the combination of ifosfamide/etoposide.44 In a previous St Jude trial of osteosarcoma, single-agent ifosfamide window therapy showed clinical and radiographic responses in 30% of patients.45 However, the tumors of approximately 30% of patients in this trial showed disease progression during the 6-week single-agent ifosfamide window. Patients who failed to respond to ifosfamide did not have a worse ultimate outcome. Subsequently, in the present trial (OS-91), we combined carboplatin with ifosfamide to determine the activity of this combination in patients with previously untreated osteosarcoma and to estimate the outcome for patients treated with a multiagent chemotherapy protocol that eliminated cisplatin. PATIENTS AND METHODS Patients Patients with previously untreated osteosarcoma, malignant fibrous histiocytoma of bone, fibrosarcoma, or multipotential sarcoma of bone and with good performance status (Eastern Cooperative Oncology Group scores 0 to 2) and normal renal, cardiac, and hepatic function were eligible for enrollment. At the time of enrollment, patients were placed into one of three strata on the basis of resectability of the primary tumor and the presence of metastatic disease at the time of diagnosis: stratum A, resectable primary tumor without evidence of metastases; stratum B, unresectable primary tumor, without evidence of metastases; and stratum C, metastatic disease at diagnosis regardless of resectability of the primary tumor. Evaluation Biopsy of the primary tumor was required for diagnosis; typically, open incisional biopsies were performed. Biopsy of metastatic lesions was not required at time of diagnosis; however, patients with metastatic pulmonary lesions underwent delayed thoracotomy unless the pulmonary lesions were judged unresectable. The initial staging workup included plain radiographs, computed tomography (CT) scan, and magnetic resonance (MR) imaging (including dynamic contrast MR imaging [DEMRI]46,47) of the primary tumor, 99TcDMP bone scanning, 210thallium scan of tumor area, plain chest radiograph and noncontrast CT scan of lungs, comprehensive chemistry panel, echocardiogram, and pure-tone audiometry. After the first cohort of patients enrolled showed no evidence of hearing loss on serial studies, routine close follow-up by serial audiometry was discontinued. Response evaluation after two and three cycles of presurgery carboplatin/ ifosfamide chemotherapy (at weeks 6 and 9, respectively) comprised plain radiographs, CT and MR imaging, and 210thallium scanning of tumor area. Known metastatic tumor sites were also evaluated for response. Complete re-evaluation of tumor status was performed at completion of therapy. Follow-up evaluation for tumor recurrence comprised plain chest radiographs every 6 weeks, CT scans of lungs every 3 months, and bone scanning every 6 months for the first 2 years. After this point, patients were evaluated only by clinical assessment and plain chest radiographs unless clinical symptoms suggested tumor recurrence. Protocol Drug Administration Patients received two cycles of carboplatin (560 mg/m2 on day 1) and ifosfamide (2.65 g/m2/d for 3 days with mesna uroprotection), followed by response assessment at week 6. Unless there was evidence of tumor progression, a third cycle of carboplatin/ifosfamide was administered. After complete clinical and radiographic assessment, patients in stratum A and those in stratum C with resectable primary tumors underwent ablative surgery of the primary tumor. Limb-salvage surgery was attempted when deemed feasible by the surgeon. All joint reconstructions used metallic custom endoprosthetic devices. For those few reconstructions that did not involve a joint, allograft bone with intramedullary rod fixation was used. Patients with unresectable primary tumors typically were removed from protocol to receive individualized local therapy, usually with local irradiation of the primary tumor. After resection of the primary tumor, patients received doxorubicin (five treatments at 75 mg/m2/dose as 72-hour continuous infusion, Downloaded from jco.ascopubs.org on August 22, 2014. For personal use only. No other uses without permission. Copyright © 2001 American Society of Clinical Oncology. All rights reserved. 173 CARBOPLATIN/IFOSFAMIDE IN OSTEOSARCOMA Fig 1. Chemotherapy Outline. The “S” box indicates scheduled surgery. For the OS-91 protocol, alternate chemotherapy for progressive disease is in hatched area. Abbreviations and dosages: i, ifosfamide 1.6 g/m2/d ⴛ 5; ic, ifosfamide 2.65 g/m2/d ⴛ 3, carboplatin 560 mg/m2/dose; m, methotrexate 12 g/m2/dose; a, doxorubicin 75 mg/m2 72-hour continuous infusion; a*, doxorubicin 90 mg/m2/cycle (OS-86); ap, doxorubicin 75 mg/m2/dose and cisplatin 100 mg/m2/dose; Rx, treatment; PD, progressive disease. usually administered in the ambulatory setting) and high-dose methotrexate (nine treatments at 12 g/m2/dose with leucovorin rescue). Patients whose tumors demonstrated clinical and radiographic progression during the 9-week presurgical phase of the protocol received cisplatin (100 mg/m2/dose) in substitution for carboplatin/ifosfamide cycles. All other patients received two additional cycles of carboplatin/ ifosfamide. The outlines of chemotherapy for OS-91 and for the historic comparison trial, OS-86, are shown in Fig 1. Response Evaluation Clinical and radiographic responses were recorded at weeks 6 and 9 of the protocol. Patients who became pain-free without analgesics and had increased peripheral rimming calcification (healing mineralization)48 or had decrease in tumor size were coded as having clinical responses. Patients with no significant radiographic change in tumor were coded as having stable disease. Patients with tumors who had measurable tumor growth were defined as having disease progression. Histologic assessment at week 9 used the four-grade system described by Huvos et al.49,50 Statistical Methods The study was designed with two primary end points: (1) estimating the response rate to presurgical chemotherapy using carboplatin/ ifosfamide, and (2) determining whether combination therapy with these two agents would significantly decrease the rate of disease progression compared with presurgical ifosfamide alone, administered in the prior St Jude OS-86 protocol (30% disease progression rate at week 6). This was the most appropriate end point for comparison with the historical control group enrolled onto OS-86 for several reasons. First, clinical and radiologic assessment of response in osteosarcoma may be difficult and somewhat subjective; therefore, comparison of the proportions responding was not feasible. However, the assessment of tumor progression, with increased tumor bulk and usually continued tumor pain during the 6-week evaluation, was not difficult to discern. Second, by study design, patients enrolled onto OS-86 had surgical ablation of the primary tumor at week 13, after exposure to high-dose methotrexate and doxorubicin. Comparison of histologic grading in this group with that of the OS-91 patients who received only carboplatin/ ifosfamide was meaningless. Evaluation of at least 40 patients was required to detect a decrease in the progressive disease rate to 10%, with a type I error rate of 5% and power of 80% (one-tailed test). When preliminary analysis showed that the estimated progressive disease rate was less than 10%, accrual to the study was continued to allow a better estimate of the outcomes of patients with resectable tumors. Fisher’s exact test was used to compare progressive disease rates during the window treatment (at week 6) for OS-86 and OS-91. Response rates were calculated as binomial proportions, and exact 95% confidence intervals were computed. The association between histologic and clinical response was investigated using the exact Jonckheere-Terpstra test,51 which considers both rows and columns as ordinal variables. Survival was defined as the interval from diagnosis to death from any cause or to last follow-up. Event-free survival was defined as the interval from diagnosis to relapse or progressive disease, second malignancy, or death from any cause. We also calculated event-free survival by excluding progressive disease in the window as an event for patients in stratum A. The rationale for this exclusion was that the protocol allowed patients in stratum A to have progressive disease in the window and to continue on the protocol with substitution of cisplatin for carboplatin/ifosfamide. Survival and event-free survival were estimated using the method of Kaplan and Meier52; associated SEs were calculated using the method of Peto et al.53 Differences in survival distributions were tested using the Mantel-Haenszel test.54 Exact log-rank tests were used when groups were small. A test developed by Mantel and Byar55 was used to compare event-free survival distributions by type of surgery (limb-salvage v amputation); this method accounts for patients transferring from one group to another. StatXact3 (CYTEL Software Corporation, Cambridge, MA) and SAS (Version 6.12, SAS Institute, Inc, Cary, NC) were used for statistical analysis. As part of a program project grant, the trial was approved by the Clinical Therapy and Evaluation Program of the National Cancer Institute and by the local institutional review boards. The trial was monitored annually by the local institutional review boards and the St Jude Children’s Research Hospital Clinical Protocol Scientific Review and Monitoring Committee. Informed consent of the patient and/or parent or legal guardian was obtained before enrollment. RESULTS Patient Characteristics Patient characteristics for the 69 OS-91 patients are listed in Table 1. Patients were enrolled onto three strata: the majority of patients (n ⫽ 47; 68%) had potentially resectable primary lesions (stratum A), five patients had unresectable primary lesions (stratum B), and the remaining 17 patients had metastatic disease at diagnosis (stratum C). The median age at diagnosis of the 69 patients was 14.1 years (range, 1.2 to 24.1 years). Thirty-five patients (51%) were male; 47 (68%) were white. The most common sites of the primary tumor were the femur (n ⫽ 36) and the tibia (n ⫽ 11), which accounted for 68% of the cases. Fifty-six patients had extremity primaries, and two had multifocal primary sites. The most common histologies at initial biopsy were osteosarcoma (not otherwise specified) (n ⫽ 41; 59%) and osteoblastic osteosarcoma (n ⫽ 12; 17%). Of Downloaded from jco.ascopubs.org on August 22, 2014. For personal use only. No other uses without permission. Copyright © 2001 American Society of Clinical Oncology. All rights reserved. 174 MEYER ET AL Table 1. Patient Characteristics Characteristic Stratum A (resectable primary lesion) B (unresectable primary lesion) C (metastatic disease at diagnosis) Race White Black Hispanic Asian Primary site Femur Tibia Humerus Other extremity Flat bones Multifocal Site of metastases at diagnosis (n ⫽ 17) Lung only Bone only Lung and bone Other No. of Patients % 47 5 17 68.1 7.2 24.6 47 14 7 1 68.1 20.3 10.1 1.4 36 11 5 4 11 2 52.2 15.9 7.2 5.8 15.9 2.9 12 2 2 1 70.6 11.8 11.8 5.9 the 17 patients with metastatic disease at diagnosis, 12 had lung metastases only, two had bone metastases only, two had both lung and bone metastases, and one had a drop metastatic lesion at the costophrenic angle. Thirty-nine patients (57%) were alive with a median follow-up of 4.8 years (range, 2.6 to 8.2 years). The characteristics of these 69 patients were compared with those of the 51 patients enrolled onto OS-86. The median age at diagnosis for OS-86 patients was 14.9 years (range, 4.8 to 23.6 years). Thirty-seven patients (73%) were placed in stratum A, two (4%) in stratum B, and 12 (24%) in stratum C. The median follow-up for the 30 survivors enrolled onto OS-86 was 11.2 years (range, 4.7 to 13.1 years); 93% of survivors had their most recent follow-up within 1 year. There was no evidence that age at diagnosis or distribution of patients by stratum differed significantly between OS-86 and OS-91 (P ⫽ .73 and P ⫽ .77, respectively). Comparison of Disease Progression Rates (week 6) One primary objective of OS-91 was to determine whether combination therapy with carboplatin/ifosfamide significantly decreased the rate of disease progression compared with ifosfamide alone (OS-86). The disease progression rates for these two protocols were compared at week 6. Forty-seven of 51 OS-86 patients and 67 of 69 OS-91 patients were assessable for this comparison of response. Of the four nonassessable patients enrolled onto OS-86, one was ineligible (wrong diagnosis) and three had resection of tumor before week 6. Two OS-91 patients were not assessable at week 6: one had surgery before the evaluation (pathologic fracture) and one refused all therapy after one carboplatin/ifosfamide cycle. Fifteen of 47 assessable OS-86 patients (31.9%) had progressive disease at week 6, whereas only six of 67 assessable OS-91 patients (9.0%) had progressive disease at week 6 (P ⫽ .003). Among patients in stratum A assessable at week 6, two (4.4%) of 45 OS-91 patients had progressive disease, whereas eight (24.2%) of 33 OS-86 patients had progressive disease (P ⫽ .015). Clinical/Radiologic Responses The other primary objective of OS-91 was to estimate the response rate to the carboplatin/ifosfamide window. Responses were coded at both weeks 6 and 9. Nine patients overall had disease progression during the window at either week 6 or week 9. Six patients had progressive disease at week 6. Three additional patients had no radiographic or clinical changes in tumor at week 6, received a third cycle of carboplatin/ifosfamide, and had tumor progression at the week-9 evaluation. One patient who had early amputation for pathologic fracture and another who refused all therapy after one carboplatin/ifosfamide cycle (noted above) were not assessable for response at weeks 6 and 9. Two other patients with unresectable tumors were treated with local irradiation after the week-6 evaluation and were not assessable for the week-9 response. All patients enrolled were included in the survival analyses discussed below, except where specifically noted. Of the remaining 56 patients with induction responses coded, 12 had stable disease throughout the window and 44 had tumors that showed clinical/ radiographic response. Twenty-two patients had tumors that showed responses at both week 6 and week 9; 13 had responding tumors at week 6 with no further change in radiographic appearance at week 9. Of the 24 patients with stable disease at week 6, nine became responders by week 9, 12 remained stable, and three experienced progression of their disease. The response rate to three cycles of carboplatin/ifosfamide window for assessable patients was 67.7% (44 of 65; 95% confidence interval, 55.0% to 78.8%). Of the 47 patients in stratum A, 31 had tumors that responded to the carboplatin/ifosfamide window, 10 had stable disease, two had disease progression at week 6, and one had progressive disease at week 9. The response rate for all patients in stratum A was 66% (31 of 47). By protocol design, the three patients with progressive disease during window carboplatin/ifosfamide received subsequent cisplatin therapy. There were only five patients in stratum B: one had progressive disease at week 6, one had a tumor that showed Downloaded from jco.ascopubs.org on August 22, 2014. For personal use only. No other uses without permission. Copyright © 2001 American Society of Clinical Oncology. All rights reserved. 175 CARBOPLATIN/IFOSFAMIDE IN OSTEOSARCOMA Table 2. Histologic and Clinical Responses Clinical and Radiographic Assessment Histologic Response All assessable patients I/II III/IV Total Assessable patients in stratum A I/II III/IV Total Progressive Disease Stable Disease Response Total 3 1 4 9 1 10 12 29 41 24 31 55 2 1 3 9 1 10 8 22 30 19 24 43 response at week 6 but was not assessable at week 9 because the patient was removed from the protocol to receive local irradiation, one had tumor that responded at both weeks 6 and 9, one had a response at week 6 and stable disease at week 9, and one had stable disease at both weeks 6 and 9. Three of the 17 patients with metastatic disease at diagnosis had disease progression at week 6. Two additional patients developed disease progression by week 9. Eleven patients had clinical response to induction; one had stable disease. The response rate for patients in stratum C was 64.7% (11 of 17). Histologic Grading Histologic grades were evaluated in 56 of 69 patients enrolled onto the study. Twelve patients did not have histologic grade because they did not undergo surgical resection of the primary tumor; one additional patient who underwent surgical resection of an ethmoid primary was not assigned a histologic grade. Grades I and II histologic necrosis were combined (⬍ 90% tumor necrosis), as were grades III and IV (ⱖ 90% tumor necrosis). Of the samples available for histologic grade, 31 were grade III or IV histologic necrosis and 25 were grade I or II. The tumors of 55 patients had coded both histologic grade and clinical responses to induction at week 9. We investigated whether there was an association between clinical/radiographic response (after three cycles of carboplatin/ifosfamide) and histologic grade among these patients. There was evidence that the distributions of week-9 responses (progressive disease, stable disease, and partial response) were significantly different by histologic grade (I/II v III/IV; P ⬍ .001). Tumors in 94% of patients with good histologic necrosis (III or IV) showed clinical response to induction, whereas only 50% of tumors with grade I or II showed clinical response (Table 2). Findings were similar for the subset of patients in stratum A (Table 2). Forty-three of 47 tumors in patients in stratum A had both histologic and clinical responses coded. Ninety- two percent of tumors (22 of 24) with good histologic grade (grade III or IV) at week 9, but only 42% (8 of 19) of tumors with poor histologic necrosis (grade I or II), showed clinical response (P ⬍ .001). Surgery Fifty-seven of 69 patients had tumor resections; twelve did not. These twelve patients included two patients in stratum A who refused therapy, including surgical resection of tumor (after one and three cycles of carboplatin/ifosfamide chemotherapy). One patient who underwent resection of her primary ethmoid tumor was not assigned a histologic response grade. Of the 57 patients who had surgery (all strata), 41 had limb-salvage procedures (five later converted to amputation), 14 had primary amputations, and two had other operations (resection of primary ethmoid tumor and resection of a rib). Among the 55 patients with definitive surgery, all received their next course of chemotherapy within 26 days of their first definitive surgery (median time, 14 days). For a comparison of survival and event-free survival between limb-salvage and amputation surgery groups, we used only patients in stratum A with extremity tumors. Of these 44 patients, two had no surgery, seven had primary amputations, and 35 had limb salvage operations, including the five patients who later had amputations. Of these five patients, none had progressive or recurrent local disease that required subsequent amputation. One had microscopic residual tumor at the bone margin of resection of limb-salvage and was converted to amputation 1.4 months after limbsalvage. The other four patients had postoperative complications (wound infections, abscesses, or chronic osteomyelitis) that required amputation after limb salvage. The median time from limb salvage to amputation for these five patients was 5.3 months (range, 1.4 to 11.2 months). No difference was noted in event-free survival by type of surgery (3-year estimates: 80 ⫾ 9 for limb salvage v 75 ⫾ 12 for amputation; P ⫽ .96) Outcomes The following analyses were performed on all patients enrolled (except where specifically noted), including those who refused therapy or definitive surgery. The 3-year estimate of survival for all patients enrolled onto OS-91 was 62.1% ⫾ 6.1%. The estimated 3-year survival for patients in stratum A was 76.4% ⫾ 6.4% (Fig 2). Event-free survival by stratum is shown in Fig 3; 3-year estimates of event-free survival were 72.3% ⫾ 6.7% for patients in stratum A, 60.0% ⫾ 21.9% for those in stratum B, and 5.9% ⫾ 4.0% for those with metastatic disease. Excluding progression of Downloaded from jco.ascopubs.org on August 22, 2014. For personal use only. No other uses without permission. Copyright © 2001 American Society of Clinical Oncology. All rights reserved. 176 MEYER ET AL Fig 2. Kaplan-Meier estimated survival for patients in stratum A enrolled onto the OS-86 and OS-91 trials. disease during the window as an event, the 3-year event-free survival for patients in stratum A was 74.5% ⫾ 6.6%. Disease progression during window. We compared survival for patients with and without progressive disease at week 6. Patients who were not assessable for response were excluded from this analysis. Six patients had progressive disease at week 6: of these, five died (three were patients in stratum C). There was a significant difference in the survival distributions among patients with progressive disease at week 6 and those with other assessable responses (stable disease or partial response; P ⫽ .050). Among patients in stratum A, only two had progression at week 6 (one died). Similarly, we compared survival among patients with progressive disease at any time during the window (week 6 or week 9) with survival among those with other assessable responses. The results were similar to the those of the previous analysis. There was a significant difference in survival among patients with and without progressive disease in the window (P ⫽ .001). Adjusted by stratum, the difference remained significant (P ⫽ .022). For the entire Fig 3. Kaplan-Meier estimated event-free survival by stratum for patients enrolled onto OS-91. Downloaded from jco.ascopubs.org on August 22, 2014. For personal use only. No other uses without permission. Copyright © 2001 American Society of Clinical Oncology. All rights reserved. 177 CARBOPLATIN/IFOSFAMIDE IN OSTEOSARCOMA Fig 4. Kaplan-Meier estimated event-free survival for patients in stratum A enrolled onto the OS-91 protocol by histologic grade. This figure excludes progressive disease at week 6 or 9 as an event as discussed in Patients and Methods. cohort entered onto OS-91, nine patients had progressive disease during the window and eight died. Three-year survival estimates were 22.2% ⫾ 11.3% for patients with progressive disease during the window and 71.2% ⫾ 6.4% for patients with other assessable responses. Only three patients in stratum A had progressive disease during the carboplatin/ifosfamide window; two of these died of disease progression. Three-year survival estimates for patients in stratum A were 33.3% ⫾ 19.2% for patients with progressive disease in the window and 82.6% ⫾ 6.1% for those with other assessable responses. Histologic necrosis grade. Survival distributions were compared for patients with good (grade III or IV) and poor (grade I or II) histologic necrosis. For all patients, 3-year survival estimates were 68.0% ⫾ 9.3% (grade I/II) and 70.2% ⫾ 8.6% (grade III/IV). There was no significant difference in survival by histologic necrosis grade (P ⫽ .81). Results were similar for patients in stratum A, with no evidence of a difference in survival distributions (P ⫽ .69). Three-year survival estimates were 80.0% ⫾ 8.9% (grades I/II) and 78.5% ⫾ 8.6% (grades III/IV). There was also no evidence of a difference in event-free survival by histologic grade among patients in stratum A (P ⫽ .90; Fig 4). This trial was designed to detect a difference in week-6 response rates compared with the historical population treated with ifosfamide only (OS-86) and not designed to establish differences in outcome by histologic necrosis grade. Hence, power to detect differences among survival distributions was limited. As discussed above, histologic necrosis was correlated with the induction response. However, although progressive disease during the window correlated with survival, histologic grade did not. Comparison With OS-86 We compared survival distributions of patients enrolled onto OS-86 and OS-91. The median follow-up periods for OS-86 and OS-91 were 11.2 years (range, 4.7 to 13.1 years) and 4.8 years (range, 2.6 to 8.2 years), respectively. Threeyear survival estimates were 74.5% ⫾ 6.0% for OS-86 and 62.1% ⫾ 6.1% for OS-91 (all strata). Four-year estimates were 66.7% ⫾ 6.5% for OS-86 and 60.4% ⫾ 6.8% for OS-91. There was no significant difference in survival distributions between the two protocols (P ⫽ .35). For patients in stratum A, 3-year survival estimates were 83.8% ⫾ 6.0% (OS-86) and 76.4% ⫾ 6.4% (OS-91) (P ⫽ .96). Patients with metastatic disease at diagnosis enrolled onto OS-91 had a worse outcome when compared with those in the OS-86 study, with 3-year survival estimates of 23.5% ⫾ 10.3% (OS-91) and 50.0% ⫾ 13.4% (OS-86; P ⫽ .062). Although disease progression rate during the first 6-week window of OS-86 was much higher than that in the OS-91 trial, there was no significant difference in survival for patients enrolled onto OS-86 with or without progressive disease at week 6 (P ⫽ .61). Toxicity Toxicity was moderate and manageable for most patients. Patients received doxorubicin infusions primarily in the ambulatory setting, carrying infusion pumps for the 72-hour infusions. All patients had external venous access devices Downloaded from jco.ascopubs.org on August 22, 2014. For personal use only. No other uses without permission. Copyright © 2001 American Society of Clinical Oncology. All rights reserved. 178 MEYER ET AL (Hickman and Broviac catheters), which eliminated concerns regarding extravasation of continuous infusion doxorubicin. The acute renal toxicities of carboplatin/ifosfamide for the initial cohort enrolled onto this trial have been reported recently.56 Fifteen percent of this cohort developed significant acute electrolyte problems during the first three cycles of carboplatin/ifosfamide. Grade 3 or 4 renal toxicity developed in only two (5%) of the 42 patients alive and assessable 1 year after completion of therapy with carboplatin and ifosfamide (OS-91) but in six (21%) of 29 patients after therapy with cisplatin and ifosfamide (OS-86). Renal tubular toxicity developed in four additional patients in the OS-86 group more than 1 year after completion of therapy. Of the 33 patients with serial audiograms enrolled onto OS-91 who did not receive cisplatin, none had hearing loss detectable by serial pure-tone audiometric assessment. Because no hearing toxicity was detected in the initial cohort enrolled onto OS-91, routine serial audiometric assessment was stopped. Of the patients whose disease progressed during the carboplatin/ifosfamide window and who subsequently received cisplatin, four of five with serial audiograms had changes in hearing thresholds detected by serial audiometry. DISCUSSION Improvements in the cure rate for osteosarcoma depend on the identification of new active agents. Such agents have been identified traditionally through phase II studies in patients with recurrent disease. Several such phase II studies demonstrated that ifosfamide was active in recurrent osteosarcoma.16,17,57,58 However, estimates of response varied; a more accurate assessment of responsiveness of osteosarcoma awaited up-front window trials conducted in previously untreated patients. Harris et al42,43 conducted a phase II window trial of single-agent ifosfamide in previously untreated patients with metastatic osteosarcoma. This study demonstrated a higher response rate than that seen in a similar phase II trial of ifosfamide conducted in previously treated patients. The OS-86 trial, conducted during the same period, showed a response rate in previously untreated patients similar to that reported by Harris et al. Of significant concern, more than 30% of patients enrolled onto the OS-86 trial developed disease progression during the 6-week window therapy, a progressive disease rate also equivalent to that reported by Harris et al. Although the overall outcomes of patients enrolled onto the OS-86 trial with or without progressive disease during the ifosfamide window were similar, such progressive disease rates are not acceptable. Consequently, we elected to combine ifosfamide with carboplatin in the next window trial. This permitted direct historic compari- son. Although assessment of response in osteosarcoma is problematic, progression of disease is a more straightforward end point. We thus selected this end point for comparison with the OS-86 trial. With the use of the combination of carboplatin and ifosfamide, the rate of disease progression at week 6 was statistically significantly lower than that observed with ifosfamide alone, which demonstrates the antitumor activity of this combination. Single-agent carboplatin has been reported to cause regression of osteosarcoma.59 However, the only single-agent phase II trial conducted in patients with recurrent tumors failed to demonstrate activity against osteosarcoma.60 The Pediatric Oncology Group conducted a single-agent trial of carboplatin in previously untreated patients with metastatic and/or unresectable osteosarcoma.61 Although antitumor responses were observed in at least one tumor site in 22% of patients, only one of the 37 patients enrolled onto this trial had an overall partial response defined by traditional criteria (more than 50% regression of tumor at all sites). Our data may suggest clinical synergy for the combination of carboplatin/ifosfamide. The hypothesis that platin analogs and oxazaphosphorines are synergistic is supported by preclinical data and similar clinical experiences with other drug combinations used in pediatric sarcomas. The combination of cisplatin and cyclophosphamide showed therapeutic synergism against advanced Ridgway osteogenic sarcoma and P388 leukemia.62 Therapeutic synergism was reported for the combination of ifosfamide and cisplatin in L1210 leukemia, with greater than four-fold increase in life span compared with ifosfamide or cisplatin used as single agents.63 The more recent clinical experience with topotecan in childhood sarcomas provides a clinical example of possible synergism. In a classic phase II trial, Nitschke et al64 reported that single-agent topotecan induced no responses in children with recurrent rhabdomyosarcoma. However, the combination of topotecan and moderate-dose cyclophosphamide is active in recurrent rhabdomyosarcoma, with responses seen in the initial phase I65 and phase II (H. Grier, personal communication, November 1998) Pediatric Oncology Group trials. The probable clinical synergism for topotecan and cyclophosphamide also is predicted from preclinical data for this combination.66,67 Finally, single-agent etoposide has minimal activity in osteosarcoma; however, when combined with ifosfamide, the drug pair has shown more activity in osteosarcoma than other agents tested by the Pediatric Oncology Group in window trials.44 Of significance, the design of our trial permitted assessment of the long-term outcome for patients with osteosarcoma who received adjuvant therapy without cisplatin. For the entire patient population enrolled onto OS-91, including Downloaded from jco.ascopubs.org on August 22, 2014. For personal use only. No other uses without permission. Copyright © 2001 American Society of Clinical Oncology. All rights reserved. 179 CARBOPLATIN/IFOSFAMIDE IN OSTEOSARCOMA patients with metastatic and unresectable tumors, the 3- and 4-year survival estimates were 62.1% and 60.4%, respectively. Most adjuvant trials for osteosarcoma enroll and report results only for those patients with nonmetastatic tumors that are located in an extremity or are resectable, the type of patients we enrolled onto stratum A. For this stratum, the 3-year event-free survival was 72.3%, similar to that of other contemporary clinical trials for this disease. This event-free survival estimate includes two patients who refused therapy before surgical resection of the primary tumor, experienced disease progression, and ultimately died of tumor. We were unable to demonstrate any difference in outcome, based on histologic response. Our study was not designed to test this end point, and power to detect differences was low. However, the outcomes for good and poor histologic necrosis were similar. This finding is not consistent with those of many other neoadjuvant chemotherapy trials in osteosarcoma; this may be a direct reflection of the study design. Patients received carboplatin/ifosfamide without other known effective agents before surgical resection of the primary tumor. The postsurgical therapy comprised primarily doxorubicin and high-dose methotrexate cycles, with only two additional cycles of carboplatin/ifosfamide. Consequently, lack of response to the experimental drug pair would not necessarily predict a poor outcome when the remainder of the therapy was predominantly with agents of known efficacy in the treatment of osteosarcoma. Two important toxicities associated with curative therapy for osteosarcoma were less prominent in patients enrolled onto the OS-91 trial. Cisplatin-based therapy for this tumor routinely results in irreversible hearing loss in some patients. We have previously reported that one half of patients treated for osteosarcoma on the OS-86 trial, using both ifosfamide and cisplatin, had hearing loss of sufficient magnitude to require amplification.68 In the OS-91 study, hearing loss developed only in the small number of patients who had tumor progression in the window and subsequently received cisplatin. No other patients had measurable changes in hearing thresholds. Similarly, late renal toxicity occurred less frequently. Although some patients who received carboplatin/ifosfamide had renal tubular toxicity during therapy, at 1 year after the end of treatment, few patients who were enrolled onto OS-91 had permanent sequelae; several patients similarly treated on the OS-86 trial required long-term electrolyte supplementation, and one had irreversible renal failure.69 Is window therapy a safe approach for patients with osteosarcoma? In both the OS-86 and OS-91 trials, the survival and disease-free survival for patients with nonmetastatic, resectable disease was similar to those obtained in other modern multiagent trials. Patients with disease progression during the single-agent ifosfamide window did not have inferior outcomes. However, the disease progression rate with single-agent ifosfamide was unacceptable for a group of patients with good long-term outcome and for whom limb salvage was a potential surgical option. This, in part, was the rationale for combining carboplatin with ifosfamide in the OS-91 study. With the addition of a second agent, the progressive disease rate dropped significantly. However, the predicted outcome for patients enrolled onto our stratum A is clearly higher than that judged acceptable by the recent National Cancer Institute Consensus Panel for pediatric window therapy trials in previously untreated patients. We believe that in appropriate circumstances and with appropriate patient safeguards and early stopping rules, continued window trials are an acceptable and efficient way to test new agents in pediatric oncology. However, they should include only those patients with osteosarcoma who have been determined to be at higher risk of treatment failure, specifically patients with unresectable primary tumors and patients with metastatic disease present at diagnosis. Additional challenges remain with window trials in osteosarcoma. Assessment of tumor response using standard clinical and radiologic techniques in patients with osteosarcoma is difficult. These tumors often do not change in size, even with effective multiagent chemotherapy; therefore, standard definitions for partial and complete response based on change in tumor size are not useful.70 Although newer radiologic techniques, such as DEMRI functional imaging46,47,71 and thallium scanning,72 are promising, they are not routinely used in multi-institutional trials. In particular, DEMRI imaging requires technical expertise and is not widely available. However, although assessing response is problematic, determining that a tumor is progressing during initial therapy is typically unequivocal. Histologic assessment of response remains the gold standard. However, this assessment also has problems: investigators have reported substantial variations in grading response50,73-75; typically, only one cross-section of tumor is assessed; assessment is available for one point in time; degree of response may be dependent on length of presurgical therapy2 and agents used; and this assessment is not applicable for tumors that cannot be resected. Assessment of response of pulmonary metastatic lesions is also imperfect. Not all pulmonary lesions seen on CT imaging are tumor. As with the primary tumor, chemotherapy may cause complete tumor necrosis without change in the size of the pulmonary lesions. Not all patients with pulmonary metastatic disease are candidates for thoracotomy and metastasectomy. Finally, survival estimates for patients with met- Downloaded from jco.ascopubs.org on August 22, 2014. For personal use only. No other uses without permission. Copyright © 2001 American Society of Clinical Oncology. All rights reserved. 180 MEYER ET AL astatic disease are more variable than outcomes reported for patients with nonmetastatic resectable tumors.6,43 Investigators who design and conduct such trials must carefully define response criteria and encourage histologic response assessment in as many patients as feasible. This trial demonstrates that the combination of carboplatin and ifosfamide has substantial antitumor activity in osteosarcoma and, in the context of multiagent therapy, produces outcomes comparable to regimens that include cisplatin with less long-term toxicity. Whether this combination will have substantial benefit, as judged by improve- ment in outcome or similar outcome with significantly less long-term toxicity than cisplatin-based treatment, requires assessment by a prospective randomized comparative study. ACKNOWLEDGMENT We are indebted to the patients and families who participated in this trial and to the staff members at St Jude Children’s Research Hospital and the University of Nebraska Medical Center who assisted in their care. We thank Mickey Cain and Loraine Avery for expert data management support. This work is dedicated to Matt and his family. REFERENCES 1. Link MP, Goorin AM, Miser AW, et al: The effect of adjuvant chemotherapy on relapse-free survival in patients with osteosarcoma of the extremity. N Engl J Med 314:1600-1606, 1986 2. Meyers PA, Heller G, Healey J, et al: Chemotherapy for nonmetastatic osteogenic sarcoma: The Memorial Sloan-Kettering experience. J Clin Oncol 10:5-15, 1992 3. Winkler K, Beron G, Delling G, et al: Neoadjuvant chemotherapy of osteosarcoma: Results of a randomized cooperative trial (COSS-82) with salvage chemotherapy based on histological tumor response. J Clin Oncol 6:329-337, 1988 4. Hudson M, Jaffe MR, Jaffe N, et al: Pediatric osteosarcoma: Therapeutic strategies, results, and prognostic factors derived from a 10-year experience. J Clin Oncol 8:1988-1997, 1990 5. Campanacci M, Bacci G, Bertoni F, et al: The treatment of osteosarcoma of the extremities: Twenty year’s experience at the Istituto Ortopedico Rizzoli. Cancer 48:1569-1581, 1981 6. Marina NM, Pratt CB, Rao BN, et al: Improved prognosis of children with osteosarcoma metastatic to the lung(s) at the time of diagnosis. Cancer 70:2722-2727, 1992 7. Meyers PA, Heller G, Healey JH, et al: Osteogenic sarcoma with clinically detectable metastasis at initial presentation. J Clin Oncol 11:449-453, 1993 8. Cortes EP, Holland JF, Wang JJ, et al: Amputation and adriamycin in primary osteosarcoma. N Engl J Med 291:998-1000, 1974 9. Pratt CB, Shanks EC: Doxorubicin in treatment of malignant solid tumors in children. Am J Dis Child 127:534-536, 1974 10. Ochs JJ, Freeman AI, Douglass HO Jr, et al: cis-Dichlorodiammineplatinum (II) in advanced osteogenic sarcoma. Cancer Treat Rep 62:239-245, 1978 11. Gasparini M, Rouesse J, van Oosterom A, et al: Phase II study of cisplatin in advanced osteogenic sarcoma. European Organization for Research on Treatment of Cancer Soft Tissue and Bone Sarcoma Group. Cancer Treat Rep 69:211-213, 1985 12. Baum ES, Gaynon P, Greenberg L, et al: Phase II trial cisplatin in refractory childhood cancer: Children’s Cancer Study Group Report. Cancer Treat Rep 65:815-822, 1981 13. Jaffe N, Frei E III, Traggis D, et al: Adjuvant methotrexate and citrovorum-factor treatment of osteogenic sarcoma. N Engl J Med 291:994-997, 1974 14. Pratt CB, Howarth C, Ransom JL, et al: High-dose methotrexate used alone and in combination for measurable primary or metastatic osteosarcoma. Cancer Treat Rep 64:11-20, 1980 15. Pratt CB, Horowitz ME, Meyer WH, et al: Phase II trial of ifosfamide in children with malignant solid tumors. Cancer Treat Rep 71:131-135, 1987 16. Gasparini M: High-dose ifosfamide alone and in combination for solid malignancies in childhood. Cancer Chemother Pharmacol 18:S18, 1986 (suppl 2) 17. de Kraker J, Voute PA: Experience with ifosfamide in paediatric tumours. Cancer Chemother Pharmacol 24:S28-S29, 1989 18. Hrushesky WJ, Shimp W, Kennedy BJ: Lack of age-dependent cisplatin nephrotoxicity. Am J Med 76:579-584, 1984 19. Helson L, Okonkwo E, Anton L, et al: cis-Platinum ototoxicity. Clin Toxicol 13:469-478, 1978 20. McHaney VA, Thibadoux G, Hayes FA, et al: Hearing loss in children receiving cisplatin chemotherapy. J Pediatr 102:314-317, 1983 21. Pasic TR, Dobie RA: cis-Platinum ototoxicity in children. Laryngoscope 101:985-991, 1991 22. Ruiz L, Gilden J, Jaffe N, et al: Auditory function in pediatric osteosarcoma patients treated with multiple doses of cis-diamminedichloroplatinum(II). Cancer Res 49:742-744, 1989 23. Jaffe N, Keifer R III, Robertson R, et al: Renal toxicity with cumulative doses of cis-diamminedichloroplatinum-II in pediatric patients with osteosarcoma. Effect on creatinine clearance and methotrexate excretion. Cancer 59:1577-1581, 1987 24. Sleijfer DT, Smit EF, Meijer S, et al: Acute and cumulative effects of carboplatin on renal function. Br J Cancer 60:116-120, 1989 25. van der Hulst RJ, Dreschler WA, Urbanus NA: High frequency audiometry in prospective clinical research of ototoxicity due to platinum derivatives. Ann Otol Rhinol Laryngol 97:133-137, 1988 26. Lokich J, Anderson N: Carboplatin versus cisplatin in solid tumors: An analysis of the literature. Ann Oncol 9:13-21, 1998 27. Bajorin DF, Sarosdy MF, Pfister DG, et al: Randomized trial of etoposide and cisplatin versus etoposide and carboplatin in patients with good-risk germ cell tumors: A multiinstitutional study. J Clin Oncol 11:598-606, 1993 28. Horwich A, Sleijfer DT, Fossa SD, et al: Randomized trial of bleomycin, etoposide, and cisplatin compared with bleomycin, etoposide, and carboplatin in good-prognosis metastatic nonseminomatous germ cell cancer: A Multiinstitutional Medical Research Council/ European Organization for Research and Treatment of Cancer trial. J Clin Oncol 15:1844-1852, 1997 29. Petrioli R, Frediani B, Manganelli A, et al: Comparison between a cisplatin-containing regimen and a carboplatin-containing regimen for recurrent or metastatic bladder cancer patients. A randomized phase II study. Cancer 77:344-351, 1996 30. Forastiere AA, Metch B, Schuller DE, et al: Randomized comparison of cisplatin plus fluorouracil and carboplatin plus fluorouracil versus methotrexate in advanced squamous-cell carcinoma of the Downloaded from jco.ascopubs.org on August 22, 2014. For personal use only. No other uses without permission. Copyright © 2001 American Society of Clinical Oncology. All rights reserved. 181 CARBOPLATIN/IFOSFAMIDE IN OSTEOSARCOMA head and neck: A Southwest Oncology Group study. J Clin Oncol 10:1245-1251, 1992 31. Robson H, Meyer S, Shalet SM, et al: Comparative evaluation of cisplatin and carboplatin sensitivity in osteosarcoma cell lines. Med Pediatr Oncol 33:167, 1999 (abstr) 32. Bergman PJ, MacEwen EG, Kurzman ID, et al: Amputation and carboplatin for treatment of dogs with osteosarcoma: 48 cases (1991 to 1993). J Vet Intern Med 10:76-81, 1996 33. Crnalic S, Bergstrom P, Lofvenberg R, et al: Antagonistic effect of amphotericin B on carboplatin antitumor activity in human osteosarcoma xenografts. Anticancer Drugs 7:489-492, 1996 34. Bacha DM, Caparros-Sison B, Allen JA, et al: Phase I study of carboplatin (CBDCA) in children with cancer. Cancer Treat Rep 70:865-869, 1986 35. Ettinger LJ, Gaynon PS, Krailo MD, et al: A phase II study of carboplatin in children with recurrent or progressive solid tumors. A report from the Childrens Cancer Group. Cancer 73:1297-1301, 1994 36. Asada N, Tsuchiya H, Ueda Y, et al: Establishment and characterization of an acquired cisplatin-resistant subline in a human osteosarcoma cell line. Anticancer Res 18:1765-1768, 1998 37. Petrilli AS, Kechichian R, Broniscer A, et al: Activity of intraarterial carboplatin as a single agent in the treatment of newly diagnosed extremity osteosarcoma. Med Pediatr Oncol 33:71-75, 1999 38. Horowitz ME, Etcubanas E, Christensen ML, et al: Phase II testing of melphalan in children with newly diagnosed rhabdomyosarcoma: A model for anticancer drug development. J Clin Oncol 6:308-314, 1988 39. Meyer WH, Houghton PJ, Horowitz ME, et al: Use of investigational drugs as initial therapy for childhood solid tumors. Hamatol Bluttransfus 32:141-148, 1989 40. Castleberry RP, Cantor AB, Green AA, et al: Phase II investigational window using carboplatin, iproplatin, ifosfamide, and epirubicin in children with untreated disseminated neuroblastoma: A Pediatric Oncology Group study. J Clin Oncol 12:1616-1620, 1994 41. Pappo AS, Etcubanas E, Santana VM, et al: A phase II trial of ifosfamide in previously untreated children and adolescents with unresectable rhabdomyosarcoma. Cancer 71:2119-2125, 1993 42. Harris MB, Cantor AB, Goorin AM, et al: Treatment of osteosarcoma with ifosfamide: Comparison of response in pediatric patients with recurrent disease versus patients previously untreated—A Pediatric Oncology Group study. Med Pediatr Oncol 24:87-92, 1995 43. Harris MB, Gieser P, Goorin AM, et al: Treatment of metastatic osteosarcoma at diagnosis: A Pediatric Oncology Group study. J Clin Oncol 16:3641-3648, 1998 44. Goorin AM, Geiser P, Ferguson W, et al: Successful phase II trial of etoposide and high dose ifosfamide in newly diagnosed metastatic osteosarcoma: A Pediatric Oncology Group trial (POG). Proc Am Soc Clin Oncol 17:535a, 1998 (abstr 2056) 45. Meyer WH, Luo X, Parham DM, et al: Ifosfamide window therapy for previously untreated osteosarcoma. Proc Annu Meet Am Assoc Cancer Res 35:241, 1994 (abstr) 46. Fletcher BD, Hanna SL, Fairclough DL, et al: Pediatric musculoskeletal tumors: Use of dynamic, contrast-enhanced MR imaging to monitor response to chemotherapy. Radiology 184:243-248, 1992 47. Reddick WE, Bhargava R, Taylor JS, et al: Dynamic contrastenhanced MR imaging evaluation of osteosarcoma response to neoadjuvant chemotherapy. J Magn Reson Imaging 5:689-694, 1995 48. Fletcher BD: Response of osteosarcoma and Ewing sarcoma to chemotherapy: Imaging evaluation. AJR Am J Roentgenol 157:825833, 1991 49. Huvos AG, Rosen G, Marcove RC: Primary osteogenic sarcoma: Pathologic aspects in 20 patients after treatment with chemotherapy en bloc resection, and prosthetic bone replacement. Arch Pathol Lab Med 101:14-18, 1977 50. Rosen G, Caparros B, Huvos AG, et al: Preoperative chemotherapy for osteogenic sarcoma: Selection of postoperative adjuvant chemotherapy based on the response of the primary tumor to preoperative chemotherapy. Cancer 49:1221-1230, 1982 51. Hollander M, Wolfe DA (eds): Nonparametric Statistical Methods. New York, NY, John Wiley & Sons, 1973, pp 120-123 52. Kaplan EL, Meier P: Nonparametric estimation from incomplete observations. J Am Stat Assoc 53:457-481, 1958 53. Peto R, Pike MC, Armitage P, et al: Design and analysis of randomized clinical trials requiring prolonged observation of each patient. II. Analysis and examples. Br J Cancer 35:1-39, 1977 54. Mantel N, Haenszel W: Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst 22:719-748, 1959 55. Mantel N, Byar DP: Evaluation of response-time data involving transient states: An illustration using heart-transplant data. J Am Stat Assoc 69:81-86, 1974 56. Marina NM, Poquette CA, Cain AM, et al: Comparative renal tubular toxicity of chemotherapy regimens including ifosfamide in patients with newly diagnosed sarcomas. J Pediatr Hematol Oncol 22:112-118, 2000 57. Marti C, Kroner T, Remagen W, et al: High-dose ifosfamide in advanced osteosarcoma. Cancer Treat Rep 69:115-117, 1985 58. Pratt CB, Meyer WH, Rao BN, et al: Osteosarcoma studies at St Jude Children’s Research Hospital from 1968 through 1990. Cancer Treat Res 62:323-326, 1993 59. Bacha DM, Caparros-Sison B, Allen JA, et al: Phase I study of carboplatin (CBDCA) in children with cancer. Cancer Treat Rep 70:865-869, 1986 60. Ettinger LJ, Gaynon PS, Krailo MD, et al: A phase II study of carboplatin in children with recurrent or progressive solid tumors. A report from the Children’s Cancer Group. Cancer 73:1297-1301, 1994 61. Ferguson W, Harris M, Link M, et al: Carboplatin in the treatment of newly-diagnosed metastatic or unresectable osteosarcoma: A Pediatric Oncology Group study. Proc Am Soc Clin Oncol 15:521a, 1996 (abstr 1678) 62. Schabel FMJ, Trader MW, Laster WRJ, et al: cis-Dichlorodiammineplatinum(II): Combination chemotherapy and cross-resistance studies with tumors of mice. Cancer Treat Rep 63:1459-1473, 1979 63. Goldin A: Ifosfamide in experimental tumor systems. Semin Oncol 9:14-23, 1982 64. Nitschke R, Parkhurst J, Sullivan J, et al: Topotecan in pediatric patients with recurrent and progressive solid tumors: A Pediatric Oncology Group phase II study. J Pediatr Hematol Oncol 20:315-318, 1998 65. Saylors RL, Stewart CF, Zamboni WC, et al: Phase I study of topotecan in combination with cyclophosphamide in pediatric patients with malignant solid tumors: A Pediatric Oncology Group study. J Clin Oncol 16:945-952, 1998 66. Janss AJ, Cnaan A, Zhao H, et al: Synergistic cytotoxicity of topoisomerase I inhibitors with alkylating agents and etoposide in human brain tumor cell lines. Anticancer Drugs 9:641-652, 1998 67. Giovanella B, Potmesil M, Wall ME, et al: Treatment of human cancer xenografts with camptothecin analogues in combination with Cytoxan or x-irradiation. Proc Annu Meet Am Assoc Cancer Res 35:454, 1994 (abstr) Downloaded from jco.ascopubs.org on August 22, 2014. For personal use only. No other uses without permission. Copyright © 2001 American Society of Clinical Oncology. All rights reserved. 182 MEYER ET AL 68. Meyer WH, Ayers D, McHaney VA, et al: Ifosfamide and exacerbation of cisplatin-induced hearing loss. Lancet 341:754-755, 1993 69. Marina NM, Poquette CA, Cain AM, et al: Comparative renal tubular toxicity of chemotherapy regimens including ifosfamide in patients with newly diagnosed sarcomas. J Pediatr Hematol Oncol 22:112-118, 2000 70. Lawrence JA, Babyn PS, Chan HS, et al: Extremity osteosarcoma in childhood: Prognostic value of radiologic imaging. Radiology 189:43-47, 1993 71. Erlemann R, Reiser MF, Peters PE, et al: Musculoskeletal neoplasms: Static and dynamic Gd-DTPA-enhanced MR imaging. Radiology 171:767-773, 1989 72. Ramanna L, Waxman A, Binney G, et al: Thallium-201 scintigraphy in bone sarcoma: Comparison with gallium-67 and technetium-MDP in the evaluation of chemotherapeutic response. J Nucl Med 31:567-572, 1990 73. Salzer-Kuntschik M, Delling G, Beron G, et al: Morphological grades of regression in osteosarcoma after polychemotherapy: Study COSS 80. J Cancer Res Clin Oncol 106:21-24, 1983 74. Rosen G, Marcove RC, Caparros B, et al: Primary osteogenic sarcoma: The rationale for preoperative chemotherapy and delayed surgery. Cancer 43:2163-2177, 1979 75. Bacci G, Springfield D, Capanna R, et al: Neoadjuvant chemotherapy for osteosarcoma of the extremity. Clin Orthop 224:268-276, 1987 Downloaded from jco.ascopubs.org on August 22, 2014. For personal use only. No other uses without permission. Copyright © 2001 American Society of Clinical Oncology. All rights reserved.