Osteosarcoma (OS) is the most common pediatric malignant bone

2447 Pattern of Disease Recurrence and Prognostic Factors in Patients with Osteosarcoma Treated with Contemporary Chemotherapy Douglas S. Hawkins, M.D. 1 Carola A. S. Arndt, M.D. 2 1 Department of Pediatrics, Children s Hospital and Regional Medical Center, Seattle, Washington. 2 Department of Pediatrics, Mayo Clinic, Rochester, Minnesota. BACKGROUND. The goal of the current study was to define the clinical features and outcome of recurrent osteosarcoma (OS) in children and young adults initially treated with contemporary chemotherapy. METHODS. The authors reviewed the clinical features, therapy, and outcome for 59 patients from the Mayo Clinic (Rochester, MN) and Children s Hospital and Regional Medical Center (Seattle, WA). They were diagnosed initially with OS between January 1990 and December 2000, received multiagent chemotherapy (most frequently cisplatin, doxorubicin, high-dose methotrexate, and ifosfamide), and developed disease recurrence after achieving an initial complete response (CR). RESULTS. The most common site of initial disease recurrence was the lung only (n 36 patients), followed by distant bone (n 8 patients), combined lung and other sites (n 7 patients), and other sites (n 8 patients). The median time to first disease recurrence was 15 months (range, 2 92 months) from the initial diagnosis. Thirty patients with isolated pulmonary recurrence achieved a second CR, either with surgery alone (n 15 patients) or surgery and salvage chemotherapy (n 15 patients). For this group, the 4-year disease-free survival (DFS) and overall survival rates were 7% (95% confidence interval [95% CI], 0 16%) and 28% (95% CI, 11 45%), respectively. For all 59 patients with recurrent OS, the 4-year DFS and overall survival rates were 6% (95% CI, 0 12%) and 23% (95% CI, 10 36%), respectively. The only factors associated with improved DFS and overall survival rates were unilateral pulmonary recurrence, solitary pulmonary nodule at recurrence, more than 24 months between the initial diagnosis and first disease recurrence, and achievement of a second CR. CONCLUSIONS. The DFS and overall survival rates for recurrent OS after contemporary therapy remained poor even for patients with isolated pulmonary recurrence. Therefore, new treatment strategies are needed. Cancer 2003;98:2447 56. 2003 American Cancer Society. This manuscript was presented in abstract form at the American Society of Clinical Oncology meeting, Orlando, Florida, May 18 21, 2002. Address for reprints: Douglas S. Hawkins, M.D., Department of Pediatrics, Children s Hospital and Regional Medical Center, MS: 6D-1, 4800 Sand Point Way, Seattle, WA 98105; Fax: (206) 987-3946; E-mail: doug.hawkins@seattlechildrens.org Received May 15, 2003; revision received August 6, 2003; accepted August 26, 2003. KEYWORDS: osteosarcoma, recurrence, chemotherapy, surgery. Osteosarcoma (OS) is the most common pediatric malignant bone tumor, with an incidence of 400 cases among children and adolescents younger than 20 years old each year in the U.S. 1 The addition of multiagent chemotherapy to surgery for OS has improved dramatically the prognosis for patients with OS. Combined modality therapy results in a disease-free survival (DFS) rate of 65 70% for children with localized OS. 2,3 Recurrence of OS remains the primary cause of treatment failure, with the lung being the most common site of disease recurrence. Multiple retrospective series have shown the curative potential of pulmonary metastatectomy, often with repeated aggressive resections. The overall survival rate after OS recurrence 2003 American Cancer Society DOI 10.1002/cncr.11799

2448 CANCER December 1, 2003 / Volume 98 / Number 11 ranges from 13% to 57%. 4 26 The benefit of secondline chemotherapy in the treatment of recurrent OS is less clear. 7 13,15,16,19,20,25,26 Several factors complicate the analysis of the published literature regarding DFS and survival after OS recurrence. First, some series represented selected cohorts of patients amenable to surgical excision, thereby excluding patients with unresectable (and therefore unfavorable) recurrent tumors. 5,8,17,18 Second, some series included patients who did not receive neoadjuvant or adjuvant therapy as part of their initial OS therapy. 4,5,9,13,16 19 Third, the chemotherapy given to patients who received neoadjuvant or adjuvant treatment often differed significantly from regimens used currently, including relatively inactive or nonintensive agents, 6 13,15 21,23,24 or were accompanied by inadequate descriptions of the chemotherapy regimens. 25 Fourth, most series reported only survival rates and only three series reported DFS or event-free survival rates after disease recurrence. 12,15,26 Those rates ranged from 16% to 27%. For these reasons, the clinical features and prognostic factors associated with recurrent OS in the contemporary chemotherapy era are not well defined. To describe the pattern of recurrent OS, therapies used to treat disease recurrence, and features associated with DFS and survival, we reviewed the experience in patients with recurrent OS treated at our two centers from 1990 to 2000. MATERIALS AND METHODS Patients Eligibility criteria for the current retrospective review included the following: 1) diagnosis of OS between January 1, 1990 and December 31, 2000; 2) treatment at the Mayo Clinic (Rochester, MN) or Children s Hospital and Regional Medical Center (Seattle, WA); 3) development of recurrent disease after achieving an initial complete remission (CR) of disease. Patients were excluded if medical records were inadequate to characterize the clinical features and treatment of recurrent disease. All patients received neoadjuvant and adjuvant chemotherapy, which most frequently consisted of doxorubicin, ifosfamide, and high-dose methotrexate with or without cisplatin according to previously reported institutional pilot regimens 27,28 or they received doxorubicin, cisplatin, and high-dose methotrexate with or without ifosfamide and with or without muramyl tripeptide phosphatidyl ethanolamine (MTP-PE) according to a previously reported cooperative group trial. 29 Individual patients received other pilot OS chemotherapy regimens: topotecan, doxorubicin, high-dose methotrexate, carboplatin, etoposide, and ifosfamide 30 ; doxorubicin, high-dose methotrexate, carboplatin, and etoposide 31 ; doxorubicin, high-dose methotrexate, cisplatin, and dexrazoxane; and doxorubicin, high-dose methotrexate, ifosfamide, cisplatin, and etoposide. One patient with small cell OS received chemotherapy consisting of vincristine, doxorubicin, cyclophosphamide, ifosfamide, and etoposide according to a previously described regimen for Ewing sarcoma. 32 Regardless of the initial treatment regimen, the doses of the most common chemotherapeutic agents were the same: doxorubicin, 75 mg/m 2 intravenously (i.v.); high-dose methotrexate, 12 g/m 2 (20 g maximum) i.v.; cisplatin, 120 mg/m 2 i.v.; and ifosfamide, 9 g/m 2 i.v.. Therapy after disease recurrence was not mandated by institutional policy. In general, aggressive surgical excision was recommended for all patients. For patients with unilateral pulmonary recurrence, the surgical approach was a unilateral thoracotomy for exploration and excision of pulmonary nodules. For patients with bilateral pulmonary recurrence, the surgical approach was sequential, bilateral thoracotomies, usually separated by 1 2 weeks. The more affected side of the lung underwent surgical resection first. Chemotherapy was generally reserved for patients with unresectable recurrent tumors or (at the discretion of the treating physician) for patients with early or multifocal disease recurrence. The most common chemotherapy regimens used for recurrent OS were ifosfamide and etoposide (IE) alone 33 or in combination with carboplatin (ICE). 34 Age, primary tumor site, sites of initial metastases, initial treatment regimen, histologic response to initial chemotherapy, pattern of disease recurrence, treatment regimen for disease recurrence, and survival were reviewed for each patient. The local institutional review boards approved medical record reviews for the current study. Statistical Analysis Statistical analyses of DFS and overall survival were performed using the Kaplan Meier method 35 for calculating survival curves using SPSS version 10.0 statistical package. Patients who did not achieve a second complete remission (CR2) were considered to have a DFS of 0 months. For patients who did achieve CR2, DFS was defined as the time from achievement of CR2 to either a second episode of disease progression or death from any cause. Survival was defined as the time from initial disease recurrence to death from any cause. The initial disease recurrence-free interval (RFI) was defined as the time from initial diagnosis to first disease progression. DFS and survival for patients who had not experienced disease recurrence or who had not died were censored from the date of last contact. Differences in DFS and survival among groups defined by patient or treatment characteristics were analyzed

Recurrent Osteosarcoma/Hawkins and Arndt 2449 TABLE 1 Patient Characteristics at Initial Diagnosis and Treatment Characteristics Value (range) or no. of patients (%) Median age (yrs) 15.3 (4.5 23) Primary site Femur 27 (46) Tibia 11 (19) Humerus 8 (13) Pelvis 6 (10) Other a 7 (10) Metastatic sites None 44 (75) Isolated lung 10 (17) Isolated distant bone 2 (3) Lung and other 2 (3) Lymph node 1 (2) Chemotherapy D/C/M/I 44 (75) D/C/M 5 (8) D/C/M/MTP-PE 3 (5) D/M/I 2 (3) D/M/other b 4 (7) VDC/IE 1 (2) Surgery Limb salvage 43 Amputation 7 Other c 9 D: doxorubicin; C: cisplatin; M: methotrexate; I: ifosfamide; MTP-PE: muramyl tripeptide phosphatidyl ethanolamine; VDC/IE: vincristine, doxorubicin, cyclophosphamide, ifosfamide, etoposide. a Other primary sites include the fibula (n 3), rib (n 1), clavicle (n 1), radius (n 1), and skull (n 1). b Other chemotherapy (one patient each): carboplatin and etoposide; cisplatin and dexrazoxane; topotecan, carboplatin, etoposide, and ifosfamide; cisplatin, etoposide, and ifosfamide. c Other surgery includes pelvic resection (n 6), rib resection (n 1), clavicle resection (n 1), and ethmoidectomy (n 1). using the log-rank test. 36 DFS and survival data were analyzed as of March 1, 2003. RESULTS Demographic, Clinical Features, and Treatment of the Study Group The clinical characteristics of 59 patients with OS at initial presentation and at disease recurrence are listed in Tables 1 and 2, respectively. Seventy-five percent of patients presented without distant metastases and 85% had extremity primary sites. All patients received dose-intensive neoadjuvant chemotherapy, which was followed by complete excision of the primary tumor (and all metastatic sites of tumor), followed by continued adjuvant chemotherapy. The patients received a variety of chemotherapeutic regimens (Table 1). However, all patients received doxorubicin and (with one exception) high-dose methotrexate. All but three patients received a platinum compound, most frequently cisplatin, and 83% TABLE 2 Patient Characteristics at Initial Disease Recurrence Characteristics Value (range) or no. of patients (%) Median initial RFI (mos) 15 (2 92) Site of initial disease recurrence Isolated lung 36 (61) Isolated distant bone 8 (14) Lung and other site a 7 (12) Isolated local 4 (7) Other b 4 (7) Achieved CR2? 40 (68) RFI: disease recurrence-free interval; CR2: second complete response. a Other sites with lung: distant bone (n 2), local (n 2), distant bone and local (n 1), distant bone and brain (n 1), distant soft tissue (n 1). b Other sites: local and distant bone (n 2), local, distant bone, and abdomen (n 1), and local and distant soft tissue (n 1). received ifosfamide. All patients achieved an initial CR after chemotherapy and surgical excision. The median time to median initial RFI was 15 months (range, 2 92 months). The lung was the most common site of initial disease recurrence, occurring in 73% of patients. Disease recurrences limited to pulmonary sites occurred in 61% of patients. Bilateral pulmonary nodules were detected in 14 of 43 (33%) of patients with pulmonary recurrence. A median of 2 pulmonary nodules were detected by computed tomography (CT) scan (range, 1 10 pulmonary nodules). A median of 2 pulmonary nodules were surgically resected (range, 1 37 pulmonary nodules). Data comparing the number of pulmonary nodules detected by CT scan with the number of pulmonary nodules found during surgery were available for 21 patients. In 13 of 21 patients, the number of pulmonary nodules detected by CT scan was identical to the number found during surgery. In 5 of 21 patients, more pulmonary nodules were found during surgery than were detected by CT scan. In 3 of 21 patients, fewer pulmonary nodules were found (because some pulmonary nodules detected by CT scan proved to be nonmalignant after resection). A summary of treatment modalities used and outcome for initial OS recurrence are listed by site of disease recurrence in Table 3. Factors Associated with Disease-Free Survival and survival for All Patients For all patients with recurrent OS, the 4-year DFS rate was estimated to be 6% (95% confidence interval [95% CI], 0 12%; Fig. 1) and the 4-year survival rate was estimated to be 23% (95% CI, 10 36%; Fig. 2). Univariate analysis of potential prognostic factors (Table 4) demonstrated that improved DFS and survival rates were both associated with an RFI of more than 24

2450 CANCER December 1, 2003 / Volume 98 / Number 11 TABLE 3 Management of First Disease Recurrence and Outcome According to Site of Recurrence Treatment/outcome Isolated lung Isolated distant bone Isolated local Lung and other sites Other sites Surgery only Total 16 1 1 0 0 CR2 15 1 1 R2 13 1 1 Chemotherapy only Total 3 2 0 4 0 CR2 0 0 0 R2 Chemotherapy before surgery Total 2 2 0 0 0 CR2 2 2 R2 2 2 Chemotherapy before and after surgery Total 2 0 0 1 1 CR2 2 0 0 R2 2 Surgery before chemotherapy Total 12 2 3 2 1 CR2 11 2 2 1 1 R2 11 1 2 0 1 Palliative treatment Total 1 1 0 0 1 CR2 0 0 0 R2 Total 36 8 4 7 4 FIGURE 1. Kaplan Meier estimated disease-free survival for all patients. CR2: second complete response; R2: second disease recurrence after achieving CR2. months. For patients with the lungs as the site of initial disease recurrence (either in isolation or combination), unilateral pulmonary recurrence was associated with improved DFS and survival rates compared with bilateral pulmonary recurrence. Also, the presence of solitary pulmonary nodules was associated with improved DFS and survival rates compared with the presence of multiple pulmonary nodules. Survival was associated strongly with the achievement of CR2 (4-year survival rate of 33% vs. 0%, P 0.001). The median survival period for patients unable to achieve CR2 was 7.4 months (range, 1.8 30 months) compared with 31 months (range, 4 120 months) for patients who did achieve CR2. The DFS was similar for patients treated with either surgery alone or chemotherapy with or without surgery. Survival was superior among patients treated with surgery alone compared with patients treated with chemotherapy and surgery (47% vs. 13%, P 0.005). There was no difference in DFS or survival in patients with pulmonary metastases isolated to the lungs compared with other sites of disease recurrence. The presence of metastatic disease FIGURE 2. Kaplan Meier estimated overall survival for all patients. at initial diagnosis, primary tumor site, initial treatment with ifosfamide, and the histologic response to initial therapy were not associated with DFS or survival (data not shown). Figure 3 shows the overall clinical course for all patients. Ultimately, 14 patients survived without disease at last contact (median time after first disease recurrence, 58 months; range, 26 120 months). Forty-three patients developed pulmonary nodules (isolated in 36 patients or combined with other sites of recurrence in 7 patients). Eleven patients did not receive complete surgical excision of pulmonary tumors due to extensive or unresectable pulmonary metastases or to progressive nonpulmonary sites of disease recurrence. Thirty-two patients received complete surgical excision of all pulmonary disease, either by unilateral thoracotomy (n 23), bilateral thoracot-

Recurrent Osteosarcoma/Hawkins and Arndt 2451 TABLE 4 Univariate Analysis of Disease-Free Survival and Overall Survival with Clinical Features for All Patients Variable No. of patients 4-yr DFS rate 4-yr S rate (95% CI) P value a (95% CI) P value a Overall 59 6 (0 12) 23 (10 36) Initial disease recurrence-free interval (mos) 24 41 2 (0 7) 0.01 17 (5 29) 0.05 24 18 15 (0 34) 38 (15 61) Bilateral lung recurrence b Yes 14 0 (0 23) 0.003 0 (0 23) 0.008 No 29 10 (0 21) 35 (17 53) No. of lung nodules c 1 14 14 (0 32) 0.05 33 (7 59) 0.06 1 24 4 (0 12) 17 (2 32) CR2 achieved Yes 40 10 (0 19) NA 33 (18 48) 0.001 No 19 0 (0 18) 0 (0 18) Site Isolated lung 35 6 (0 13) 0.44 24 (10 48) 0.34 Other sites lung 24 9 (0 20) 21 (3 39) Treatment for first recurrence Surgery alone 18 11 (0 26) 0.19 47 (22 73) 0.005 Chemotherapy surgery 41 5 (0 12) 13 (0 23) DFS: disease-free survival; S: overall survival; CI: 95% confidence interval; CR2: second complete response; NA: not applicable. a P values were determined by log-rank comparison. b Analysis of prognostic significance of unilateral versus bilateral lung recurrence was restricted to 43 patients with pulmonary recurrence. c Only 38 patients with pulmonary recurrence and number of lung nodules were included. omy (n 9), or median sternotomy (n 1). Bilateral thoracotomy was only performed in one patient with radiographic evidence of unilateral pulmonary nodules. Of the 23 patients undergoing unilateral thoracotomy, 11 developed isolated ispilateral second pulmonary recurrence, 2 developed contralateral pulmonary recurrence (one in combination with other sites of disease recurrence), 4 developed nonpulmonary sites of second disease recurrence, and 3 had no further disease recurrence. Of the patients who underwent bilateral thoracotomy, five experienced a bilateral second pulmonary recurrence and three experienced a unilateral second pulmonary recurrence. The one patient who underwent a median sternotomy experienced a unilateral pulmonary recurrence. The one patient who received a bilateral thoracotomy for unilateral pulmonary nodules had no surgically identified nodules on the radiographically negative lung. However, he developed contralateral pulmonary recurrence. All five surviving patients with nonpulmonary sites of initial disease recurrence were treated with surgical resection in combination with chemotherapy. Two of these patients also received samarium-153 ethylene diamine tetramethylene phosphonate ( 153 Sm- EDTMP). Twenty patients developed disease recurrence in distant bone: 15 at first disease recurrence, 4 at second disease recurrence, and 1 at third disease recurrence. Four of 20 patients with distant bone recurrence survived without disease at last contact. The median time after distant bone recurrence was 51 months (range, 36 73 months). These four patients were treated with surgery only (n 1), surgery and chemotherapy (n 2), and 153 Sm-EDTMP (n 1). Patients with Isolated Pulmonary Recurrence Who Achieved a Second Complete Remission of Disease Of 36 patients, 30 (83%) with isolated pulmonary recurrence achieved a CR2. This subgroup was the largest in the study population and was analyzed separately. The 4-year DFS rate for this subset of patients was estimated to be 7% (95% CI, 0 16%; Fig. 4) and the 4-year survival rate was estimated to be 28% (95% CI, 11 45%; Fig. 5). The median duration of DFS for patients who developed a second disease recurrence was 8.8 months (range, 2 20 months). Univariate analysis of potential prognostic factors (Table 5) demonstrated that improved DFS was associated with an RFI longer than 24 months. A nonsignificant trend towards improved survival was also observed for an RFI longer than 24 months (48% vs. 19%, P 0.18). Compared with bilateral pulmonary recurrence, unilateral disease recurrence was associated strongly with improved DFS and survival rates. As observed in the

2452 CANCER December 1, 2003 / Volume 98 / Number 11 FIGURE 4. Kaplan Meier estimated disease-free survival for patients with isolated pulmonary recurrence who achieved a second complete response. FIGURE 5. Kaplan Meier estimated overall survival for patients with isolated pulmonary recurrence who achieved a second complete response. FIGURE 3. Clinical course for all patients. DOD: dead of disease; AWD: alive with disease; CR: complete response; NED: no evidence of disease. entire study population, the DFS rate was similar for patients treated with either surgery alone or surgery with chemotherapy. A trend toward improved survival rates for patients treated with surgery alone approached statistical significance (45% vs. 13%, P 0.08). An improved DFS rate (14 % vs. 0%) and a trend toward improved survival rates (40% vs. 23%) for patients with solitary versus multiple pulmonary nodules were observed. The presence of metastatic disease at initial diagnosis, primary tumor site, initial treatment with ifosfamide, and histologic response to initial therapy were not associated with DFS or survival (data not shown). The treatment of isolated pulmonary recurrence varied, most frequently involving

Recurrent Osteosarcoma/Hawkins and Arndt 2453 TABLE 5 Univariate Analysis of Disease-Free Survival and Overall Survival with Clinical Features for Patients with Isolated Lung Recurrence who Achieved a Second Complete Response Variables no. of patients 4-yr DFS rate 4-yr S rate (95% CI) P value a (95% CI) P value a Overall 30 7 (0 16) 28 (11 45) Initial recurrence-free interval (mos) 24 20 0 (0 17) 0.05 19 (0 37) 0.18 24 10 20 (0 45) 48 (15 81) Bilateral lung recurrence Yes 8 0 (0 41) 0.006 0 (0 41) 0.001 No 22 9 (0 21) 37 (16 58) No. of lung nodules b 1 14 14 (0 33) 0.05 40 (13 67) 0.14 1 13 0 (0 25) 23 (0 46) Treatment for first recurrence Surgery alone 15 13 (0 30) 0.6 45 (19 71) 0.08 Chemotherapy and surgery 15 0 (0 22) 13 (0 32) DFS: disease-free survival; S: overall survival; CR2: second complete response; CI: 95% confidence interval. a P values are by log-rank comparison. b Only 27 patients with data on number of pulmonary nodules were included. surgical excision alone (n 15 patients) or in combination with chemotherapy (n 15 patients). Eight patients received IE and five received ICE. In addition, five patients received other multiagent chemotherapy regimens and three received single agents as part of Phase I or Phase II studies. Four patients received two or more different chemotherapy regimens. Eighteen of 28 (64%) second disease recurrences were isolated to the lungs. The treatment of second disease recurrence was highly variable, most frequently involving surgical excision alone (n 11 patients), surgery in combination with chemotherapy or radiotherapy (n 13 patients), or palliative therapy (n 4 patients). Ultimately, nine patients survived without disease at last contact (median time after first disease recurrence, 70 months; range, 33 120 months). Six of the nine patients surviving without disease at last contact were treated exclusively with surgical resection of recurrent tumors (median of two resections; range, 1 3 resections). DISCUSSION The current retrospective evaluation of the pattern of disease recurrence, treatment, and outcome of recurrent OS at two large institutions included only patients receiving contemporary, intensive initial chemotherapy and aggressive surgery. In addition, most patients who received chemotherapy for disease recurrence were treated with IE or ICE. The reported response rates in recurrent OS for IE and ICE were 38% 33 and 30%, 34 respectively. Despite the frequent use of the most active chemotherapeutic regimens for recurrent OS, DFS and survival in the current series were poor. With careful evaluation of the pattern of disease recurrence and treatment strategies used at disease recurrence, the current study demonstrated that prolonged survival was possible, although usually at the cost of repeated surgical procedures with or without further adjuvant therapy. The predominance of isolated pulmonary recurrence in the current series (i.e., 61%) was similar to that reported by others (i.e., 48 93%). 7,9 13,15,19,21,22,26 Most patients with pulmonary recurrence (55%) had 2 or fewer lung nodules. A reduction in the number of pulmonary nodules at initial disease recurrence has been reported since the introduction of adjuvant therapy for OS. 12 The rate of bilateral pulmonary recurrence (i.e., 33%) was similar to that reported by others (i.e., 22 73%). 4 7,9 10,12,13,16,19,26 The majority of patients with isolated pulmonary recurrences achieved a CR2 rate of 83%, which is also similar to the CR2 rate reported by others (i.e., 34 100%). 7,9 13,15,17,19,21,26 The clinical factors associated with improved survival after OS recurrence in the current series have been reported previously. The most important factor associated with prolonged survival was achievement of CR2 with complete surgical excision. Inability to achieve CR2 has been associated with an exceptionally poor outcome, with overall survival rates ranging from 0% to 8%. 8,9,11,13,19,20,25,26 Goorin et al. 12 reported a survival rate of 32% (7 of 22 patients) among patients with unresectable recurrent OS, although the duration of follow-up was not reported. Improved survival has also been associated with prolonged RFI, variably de-

2454 CANCER December 1, 2003 / Volume 98 / Number 11 fined as longer than 6 months, 9 longer than 8 months, 16 and longer than 24 months. 20,26 Other series did not observe an association between survival and RFI. 10 13,15,17,21,24 Saeter et al. reported an improved survival associated with an RFI longer than 21 months with univariate analysis, which lost significance in a multivariate analysis controlling for complete excision, solitary metastasis, and adequate salvage chemotherapy (p. 1090). 19 Bilateral pulmonary nodules have been associated with inferior survival in one report, 16 but not in others. 9,11,12,26 Improved survival has been associated with fewer pulmonary nodules, defined as 1, 19 1 2, 26 or 1 3, 9,16,25 but not in some series. 10 13,15,17 The variability among prognostic factors in these series may be due to the small study populations, changing radiographic evaluations, or differences in surgical approaches for recurrent OS. The role of second-line chemotherapy in the treatment of recurrent OS is controversial. Although the benefit of adjuvant chemotherapy in the initial therapy of OS has been proven in a randomized trial, 2 no randomized study evaluating the benefit of chemotherapy after OS recurrence has been performed. In many series, the number of patients treated with second-line chemotherapy was too small to draw conclusions regarding the impact on survival. 7,21 Pastorino et al. 13 and Tabone et al. 15 observed no difference in survival with or without second-line chemotherapy. However, chemotherapy was given selectively to patients with multiple pulmonary nodules or with an RFI shorter than 1 year RFI 13 or to patients with multiple recurrent tumors or tumors that were difficult to resect. 15 In contrast, Saeter et al. 19 concluded that the delivery of modern intensive chemotherapy (p. 1090) improved overall survival when compared with patients treated with no or inadequate chemotherapy. In two more recent series, a benefit for second-line chemotherapy was not demonstrated. Tsuchiya et al. 25 reported similar overall survival for patients with pulmonary recurrence treated with surgery alone or with second-line chemotherapy and surgery.ferrari et al. 26 reported no difference in overall survival between patients treated with or without second-line chemotherapy after first disease recurrence-.for patients with completely excised isolated pulmonary recurrence, there was no difference in second event-free survival with the addition of chemotherapy. A subgroup analysis of patients with completely excised isolated pulmonary recurrence showed a superior 5-year survival with surgery alone compared with surgery and second-line chemotherapy. Patients with a short RFI had a 5-year survival rate of 30% with surgery alone versus 6% with surgery and second-line chemotherapy (P 0.001). Patients with one or two pulmonary nodules had a 5-year survival rate of 65% with surgery alone versus 30% with surgery and second-line chemotherapy (P 0.02). The superior survival rates observed for patients treated with surgery alone were likely due to the biased use of chemotherapy in patients with marginally excised, early, or multiple recurrences. Because of these limitations, an unambiguous conclusion regarding the role of secondline chemotherapy could not be drawn. However, the current series did not provide evidence supporting the routine use of second-line chemotherapy in the treatment of recurrent OS. Distant bone recurrence of OS reportedly has an exceptionally poor prognosis. Tabone et al. 15 reported that only one of seven patients with distant bone recurrence survived disease free with short-term follow-up. Ward et al. 16 reported a 4-year postrecurrence survival rate of 0% n 17 patients with distant bone recurrence. Ferrari et al. 20 reported no surviving patients among four patients with bone recurrence. Analyzing a subsequent cohort of patients, Ferrari et al. 26 reported a 5-year postrecurrence survival rate of 11% in 37 patients with extrapulmonary recurrence (of whom 31 of 37 had distant bone recurrence alone or in combination with other sites of recurrence). In contrast to these results, 20% of patients with distant bone recurrence (either with initial or subsequent recurrence) survived without active disease in the current series. Multimodality therapy may have contributed to the survival of these four patients, including the use of high-dose 153 Sm-EDTMP in one case. 37 153 Sm-EDTMP is a bone-seeking radiopharmaceutical with a biodistribution similar to technetium-99m methylene diphosphonate, with rapid uptake and avid retention in bone. Because 153 Sm emits a medium energy beta particle, 153 Sm-EDTMP can deliver a relatively high dose of radiation targeted to bone, with preferential uptake in sites of increased bone turnover. Others have reported therapeutic activity of 153 Sm-EDTMP in recurrent or unresectable OS. 38,39 Further investigation of this novel form of biophysically targeted therapy is warranted. An unusual aspect of the current series was the evaluation of DFS after first OS recurrence. Most published series reported only overall survival rates, which ranged from 13% to 58%, 4 13,15 17,19 21,25,26 whereas only three series reported DFS or event-free survival rates after disease recurrence. 12,15,26 The discrepancy between DFS and survival rates for the entire study population (6% vs. 23%) and for patients with isolated pulmonary recurrence achieving CR2 (7% vs. 28%) showed the high probability of second disease recurrence. Despite the ability to achieve CR2, most patients with recurrent OS developed other disease re-

Recurrent Osteosarcoma/Hawkins and Arndt 2455 currence (Fig. 3). DFS analysis also demonstrated that the time to second disease recurrence was rapid, occurring at a median of 8.8 months after achieving CR2 in patients with isolated pulmonary recurrence. The short CR2 duration and high probability of a second OS recurrence provide an opportunity for a novel study design to evaluate a therapeutic intervention. The traditional Phase II study design assesses response radiographically in patients with measurable disease. This assessment of efficacy is difficult in patients with recurrent OS for three reasons. First, pathologically responding nodules of OS may not change in size radiographically despite efficacious treatment, leading the investigational treatment to be deemed ineffective in error. Second, many patients with recurrent OS may not be considered candidates for Phase II agents if all sites of disease recurrence are amenable to surgical excision. Patients with resectable recurrent tumors frequently have complete excision without a trial of second-line chemotherapy, thus excluding them from enrollment in a classic Phase II trial and limiting the number of patients available to investigate novel treatment strategies. Given the poor DFS and survival observed in patients in CR2, the risk of investigational therapy appears to be warranted even in a patient with resectable recurrent tumors. Third, some novel biologic therapies are unlikely to be effective in patients with extensive disease but might improve outcome for patients with minimal residual disease after surgical excision. This rationale supported the adjuvant use of MTP-PE in the adjuvant therapy of surgically resected OS. 40 Similarly, this rationale would support the adjuvant use of aerosolized granulocyte-macrophage colony-stimulating factor as a novel, pulmonary immunotherapy to prevent subsequent disease recurrence in the lung. 41 Despite remarkable improvements in the DFS and survival for patients with newly diagnosed OS, the prognosis after OS recurrence remains poor. Bilateral pulmonary nodules, more than one pulmonary nodule, and short duration from initial diagnosis to first disease recurrence are predictive of poor DFS and survival. Repeated surgical excision may be necessary to achieve prolonged survival and the role of chemotherapy to treat recurrent OS is unclear. Given the poor DFS and relatively short time to second disease recurrence (even for patients with isolated pulmonary recurrent tumors amenable to surgical excision), time to next progression after complete excision of pulmonary disease could be used as an end point to assess the efficacy of a therapeutic intervention. New treatment strategies, including novel chemotherapy or biologic modulators, are needed to improve the prognosis of recurrent OS. REFERENCES 1. Gurney JG, Swensen AR, Bulterys M. Malignant bone tumors. In: Ries LAG, Smith MA, Gurney JG, et al., editors. Cancer incidence and survival among children and adolescents: United States SEER program 1975-1995. National Cancer Institute, SEER Program. NIH Pub. No. 99-4649. Besthesda, MD: National Cancer Institute, 1999:99 110. 2. 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. 1986;314:1600 1606. 3. Meyers PA, Gorlick R. Osteosarcoma. Pediatr Clin North Am. 1997;44:973 989. 4. Martini N, Huvos AG, Mike V, Marcove RC, Beattie EJ Jr. Multiple pulmonary resections in the treatment of osteogenic sarcoma. Ann Thorac Surg. 1971;12:271 280. 5. Spanos PK, Payne WS, Ivins JC, Pritchard DJ. Pulmonary resection for metastatic osteogenic sarcoma. J Bone Joint Surg [A]. 1976;58:624 628. 6. Telander RL, Pairolero PC, Pritchard DJ, Sim FH, Gilchrist GS. Resection of pulmonary metastatic osteogenic sarcoma in children. Surgery. 1978;84:335 341. 7. Rosenberg SA, Flye MW, Conkle D, Seipp CA, Levine AS, Simon RM. Treatment of osteogenic sarcoma. II. Aggressive resection of pulmonary metastases. Cancer Treat Rep. 1979; 63:753 756. 8. Schaller RT, Haas J, Schaller J, Morgan A, Bleyer A. Improved survival in children with osteosarcoma following resection of pulmonary metastases. J Pediatr Surg. 1982;17:546 550. 9. Putnam JB, Roth JA, Wesley MN, Johnston MR, Rosenberg SA. Survival following aggressive resection of pulmonary metastases from osteogenic sarcoma: analysis of prognostic factors. Ann Thorac Surg. 1983;36:516 523. 10. Goorin AM, Delorey MJ, Lack EE, et al. Prognostic significance of complete surgical resection of pulmonary metastases in patients with osteogenic sarcoma: analysis of 32 patients. J Clin Oncol. 1984;2:425 431. 11. Carter SR, Grimer RJ, Sneath RS, Matthews HR. Results of thoracotomy in osteogenic sarcoma with pulmonary metastases. Thorax. 1991;46:727 731. 12. Goorin AM, Shuster JJ, Baker A, Horowitz ME, Meyer WH, Link MP. Changing pattern of pulmonary metastases with adjuvant chemotherapy in patients with osteosarcoma: results from the Multiinstitutional Osteosarcoma Study. J Clin Oncol. 1991;9:600 605. 13. Pastorino U, Gasparini M, Tavecchio L, et al. The contribution of salvage surgery to the management of childhood osteosarcoma. J Clin Oncol. 1991;9:1357 1362. 14. La Quaglia MP. The surgical management of metastases in pediatric cancer. Semin Pediatr Surg. 1993;2:75 82. 15. Tabone M-D, Kalifa C, Rodary C, Raquin M, Valteau-Couranet D, Lemerle J. Osteosarcoma recurrences in pediatric patients previously treated with intensive chemotherapy. J Clin Oncol. 1994;12:2614 2620. 16. Ward WG, Mikaelian K, Dorey F, et al. Pulmonary metastases of stage IIB extremity osteosarcoma and subsequent pulmonary metastases. J Clin Oncol. 1994;12:1849 1858. 17. Heij HA, Vos A, de Kraker J, Voûte PA. Prognostic factors in surgery for pulmonary metastases in children. Surgery. 1994; 115:687 693. 18. Temeck BK, Wexler LH, Steinberg SM, McClure LL, Horowitz M, Pass HI. Metastasectomy for sarcomatous pediatric histologies: results and prognostic factors. Ann Thorac Surg. 1995;59:1385 1390.

2456 CANCER December 1, 2003 / Volume 98 / Number 11 19. Saeter G, Høie J, Stenwig AE, Johannsson AK, Hannisdal E, Solheim Ø. Systemic relapse of patients with osteogenic sarcoma: prognostic factors in long term survival. Cancer. 1995;75:1084 1093. 20. Ferrari S, Bacci G, Picci P, et al. Long-term follow-up and post-relapse survival in patients with non-metastatic osteosarcoma of the extremity treated with neoadjuvant chemotherapy. Ann Oncol. 1997;8:765 771. 21. Ellis PM, Tattersall MHN, McCaughan B, Stalley P. Osteosarcoma and pulmonary metastases: 15-year experience from a single institution. AustNZJSurg. 1997;67:625 629. 22. Bacci G, Ferrari S, Longhi A, et al. Pattern of relapse in patients with osteosarcoma of the extremities treated with neoadjuvant chemotherapy. Eur J Cancer. 2001;37:32 38. 23. Weeden S, Grimmer RJ, Cannon SR, Taminiau AH, Uscinska BM. The effect of local recurrence on survival of resected osteosarcoma. Eur J Cancer. 2001;37:39 46. 24. McCarville MB, Kaste SC, Cain AM, Goloubeva O, Rao BN, Pratt CB. Prognostic factors and imaging patterns of recurrent pulmonary nodules after thoracotomy in children with osteosarcoma. Cancer. 2001;91:1170 1176. 25. Tsuchiya H, Kanazawa Y, Abdel-Wanis ME, et al. Effect of timing of pulmonary metastases identification on prognosis of patients with osteosarcoma: the Japanese Musculoskeletal Oncology Group Study. J Clin Oncol. 2002;20:3470 3477. 26. Ferrari S, Briccoli A, Mercuri M, et al. Post-relapse survival in osteosarcoma of the extremities: prognostic factors for longterm survival. J Clin Oncol. 2003;21:710 715. 27. Miser J, Arndt C, Smithson W, et al. Treatment of high-grade osteosarcoma (OGS) with ifosfamide (IFOS), MESNA (MES), Adriamycin (ADR), high-dose methotrexate (HDMTX), with or without cisplatin (CDDP); results of two pilot trials [abstract]. Proc Am Soc Clin Oncol. 1994;13:421a. 28. Arndt C, Miser J, Prichard D, et al. Treatment of high-grade osteosarcoma (OGS) with ifosfamide (IFOS), MESNA (MES), Adriamycin (ADR), high-dose methotrexate (HDMTX), and cisplatin (CDDP) [abstract]. Med Pediatr Oncol. 1996;27: 227a. 29. Meyers P, Schwartz C, Bernstein M, et al. Addition of ifosfamide and muramyl tripeptide to cisplatin, doxorubicin, and high-dose methotrexate improves event free survival (EFS) in localized osteosarcoma (OS) [abstract]. Proc Am Soc Clin Oncol. 2001;20:367a. 30. Seibel N, Krailo M, Sato J, et al. Phase II window of topotecan in newly diagnosed metastatic osteosarcoma-ccg 7943 [abstract]. Proc Am Soc Clin Oncol. 2001;20:378a. 31. Nachman J, Simon M, Rubin C, Peabody T. Intensive preoperative therapy with escalating doses of ifosfamide and carboplatin in conjunction with Adriamycin for patients with localized osteosarcoma [abstract]. Med Pediatr Oncol. 2001;37:176a. 32. Womer RB, Daller RT, Fenton JG, Miser JS. Granulocyte colony stimulating factor permits dose intensification by interval compression in the treatment of Ewing s sarcomas and soft tissue sarcomas in children. Eur J Cancer. 2000;36: 87 94. 33. Miser JS, Kinsella TJ, Triche TJ, et al. Ifosfamide with mesna uroprotection and etoposide: an effective regimen in the treatment of recurrent sarcomas and other tumors of children and young adults. J Clin Oncol. 1987;5:1191 1198. 34. Cairo MS, Shen V, Krailo MD, et al. Prospective randomized trial between two doses of granulocyte colony-stimulating factor after ifosfamide, carboplatin, and etoposide in children with recurrent or refractory solid tumors: a Children s Cancer Group report. J Pediatr Hematol Oncol. 2001;23:30 38. 35. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc. 1958;53:457 481. 36. Peto R, Pike MC, Armitage P, et al. Design and analysis of randomized clinical trials requiring prolonged observation of each patient: I-analysis and examples. Br J Cancer. 1977; 35:1 39. 37. Anderson PM, Wiseman GA, Dispenzieri A, et al. High-dose samarium-153 ethylene diamine tetramethylene phosphonate: low toxicity of skeletal irradiation in patients with osteosarcoma and bone metastases. J Clin Oncol. 2002;20: 189 196. 38. Franzius C, Bielack S, Sciuk J, Vollet B, Jürgens H, Schober O. High activity samarium-152 EDTMP therapy in unresectable osteosarcoma. Nuklearmedizin. 1999;38:337 340. 39. Welsh JS, Thurman SA, Borzillary S, Schwartz CL, Howard SP. Sm-153 EDTMP for recurrent metastatic osteosarcoma. Clin Adv Hematol Oncol. 2003;1:181 183. 40. Kleinerman ES. Biologic therapy for osteosarcoma using liposome-encapsulated muramyl tripeptide. Hematol Oncol Clin North Am. 1995;9:927 938. 41. Anderson PM, Markovic SN, Sloan JA, et al. Aerosol granulocyte macrophage-colony stimulating factor: a low toxicity, lung-specific biological therapy in patients with lung metastases. Clin Cancer Res. 1999;5:2316 2323.