Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Introduction

Most data on osteosarcoma is derived from pediatric studies. Although the majority of adult patients with osteosarcoma are young adults, who might be treated in a similar fashion, experience derived from a slightly older population is helpful in directing therapy. We describe a consecutive series of patients with primary, high-grade osteosarcoma of the extremities treated at the Department of Melanoma/Sarcoma Medical Oncology (now the Department of Sarcoma Medical Oncology) at the University of Texas M.D. Anderson Cancer Center between May, 1980 and October, 1991, using systemic adriamycin and intra-arterial cisplatin as primary chemotherapy. The series is divided into three groups of patients based on the postoperative chemotherapy given. Preliminary reports of the first two groups of patients have been published1–5; but the third group has been reported only at meetings.6–9 Taken together, the three groups illustrate the advantages of the neoadjuvant strategy.

Except for four, all patients older than 16 with primary, high-grade osteosarcoma of the extremities and no demonstrable metastatic disease treated with preoperative chemotherapy at the University of Texas, M.D. Anderson Cancer Center between May, 1980 and October, 1991 are the subjects of this report. Four patients who declined initial surgery but underwent delayed surgery (including two long-term, disease free survivors) were excluded from the analysis. Otherwise, this represents a consecutive series of 123 patients. They are further divided into three groups depending on the time period in which they were treated and the approach to postoperative chemotherapy.

Age, sex, and skeletal distribution were typical for osteosarcoma, except that patients under age 16, who were treated on our Pediatric service with a different regimen, are excluded. Males outnumbered females by about 3:2; 77% of the patients were below the age of 30; the most commonly involved bone was the femur; and three quarters of the tumors were located around the knee. No significant differences in the demographics of the three groups were detected. Most patients had conventional osteosarcoma (79%) and osteoblastic osteosarcoma was the most frequent subtype. There was an increased proportion of fibroblastic osteosarcoma in the third group.

After informed consent was obtained, all the patients were treated with adriamycin, 90 mg/m2 by continuous 96-h ambulatory intravenous infusion through a percutaneous silicone elastomer central venous catheter, starting on day 1. At the end of the infusion (day 5), they were admitted to the hospital, and on day 6, underwent an arteriogram with subtraction images and catheter placement. After verification of correct catheter position to infuse the tumor by nuclear flow study, patients received intra-arterial cisplatin.

Group 1 was treated at a dose of 120 mg/m2, infused over 2 h. Groups 2 and 3 were treated at a dose of 160 mg/m2 infused over 24 h. All the patients received intensive intravenous hydration (≥250 ml/h) and mannitol diuresis. Intake and output was balanced as needed by infusion of furosemides. After hypomagnesemia was noted in the initial patients, magnesium sulfate (MgSO4) was routinely added to the intravenous fluids. The cisplatin was initially infused in normal saline, but later, it was infused in 3% saline. Chemotherapy cycles were repeated at 4-week intervals. Patients in group 1 received a median of three courses of preoperative chemotherapy, the time required to obtain a custom prosthesis for most patients. Preoperative chemotherapy was stopped early in the case of disease progression and was extended if the prosthesis could not be obtained by the end of the third course. Postoperatively, patients with ≥60% tumor necrosis continued the same chemotherapy intravenously until the development of grade 1–2 peripheral neuropathy. Thereafter, adriamycin was continued and cisplatin was replaced with dacarbazine 750 mg/m2 as a 96-h infusion (ADIC). Twelve cycles of chemotherapy were administered. Drug doses were decreased for febrile neutropenia with morbidity or documented infection. The doses of cisplatin and dacarbazine were selectively decreased for thrombocytopenia or delayed granulocyte recovery (4 weeks). Hematopoietic growth factors were not used. Grade 2 mucositis was an indication to decrease the duration of the adriamycin infusion to 48 h (and in the first group to 24 h). Only if mucositis≥grade 2 persisted after the shorter infusion was the dose decreased to 75 and 60 mg/m2, but never to <60 mg/m2. Patients in the first group also participated in our studies assessing the cardiac toxicity of continuous-infusion adriamycin and were monitored with endomyocardial biopsies every four courses after a cumulative adriamycin dose of 450 mg/m2. The five patients with <60% tumor necrosis were felt to have suboptimal response to primary chemotherapy and were allowed to receive alternative treatment at the discretion of their primary physicians. Two received adriamycin and dacarbazine (ADIC), two received high-dose methotrexate, and one received no further chemotherapy.

After the analysis of the patients in group 1 indicated the prognostic importance of obtaining a good response defined as ≥90% necrosis in the resected specimen, processed and analyzed by the method of Raymond et al,2 three modifications were made in the treatment program for group 2. The dose of cisplatin was increased to 160 mg/m2, the duration of preoperative therapy was increased to four courses, and the postoperative therapy was modified for poor responders (<90% necrosis). Such patients received alternating chemotherapy with four courses of high-dose methotrexate, 8 gm/m2 with leucovorin rescue repeated every 2 weeks, followed by two courses of ADIC, and then, by two courses of the combination of bleomycin, cyclophosphamide, and actinomycin-D (BCD)10 at 3-week intervals. The entire cycle of methotrexate, ADIC, and BCD was repeated twice. Good responders continued on adriamycin-cisplatin/ADIC for only three courses. Group 3 received identical primary chemotherapy to that given to group 2. Postoperatively, good responders received three cycles of adriamycin with cisplatin or dacarbazine. Thereafter, they were treated with three cycles of high-dose methotrexate. Poor responders were also treated in the same way as those in group 2; however, ifosfamide 2 g/m2 given as a 2-h infusion daily for 5 days with continuous infusion mesna 1,200 mg/m2 qd×5 after a loading dose of 400 mg/m2 on day 1, replaced BCD.

The median follow-up time for censored patients in groups 1, 2, and 3 respectively as shown in the figures are 134, 91, and 55 months. With the exception of three patients lost to follow-up, the minimum follow-up is 65, 37, and 35 months, respectively. Extending the median and minimum follow-up times to 140, 141, and 120 months and 68, 41, and 51 months (data not shown) did not alter any of the outcomes.

Response to chemotherapy is shown in Table 1. One patient died of pulmonary embolism prior to resection. She is included in the overall relapse-free survival statistics but not in the response rate calculations or the relapse-free survival analyses when stratified by response group. Sixty percent of patients achieved a good response. Of the poor responders, 29% had 60–89% tumor necrosis, and 11% had <60% tumor necrosis. There was no significant difference in the rate of ≥90% necrosis or <60% necrosis between the three groups or between patients who received 120 mg/m2 and 160 mg/m2 of cisplatin.

Table 1 Response to therapy
Full size table

Continuous relapse-free survival for the good responders (>90% necrosis) is shown in Fig. 1. There was no significant difference between any of the three groups. If anything, there was a suggestion that the first group, who got the longest postoperative adriamycin therapy, had the best relapse-free survival. There was no evidence that the addition of methotrexate to group 3 improved the relapse-free survival over that of group 2.

Fig. 1
figure 1_19

Continuous relapse-free survival of good responders (>90% necrosis) by treatment group. There are no significant differences between the groups, but there is a suggestion of improvement in the first group

Full size image

Continuous relapse-free survival for poor responders (<90% necrosis) is shown in Fig. 2. In group 1, only 13% of poor responders have not relapsed, and their continuous relapse-free survival is no different from our historical control series treated with surgery alone (Fig. 3, “historical control”). The addition of methotrexate and BCD in group 2 led to a small but statistically significant improvement in continuous relapse-free survival of 34% (p=0.04). The substitution of ifosfamide for BCD in group 3 led to a further statistically significant improvement in continuous relapse-free survival of 67% (p=0.01).

Fig. 2
figure 2_19

Continuous relapse-free survival of poor responders (<90% necrosis) by treatment group. Each successive group shows significant improvement

Full size image
Fig. 3
figure 3_19

Continuous relapse-free survival of good and poor responders compared with a historical control (surgery only). The only benefit is in good responders

Full size image

Continuous relapse-free survival for all patients in groups 1, 2, and 3 is shown in Fig. 4. As a result of the improved continuous relapse-free survival for poor responders in group 3, all the patients in group 3 have a superior continuous relapse-free survival when compared to patients in group 1 (p=0.04). The five-year relapse-free survival for patients in groups 1, 2, and 3 is 54%, 61%, and 70%, respectively.

Fig. 4
figure 4_19

Continuous relapse-free survival of all treated patients by treatment group. Each successive group shows significant improvement

Full size image

This series contrasts with other published and unpublished series using alternating chemotherapy with more drugs in the preoperative chemotherapy regimen. With two drugs for induction therapy, we can demonstrate that changing therapy for poor response, a keystone of the neoadjuvant strategy, was indeed effective confirming the original observation by Rosen11 and the findings of Bacci12 using an induction chemotherapy similar to ours but with the addition of high-dose methotrexate.

Although the majority of patients in this series were young enough to qualify for pediatric studies, there are some differences seen in the adult population. Patients with secondary osteosarcomas, most commonly post-radiation sarcomas, but also those arising in pre-existing bone disease such as fibrous dysplasia, bone infarcts, or Paget’s disease, have a worse prognosis and were excluded from the previous series. Similarly, osteosarcoma of the jaw which is more common in adults and which tends to be locally recurrent rather than metastatic and to have a better prognosis, was also excluded.

Not excluded were more unusual variants of osteosarcoma, which are more common in adults than in children. With the exception of telangiectatic osteosarcoma, which responded in a similar fashion to the conventional subtypes,13 other high-grade variants fared significantly worse.14,15 These included dedifferentiated parosteal osteosarcoma, dedifferentiated well-differentiated-intraosseous osteosarcoma, small cell osteosarcoma, and high-grade surface osteosarcoma (Fig. 5).

Fig. 5
figure 5_19

Continuous relapse-free survival of all treated patients by histologic group. Telangiectatic osteosarcoma has similar relapse-free survival to the conventional subtypes. Other subtypes have significantly inferior prognosis

Full size image

In an attempt to improve the prognosis of high-risk patients, we imitated a study using full doses of adriamycin, ifosfamide, and cisplatin. This could not be accomplished without stem cell support. Full details of this group of patients are beyond the scope of this manuscript but are published elsewhere.16 Suffice it to say, while initial results were promising based on initial response to chemotherapy, patients were left with seriously impaired bone marrow reserve after induction therapy, despite stem cell support, and thus could not tolerate sufficient postoperative therapy; so, median relapse-free survival was only 19 months.

This communication also recognizes a recent report in which a worse prognosis was described in patients over 65 years when compared to that in younger patients.17 The older age group was characterized by a longer time lapse from the onset of symptoms to diagnosis, a larger number of metastatic cases, less use of limb salvage and a reduced number of patients treated with chemotherapy (compare this experience with the above report of impaired bone marrow reserve), and more patients excluded from clinical trials as opposed to the younger age group. Only one of out patients was older than 65, one was 65, and 4 were 57-63. All received chemotherapy.

Conspectus

New drugs are needed to salvage the small proportion of patients who do poorly with our current regimens, and studying these new approaches in patients with poor prognostic characteristics is warranted.