In vitro radiation studies on Ewing's sarcoma cell lines and human bone marrow: application to the clinical use of total body irradiation (TBI)

Does consolidation with autologous stem cell transplantation improve the outcome of children with metastatic or relapsed Ewing sarcoma?

OBJECTIVE

To evaluate the role of high-dose chemotherapy (HDC) and autologous stem cell transplantation (ASCT) as consolidation therapy for children with high-risk Ewing sarcoma (ES) treated at The Hospital for Sick Children (SickKids), Toronto.

PATIENTS AND METHODS

The charts of children treated for high-risk ES (defined as metastatic at diagnosis or relapsed) between 1990 and 2005 at SickKids were reviewed. Forty-five children were identified. Twenty patients received ASCT after induction with vincristine, doxorubicin, ifosfamide, cyclophosphamide, and etoposide. Patients with resectable tumor or lung metastases underwent surgery and those with non-resectable tumors were treated with irradiation. Twenty-five patients were treated with conventional chemotherapy (CC). Primary metastatic patients were treated with either a local protocol or as per POG 9354. At relapse, patients were treated with topotecan, cyclophosphamide, then ifosfamide, carboplatin, and etoposide (ICE). Local control was attained with surgery and/or irradiation.

RESULTS

Ten of the 20 patients treated with ASCT are alive (median follow-up 6 years), with 8/10 being in remission more than 5 years from diagnosis. The 3-year overall survival (OS) for ASCT was 59%, (95% CI: 36%, 81%) compared to 34% (14%, 53%) for patients treated with CC (P-value = 0.06). The 3-year event-free survival (EFS) for the ASCT was 39% (17%, 60%) compared to 32% (13%, 50%) in the CC group (P = 0.08).

CONCLUSION

ASCT appears to add some benefit to conventional multimodality therapy for children with high-risk ES. Randomized controlled trials are warranted.

High-dose chemoradiotherapy (HDC) in the Ewing family of tumors (EFT)

EFT is defined by the expression of ews/ets fusion genes. The type of the fusion transcript impacts on the clinical biology. EFT requires risk adapted treatment. A risk-adapted treatment is determined by tumor localisation, tumor stage and volume. For metastatic and relapsed disease the pattern of spread and the time of relapse are the determinants of risk stratification. Staging of Ewing tumors has been considerably improved by magnetic resonance imaging and modern isotope scanning techniques. However, the determination of the extent of the metastatic spread in particular number of involved bones remains an unresolved issue. The prognosis for high-risk Ewing tumors has been improved by multimodal and high-dose radio/chemotherapy (HDC). The concepts for high-dose therapy in Ewing tumors are based on dose response and dose intensity relationships. In single agent HDC most experience exists with Melphalan. Several chemotherapeutic agents have been used in combination HDC with or without TBI such as Adriamycin, BCNU, Busulphan, Carboplatin, Cyclophosphamide, Etoposide, Melphalan, Thiotepa Procarbazin and Vincristine. To date, superiority of any high-dose chemotherapy regimen has not been established. However, the clinical biology, the pattern of spread and the time of relapse determine the prognosis of patient who are eligible for HDC. In particular, patients with multifocal bone or bone marrow metastases have a poorer prognosis than patients with lung metastases. In addition, patients with a relapse within 24 months have a poorer prognosis than patients with a relapse later than 24 months after diagnosis. This review will analyze the results of single- and multi-agent chemotherapy with respect to agent combination, dose and risk stratum of patient population. Future therapeutic modalities for the treatment of EFT might encompass immunotherapeutic and genetic strategies including allogeneic stem cell transplantation.

Busulfan, melphalan, and thiotepa with or without total marrow irradiation with hematopoietic stem cell rescue for poor-risk Ewing-Sarcoma-Family tumors

BACKGROUND

Survival following metastatic or recurrent Ewing sarcoma family tumors (ESFT) remains <25%. Myeloablative therapy with hematopoietic stem cell transplantation (HSCT) may improve survival for poor-risk ESFT. We describe the toxicity and efficacy of a myeloablative chemotherapy regimen, followed by a second myeloablative radiotherapy regimen as consolidation treatment for poor-risk ESFT.

PROCEDURE

Sixteen patients with poor-risk ESFT were treated with myeloablative therapy followed by HSCT. All patients received busulfan, melphalan, and thiotepa (BuMelTT) as chemotherapy conditioning. Nine patients received total marrow irradiation (TMI) as a second myeloablative therapy, also followed by HSCT. Seven patients were excluded from TMI because of inadequate peripheral blood stem cell harvest, extensive prior radiation therapy, early disease progression, orpatient refusal. The disease status prior to my eloablative therapy was first complete response (CR1) in three patients, CR2 in nine, second partial response (PR2) in one, CR3 in one, and progressive disease (PD) in two.

RESULTS

One patient died of regimen-related toxicity, one from late pulmonary toxicity, and one following allogeneic transplantation for myelodysplasia. Eight developed recurrent disease (median time to progression 6.8 months). Six survive without relapse from 27 to 66 months following BuMelTT (median follow-up 42 months), all of whom received both BuMelTT and TMI patients (3-year event-free survival 36%).

CONCLUSIONS

Dual myeloablative therapy with BuMelTT and TMI was a feasible and promising treatment approach for patients with poor-risk ESFT. Inability to collect sufficient PBSC and extensive previous radiation therapy limit the ability to deliver TMI as a second HSCT conditioning regimen.

A review of human cell radiosensitivity in vitro

The survival curves of 694 human cell lines irradiated in exponentially growing phase in vitro were collected from the literature. Among them, 271 were derived from tumors, 423 were nontransformed fibroblasts and other normal cell strains from healthy people or people with some genetic disorders. Seventy-six different cell types are identified, and a specific radiosensitivity could be associated with each, using D and surviving fraction at 2 Gy. Technical factors such as culture medium, feeder cells, and scoring method were found to affect intrinsic radiosensitivity. In particular, the cell type is not a discriminating factor when cells are studied in agar. Results obtained with cells irradiated in agar must be used cautiously, depending on how the cells were prepared for the experiments. The use of feeder cells narrows the range of radiosensitivity of human cells. For cells irradiated as monolayer, it was possible to build a scale of radiosensitivity according to cell type, ranging, in terms of D from 0.6 Gy for the most sensitive cell lines to more than 4 Gy for the most resistant. Considering that, in most cases, we could estimate the variation of radiosensitivity within each cell type, our classification among cell types can be used by researchers to place their results in the context of the literature.

Bone sarcomas

The literature on radiotherapy for sarcomas originating in bone is limited, and the disease is uncommon. Consequently, this review is based on only 28 scientific articles, including 6 randomized studies, I prospective study, and II retrospective studies. These studies involve 1,394 patients. Radiotherapy is of limited value for treating the primary tumors of osteosarcoma. The effects of prophylactic lung irradiation for treating osteosarcoma are unconfirmed. Radiotherapy represents an integrated component of primary treatment for Ewing's sarcoma, but different fractionation schedules should be assessed. Radiotherapy is of limited value in the treatment of chondrosarcoma.

Radiation sensitivity of human B-lineage lymphoid precursor cells

We studied the radiation sensitivity of eight immunophenotypically distinct B-lineage lymphoid precursor cell (LPC) lines of acute lymphoblastic leukemia (ALL) or fetal liver origin corresponding to discrete developmental stages of human B-cell ontogeny. The radiation sensitivity of B-lineage LPC showed a temporal association with the distinct stages of development. FL112 and FL114 fetal liver pro-B cells (Stage 0 B-lineage LPC) with germline immunoglobulin heavy chain (IgH) genes but rearranged T-cell receptor gamma (T gamma) genes (DO of FL112 = 80.3 cGy, DO of FL114 = 50.2 cGy), REH ALL pre-pre-B cells (Stage I B-lineage LPC) with rearranged IgH and T gamma genes (DO = 66.1 cGy), and NALM-6 ALL pre-pre-B/pre-B cells (Stage II B-lineage LPC) (DO = 50.5 cGy) corresponding to the earliest three stages of human B-lymphocyte development were the most radiation sensitive B-lineage LPC populations. By comparison, KM-3 ALL pre-B (Stage III B-lineage LPC) (DO = 194.7 cGy), HPB-NULL ALL pre-B (Stage IV B-lineage LPC) (DO = 134.6 cGy), and sIgM+ RAJI/NAMALWA early B (Stage Va/b B-lineage LPC) cell lines (DO of RAJI = 144.0 cGy, DO of NAMALWA = 165.5 cGy) corresponding to the later stages of human B-lymphocyte development were much more radiation resistant. These results indicate that the radiation sensitivity of B-lineage LPC decreases during maturation within the B-lineage lymphoid precursor pathway. By comparison, the S-phase index (% of S-phase cells as determined by DNA flow cytometry) or proliferation index (% S + G2M), cellular protein content, intracellular glutathione (GSH) level, glutathione-S-transferase (GST) activity, intracellular pH, or free cytoplasmic calcium concentration did not correlate with the radiation sensitivity of the B-lineage LPC.

Radiobiological features of human pluripotent bone marrow progenitor cells (CFU-GEMM)

The purpose of this study was to evaluate the radiobiologic features of human pluripotent bone marrow progenitor cells (CFU-GEMM; colony forming unit-granulocyte-erythroid-macrophage/monocyte-megakaryocyte). Experiments were performed using fresh bone marrow cells as well as bone marrow cells stimulated with recombinant granulocyte-macrophage-colony stimulating factor (rGM-CSF) to increase the CFU-GEMM pool. The D0 values for CFU-GEMM in normal bone marrow samples (n = 9) ranged from 30.9 cGy to 85.7 cGy (mean +/- SE = 54.4 cGy +/- 6.2 cGy) and the D0 value of the composite radiation survival curve was 56.9 cGy, indicating that CFU-GEMM were acutely sensitive to the lethal effects of ionizing radiation. There was no distinct shoulder on the single dose radiation survival curves with Dq values ranging from -29.6 cGy to 4.4 cGy, and no increase in CFU-GEMM survival was observed when the radiation was fractionated. Hence, CFU-GEMM were unable to repair sublethal radiation damage. These findings confirm and extend previous studies on the radiobiologic features of human hematopoietic progenitor cell populations.

Effects of recombinant growth factors on radiation survival of human bone marrow progenitor cells

The purpose of this study was to evaluate the individual radioprotective effects of 4 distinct purified recombinant human hematopoietic growth factors, namely recombinant human granulocyte-macrophage colony stimulating factor (rGM-CSF), recombinant human granulocyte colony stimulating factor (rG-CSF), recombinant human interleukin 1 (rIL-1), and recombinant human interleukin 2 (rIL-2) on human myeloid (CFU-GM) and erythroid (BFU-E) bone marrow progenitor cells. We demonstrate that (a) preconditioning with rGM-CSF, rG-CSF, or rIL-1 enables CFU-GM to repair sublethal radiation damage and renders CFU-GM less radiosensitive, (b) preconditioning with rGM-CSF or rIL-1 enables BFU-E to repair sublethal radiation damage, and (c) preconditioning with rIL-2 does not increase the radiation survival of CFU-GM or BFU-E. The effects of recombinant growth factors, in particular rGM-CSF, on the radiation damage repair, radiosensitivity, and proliferative activity of bone marrow progenitor cells resulted in a substantial increase in the mean numbers of progenitor cell-derived hematopoietic colonies in irradiated marrow samples. The effects of rGM-CSF on the radiation response of CFU-GM and BFU-E, and the effects of rG-CSF as well as rIL-1 on the radiation response of CFU-GM did not appear to require the presence of T-cells/T-cell precursors, NK-cells, B-cells/B-cell precursors, monocytes, macrophages, MY8 antigen positive non-CFU-GM myeloblasts, promyelocytes, myelocytes, metamyelocytes, granulocytes, or glycophorin A positive erythroid cells since virtually identical results were obtained with unsorted marrow samples or highly purified fluorescence activated cell sorter (FACS) isolated progenitor cell suspensions. To our knowledge, this report represents the first study on recombinant human growth factor-induced modulation of the radiation responses of normal human bone marrow progenitor cells.

Treatment of sarcomas of the chest wall using intensive combined modality therapy

As part of two sequential protocols using intensive combined modality treatment in pediatric and adolescent sarcomas, 31 consecutive patients with primary chest wall tumors were treated between November 1977 and March 1986. This group included 13 patients with peripheral neuroepithelioma (Askin's tumor), 11 patients with Ewing's sarcoma, 3 patients with rhabdomyosarcoma, and 4 patients with undifferentiated sarcomas. Following complete work-up, 17 patients presented with localized disease and 14 patients presented with metastases. Patients received intensive combined modality treatment with combination chemotherapy (vincristine, cyclophosphamide, Adriamycin, +/- actinomycin-D and DTIC) and high-dose conventionally fractionated radiation therapy to the primary (55-60 Gy) and non-pulmonary metastases (45-50 Gy). Radiation techniques used for the primary chest wall tumor varied with the clinical presentation. Patients achieving a complete response received either low-dose fractionated TBI (1.5 Gy/0.15 Gy fx/5 weeks) or high-dose TBI (8 Gy/4 Gy fx/2 days) and an intensive cycle of chemotherapy followed by autologous bone marrow transplantation. Twenty-five of 31 patients were judged to have a complete response (including 1 patient with complete resection). With minimum follow-up of 6 months and median follow-up of 36 months from completion of treatment, 14 patients remain disease-free with 2 additional patients alive in second remission after relapse. Patients with localized disease at presentation have improved disease-free survival and overall survival compared to patients with metastases at presentation. All 17 localized patients achieved a CR and 11 are NED compared to 8 of 14 metastatic patients achieving a CR and only 3 are NED. There have been 5 loco-regional recurrences with 3 "in-field" failures and 2 failures in the regional pleura. There were no treatment-related deaths and no clinically significant cases of pneumonitis. To date, 2 patients have significant treatment related morbidity, including 1 patient with scoliosis requiring surgery and 1 patient with acute leukemia developing 42 months after the start of therapy (presently in remission). We conclude that this intensive combined modality therapy results in a high CR rate and good local control with acceptable morbidity. Patients with metastatic disease at presentation remain a therapeutic challenge.

In vitro chemosensitivity of two Ewing's sarcoma cell lines: implication for autologous bone marrow transplantation

Intensive chemotherapy requiring rescue with autologous bone marrow may be an encouraging mode of treatment for poor prognosis Ewing's sarcoma (ES). In vitro chemosensitivity test may be useful for establishing an effective purging condition. We studied the in vitro effects of a variety of chemotherapeutic agents on two established ES cell lines (ES-5838 and ES-A4573) and marrow colony-forming unit-granulocyte macrophage (CFU-GM). 4-Hydroperoxycyclophosphamide (4-HC), at 100 microM produced complete inhibition (greater than 5 log) of clonogenic growth of both ES cell lines and spared 6.9% of normal CFU-GM growth. Etoposide (VP-16), at 100 microM produced 3-3.5 log inhibition of ES cell lines and complete inhibition of CFU-GM growth. Adriamycin (ADR) and vincristine (VCR) were more cytotoxic to ES-5838 cells than ES-A4573 cells. ADR at 1 microM produced 99.7% inhibition of ES-5838 cells, 92.2% of ES-A4573 cells, and 86% inhibition of CFU-GM. VCR at 1 microM produced 98.6% inhibition of ES-5838 cells, only 43.7% of ES-A4573 cells, and 75% inhibition of CFU-GM growth. Addition of verapamil did not enhance VCR cytotoxicity of ES cell lines. These studies indicate that 4-HC may be a useful agent for purging metastatic ES cells from the bone marrow for autologous marrow transplantation.

Small-cell osteosarcoma: correlation of in vitro and clinical radiation response

Small-cell osteosarcoma is an entity which shares some clinical and pathological features with both classic osteosarcoma and Ewing's sarcoma of bone. While noted to be "not radiosensitive" when first described, a retrospective review the National Cancer Institute experience of five patients with small-cell osteosarcoma treated with radiation therapy following biopsy (three pts) or limited excision (two pts) showed local control in all five patients with two long-term disease-free survivors (12, 18 years). This compares to three patients treated with surgery alone where one patient failed locally and one patient is a long-term disease-free survivor (7 years). We have studied the in vitro radiation response of a recently established small-cell osteosarcoma cell line (TC-252) and compared its response with that of a classic osteosarcoma cell line (U2-OS) and an Ewing's sarcoma cell line (5838). The small-cell osteosarcoma line responded with a similar Do and extent of PLDR compared to the Ewing's line and was different from the in vitro radiation response of classic osteosarcoma. Based on this small clinical series and the in vitro radiation studies, we conclude that small-cell osteosarcoma is a radioresponsive tumor. Definitive radiation therapy or conservative surgery plus radiation therapy are effective alternative therapeutic options, compared to ablative surgery, for the local treatment of this uncommon bone tumor of children and young adults.

In vitro radiobiological parameters of human sarcoma cell lines

In vitro radiobiologic survival parameters have been determined for 7 human osteosarcoma, 5 human soft tissue and bone sarcomas, and 4 Ewing's sarcoma cell lines. The mean D0 values were 99.5 +/- 11.6 cGy, 90.5 +/- 7.7 cGy and 95.8 +/- 7.9 cGy for osteosarcomas, soft tissue and bone sarcomas and Ewing's sarcomas, respectively. These in vitro survival data do not predict the clinical radiation resistance generally attributed to osteosarcomas and soft tissue and bone sarcomas, and do not differ substantially from the results obtained with the clinically radioresponsive Ewing's sarcomas.

High-risk Ewing's sarcoma: end-intensification using autologous bone marrow transplantation

Because of retrospective analysis showing survival to be related to primary tumor size, in February 1982 a study to test this hypothesis prospectively was begun at the University of Florida. Patients with primary tumors 8 cm or less in maximum diameter and no metastases received adjuvant chemotherapy consisting of vincristine, cyclophosphamide, doxorubicin, and dactinomycin plus radiotherapy or surgery (standard-risk protocol). All others received a similar regimen followed by end-intensification with high-dose melphalan and autologous bone marrow transplantation (Protocol HR-2). Because of poor results of HR-2, another high-risk protocol (HR-3) was initiated in January 1985. Patients on HR-3 received 2 cycles of chemotherapy containing vincristine, cyclophosphamide, and doxorubicin followed by local radiation therapy and maintenance chemotherapy. At the end of this therapy, autologous bone marrow transplantation (ABMT) was performed, using a preparatory regimen of total body irradiation and intensive chemotherapy. The 2-year disease-free survival rate was 70% for the standard-risk protocol, 20% for HR-2, and 80% for HR-3. The follow-up on HR-3 is still short, but the results are promising enough to warrant further clinical trials.

Treatment of pelvic sarcomas in adolescents and young adults with intensive combined modality therapy

Adolescent and young adult patients with pelvic sarcomas continue to have a poor prognosis with standard combination chemotherapy and local irradiation. In addition to a significant risk of local failure, these patients are at high risk for systemic relapse. Twenty-three consecutive patients with Ewing's sarcoma, alveolar rhabdomyosarcoma, undifferentiated sarcoma, or malignant peripheral neuroepithelioma originating in the pelvis were treated with short, intensive combined modality therapy. This approach integrates 5 cycles of VADRIAC chemotherapy (Vincristine, Adriamycin, Cyclophosphamide) with high dose irradiation to the primary lesion (55-60 Gy) and sites of gross metastatic disease (45-50 Gy). Following achievement of a complete response, intensification therapy consisting of total body irradiation (TBI) (8.0 Gy), high dose VADRIAC chemotherapy, and autologous bone marow transplantation is given. All therapy is completed within 6-7 months. No maintenance chemotherapy is given; no surgery is intended. Of the twenty-three patients with pelvic sarcomas treated on this combined modality protocol, 22 achieved a complete remission. Local control was achieved and maintained in all twenty-three patients. With a median follow-up of 21 months since initiation of treatment, there have been nine relapses (all systemic). Seven relapses occurred among the thirteen patients who presented with overt metastatic disease and the other two relapses were among the ten patients with localized disease at presentation. All seven metastatic patients who relapsed have died, whereas both of the relapsed localized patients remain alive. Acute and late toxicities have been acceptable using this aggressive combined modality approach. Induction chemotherapy had a significant impact on reduction of the typically large (greater than 10 cm diameter) soft tissue mass associated with these pelvic tumors, thus facilitating achievement of local control by high dose irradiation. Of 18 patients with measureable soft tissue tumor, all experienced a partial response (greater than 50% reduction in size) following the initial two cycles of chemotherapy given prior to local irradiation. In conclusion, this short, intensive chemoradiotherapeutic regimen is highly effective in controlling the primary lesion (100% local control) and inducing a complete response in a high proportion (96%) of these high risk pediatric and young adult patients with pelvic sarcomas. The role of TBI as "systemic" adjuvant therapy to control micrometastatic disease is discussed as still under investigation.

The maximum recovery potential of human tumor cells may predict clinical outcome in radiotherapy

We studied inherent radiosensitivity/resistance (D0), ability to accumulate sublethal damage (n) and repair of potentially lethal damage (PLDR) in established human tumor cell lines as well as early passage human tumor cell lines derived from patients with known outcome following radiotherapy. Survival 24 hrs after treatment of human tumor cells with X rays in plateau phase cultures is a function of initial damage (D0, n), as well as recovery over 24 hrs (PLDR). A surviving fraction greater than .1 24 hrs following treatment with 7 Gy in plateau phase cultures is associated with tumor cell types (melanoma, osteosarcoma) with a high probability of radiotherapy failure or tumor cells derived from patients who actually failed radiotherapy. Therefore, total cellular recovery following radiation may be an important determinant or radiocurability. Accurate assays of radiotherapy outcome may need to account for all these radiobiological parameters.

Fractionated versus low dose-rate total body irradiation. Radiobiological considerations in the selection of regimes

Total body irradiation (TBI) followed by bone marrow rescue is being increasingly used in the systemic treatment of acute leukaemia and some solid tumours such as neuroblastoma. Typically, these neoplasms are radiosensitive with little or no shoulder on the in vitro survival curve (n approximately equal to 1.0, Do approximately equal to 1.0 Gy). In such cases, fractionated or low-dose-rate TBI should allow preferential sparing of normal tissues. With the appropriate choice of dose rate, low-dose-rate TBI should, in principle, be radiobiologically equivalent to fractionated TBI. Calculations based on an extension to the linear quadratic model suggest that extremely low dose rates (e.g., approximately equal to 0.5 Gy h-1) might be required for equivalence to conventionally fractionated schedules. Such low dose rates would require very long treatment times (e.g., approximately equal to 24 h), which renders them impractical. For cell survival parameters of typical radiosensitive neoplasms the effects of proliferation do not alter this conclusion. These studies suggest that fractionated TBI (with high dose rates) is preferable to low-dose-rate therapy for neoplasms such as leukaemia and neuroblastoma.

Differential response to gamma radiation of human stomach cancer cells in vitro

In vitro effects of radiation were studied in two permanent cell lines (AGS and SII) from two patients with adenocarcinoma of the stomach and three permanent sublines from each cell line. Radiation survival parameters for AGS and SII parent cell lines and sublines were determined after in vitro irradiation of their cells with 0.5 to 10 Gy of 60Co gamma rays. The AGS and SII cell lines had different growth properties, DNA contents and radiation survival curves. Surviving fractions of SII parent cells (76 chromosomes) after 2.0 and 10 Gy were 1.22 and 17.8 times greater, respectively, than values for AGS parent cells (47 chromosomes). Sensitivities (D0) were 1.08 and 1.45 Gy for AGS and SII parent lines, respectively. The D0 values for AGS parent cells and sublines were similar (1.01 to 1.08 Gy), but SII parent cells and sublines had D0 values of 1.45, 1.36, 1.37 and 1.12 Gy (for SII-A). Also, the SII parent cells had survival fractions after 2.0 and 10 Gy that were 1.3 and 11.3 times greater, respectively, than values for the SII-A cells. These data show differences in radiation responses among stomach cancer cell lines and sublines that may relate to DNA content, but there was no consistent correlation between radiation response and a particular cell characteristic.

Advances in pediatric radiotherapy in the last ten years and future proposals

Advances made can be divided into five main categories. Firstly, the problem of geographic miss which has been reduced by delivering effective radiation doses with greater precision. This has been accomplished with more sophisticated diagnostic and therapeutic equipment, immobilization techniques and computerized treatment planning. Second is the recognition of the interplay of radiation and chemotherapy on normal tissue tolerance and local tumor control. This interaction has necessitated reduction in both dose and volume of irradiation. Third is the use of wide field irradiation as a systemic treatment. Fourthly, the utilization of cooperative group trials to define the role of irradiation. Finally, with the improvement in survival has come the recognition of late effects of irradiation in the growing child and the means of reducing such effects. The current role of radiation therapy in childhood malignancies is summarized, controversies are identified, and future prospects explored.

Ewing's sarcoma: a trial of adjuvant chemotherapy and sequential half-body irradiation

The results of a pilot study using adjuvant chemotherapy and sequential half-body irradiation (HBI) for nonmetastatic Ewing's sarcoma are presented. Seventeen patients received Cyclophosphamide, Vincristine, and Adriamycin (8 cycles), followed by sequential radiation treatment of the upper (500 cGy) and lower (600 cGy) half body. Survival at 3 years was 49%. These results are contrasted with those for 18 concurrently treated patients who received standard adjuvant therapy. Overall 5-year survival and relapse-free survival for these 35 consecutive patients was 61 and 53%. The pilot protocol was given on an out-patient basis with limited and acceptable acute toxicology. Further study is necessary to determine the value of the pilot protocol.