Significance of Surgical Margin on the Prognosis of Patients with Ewing’s Sarcoma A Report from the Cooperative Ewing ’s Sarcoma Study Toshifumi Ozaki, M.D.’ Axel Hillmann, M.D.’ Christiane Hoffmann, M.D? Christian Rube, M.D? Sebastian Btasius, M . D . ~ Jiirgen Dunst, M.O? Herbert Jiirgens, M.D? Winfried Winkelmann, M.o.’ ’ Department of Orthopaedics, Westfalische Wilhelms-University, Munster, Germany. Department of Pediatric Hematology and Oncology, Westfalische Wilhelms-University, Munster, Germany. Department of Radiation Oncology, Westfalische Wilhelms-University, Munster, Germany. Department of Pathology, Westfalische Wilhelms-University, Munster, Germany. Department of Radiotherapy, Martin Luther University Halle-Wittenberg, Halle, Germany. Presented in part at the Joint Meeting of the EMSOS [European Musculo-Skeletal Oncology Society) and AMSTS [American Musculo-Skeletal Tumor Society), Florence, Italy, May 8-9, 1995. Supported in part by a grant from Deutsche Krebshilfe (grant # M43/92/Ju2), EC BlOMEDl (grant # BMHl-CT92-1341). The authors thank Mrs. S. Jabar for the statistical analysis and her help in preparing the article. Address for reprints: Winfried Winkelmann, M.D., Department of Orthopaedics, Westfalische Wilhelms-University, Albert-SchweitzerStr. 33, 48129 Munster, Germany. Received March 11,1996; accepted April 26,1996. 0 1996 American Cancer Society BACKGROUND. There is little information reg,arding an adequate surgical margin for local control of Ewing’s sarcoma. METHODS. Two hundred and forty-four patients (PTS) with Ewing’s sarcoma who were registered in the Cooperative Ewing’s Sarcoma Studies underwent surgical treatment. Ninety-four PTS underwent definirive surgery (surgery alone), 131 PTS received postoperative irradiation, and 19 PTS received preoperative irradiation. The surgical margins were distributed as follows: radical, 29 PTS; wide, 148 PTS; marginal, 39 PTS; and intralesional, 28 PTS. The impact of the surgical margin on the treatment outcome of PTS was analyzed statistically. RESULTS. The local or combined (local recurrence and systemic metastasis) relapse rate after surgery with or without irradiation was significantly lower compared with that after definitive irradiation (irradiation alone) (7% vs. 31%, P < 0.0001). The local or combined relapse rate after complete resection (radical or wide margin) with or without irradiation was less compared with that after incomplete resection (marginal or intralesional margin) with or without irradiation (5% vs. 12% P = 0.0455). The local or combined relapse rate did not greatly decreased after irradiation after incomplete surgery (from 14% to 12%). In both groups of good (viable tumor cells <lo%) and poor (viable cells 210%) histologic response, the difference in systemic or combined relapse rate between patients undergoing complete and incomplete surgery was not significant. The 10-year overall survival of the PTS for each of the margins was distributed as follows: radical, 58%;wide, 65%; marginal, 61%; and intralesional, 71% (,P = not significant). CONCLUSIONS. Surgery in patients with Ewing’s sarcoma adds to the safety of local control. Under the current treatment regimen with intensive chemotherapy and irradiation, complete resection of the tumor appears capable of decreasing the risk of local recurrence. Cancer 1996; 78:892-900.0 1996American Cancer Society. KEYWORDS bone, neoplasms, malignant, Ewing’s sarcoma, surgery, therapy, surgical margin. E wing’s sarcoma is a malignant bone tumor comprised of small round cells. The expression of neural markers distinguishes conventional Ewing’s sarcoma from malignant primitive neuroectoderma1 tumor (PNET).’ The treatment outcome of Ewing’s sarcoma has rapidly improved with the increasing availability of adjuvant theraWith the CESS 81 protocol, the first Cooperative Ewing’s Sarcoma Study (CESS) was initiated in 1981 by the German Society of Pediatric Oncology and Hematology (GPOHI4and followed consecutively by other trials. To date, now, more than 1000 patients (PTS) have been treated according to the CESS trials. Although the impact of surgery for PTS with Ewing’s sarcoma Surgical Margin of Ewing's Sarcoma lOzaki et al. . , 0 10 20 30 40 . . . . . 50 60 70 80 90 . 100 ($6) FIGURE 1. Distribution of surgical margin. Three patients whose tumor volume was not clear were excluded. Numbers in each of the values indicate the number of patients. is sometimes controversial, surgery seems to increase the local control rate and survival rate of PTS with Ewing's ~ a r c o r n a . ~ Because ' ~ - ' ~ of the higher local relapse rate after irradiation alone, surgery is regarded as an important modality for local treatment in the CESS trials. In recent years, the surgical procedures have tended to be more conservative to acquire better function with salvaged limbs compared with amputation and hemipelvectomy. A low local relapse rate was reported even after surgery with a marginal or intralesional margin.2 However, an adequate surgical margin to prevent local relapse has never been satisfactorily addressed. Moreover, the effect of postoperative irradiation for the local control after surgery with incomplete resection remains unclear. The relationship between histologic response and optimal surgical margin is also obscure. In this study, the influence of surgery and surgical margins on the local control and overall survival of PTS with Ewing's sarcoma were investigated. The significance of irradiation after surgery and the correlation between treatment outcome and surgical margins according to the histologic response of preoperative treatment were also analyzed. PATIENTS AND METHODS Characteristics and Chemotherapy Protocol PTS registered for the consecutive CESS 81, 86, and 91P trials were investigated. Requirement for classification as protocol PTS were: absence of visible metastases at diagnosis, initiation of treatment within 3 weeks after biopsy, neither pretreatment nor primary local treatment, and the enrollment occurring within 6 weeks after the start of therapy. The age for protocol PTS was 25 years or younger in CESS 81 and 86 and 35 years or younger in CESS 91P. The chemotherapy for the CESS 81 trial was comprised of the four-drug regimen vincristine, actinomycin-D, cyclophosphamide, and doxorubicin (VACA).4The CESS 86P trial 893 was the pilot study of CESS 86. In the CESS 86 trial, PTS with standard risk (small tumors in the proximal or distal region) received the VACA regimen and those with high risk (tumors in the central region or of large volume) received the VAIA regimen, in which ifosfamide replaced cyclophosphamide. The CESS 86P and CESS 86 trials were regarded as the same protocol. In the CESS 91P trial, a pilot phase of the ongoing European Intergroup Cooperative Ewing's Sarcoma Study (EICESS) 92 trial, standard risk PTS (same definition as CESS 86) received the VACA regimen and high risk PTS received either VAIA alone or with additional etoposide as a fifth agent (EVAIA) by randomization. Theses protocols have been published in detail elsewhere.*,5The regular period of chemotherapy prior to definitive local treatment was 18 weeks in PTS treated according to the CESS 81 trial, 9 weeks in CESS 86, and 12 weeks in CESS 91P. Of 255 protocol PTS without primary metastases who underwent surgery, 244 PTS for whom information regarding the surgical margin was available were reviewed for this study. Fifty-nine, 23, 126, and 36 PTS were treated according to the CESS 81, 86P, 86, and CESS 91P trials, respectively (Fig. 1).Their ages ranged between 2 and 25 years, with a median age of 14 years. The male to female ratio was 150 to 94 (1.6 to 1). According to the histologic diagnosis, 214 PTS had Ewing's sarcoma and 30 had PNET. All belonged to Enneking's Stage IIB.I3 In this study, the local control with surgical treatment according to each surgical margin is emphasized. Although there are slight differences among the protocols, all PTS were analyzed together. Ninety-five tumors originated in the central region: pelvis, 42 tumors; rib, 26 tumors; scapula, 17 tumors; clavicle, 6 tumors; spine, 2 tumors; sternum, 1 tumor; and skull, one tumor. Seventy-three tumors originated in the proximal region: femur, 57 tumors; and humerus, 16 tumors. Seventy-six tumors were located in the distal region: fibula, 33 tumors; tibia, 29 tumors; foot, 9 tumors; ulna, 4 tumors and radius, 1 tumor. Tumor volume was evaluated according to the method described by Gobel et aL1*A tumor volume of 100 mL or higher was defined as a large tumor (150 PTS), and that below 100 mL as a small tumor (91 PTS). In three PTS, the information about tumor size was not available. The follow-up period ranged from 4.7 to 159.1 months (median, 61.1 months). In surviving PTS, the shortest follow-up time was 24.4 months. Modalities of Local Treatment Of these 244 PTS, 94 underwent surgery alone (definitive surgery). One hundred thirty one PTS underwent 894 CANCER August 15,1996 / Volume 78 / Number 4 TABLE 1 Local or Combined Failures According to Surgical Margin and Type of Local Treatment Adequate Radical Wide Inadequate Marginal Intralesional Surgery (%I Surgery t postop RT (%) Preop RT t surgery (%) Total (%I 0/27 (0) 2/53 (4) 0/2 (0) - 6/80 (8) 0115 (0) 0129 (0) 81148 (5) 1/ 10 (10) 1/4 (25) 2/26 (8) 4/23 (17) 013 (0) Oil (0) 3/39 (8) 5/28 (181 postop: postoperative; R T radiation therapy; preop: preoperative. surgery with postoperative irradiation. Nineteen PTS received preoperative irradiation followed by surgery. Surgery All surgical procedures were classified according to Enneking’s criteria.I3 The information regarding the surgical margin was reported to the CESS trial office by the surgeons and pathologists who carefully examined the resected specimen. Definitions were as follows: radical: the whole tumor-bearing compartment is removed en bloc; wide: the tumor and its pseudocapsule were removed en bloc surrounded by healthy tissue within the tumor-bearing compartment; marginal: the tumor was removed en bloc, but the line of resection ran through the pseudocapsule or reactive area of the tumor, and microscopic residual disease was likely; and intralesional: the tumor was opened during surgery, the surgical field was contaminated, and microscopic or macroscopic residual disease remained. Radical and wide margins were termed adequate and marginal and intralesional margins were termed inadequate. Of 29 PTS who underwent surgery with a radical margin, 27 PTS underwent definitive surgery and 2 PTS received postoperative irradiation (Table 1). Of 148 PTS who underwent surgery with a wide margin, 53 PTS underwent definitive surgery, 80 PTS received postoperative irradiation, and 15 PTS received preoperative irradiation. Of 39 PTS with a marginal margin, 10 PTS underwent definitive surgery, 26 PTS received postoperative irradiation, and 3 PTS received preoperative irradiation. Of 28 PTS with an intralesional margin, 4 PTS underwent definitive surgery, 23 PTS received postoperative irradiation, and 1 patient received preoperative irradiation. Irradiation Doses of irradiation ranged between 22.4 and 76 Gray (Gy) (median; 46 Gy). The recommended protocol dose of postoperative irradiation was 36 Gy in CESS 81 and 44.8 Gy in CESS 86.Is In CESS 91P, preoperative irradiation was recommended for PTS with a less than 50% reduction of evaluable soft tissue mass after 2 courses of chemotherapy. The dose of the preoperative irradiation in 19 PTS ranged from 36 to 63.2 Gy (median, 51.8 Gy). Histologic Response After surgery, the surgical specimen was examined histologically and classified into 6 grades according to the criteria of regression by Salzer-Kuntschik et a1.,16 Grade 1: no viable tumor cells detectable; Grade 2: 1 focus of the viable cells or a tumor island smaller than 0.5 cm remained; Grade 3: less than 10% of viable tumor cells remained; Grade 4: 10-50% of the viable tumor cells remained; Grade 5: greater than 50% of the viable tumor cells remained; and Grade 6: no histologic response to chemotherapy. The breakpoint between good and poor response by definition was located between Grades 3 and 4. Hence, Grades 1, 2, and 3 were termed good, and Grades 4, 5 , and 6 were termed poor responses. Relapses Relapses were classified in three types: local (only local recurrence), systemic (only metastases), and combined (local recurrence and systemic metastases), at the time of first relapse. Statistical Analysis Statistical significance of the differences between proportions was evaluated by the chi-square test with Fisher’s correction with the criterion of P < 0.05.’’ The cumulative probability of relapse free survival was calculated by the Kaplan-Meier method.18 Univariate log rank analysis was used to compare the c u r ~ e s . ’ ~ TREATMENT AND RESULTS Treatment Modalities and Failures First, the general impact of surgery on treatment outcome was evaluated, and thereafter the influence of Surgical Margin of Ewing's Sarcoma /Ozaki et al. RAD 895 Radical n=29 n=lO2 0P n=94 OP+RAD n=l31 Local Failure Marginal Local Failure Combined Failure n=39 u Combined Failure R A D+ 0 P n=l9 Systemic Failure Remission Systemic Failure Remission 0 FIGURE 2. Relapse pattern according to the modalities of the treatment. Relapses were classified as local, combined (local and systemic), and systemic. Numbers in each of the values indicate the number of the patients. the different surgical margins was analyzed. One hundred two PTS who received irradiation alone (definitive irradiation) were compared with 244 PTS who underwent surgery. Of the 102 PTS who received definitive irradiation as alocal treatment, 15 (15%), 17 (17%), and 15 (15%) PTS developed local, combined, and systemic relapses, respectively (Fig. 2). Of 94 PTS who underwent definitive surgery without additional irradiation, 3 (3%), 1 (l%),and 27 (29%) PTS developed local, combined, and systemic relapses, respectively. Of 1.31 PTS who received surgery and postoperative irradiation, 7 (5%),5 (4%),and 35 (27%)PTS developed local, combined and systemic failures, respectively. Of 19 PTS who received preoperative irradiation and surgery, 5 (26%) PTS developed systemic failures. Impact of Surgery on Local Failure The local relapse rate of PTS from the above mentioned 3 groups with surgery (10 of 244) was significantly lower compared with that of the PTS with definitive irradiation (15 of 102) (4%vs. 15%; P = 0.0011). The combined relapse rate of PTS of the 3 groups including surgery (6 of 244) was also significantly lower compared with that of the PTS with definitive irradiation (17 of 102) (2% vs. 17%; P < 0.0001). If local and combined relapses were analyzed together with respect to local control, the local or combined relapse rate of the 3 groups including surgery (16 of 244) was significantly lower compared with that of the PTS with definitive irradiation (32 of 102) (7% vs. 31%; P < 20 40 60 80 100 (%) FIGURE 3. Relapse pattern according to the surgical margin. Relapses were classified as local, combined (local and systemic), and systemic. Numbers in each of the values indicate the number of the Patients. diation (67 of 244) was significantly higher than that of the group with definitive irradiation (15 of 102) (27% vs. 15%;P = 0.0123). If systemic and combined failures were analyzed together with respect to systemic metastases, the difference of systemic or combined failure between the 3 groups including surgery (73 of 244; 30%) and PTS with definitive irradiation (31%)was not significant. Impact of Treatment Modalities on Overall Survival The 10-year overall survival of the 244 PTS who underwent surgery with or without irradiation was significantly superior to that of the 102 PTS with definitive irradiation (65% vs. 50%; P = 0.0088). The 10-year overall survival of the PTS who underwent definitive irradiation definitive surgery, postoperative irradiation, and preoperative irradiation were 50% 60%, 68%, and 79%, respectively ( P = 0.0457). If the I0-year overall survival of 19 PTS who received preoperative irradiation was separated from that of the 225 PTS without preoperative irradiation, 79% and 65% of the 10-year survival were obtained ( P = not significant). 0.0001). Factors Influencing the Surgical Margin Tumors that originated in the central region were resected with inadequate surgical margins more often than those in the proximal or distal region (44 of 95, 46% vs. 23 of 149, 15%; P < 0.0001) (Fig. 1). There were no significant differences of the distribution of the surgical margins (adequate or inadequate) between or among the variables of sex, tumor size, protocol, treatment with or without preoperative irradiation, and histologic response. Impact of Surgery on Systemic Failure Conversely, the systemic failure rate of the PTS of the 3 groups that underwent surgery with or without irra- Surgical Margins and Relapse Pattern Of 29 PTS who underwent surgery with a radical margin, no PTS developed local or combined failures; 896 CANCER August 15,1996 / Volume 78 I Number 4 “t 40 20 0 - ___ -_- - Marginal (n=39) Radical (n=29) 1 I 0 40 80 120 160 (mon.) FIGURE 4. Overall survival according to the type of the surgical margin. The differences among the surgical margins were not significant. however, 11 PTS (38%) developed systemic failures (Fig. 3). Of 148 PTS who underwent surgery with a wide margin, 7 (5%), 1 (1%),and 38 (26%) PTS developed local, combined, and systemic relapses, respectively. Of 39 PTS who underwent surgery with a marginal margin, l (3%),2 (5%),and 11 (28%) PTS developed local, combined, and systemic relapses, respectively. Of 28 PTS who underwent surgery with , 7 (25%) an intralesional margin, 2 (7%),3 ( l l % )and PTS developed local, combined, and systemic relapses, respectively. The local or combined failure rate had a tendency to increase as the resection line got closer to the tumor. There was no significant difference in the local relapse rate after surgery between adequate (radical and wide) (7 of 177; 4%) and inadequate (marginal and intralesional) margins (3 of 67; 4%). However, the combined relapse rate after surgery with adequate surgical margins (1 of 177) was significantly lower compared with that after surgery with inadequate surgical margins (5/67) (1% vs. 7%; P = 0.0066). If local and combined relapses are analyzed together with respect to local control, the local or combined relapse rate after surgery with adequate surgical margins (8 of 177) was significantly lower compared with that after surgery with inadequate margins (8/671 (5% vs. 12%; P = 0.0455). There was no significant difference in the systemic failure rate between adequate (49 of 177; 28%) and inadequate (18 of 67; 27%) margins. Even if systemic or combined relapses were analyzed together with respect of systemic metastasis, the difference in systemic or combined failure rate between adequate (50 of 177; 28%) and inadequate margins (23 of 67; 34%) was not significant. Surgical Margin and Overall Survival If all 244 patients were analyzed together, the 10-year overall survival of PTS with a radical margin was 58% (12 of 29 failed); wide margin, 65% (41 of 148 failed); marginal margin, 61% (14 of 39 failed); and intralesional margin, 71% (8 of 28 failed) (Fig. 4). There was no significant difference among these four groups. The 10-year overall survivals of 225 PTS who did not receive preoperative irradiation were 58%with a radical margin (12 of 29 failed), 66% with a wide margin (38 of 133 failed), 61% with a marginal margin (13 of 36 failed), and 70% with an intralesional margin (8 of 27 failed) (P-not significant). The 10-year overall survivals of 19 PTS who received preoperative irradiation were as follows; wide: 80% (3 of 15 failed); marginal: 67% (1 of 3 failed); and intralesional: 100% (0 of 1 failed) (P-not significant). The survival rate after relapse of 16 PTS who developed local or combined relapses was poor (27%). Analysis of Local or Combined Failures According to the Surgical Margin and Type of Local Treatment The frequency of local failures in each surgical margin was analyzed according to the difference of the treatment modalities. In PTS who underwent surgery with or without postoperative irradiation, local or combined relapse rates increased according to the transition of the surgical margins from radical to intrale- Surgical Margin of Ewing’s Sarcoma /Ozaki et al. TABLE 2 Local or Combined Failures according to Tumor Location and Surgical Margin Adequate Central Proximal Distal = 0.0077). Tumors that relapsed locally after surgery with adequate margins were distributed as follows: three in the rib, one in the clavicle, one in the pelvis, one in the scapula, one in the femur, and one in the fibula. Inadequate R (%) W (%I M (%) IL(%) Total (%) 010 (0) 017 (0) 0122 (01 6/51 (12) 1/56 (18) 1/41 (2) 1122 (51 118 (13) 119 (Ill 5/22 (23) 012 (01 014 (01 12/95 (13) 2/73 (31 2/76 (3) H: radical; Cz‘: wide: M: marginal: I 1 inlralesional TABLE 3 Failures ,4ccording to Histologic Response and Adequacy of Surgical Margin Good (1-3) Adequate Inadequate Poor (4-6) Adequate lnadeauate Local and systemic (%) No. Local (%I 121 39 5 (4) 2 (5) 0 (0) 47 27 2 (4) 0 I01 897 Systemic (%I Total (%I 4 (10) 26 (22) 6 1151 31 (26) 12 (31) 1 (2) 1 (41 22 (471 12 1441 25 (53) 13 1481 sional (Table 1). In PTS who underwent surgery with adequate margins, the local or combined relapse rate did not decrease after postoperative irradiation (6 of 82; 7%) compared with that after definitive surgery (2 of 80; 3%). In PTS who underwent surgery with inadequate margins, the local or combined relapse rate decreased after postoperative irradiation (6 of 49; 12%)compared with that after definitive surgery, but the difference was not significant (2 of 14; 14%). No local relapses developed in PTS who underwent preoperative irradiation. Analysis of Local or Combined Failures According to the Surgical Margin and Tumor Location The influence of surgical margin on local control may change with the tumor location. The local or combined relapse rate of the tumors in the central region (12 of 95) was higher than that of tumors in the proximal or distal region (4 of 149) (13%vs. 3%;P = 0.0031) (Table 2). After surgery with inadequate surgical margins, the local or combined relapse rate of the tumors in the central region was 14% (6 of 44) and that of the tumors in the proximal or distal region was 9% (2 of 23) (P-not significant). After surgery with adequate margins the local or combined relapse rate of the tumors in the central region was 12% (6 of 51) and that in the proximal or distal region was 2% (2 of 126) (P Histologic Response and Relapse Pattern The local failure rate for PTS with a good histologic response was 4% (7 of 160) and that for PTS with poor response was 3% (2 of 74) (Table 3). The combined failure rate for PTS with good response was 3% (4 of 160) and that for PTS with poor response was 3% (2 of 74). The local or combined failure rate for PTS with good histologic response was 7% (11 of 160) and that for PTS with poor response was 5% (4 of 74) (P-not significant). In PTS with good response, the local or combined relapse rate after surgery with inadequate margins (6 of 39) was significantly higher than with adequate margins (5 of 121) (15% vs. 4%; P = 0.0258). In the poor response group, the local or combined relapse rate after adequate surgical margins (3 of 47) was not lower than that after surgery with inadequate margins (1 of 27) (6% vs. 4%; P-not significant). The systemic failure rate for PTS with good response was 20% (32 of 160) and that for PTS with poor response was 46% (34 of 74) ( P < 0.0001). The systemic or combined failure rate for PTS with good response was 23% (36 of 160) and that for PTS with poor response was 49% (36 of 74) (P = 0.0001). In both groups with good and poor response, the difference of systemic or combined relapse rate between adequate and inadequate surgical margins was not significant. DISCUSSION The development of the surgical techniques and limbsparing surgery have made surgical treatment an attractive choice for local treatment of Ewing’s sarcoma. In a large number of PTS, a good histologic response is obtained after preoperative chemotherapy; however, viable tumor cells often remain in the resected specim e n ~Thus, . ~ PTS who underwent surgical excision of the primary tumor had a better local control and survival than those who received definitive irradiation.’,6-10 In 1980, Marcove and Rosen12 reported the excellent local control rate of PTS with Ewing’s sarcoma who underwent en bloc surgical resection (clear margin) with postoperative irradiation. Although surgery is currently the most important modality of local treatment for PTS with Ewing’s sarcoma, there is little information about the treatment results for the different surgical margins in PTS with Ewing’s sarcoma. Most of the local relapses are imputed to an inadequate excision that leaves viable tumor cells behind. Conversely, local relapses do not necessarily occur 898 CANCER August 15, 1996 I Volume 78 I Number 4 after surgery with an inadequate margin.2 Under the current treatment with intensive chemotherapy and irradiation, a high local control rate may be expected even after inadequate surgical margins. However, the systemic failure rate may be increased after surgery with an inadequate margin because of the possibility of intravascular dissemination of the viable tumor cells during surgery. However, this hypothesis has never been proven. In this reported series, the combined or systemic failure rate after irradiation is 31% compared with 30% after surgery, 28% after surgery with radical or wide (adequate) margins, and 34% after surgery with marginal or intralesional (inadequate) margins. The local or combined failure rate of PTS who underwent surgery (7%) was significantly lower compared with that of PTS with definitive irradiation (31%) (Fig. 2). PTS who received definitive irradiation might include a considerable number of the PTS with inoperable lesions, in contrast to those who underwent surgery. In the surgically inaccessible tumors, the local control rate naturally becomes poor. The high local or combined failure rate in PTS who received definitive irradiation would lead to the worst prognosis. Apart from the inoperable lesions, definitive irradiation would be the treatment of choice in PTS for whom surgical resection of the tumor with adequate margins is considered either to be difficult or believed to result in severe functional deficit.15Even after chemotherapy and 45-54 Gy preoperative irradiation, viable tumor cells remained in approximately 50% of the PTS.20If tumor is surgically accessible, surgery will add to the safety of local control. In general, tumor size and site are thought to affect the distribution of the acquired surgical margin. The effect of preoperative treatment may also closely influence the acquired surgical margins because PTS who undergo surgical resection tend to show good response to preoperative treatment.6 In this study, primary tumor site affected the acquired surgical margin; inadequate surgical margins were obtained in the PTS with tumor in the central region (46%) more often than in the proximal or distal regions (15%) (Fig. 1). Tumors in the central area are close to important organs. In the clavicle, rib, or spine, tumors are likely to infiltrate into the neurovascular bundles, the vertebral bodies, or the spinal cord, respectively. Because of the close distance to essential organs in the pelvic area, acquiring adequate margins becomes difficult. Due to the three-dimensional anatomy of the pelvis," the evaluation of the exact tumor extent is also complicating. In a report from Mayo Clinic, PTS who underwent complete excision of the tumor did not develop local relapses and survival was significantlyimproved. How- ever, the survival rate of PTS with debulking surgery was not significantly different from that of PTS who underwent no ~urgery.'~'~ Conversely, in the report from the Intergroup Ewing's Sarcoma Study (IESS) group, no significant differences were noted in local failure rates in PTS with pelvic tumors treated with either complete or incomplete resection followed by irradiation." In PTS divided into groups based on each surgical margin, local or combined relapse rates increased from radical to intralesional margins (Fig. 3). The statistical difference of the local or combined relapse rate was noted between adequate (4%) and inadequate margins (12%).The PTS who developed local relapses had a poor prognosis, so it is advisable to choose adequate surgical margins to prevent local or combined failures. Systemic metastasis does not seem to be affected by the adequacy of surgical margins. Although similar margins were acquired, the local failure rate was different according to tumor location. In this study, after surgery in central tumors (13%), local relapses developed more often than in proximal or distal tumors (3%)(Table I). This tendency was also noted after surgery with adequate margins. Three of six central tumors that relapsed locally after surgery with adequate margins originated in the ribs. The evaluation of surgical margins in rib tumors is difficult because of the pleural dissemination or implantation of the tumor cells. The criteria of the surgical margin used in this study would not be suitable for evaluation of the surgical adequacy of the chest wall tumors. If the tumor is violated or the surgical margin is contaminated, it is generally believed that postoperative irradiation becomes necessary." In this series, there was a tendency for the local or combined relapse rate after intralesional surgery to decrease from 25% to 17% with the addition of postoperative irradiation (P-not significant) (Table 2). After marginal resection, the local or combined relapse rate decreased from 10% to 8% after postoperative irradiation. The effect of postoperative irradial.ion for preventing local relapses after surgery with an inadequate margin was not satisfactory, so acquiring adequate margins becomes more important for the local control than additional postoperative irradiation afi er surgery with inadequate margins. In PTS who underwent surgery with a wide margin, the local or combined relapse rate did not decrease after postoperative irradiation. The tumor locations of six PTS who had local relapse after surgery with a wide margin and postoperative irradiation were distributed as follow!;: three in the rib, one in the scapula, one in the clavicle, and one in the pelvis. This result would also be affected by centrally located tumors, especially by the rib tumors as mentioned above. Surgical Margin of Ewing’s Sarcoma lOzaki et al. PTS with a histologic good response tended to have better relapse free survival than those with poor respo~ise.~ This result is attributed to the low relapse rates in PTS with a good histologic response (Table 3). Even if good response was observed after preoperative treatment, the local or combined relapse rate after surgery with inadequate margins was significantly higher than that after surgery with adequate margins. In PTS with poor histologic response, a contradicted trend between adequacy of the surgical margin and the local or combined relapse was noted. This might be because the number of PTS is fairly small and the prognosis of the PTS with poor response was poor so that survival was not long enough to develop local relapses. From these results, at least, surgery with inadequate margins, even after good histologic response, does not become pertinent for local control of Ewing’s sarcoma. Indeed the PTS with poor response suffered systemic or combined relapses more often than those with good response, but surgical margin did not influence systemic or combined relapse rate in either the good or poor response groups. Although it was difficult to determine the relationship between surgical margin and overall survival, the inadequate margins led to local and combined failure more often than the adequate margins. The number of PTS who developed local relapses is not large enough to show a definitive influence on overall survival. But most PTS who developed local or combined relapses had a poor prognosis compared with those without relapses. From these results, there are no proper reasons to justify an inadequate surgical margin. Although adequate margins should be the objective for good local control, inadequate margins sometimes cannot be avoided because the other options would be amputation or procedures leading to significant restrictions of the extremital functions. Furthermore, in some PTS, acquiring an adequate margin is difficult because of anatomic tumor locations. For PTS who suffered local relapses, further limb-salvage surgery becomes difficult and the risk of rapid systemic dissemination is high. If PTS undergo surgery with inadequate margins, additional therapy to prevent local relapses should be considered. Options include amputation, additional wider resection, brachytherapy, and postoperative irradiati~n.’~ In recent years, brachytherapy has been introduced to CESS trials, mf the margin is anticipated to be inadequate or actually contaminated. No severe complications and no local relapses have been reported in a series of 22 PTS after bra~hytherapy.~~ A prospective randomized trial on soft tissue sarcoma also demonstrated improvement of the local control rate even though the overall survival did not i m p r ~ v e . ’ ~Brachytherapy ~‘~ 899 may be the current method of choice for local control after an inadequate surgery because it can increase the local control without additional surgical invasion. High dose irradiation to the lower extremities led to unacceptable functional morbidity, so primary amputation was enco~raged.’~ Because preoperative irradiation did not raise the complication rate of surgery in Ewing’s sarcoma,28with the initiation of CESS 91P, preoperative irradiation was incorporated in the treatment schedule to further sterilize the tumor-bearing compartment before surgery. Resection of the involved bone after preoperative irradiation may not only lessen the risk of postirradiation sarcoma,6’21 but may also be useful for preventing postirradiative pathologic fractures and local relapses. In 19 PTS who received preoperative irradiation, none developed local relapses, even after limited surgery. The number of PTS who received preoperative irradiation was still quite small; nevertheless, good local control by the preoperative irradiation could be expected from this preliminary result. For further improvement of the overall survival in PTS with Ewing’s sarcoma, not only local but also systemic failures must be controlled more effectively. The elevation of the chemotherapeutic dose intensity under the current supportive care to avoid side effects and the selection of antitumor drugs for Ewing’s sarcoma are relatively limited. Myeloablative therapy with high dose anticancer agents was reported to improve the prognosis of PTS with multifocal primary disease or multiple relapse^.'^ However, this treatment is currently not justified for all PTS with Ewing’s sarcoma without primary metastasis but is restricted to high risk subgroups. The influence of cytokine regulation of Ewing’s sarcoma cells in vitro has recently been dem~nstrated.~’ Although these results are still in vitro, the response of Ewing’s sarcoma cells to cytokines might lead to a new modality of systemic treatment. In conclusion, the local or combined relapse rate after surgery with radical or wide margin was significantly lower compared with that after surgery with marginal or intralesional margins. The local or combined relapse rate did not significantly decrease after irradiation after inadequate surgery. The difference in 10-year overall survival of the PTS for each of the margins was also not significant; however, PTS who developed local or combined relapses had a very poor prognosis. Therefore, acquiring adequate surgical margins appears to remain beneficial. REFERENCES 1. Roessner A, Jurgens H. Round cell tumors of bone. Path01 Res Pract 1993;189:1111-36. 900 2. 3. 4. 5. 6. 7. 8. 9. 10 11 12 13. 14. 15. 16. 17. CANCER August 15,1996 / Volume 78 / Number 4 Bacci G, Toni A, Avella M, Manrfini M, Sudanese A, Ciaroni D, et al. Long-term results in 144 localized Ewing’s sarcoma patients treated with combined therapy. Cancer 1989; 63:1477-86. Jurgens HF. Ewing’s sarcoma and peripheral primitive neuroectodermal tumor. Curr Opin Oncol 1994;6:391-6. Jurgens H, Exner U, Gadner H, Harms D, Michaelis J, Sauer R, et al. Multidisciplinary treatment of primary Ewing’s sarcoma of bone. A 6-year experience of European cooperative trial. Cancer 1988;61:23-31. Jurgens H, Treuner J, Miinkler K, Gobel U. Ifosfamide in pediatric malignancies. Semin Oncol 1989; 16:46-50. Sailer SL, Harmon DC, Mankin HJ, Truman IT, Suit HD. Ewing’s sarcoma: surgical resection as a prognostic factor. Int J Radiat Oncol Biol Phys 1988;15:43-52. Wilkins RM, Pritchard DJ, Burgert EO Jr, Unni KK. Ewing’s sarcoma of bone. Experience with 140 patients. Cancer 1986;5812551-5. Rosen G, Caparros B, Nirenberg A, Marcove RC, Huvos AG, Kosloff C, et al. Ewing’s sarcoma: ten-year experience with adjuvant chemotherapy. Cancer 1980;47:2204- 13. Burgert EO, Nesbit ME, Garnsey IA. Multimodal therapy for the nonpelvic localized Ewing’s sarcoma of bone: intergroup study IESS-11. J Clin Oncol 1990;8:1514-24. Neff JR. Nonmetastatic Ewing’s sarcoma of bone: the role of surgical therapy. Clin Orthop 1986;204:lll-7. Pritchard DJ, Dahlin DC, Dauphine RT, Taylor WF, Beabout JW. Ewing‘s sarcoma. A clinicopathological and statistical analysis of patients surviving five years or longer. J Bone Joint Surg Am 1975;57:10-6. Marcove RC, Rosen G. Radical en bloc excision of Ewing’s sarcoma. Clin Orthor, 1980;153:86-91. Enneking WF, Spanler SS, Goodman MA. A system for the surgical staging of musculoskeletal sarcoma. Clin Ortkop 1980; 153:106-21. Gobel V, Jurgens H, Etspuler G, Kemperdick H, Jungblut RM, Steinen U, et al. Prognostic significance of tumor volume in localized Ewing’s sarcoma of bone in children and adolescents. J Cancer Res Clin Oncol 1987; 113:187-91. Dunst J, Burgers JMV, Hawliczek R, Rurten R, Winkelmann W, Salzer-Kuntschik M, et al. Radiation therapy as local treatment in Ewing’s sarcoma. Cancer 1991;67:2818-25. Salzer-Kuntschik M, Delling D, Beron G, Sigmund R. Morphological grades of regression on osteosarcoma after polychemotherapy: Study COSS 80. J Cancer Res Clin Oncol 1983;106121-4. Fleiss JL. Statistical methods for rates and proportions. 2nd edition. New York: John Wiley and Sons, 1981:14-22. 18. Kaplan EI, Meier P. Non-parametric estimation for incomplete observations. J A m Stat Assoc 1958;52457-81. 19. Pet0 R, Peto J. Asymptotically efficient rank in variant test procedures. J R Stat Soc 1972; 135A185-98. 20. Jurgens H, Hoffmann CH, Braun-Munzinger G, Blasius S, Salzer-Kuntschik M, Winkelmann W, et al. Histological response following preoperative chemoradiotherapy: the CESS experience [abstract]. Joint Meeting EMSOS-AMSTS, Florence, Italy, May 8-9, 1995:63. 21. Frassica FJ, Frassica DA, Pritchard DJ, Schomberg PL, Wold LE, Sim FH. Ewing’s sarcoma of the pelvis. CLinicopathological features and treatment. J Bone Joint Surg A m 1993; 75~1457-65. 22. Pritchard DJ. Indications for surgical treatment of localized Ewing’s sarcoma of bone. Clin Orthop 1980 153:39-43. 23. Evans RG, Nesbit ME, Gehan Eh, Garnsey LA, Burgert 0, Vir:tti TJ, et al. Multimodal therapy for the management of localized Ewing’s sarcoma of pelvic and sacral bones: a report from the second intergroup study. J Clin Oncol 19!3 1; 9: 1173-80. 24. Tanabe KK, Pollock RE, Ellis LM, Murphy A, Sherman N, Romsdahl M. Influence of surgical margins on outcome in patients with preoperatively irradiated extremity soft tissue sarcomas. Cancer 1993;i3:1652-9. 25. Hillmann A, Rubee CH, Rod1 R, Lindner N, Hoffmann CH, Schuck A, et al. Intraoperative brachytherapy of Ewing’ssarcoma [abstract] Joint Meeting EMSOS-AMSTS, Florence, Italy, May 8-9, 1995:94. 26. Brennan MF, Casper ES, Harrison LB, Shiu MH, Gaynor J, Hajdu S. The role of multimodality therapy in soft-tissue sarcoma. Ann Surg 1991;214:328-38. 27. Lewis RJ, Marcove RC, Rosen G. Ewing’s sarcoma-functioiial effects of radiatio’n therapy. J Bone Joint Surg A m 1977;59:325-31. 28. Hillmann A, Rube CH, Lindner N, Rod1 R, Hoffmann C, Jurgens H, et al. Does preoperative radiation therapy (RT) of Ewing’s sarcoma increase the complication rate? A review of the cases at the bone tumor center in Munster [abstract]. European Musculo-skeletal Oncology Society, 7th Meeting, Amsterdam, October 6-7, 1994. 29. Burdach S, Jurgens H, Peters C, Nurnberger W, Mauz-Korholz C, Korholz D, et al. Myeloablative radiochemotherapy and hematopoietic stem-cell rescue in poor-prognosis Ewing’s sarcoma. J Clin Oncol 1993;11:1482-8. 30. van Valen F, Winkelmann W, Burdach S, Gobel U, Jurgens H. Interferon y and tumor necrosis factor cy induce a synergistic antiproliferative response in human Ewing’s satrcoma cells in vitro. J Cancer Res Clin Oncol 1993;119:615-21.