Abstract
We aimed to identify pretreatment and treatment factors that may influence the outcome of Ewing's sarcoma family tumors of bone and enable customized therapy for future studies with a retrospective analysis of 579 patients with non-metastatic Ewing's sarcoma treated with combined adjuvant or neoadjuvant chemotherapy at a single institution between 1972 and 1998. We evaluated the prognostic significance of gender, age, site and volume of tumor, serum level of LDH, type of local treatment, type of chemotherapy and histologic response to preoperative treatment. The 5- and 10-year disease-free survival rates were 56.9% and 49.2% respectively. Multivariate analyses showed that all the evaluated factors, with exclusion of the tumor site, were significantly correlated with the 5-year disease-free survival. We concluded that the outcome of non-metastatic ESF of bone tumors is influenced by many clinical and treatment-correlated variables. In order to gain the greatest benefit from treatment, while reducing the morbidity, appropriate therapeutic strategies for different risk groups of patients should be selected. Criteria to stratify patients according to the risk of local or systemic relapse should not be based on a single prognostic factor, but should include all the variables that showed prognostic significance.
Surgery, radiotherapy and adjuvant or neoadjuvant chemotherapy can cure the majority of Ewing's sarcoma family tumors (ESFT). Nevertheless, up to 30–40% of patients with non-metastatic disease at presentation still relapse or do not have an initial remission, and ultimately die of tumor progression. Intensified first line chemotherapy regimens (i.e. megatherapies with stem-cell support) might improve the prognosis for some patients who actually relapse, but this would mean an over-treatment for the many patients who could benefit from less aggressive regimens. Therefore in ESFT of bone, the identification of risk factors for relapse would be of major importance in the development of new, more rational risk-adapted strategies of treatment. The aim of this study was to assess the influence of a series of clinical, laboratory, histologic and treatment parameters on disease-free survival in a continuous series of 596 patients with non-metastatic Ewing's sarcoma of bone treated at a single institution over a 26-year period. The data concerning the prognostic factors of the first 359 of these patients, previously published Citation[1], have been updated in this study.
Patients and methods
Patient selection
Data on 596 patients with non-metastatic Ewing's sarcoma of bone treated with five protocols of adjuvant or neoadjuvant chemotherapy, between March 1972 and March 1998 at the Istituti Ortopedici Rizzoli were retrospectively reviewed ().
Table I. Patients’ characteristics and cumulative probability of 5-year Disease Free Survival (DFS).
Eligibility criteria for entering in this review were: diagnosis of ESFT of bone, age younger than 40 years, absence of metastases at diagnosis, no previous treatment, and a less than 4-week interval between biopsy and beginning of therapy. One-hundred-sixty (160) patients with ESFT of bone seen at the authors’ institution over the same period were excluded for the following reasons: metastases at presentation (129), age over 40 (34), previous treatment (5), interval between biopsy and start of treatment longer than 14 weeks (4). Twelve patients had more than one cause of exclusion.
Pre-treatment evaluation
Diagnosis of Ewing's sarcoma was made on representative specimens obtained from an open biopsy. In all cases standard histologic investigations were performed, and in 424 patients seen from 1984, immunohistochemistry studies were also carried out. Histopathologic diagnosis was based on the presence of a small round cell tumor occurring in the bone with no histologic, cytologic, or, in more recent cases, immunohistochemical features of lymphoma, rhabdomyosarcoma, or neuroblastoma. No attempts were made to differentiate Ewing's sarcoma from peripheral malignant neuroectodermal tumor. A complete history, a thorough physical examination, and several chemical laboratory tests, including serum LDH, were performed in all patients. Diagnostic imaging varied reflecting the changes in radiological techniques during the 26-year period covered by the study. Local imaging of the primary lesion included radiographs in all patients, computed tomography (CT) in 446 and MRI in 220 cases. In patients treated with neoadjuvant chemotherapy, the radiological examinations were repeated before surgery. The presence of metastases was investigated by total bone scan, whereas standard chest radiographs and CT scan of the chest were used to exclude lung metastases in the 446 patients treated after 1984, and by standard radiographs and full chest tomography for the 150 patients treated before. Tumor size was estimated by CT scan measures of the three diameters of the lesion and calculated according to the method used by Gobel et al. Citation[2].
Treatment
Chemotherapy was administered according to five consecutive adjuvant (REA-1 and REA-2), and neoadjuvant (REN-1, REN-2, REN-3) protocols () sequentially activated and extensively reported in previous papers Citation[3–6].
Table II. Chemotherapy protocols for ESFT at our Institution between 1972 and 1999.
Local treatment, individually planned for each patient after consultation among the radiotherapist, surgeon and medical oncologist, consisted of surgery only, or surgery followed by radiation therapy or radiation therapy only. The choice of local treatment was based on the age of patients, the site and size of the tumor, and the presence or not of pathologic fractures. A complete local control was also associated to the need to save the best function of the tumor-affected site.
Strategies for local control have certainly changed over the years. While in the first two protocols surgery was performed only on tumors located in expendable bones, it then became the first option for local treatment, also for lesions located in sites requiring reconstruction with prostheses, allograft or autograft after the resection of the tumor. Amputation was performed only in patients in whom limb salvage with adequate surgical margins was considered impossible, and radiation therapy, due to age or tumor site, would have probably caused worse functional results than the ablative operation. Radiotherapy was used alone in case of unresectable tumors or refusal to undergo surgery (usually amputation in adults), and when adequate surgical margins would have led to poor functional results, and or the delaying of postsurgical chemotherapy for a long time, or when the use of radiotherapy after surgery with inadequate margins would have been impeded. For local control radiation therapy alone, was administered at a dose of 55–60 Gy. Before 1991 patients received conventionally fractionated irradiation, whereas after that hyperfractionated irradiation was used Citation[7]. Patients who had had prior surgical treatment with inadequate surgical margins were treated with additional radiation therapy at a dose of 40–44 Gy.
Histologic response to chemotherapy
After surgery, all specimens were carefully observed, and surface-labeled histologic sections were taken. Surgical margins were evaluated according to Enneking et al. Citation[8]. The response to chemotherapy was determined by a thorough histologic examination of an entire coronal section of the tumor according to a previously-reported method Citation[9] and were classified as Grade I response (evidence of macroscopic foci of viable tumor cells), Grade II response (only isolated microscopic nodules of viable tumor cells), and Grade III response (no viable tumor cells left).
Follow-up
During and after the combined treatment, patients were followed-up by a physical check-up and standard radiographs (CT scan after 1990) of the chest and of the involved bone. Additional studies, including biopsies if necessary, were done if indicated by clinical and radiological examination. These tests were carried out every three months for four years, and then twice a year up to ten years. After this time, for this review, patients were contacted by phone or mail.
Statistics
Due to a lack of uniformity in the therapeutic regimen performed after relapse, and considering that very few patients who relapsed are still alive, the prognostic significance of the variables investigated was evaluated only on disease-free survival (DFS). Three patients who died of toxicity were excluded from the analysis, while the 13 patients who developed a second neoplasm were censored at the time of the second tumor. DFS was established from the date of diagnosis to the date of recurrence (local recurrence or metastases) and the relative curves were calculated according to Kaplan-Meier and compared by means of the log-rank test. Cox multivariate analysis was used to identify factors predictive of DFS. In the multivariate analyses only the factors significant by the univariate analysis were investigated. The following pretreatment variables were considered (and cut -off set to allow comparisons with other authors): gender, age (>13 years vs ≤13 years) Citation[10], tumor volume (>150 ml vs ≤150 ml) Citation[11], LDH serum level [normal (≤460 U/L) vs elevated (>460 U/L)], site of the tumor (extremity vs axial tumors). Further variables related to the type of treatment were: type of local therapy (surgery vs radiation therapy or surgery followed by radiation therapy), chemotherapy regimen (REA-1 vs REA-2 vs REN-1 vs REN-2 vs REN-3), number of drugs used in the regimen of chemotherapy (3, 4 or 6) and, in patients locally treated with neoadjuvant chemotherapy and locally by surgery, histologic response to chemotherapy (Grade I, Grade II and Grade III). Data on time to recurrence were compared by means of unpaired t-test.
Results
With a follow-up ranging between 6 and 33 years (mean: 15 years) 289 patients (49.9%) remained continuously free of disease, and 290 (48.6%) relapsed. Four patients died of chemotherapy related toxicity and 13 patients developed a second neoplasm (7 radiation induced osteosarcoma, 3 leukemia, 2 lung cancer and 1 melanoma). These 17 patients were not considered in further analyses due to their completely different treatments received.
The interval between the start of treatment and the development of the second tumor ranged from four to 20 years (mean 6.5 years). The median time to relapse was 2.2 years (2–21 years). The 5 and 10-year cumulative probabilities of DFS were 56.9% (95% CI: 40%–74%) and 49.2% (95% CI: 37%–61%) respectively. 95 patients developed local recurrence (15.9%) that in all but five cases was associated with metastases. In 33 patients LR was the first adverse event, in 45 patients local and systemic relapse were contemporary, and in 12 patients local recurrence occurred after metastasis had been diagnosed. With regards to the type of local treatment, the rate of local recurrence was significantly higher in 234 patients locally treated only by radiotherapy (24.7%) than in the 199 patients treated by surgery alone (17/199 = 8.5%) or the 163 patients treated by surgery followed by radiotherapy (20/163 = 12.3%). The difference is highly significant comparing only radiotherapy to surgery (p = 0.0001) or to surgery plus radiotherapy (p = 0.004), but is not significant comparing surgery alone to surgery followed by radiotherapy (p = 0.32). In surgically treated patients with regards to surgical margins the rate of local recurrence was 18.6% for the 75 patients with inadequate surgical margins and 7.6% for the 286 patients with adequate surgical margins (p = 0.009). The rate of local recurrence according to type of local treatment and surgical margins is reported in details in . It is interesting to note that patients with inadequate margins treated by surgery or surgery and radiotherapy did not show different rates of LR: 22.2% vs. 19.2, p = 0.92. Nonetheless, the rate of inadequate margins in the group of patients treated by surgery alone was only 9%, while it was 34.9% for those treated by surgery followed by radiotherapy; This difference was highly significant (p = 0.0001) and demonstrated that in case of inadequate surgical margins radiotherapy at reduced doses is not able to compensate for the risk of LR. 285 patients developed metastases. The first metastases were located in the lung 132 patients (46.3%), in other bones in 99 (34.7%), in 46 in both lung and bone (16.1%), and in eight patients in other sites. With regards to post-relapse outcome for the 290 patients who relapsed, 20 are alive and free of disease from 0.5 to 20 years from the last treatment, seven are alive with uncontrolled disease and 280 died of their tumor. In these 280 patients the mean time to death from the beginning of treatment was 3.8 years (0.3–29 years).
Table III. Local recurrence according to local treatment and surgical margins.
Univariate analysis
The patient-related parameters and their prognostic significance on DFS are shown in , treatment-related variable and DFS are reported in . In the type of local treatment by tumor site is accounted.
Table IV. 5-year Disease Free Survival (DFS) according to treatment features (579 patients)*.
Table V. Type of local treatment by tumor site (579 patients)*.
The rate of patients locally treated with surgery increased from 32.1% (27/34) of the first protocol to 72.2% of the last protocol (190/263), while the rate of patients locally treated only by radiotherapy decreased from 67.8% (57/84) to 27.7% (73/263) (p < 0.0001). The mean age of patients treated surgically was 16.2 years (1.5–40) and for those treated by radiotherapy alone 17.6 years (3–40). Surgical margins were adequate (radical or wide) in 281 cases (79.2%) and inadequate (marginal, intralesional or contaminated in 74 (20.8%). There were no differences in terms of adequate surgical margins between patients treated with the first protocol and patients treated with the last protocol (74% vs 80%, p.89).
As regards the histologic response to preoperative treatment, it was also correlated to time to relapse. It was significantly longer for patients with a grade III response in comparison with patients with a grade I response (35.2 months vs. 22.6 months, p = 028), while there were no differences between the time to relapse of grade II (27.5 months) and Grade III response (p = 0.41), and grade II and grade I response.
Multivariate analyses
We considered all cases () according to local treatment, patients’ gender and age, serum LDH at presentation and the number of drugs employed. All variables maintained their independent prognostic values. Histologic response to chemotherapy, whose data is missing in 301 patients (137 had an adjuvant chemotherapy, and 164 a neoadjuvant treatment) and tumor volume (data lacking in 132 pts) where considered in a separate multivariate analysis on 278 cases ().
Table VI. Multivariate Cox Regression Analysis with Backward Wald Statistics Method (579 Patients).
Table VII. Multivariate Cox Regression Analysis with Backward Wald Statistics Method (278 Patients)*.
Discussion
Although there are no doubts that the presence of metastatic disease at presentation is the main negative prognostic factors, for patients with localized disease, the prognostic significance of several factors is still uncertain. In patients with localized disease a number of clinical and pathologic features have been reported to have a prognostic significance, including the site and size of the tumor Citation[12–14], the age Citation[11], Citation[14] and gender Citation[12] of patients, the serum LDH levels Citation[12], Citation[14–16], the presence of anemia Citation[11], the interval between the onset of symptoms and the diagnosis Citation[17], the number and type of drugs used in chemotherapy Citation[12], Citation[18], the type of local treatment Citation[10–12], Citation[19], and the histologic response to preoperative chemotherapy in patients who received a neoadjuvant chemotherapy and were treated locally by surgery Citation[10], Citation[11], Citation[19]. However, many of these studies have some important limits. First in the series from single institutions, in which the selection and treatment of patients is more homogeneous, the number of cases is generally too small to have reliable statistical significance. On the other hand, the large series of multidisciplinary studies are much less homogeneous and many patients have been treated in hospitals that have little experience in bone sarcomas And, as it happens in other tumors, the centralization of surgery in specialist hospitals have a significant influence on outcome. This could be particularly true in ES where a significant correlation between outcome and surgical margins has been reported Citation[20]. In ESFT of bone the same could be true for radiotherapy. Dunst et al. Citation[21] reported that in the CESS-86 study most of the participating radiation therapy departments treated an average of less than one patient per year. Optimal radiation treatment, however, is not only a question of available technical equipment, but requires personal experience with the treatment of the disease, especially if radiation therapy is part of a multimodality treatment. Second, the prognostic factors evaluated in different studies have not been standardized, with the consequence that the results must be interpreted in the context of only the variables included in the selected studies. Our present analysis evaluated a large number of patients for all the previously cited variables. The main strength of our study is that patients had been all treated at the same institution by the same team of doctors, and that the data concerning the variables evaluated are available for almost all patients. The main shortcoming is that we elaborated results not achieved in randomized studies, but concerning patients treated with protocols activated in succession over a long period or time of 27 years. Nonetheless, our study seems to demonstrate that gender, age, serum levels of LDH, tumor volume, number of chemotherapy drugs type of local treatment and histological grade of response to preoperative treatment had an independent influence on DFS of patients with non-metastatic ESFT of bone treated with combined chemotherapy. These data essentially confirm the results of our previous report Citation[1] with the only exception that the adverse prognostic effect of tumor location outside the extremity that was an independent prognostic factor lost its independent value in the present study
On the basis of our results, the criteria used in previous studies Citation[10], Citation[22], Citation[23] to classify high risk-patients seem to be questionable. In future clinical trials, the criteria to stratify patients according to the risk of relapse should not be based on one or two prognostic factors, but include all the variables with prognostic significance. Furthermore, the possibility to use molecular analysis to define different targets for therapeutic intervention is still unsure. The last consideration concerns the role of surgery in local treatment. As reported, there was an increased use of surgery over the period of our study. This seems to be associated with a marked decrease in the local relapse rate, and probably as a consequence, the DFS increase. These data confirm the results reported by Cotteril et al. Citation[13] who found the increasing use of surgery in the more recent cases to lead to a decrease in the rate of local recurrence from 24% to 4% in extremity tumors, and from 31% to 15% for high-risk axial tumors. The authors however concluded that “outside a randomized controlled trial it will never be possible to determine which modality of local therapy is optimal for ESFT of bone”. We believe, however, that such a study should not include too many centers, as it is happening today, with different surgical and radiotherapeutic experience in treating bone tumors. This could introduce a bias that might be greater than that of our retrospective studies.
References
- Bacci G, Ferrari S, Bertoni F, Rimondini S, Longhi A, Bacchini P, et al. Prognostic factors in nonmetastatic Ewing's sarcoma of bone treated with adjuvant chemotherapy: Analysis of 359 patients at the Istituto Ortopedico Rizzoli. J Clin Onco1 2000; 81: 4–11
- Gobel V, Jurgens H, Etspuler G, Kemperdick H, Junblut RM, Stienen 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
- Bacci G, Campanacci M, Pagani PA. Adjuvant chemotherapy in the treatment of clinically localised Ewing's sarcoma. J Bone Joint Surg [Br] 1978; 60: 567–74
- Bacci G, Toni A, Avella M, Manfrini 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
- Bacci G, Picci P, Ferrari S, Mercuri M, Brach del Prever A, Rosito P, et al. Neoadjuvant chemotherapy for Ewing's sarcoma of bone: No benefit observed after adding ifosfamide and etoposide to vincristine, actinomycin, cyclophosphamide, and doxorubicin in the maintenance phase: Results of two sequential studies. Cancer 1998; 82: 1174–83
- Bacci G, Mercuri M, Longhi A, Bertoni F, Barbieri E, Donati D, et al. Neoadjuvant chemotherapy for Ewing's tumour of bone: Recent experience at the Rizzoli Orthopaedic Institute. Eur J Cancer 2002; 38: 2243–51
- Barbieri E, Frezza G, Martelli O, Neri S, Mercuri M, Gherlinzoni F, et al. Non conventional fractionation in radiotherapy of the musculo-skeletal sarcomas. Tumori 1998; 84: 167–70
- Enneking W, Dunham W, Gebhardt M, Malawad M, Pritchard D. A system for the classification of skeletal resections. Chir Organi Mov 1990; 75: 217–40
- Picci P, Rougraff BT, Bacci G, Neff JR, Sangiorgi L, Cazzola A, et al. Prognostic significance of histopathologic response to chemotherapy in nonmetastatic Ewing's sarcoma of the extremities. J Clin Oncol 1993; 11: 1763–9
- Paulussen M, Ahrens S, Dunst J, Winkelmann W, Exner GU, Kotz R, et al. Localized Ewing tumor of bone: Final results of the cooperative Ewing's Sarcoma Study CESS 86. J Clin Oncol 2001; 19: 1818–29
- Delepine N, Delepine G, Cornille H, Voisin MC, Brun B, Desbois JC, et al. Prognostic factors in patients with localized Ewing's sarcoma: The effect on survival of actual received drug dose intensity and of histologic response to induction therapy. J Chemother 1997; 9: 352–63
- 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 a European Cooperative Trial. Cancer 1988; 61: 23–32
- Cotterill SJ, Ahrens S, Paulussen M, Jurgens HF, Voute PA, Gadner H, et al. Prognostic factors in Ewing's tumor of bone: analysis of 975 patients from the European Intergroup Cooperative Ewing's Sarcoma Study Group. J Clin Oncol 2000; 18: 3108–14
- Smeland S, Wiebe T, Brosjo O,Bohling, T Alvegard TA. Chemotherapy in Ewing's sarcoma. The Scandinavian Sarcoma Group experience. Acta Orthop Scand 2004;75:S87–S91 (Suppl).
- Bacci G, Ferrari S, Longhi A, Rimondini S, Versari M, Zanone A, et al. Prognostic significance of serum LDH in Ewing's sarcoma of bone. Oncol Rep 1999; 6: 807–11
- Farley FA, Healey JH, Caparros-Sison B, Godbold J, Lane JM, Glasser BD, et al. Lactase dehydrogenase as a tumor marker for recurrent disease in Ewing's sarcoma. Cancer 1987; 59: 1245–8
- Gehan EA, Nesbit ME, Jr, Burgert EO, Jr, Burgert EO, Vietti J, Teft M, Perez CA, et al. Prognostic factors in children with Ewing's sarcoma. Natl Cancer Inst Monogr 1981; 56: 273–8
- Oberlin O, Deley MC, Bui BN, Gentet JC, Philip T, Terrier P, et al. Prognostic factors in localized Ewing's tumours and peripheral neuroectodermal tumours: The third study of the French Society of Paediatric Oncology (EW88 study). Br J Cancer 2001; 85: 1646–54
- Grier HE, Krailo MD, Tarbell NJ, Link MP, Fryer CI, Priychard DJ, et al. Addition of ifosfamide and etoposide to standard chemotherapy for Ewing's sarcoma and primitive neuroectodermal tumor of bone. N Engl J Med 2003; 348: 694–701
- Ozaki T, Hillmann A, Hoffmann C, Rube C, Blasius S, Dunst J, et al. Significance of surgical margin on the prognosis of patients with Ewing's sarcoma. A report from the Cooperative Ewing's Sarcoma Study. Cancer 1996; 78: 892–900
- Dunst J, Jurgens H, Sauer R, Pape H, Paulussen M, Winkelkmann W, et al. Radiation therapy in Ewing's sarcoma: An update of the CESS 86 trial. Int J Radiat Oncol Biol Phys 1995; 32: 919–30
- Burghert EO, Jr, Nesbit ME, Garnsey LA, Gehan EA, Herrmann J, Vietti TJ, et al. Multimodal therapy for the management of nonpelvic, localized Ewing's sarcoma of bone: Intergroup study IESS-II. J Clin Oncol 1990; 8: 1514–24
- Oberlin O, Habrand JL. Ewing's sarcoma: Towards a common protocol for adults and children. Cancer Radiother 2000;4:S141–S144 (Suppl).