Annals of Oncology 3 (Suppl. 2): S23-S27, 1992. 1992 Kluwer Academic Publishers. Printed in the Netherlands. Original article Limb sparing versus amputation in osteosarcoma Correlation between local control, surgical margins and tumor necrosis: Istituto Rizzoli experience F. Gherlinzoni, P. Picci, G. Bacci & D. Campanacci Bone Tumor Center and Orthopaedic Clinic of the University, Istituto Ortopedico Rizzoli, Bologna, Italy Summary. Three hundred fifty-five patients with osteosarcoma were surgically managed at the Istituto Ortopedico Rizzoli between 1983 and 1988. Three hundred two patients performed two cycles of neoadjuvant chemotherapy according to two different protocols consecutively activated; the remaining 53 patients were immediately operated on without neoadjuvant treatment. Follow-up ranged from 19 to 88 months, with a mean of 51 months. Demolitive surgery was performed in 116 patients (33%), while conservative in 239 patients (67%). A radical margin was achieved only in patients undergoing amputations: no radical surgical margins were achieved by local resection: 174 resected patients (73%) had wide margins, while 34 (14%) had marginal margins; 20 patients had contaminated margins and 11 patients (5%) had Introduction Neoadjuvant Chemotherapy has radically altered management and outcome of patients with osteosarcoma. Prior to 1970 all patients were treated with immediate amputation and died of pulmonary metastasis in the majority of cases [1]. The proven efficacy of adjuvant chemotherapy has improved the prognosis of osteosarcoma in the seventies, but the only surgical procedure remained amputation [1-4]. In the last ten years chemotherapy was extended in the preoperative period, producing a radical and dramatic change in the surgical management of osteosarcoma: a further increase in prognosis with limb salvage surgery in the majority of cases [5-21]. Preoperative chemotherapy introduces a new factor which is likely to influence the final outcome and the local control of the disease: the tumor necrosis. To better recognize the influence of the tumor response on the surgical margins and on the final outcome, we have retrospectively studied all patients with osteosarcoma treated at the Rizzoli Institute in the last 7 years. This effort was done to provide information and possibly answers to the following issues: 1. The role of preoperative chemotherapy in reducing the number of amputations. 2. The effect of preoperative chemotherapy on tumor aggressiveness towards the surrounding intralesional margins. Wide or marginal margins were achieved in 208 (87%) of the patients managed by local resection. Local recurrence was identified in 24 of 355 patients (7%). Incidence of local recurrence is strictly related to surgical margin, but above all, to tumor necrosis induced by neoadjuvant chemotherapy. In fact, neoadjuvant chemotherapy makes limb salvage feasible in the majority of cases without increasing the risk of local recurrence: in the group of patients immediately operated on without neoadjuvant chemotherapy, the incidence of local recurrence is dramatically high even when the surgical margins are adequate. Key words: local recurrence, neoadjuvant chemotherapy, osteosarcoma, surgical margin normal tissues, in other words its effect on the stage of the tumor. 3. The surgical margins effectively achieved in the limb salvage procedures after preoperative chemotherapy. 4. A careful study of all the areas of marginality in each resected specimen to better assess the dangerous sites for the surgeon. 5. The minimal surgical margins required for adequate local control in limb salvage procedures. Materials and methods From March 1983 to December 1988, the Istituto Ortopedico Rizzoli managed 355 patients with biopsy proven osteosarcoma. The present study reports the overall surgical experience of the Istituto Rizzoli in osteosarcoma management; for this reason it comprises all osteosarcoma patients, including patients with histological varieties, advanced cases and patients not eligible for chemotherapy for different reasons. Between March 1983 and September 1986, 171 patients (protocol 1) were randomized between high dose (7.5 gm/m 2 ) methotrexate (HD-MTX) and low dose (750 mg/m 2 ) methotrexate (LD- MTX) and two cycles of intra-arterial cis-diaminedichloroplatinum (CDP) (120-150 mg/m 2 ). Analysis of the data of the patients included in this protocol revealed that 52% (88 pts.) had greater than 90% necrosis [11. Since necrosis had been shown to correlate with prognosis, the protocol was modified in September 1986 by the addition of adriamycin (ADM) and discontinuing low dose MTX in
24 Histologically, in a good chemotherapy response the pseudocapsule appeared to have matured in a nearly true capsule consisting of mature collagen bundles or in mature bone: in other words this surrounding mature collagenized tissue layer appeared to encapsulate the necrotic underlying tumor. The tumor response increased with the addition of ADM and the use of HD-MTX: patients in protocol 1 presented a good response in 51.7% (88 out of 171), compared to 69.4% (91 out of 131) achieved with protocol 2. These results were statistically significant when comparing good responses with combined fair and poor together (p = 0.029) or when comparing the three groups individually (p = 0.038). Of the 355 patients in the study, 107 patients (30%) Fig. la, b. Osteosarcoma of the distal femur in a 11-year-old male with a good necrosis. Antero-posterior view before (a) and after (b) neoadjuvant chemotherapy. In (b) the tumor is better delimitated towards the surrounding soft tissue. Results Follow-up ranged from 19 to 88 months with a mean of 51 months: follow-up of patients evaluated from the first study ranged from 45 to 88 months, while those patients from the second study were followed-up from 19 to 44 months, with a mean of 64 months; patients operated without chemotherapy had a mean follow-up of 64 months (19-85). Clinically, all patients, after two cycles of neoadjuvant chemotherapy, presented a generalized reduction in local pain, swelling and heat of the skin; the soft tissues surrounding the tumor became less inflamed and the contiguous joints restored a normal range of motion. The radiographics changes were correlated with the degree of tumor necrosis: a good response was frequently evident with the development of a rim of reactive bone around the tumor mass as seen in plane x-rays (Fig. la, b). CT scan and MRI demonstrated a marked decrease in soft tissue edema with better defined tumor margins consisting in a thicker and continuous rim dividing the tumor mass from the surrounding healthy tissues (Fig. 2a-d). Fig. 2a-d. Same patient of Fig. 1. CT. scan (a, b) and M.R.I, (c, d) before and after neoadjuvant chemotherapy. A decrease in the edema in the surrounding soft tissues and a development of a shell around the tumor are evident. After neoadjuvant chemotherapy the tumor is better divided from the surrounding healthy tissue. an attempt to improve the preoperative tumor necrosis response. The chemotherapy treatment (protocol 2) was initiated by the intravenous administration of 8 gm/m2 MTX over 6 h; all patients had MTX serum levels determined every 12 h until the level was below 0.2 M along with hydratation and alkalinization of the urine. The patients then received 120 mg/m2 of CDP intra-arterially over 72 h in day 7. ADM (60 mg/m2) was slowly infused intravenously over 8 h. A second preoperative cycle was repeated after 27 days and surgery was performed 55 to 65 days after initiation of preoperative chemotherapy. Between September 1986 and December 1988, 131 patients have been managed with the second protocol. In the same period of time of protocols 1 and 2, 53 additional patients not eligible for or refusing chemotherapy, were immediately surgically treated without any neoadjuvant treatment. After surgery the entire surgical specimen of all patients was carefully evaluated in the pathology laboratory by inspection, and all the areas thought to be in question for marginality were marked with India ink and evaluated microscopically; also the materials from the medullary canal were similarly studied. The obtained surgical margin was defined as the single least adequate margin microscopically identified on the entire surgical specimen. According to the Enneking criteria [22] we divide the surgical margins into: intralesional, marginal, wide, radical and contaminated. The resected specimen was then bi-valved in a longitudinal plane, photographed, blocked into sections and numbered [23]; 5 to 28 blocks resulted from each specimen; these were histopathologically reviewed to determine the percentage of tumor necrosis (2325). The histological chemotherapy response was divided microscopically into three grades, defined as follows: 1. Good response - 90% or greater microscopic evidence of tumor necrosis. 2. Fair response - between 60% and 89% evidence of tumor necrosis. 3. Poor response - less than 60% tumor necrosis. Preferential sites of viable tumor were then obtained by sequential mapping of the entire specimen [23]; with this procedure all areas of viable tumor were identified and the anatomic areas were also determined where wide surgical margins are surgically difficult to obtain, correlating the surgical margins with the possible development of local recurrence.
25 the margin on 8 occasions, requiring the resection of an additional segment of the shaft of the involved bone. Re-excisions of the margins were also necessary in 6 patients, when tumor thrombi were identified by frozen section in the vein draining and surrounding the tumor (Fig. 3a, b). Joint structures were found to be involved on 4 occasions, requiring additional excision of soft tissue. Three patients had biopsy tracts injudiciously positioned such that an en bloc resection of the biopsy tract could not be achieved through the operative incision. Local recurrence was identified in 24 of 355 patients (7%). All local recurrences were clinically and radiographically apparent from 2 to 56 months (mean 12) after the operative procedure. Twelve of 179 patients (7%) were identified from the first protocol and 4 of 131 (3%) from the second protocol; the last 8 patients with local recurrence belong to a group which did not receive any neoadjuvant treatment (8/53,15%). When evaluated as to the type of surgical procedure, 3 Fig. 3a, b. C.T. scan of a 14-year-old patient with an osteosarcoma treated with neoadjuvant chemotherapy. A thrombus in the popliteal vein is clearly evident (arrows) also in the sagittal reconstruction (b). In this case the tumor must be resected together with the vessel. were treated by amputation, while 239 (67%) were managed by local resection; rotation plasty was performed in 9 patients (3%). Resection rate increased from 70% to 79% in patients treated with the first and with the second protocol, while patients operated on without chemotherapy were resected only in 22%; on the contrary, amputation rate was 28%, 17% and 78%, respectively, in patients treated with the first and the second protocol and in patients operated on without pre-operative chemotherapy. Rotation plasty rate was low in both the chemotherapy groups (2%-4%). When comparing the two protocols and patients operated on without chemotherapy it was apparent that the improved tumor response to chemotherapy (especially in the second protocol) was associated with fewer amputations. Regarding surgical margins we found that a radical margin was achieved only in patients undergoing amputations (107 pts.): in fact, combining together amputations and rotation plasties (116 pts.) we found 58 patients (50%) with radical margin and 56 patients (48.2%) with wide margins; globally 114 patients (98.2%) achieved either a radical or a wide margin when amputation or rotation plasty was adopted: one patient had contaminated margin and another patient had an intralesional margin. No radical surgical margins were achieved by local resection performed in 239 patients: 174 patients (73%) had wide margins, while 34 (14%) had marginal margins; 20 patients had contaminated margins and 11 patients (5%) had intralesional margins. Wide or marginal margins were achieved in 208 (87%) of the patients managed by local resection. Marginal margins were found in 37 areas of 34 patients: the most dangerous site for a risk of marginality is the posterior aspect of the knee region when performing a resection of the distal femur or of a proximal tibia (20 cases). Vessels are spread in a fatty tissue without a barrier with the tumor which often bulges and displaces the neurovascular bundle. In 10 cases the marginal margin was found inside the joint (cruciate or teres ligaments) or in the articular capsule when performing an intraarticular resection. Tumor in the medullary canal was marginal at the level of resection in 4 cases, and in the 3 latter cases the surgical margin was marginal in the soft tissue around the tumor. Twelve of 355 total patients (3.4%) had intralesional margins not converted intraoperatively to contaminated margins: the sites of contamination in these patients revealed 8 in which the medullary canal margin was shown to be involved, 2 where soft tissue margin was involved, one where the cruciate ligaments were involved (when performing a transarticular resection) and one where the periosteal reaction was still present at the level of resection. Contaminated margins were identified in 22 areas in 21 patients and occurred in 20 patients undergoing a local resection and in one patient with a rotation plasty. Tumor within the medullary canal was discovered at
26 patients having an amputation (3%) and 21 patients having a resection (9%) developed local recurrence. Discussion The goal of this paper is to establish the influence of chemotherapy-induced tumor necrosis and of surgical margins on the local control of osteosarcoma: for this evaluation all our data must be further analyzed. In fact, to better ascertain the influence of tumor necrosis and surgical margins on the development of local recurrence, those patients managed by amputation or rotation plasty (116) and those not surviving a minimum of 18 months (35) were excluded. The remaining 204 patients all had resection of their primary tumor and all were observed for a minimum of 19 months. Twenty-one of these patients (10%) developed a local recurrence. Correlating necrosis and recurrences, only 3 patients of 127 (2%) having good necrosis developed a local recurrence while 8 of 54 (15%) with fair necrosis and 4 of 14 (29%) with poor necrosis; 6 patients of 9 (67%) operated on without chemotherapy developed a local recurrence. When correlating surgical margins and local recurrence, 6 patients of 147 (4%) with wide surgical margins developed a local recurrence compared to 9 of 31 (29%) with a marginal margin and 3 of 16 (19%) with a contaminated margin. Of the 10 patients with histopathologically proven intralesional margins, 3 patients (30%) developed a local recurrence. The combination of good necrosis and wide margins led to only one local recurrence in 93 patients (1%); furthermore, only 2 local recurrences (6%) were observed in those patients with good necrosis and either marginal or contaminated or intralesional margins. In contrast, those patients with fair necrosis (54 pts.) developed only one local recurrence with a wide margin (3%); however, 5 of 10 (50%) patients with marginal margins and 1 of 4 patients (25%) with contaminated margins developed local recurrence. Moreover, those patients with poor tumor necrosis (14 pts.) developed local recurrence in 28.5% of the cases, even with wide surgical margins. In patients operated on without neoadjuvant treatment (9 pts.) local recurrence presented in 6 patients (67%): considering the surgical margins, local recurrence appeared in 40% of patients operated with wide margins (2/5) and in 100% of patients operated with marginal, contaminated or intralesional margins (4/4). The data of our retrospective study suggest that when using neoadjuvant chemotherapy, the quality of the surgical margin is directly related to the grade of tumor necrosis and not necessarily related to the width of tissue included within the margin. When the tumor response is good, a wide margin should be achieved where possible; however, the data would suggest that marginal margins may be adequate when necessary. With fair necrosis, a wide margin is necessary for adequate control and marginal margins must be avoided. With poor tumor necrosis wide margins are often not adequate and radical margins, usually to be achieved by amputations, should be considered. Conclusions 1. From our experience neoadjuvant chemotherapy makes limb salvage feasible in the majority of cases without increasing the risk of local recurrence: in the group of patients immediately operated without neoadjuvant chemotherapy, the incidence of local recurrence is dramatically high, even when the surgical margins are adequate. 2. When the response to preoperative chemotherapy is good, the pseudocapsule around the tumor becomes a mature capsule, making limb salvage procedure more feasible and safe. 3. With improved experience in imaging interpretation and surgical planification and technique, wide non-contaminated margins can be achieved in more than 80% of the limb salvage procedures. 4. The surgeon must be aware of marginality when performing limb salvage operation especially in the most dangerous sites: popliteal fossa, articular joints, vessels and medullary canal. 5. Local control is a function of the surgical margins obtained, but with preoperative chemotherapy it is also dependent upon the grade of tumor necrosis. References 1. Campanacci M, Bacci G, Bertoni F et al. The treatment of osteosarcoma of the extremities: twenty years experience at the IstitutoOrtopedico Rizzoli.Cancer 1981;48: 1569-81. 2. Burgers JMV, Van Glabbeke M, Busson A et al. Osteosarcoma of the limbs. Report of the EORTC-SIOP trial investigating the value of adjuvant treatment with chemotherapy and/or prophylactic lung irradiation. Cancer 1988; 61:1024-31. 3. Eilber F, Giuliano A, Eckardt J et al. Adjuvant chemotherapy for osteosarcoma: a randomized prospective trial. J Clin Oncol 1987; 21-6. 4. Link MP, Goorin AM, Miser AW et al. The effect of adjuvant chemotherapy on relapse-free survival in patients with osteosarcoma of the extremity. New Engl J Med 1986; 314: 1600-6. 5. Benjamin RS, Chawla SP, Murray JA et al. Response to preoperative chemotherapy of osteosarcoma improves diseasefree survival and the chance of limb-salvage? Proc Am Soc Cancer Res 1985; 26: 174. 6. Dubousset J, Missenard G, Genin J. Traitement chirurgical conservateur des sarcomes osteogeniques des membres. Techniques et resultats functionnels. Revue de Chir Orthop 1985; 71:435-450. 7. Jaffe N, Knapp J, Chuang W et al. Osteosarcoma: Intra-arterial treatment of the primary tumor with cisdiamine-dichloroplatinum (CDP). Cancer 1983; 51:402-10. 8. Jaffe N, Prudich J, Knapp J et al. Treatment of primary osteosarcoma with intra-arterial and intra-venous high-dose metothrexate. J Clin Oncol 1983; 1:428-437. 9. Makley JT, Krado M, Ertel IJ et al. The relationship of various aspects of surgical management to outcome in childhood non-
metastatic osteosarcoma: A report from the Childrens Cancer Study Group. J Ped Surg 1988; 23: 146-51. 10. Marcove RC, Huvos AG, Meyers PA et al. En bloc surgery for osteogenic sarcoma: Analysis and review. Compr Ther 1988; 14 (5): 3-6. 11. Picci P, Bacci G, Capanna R et al. Neoadjuvant chemotherapy for osteosarcoma. Results in 126 consecutive patients. In Yamamuro T (ed): New developments for limb-salvage in musculoskeletal tumors. Springer-Verlag: Tokyo 1989; pp. 183-7. 12. Picci P, Bacci G, Capanna R et al. Neoadjuvant chemotherapy for localized osteosarcoma of the extremities. Experience related to 112 cases treated between March 1983 and June 1986. In Salmon S and De Vita V (eds): Adjuvant therapy of cancer. Grune and Stratton 1987; pp. 711-7. 13. Rosen G, Marcove R, Caparros B et al. Primary osteogenic sarcoma. The rationale for preoperative chemotherapy and delayed surgery. Cancer 1979; 43: 2163-75. 14. Rosen G, Caparros B, Huvos AG et al. Preoperative chemotherapy for osteogenic sarcoma: Selection of post-operative adjuvant chemotherapy based on the response of the primary tumor to preoperative chemotherapy. Cancer 1982; 49: 1221-30. 15. Rosen G. Preoperative (neoadjuvant) chemotherapy for osteogenic sarcoma: a ten years experience. Orthopaedics 1985; 8: 659-70. 16. Simon MA, Aschliman MA, Thomas N et al. Limb-salvage treatment versus amputation for osteosarcoma of the distal end of the femur. J Bone Joint Surg 1986; 68A: 1331-7. 17. Simon MA, Nachman J. Current concepts review. The clinical utility of preoperative therapy for sarcomas. J Bone Joint Surg 1986; 68A, 1458-63. 18. Simon MA. Current concepts review. Limb-salvage for osteosarcoma. J Bone Joint Surg 1988; 70A: 307-10. 27 19. Springfield DS, Schmidt R, Graham-Pole J et al. Surgical treatment for osteosarcoma. J Bone Joint Surg 1988; 70A: 1124-30. 20. Winkler K, Beron G, Kotz R et al. Neoadjuvant chemotherapy for osteogenic sarcoma: Results of a cooperative German/ Austrian study. J Clin Oncol 1984; 2:617-26. 21. Winkler K, Beron G, Delling G et al. Neoadjuvant chemotherapy for osteosarcoma: Results of a randomized cooperative trial (COSS-82) with salvage chemotherapy based on histological tumor response. J Clin Oncol 1988; 6: 329-37. 22. Enneking WF, Spanier SS, Goodman MA. A system for the surgical staging of musculoskeletal sarcoma. Clin Orthop 1980; 153: 106-20. 23. Picci P, Bacci G, Campanacci M et al. Histologic evaluation of necrosis induced by chemotherapy. Regional mapping of viable and non viable tissue. Cancer 1985; 56:1515-21. 24. Picci P, Campanacci M, Bacci G et al. Preoperative chemotherapy in osteosarcoma. Local results and histologic evaluation of necrosis in a study of 97 patients. In Enneking WF (eds): Limb-salvage in musculoskeletal oncology. Churchill Livingstone: New York 1987; pp. 294-6. 25. Picci P, Bacci G, Capanna R et al. The role of the pathologist in the treatment of osteosarcoma. In Ryan JR and Baker LH (eds): Recent concepts in sarcoma treatment. Kluwer Academic Publishers: Dordrecht, The Netherlands 1988; pp. 27-30. Correspondence to: Franco Gherlinzoni M.D. Bone Tumor Center Istituto Ortopedico Rizzoli Via Pupilli 1 40136 Bologna, Italy