1~~ ~~

1~~--------~~-------- Number 270 September. mpact of Two Cycles of Preoperative Chemotherapy With ntraarterial Cisplatin and ntravenous Doxorubicin on the Choice of Surgical Procedure for High-Grade Bone Sarcomas From the Depanmerns of Orthopedic Surgery. Hematology and Oncology. Radiology. and Pathology. Children's National Medical Center. The George Washington School of Medicine and Health Sciences. and the Washington Hospital Center. Washington. D.C Presented in pan at the Annual American Meeting of the Society of Clinical Oncology. San Francisco. California. May 20-24. 1989. Reprint requests to Martin Malawcr. M.D.. Department of Orthopedics. Children's National Medical Center. Michigan Ave.. N.W.. Washington. D.C. ~OOlO. Received: November 29. 1990 of the Extremities MARTN MALAWER, MD.. RCHARD BUCH, M.D., GREGORY REAMAN. M.D., DENNS PREBAT, MD., BARRY POTTER. M.D., JASHVR KHURANA. M.D.. BARRY SHMOOKLER, M.D., KATHLEEN PATTERSON, M.D.. AND RCHARD SCHULOF, MoO.. PHD. The authors assessed the impact of two cycles of preoperative chemotherapy (POCT) with intraarterial cisplatin (120 mg/m2) and continuous intravenous doxorubicin hydrochloride (Adriamycin; 20 mgfm 2 fday X 3 days) on the decision 10 perform a limb-sparing procedure (LSP) or amputation in 22 patients with high-grade bone sarcomas of the extremities. The tumor types were osteosarcoma (17), malignant fibrous histiocytoma (three), leiomyosarcoma (one), and malignant schwannoma (one). Surgical stages were la (three), lib (17), and B (two). The prechemetherapy surgical options chosen were 12 amputations (55% of patients) and ten LSPs (45%). The initial decisions to amputate were based on a combination of the following: improper biopsy (five cases). large tumors (ten) and those with neurovascular encroachment (six), and pathological fracture tone}, Following chemotherapy, 18 LSPs (81%) and four amputations (J9%) were performed. Nine of J2 patients (75%) initially deemed unresccrabte were converted to l.sp. The median tumor response (necrosis; range, 0%-100%) was 70%; ten of 22 specimens had necrosis> 95%. Median turner necrosis for the patients treated by amputation and LSPs was 45% and 88%. respectively. Following surgery, all patients received four additional cycles of cisplatin and doxorubicin. The median follow-up period is 30 months; six patients hn ve developed metastatic disease, with a median disease-free interval of 16.6 months. The rate of local tumor control is 95% (21 of 22 patients). The authors conclude that () two cycles of POC" significant )' increase the percentage of patients with highgrade bone sarcomas of the extremities who can be safe }' treated with a LSP. (2) a response to pocr has a strong influence on the surgical decisionmaking process, and (3) primary amputations arc not recommended before a trial or -Ocr. Postoperative adjuvant chemotherapy. in conjunction with surgery. ha- altered the outcome of patients with osteosarcomas. 15.] l.l).:!] Elimination or s~stcrnic micrometastases appears to be an important mechanism in the success of this for.u of therapr Two rec bone tur jllvant) lengths, tive ehe istered i gery, du period f from tv learned cause tu tivity te. in conve amenab also invr This s to evalu: cycles) ( perform sus amp pact of, vantche all patiei impact t procedu longer f, latter qu Twenf of the ext vis. partie and f, (range, se noses We] histiocytr nant schv loskeletat and 118 ability ir graphs, bi phy (CT (MR) nf STUD) Preope des or i] hours <!:-:'G drochlori. ~ 120 mg 214,

Number 270 September. 1991 Preoperative Chemotherapy 215 'perative :isplatin and lice of Surgical Sarcomas RY REAMAN, M.D., KHURANA, M.D., ON,M.D., ological fracture cne). Follow- 18 LSPs (81%) and four ampuperformed. Nine of 12 patients.med ~ectable were conrncdir or response (necro- 'Yo)W~ Jb; ten of 22 speci- > 95%. Median tumor necrosis at cd by amputation and LSPs respectively. Following surceivcd four additional cycles of -ubicin. The median follow-up s; six patients have developed with a median disease-free in- 15. The rate of local tumor conn patients). The authors concycles of poer significant )' ntage of patients with high- LS of the extremities who can be a LSP, (2) a response to pocr ence on the surgical decision-!d (3) primary amputations arc sefcre a trial of POCr. adjuvant chemotherapy. in 1 surgery. has altered the atients with osteosarconination of systemic minors to be an important mechaess of this form of therapy. Two recent developments in management of bone tumors have been preoperative (neoadjuvant) chemotherapy and shorter overall lengths of adjuvant treatment. 3 7 21 Preoperative chemotherapy (POCT) was first administered in the early days of limb-sparing surgery, during the mid-1970s, when the waiting period for a custom-made prosthesis ranged from two to three months.' 12.1) Surgeons learned that POCT could reduce tumor size, cause tumor necrosis, and allow in vivo sensitivity testing. n addition, the role of POCT in conversion of un resectable tumors to those amenable to limb-salvage procedures was also investigated. 3,4.6.10.15.16,20,21 This study had two primary objectives: () to evaluate the impact of a short course (two cycles) of POCT on the surgical decision to perform a limb-sparing procedure (LSP) versus amputation, and (2) to evaluate the impact of a short duration (six cycles) of adjuvant chemotherapy on relapse-free and overall patient survival. This report describes the impact of POCT on the choice of surgical procedure; a subsequent report (awaiting longer follow-up intervals) will address the latter question. MATERALS AND METHODS Twenty-two patients with high-grade sarcomas of the extremities, including the shoulder and pelvis, participated in the study. There were 1\ males and females. The median age was 19.5 years (range, seven to 69 years). The histogenic diagnoses were osteosarcoma (17), malignant fibrous histiocytoma (3), leiomyosarcoma (), and malignant schwan noma (1). The surgical stages (Musculoskeletal Tumor Society) were A (3), lib (17), and 1118 (2). Staging studies to dctennine resectability included bone scans, plain roentgenographs, biplane arteriograms. computed tomography (CT) and/or magnetic resonance imaging (MR) of the extremity, and chest CT. STUDY DESGN AND TREATMENT REGMEN Preoperati ve chemotherapy consisted OflWO cycles of intraarterial cisplatin (DOP) over four hours and continuous intravenous doxorubicin hydrochloride (Adriamycin [ADR]; dosage, DDP ~ 120 mg/m 2 and ADR ~ 120 mg/m". day"! X 3 days). Postoperatively. four cycles of DDP and AOR were given intravenously. The regimens were not altered on the basis of histologic tumor response (necrosis). Before POCT and after completion of the initial staging studies, a preliminary surgical decision (amputation or LSP) was recorded for each patient. Following POCT, all tumors were restaged, and a second, and final, surgical decision was made based upon the same criteria that had been used for the preliminary decision. ANG(XjRAPHC TECHNlQUE ntravenous prehydration was started eight hours before cisplatinum therapy. Catheterization, with 5 French in adults and 4 Fren'ch in children. was done via the femoral route using the standard Seldinger technique. t7 Catheters were placed in the major vessels of the extremities: the common femoral, subclavian. and iliac arteries for distal femoral, proximal humeral, and pelvic tumors respectively. The superficial femoral artery was utilized for proximal tibial tumors. Selective catheterization was not performed. An arteriogram was performed and necessary adjustments of the catheter were made. The catheter was then attached to an arteriaj infusion pump. and the patient was adequately heparinized to prolong partial thromboplastin time to.s to two times the normal. Mannitol (12.5 g) was given intravenously S minutes before cisplatin. SURGCAL CONSDERATONS The decision to proceed with a limb-sparing procedure was based on six key criteria: () The major neurovascular bundle should be free of tumor. (2) Wide resection with a normal cuff of tissue in all directions should be possible. (3) En bloc removal of potential sites of contamination, including biopsy sites, must be feasible. (4) Bone must be resectable 5 em to 6 cm beyond the level indicated by preoperative imaging studies. (5) The adjacent joint and capsule must be resectable (distal femoral tumors were treated by an intraarticular procedure, whereas all proximal humeral reseclions were extraarticular). (6) Adequate muscle must be present for both motor power and coverage of the prosthetic implant. The contra indications to limb-sparing resection used in this protocol were as follows: () major neurovascular involvement, (2) inappropriate biopsy sites with extensive soft-tissue contamination. (3) pathologic fracture, (4) large tumor size or absence of adequate muscle groups to allow a functional extremity, and (5) local infection. These contraindications were recorded for each pa-

216 Malawer et al. tient (Table ). mmature skeletal age was not considered a contraindication. PATHOLOGC EVALUATON The gross specimen was evaluated to determine the final surgical margin. Tumor necrosis was determined by the grid method." The percent tumor necrosis was recorded. "Good" and "poor" respondents were defined as greater than and less than 90% tumor necrosis respectively. ANALYSS There were four possible options with respect to the preliminary and final surgical decision: LSPlO-LSP. LSP-to-amputation, amputation-to-amputation. and amputation-to-lsp. The relationship of the initial and final surgical decisions were evaluated and compared to the degree of tumor necrosis and to the various surgical groups. Rates of local recurrence and disease-free survival were determined. RESULTS TYPE OF SURGCAL PROCEDURES nitial surgical decision (Table ). Before POCT, 12 amputations and ten LSPs were chosen, representing a 45% limb-sparing rate. The initial decisions to perform amputation were based on one or a combination of the following considerations: poor biopsy with resultant contamination (5 patients), a large Clln.cal Onl"l0; eleoocs and Related R,- ;earch tumor( 10 patients), a large tumor with neuro. vascular encroachment in which there was deemed no safe margin (7 patients). and pathologic fracture ( patient). Definitive surgical decision. The definitive procedures performed following POCT were 18 LSPs and four amputations. representing an 81% limb-sparing rate. The types of procedures performed are presented in Table 2. Relationship of pre- versus post- POl. T decision (Fig. ). Nine or 12 patients (7S',) initially deemed un resectable were converted to a limb-sparing procedure: the remaining three patients received amputations. Nine of ten patients deemed initially resecta t- le received a LSP: the remaining patient hadlocal tumor progression with a secondary fracture and required a high above-knee am put lion. Relationship of type of surgery to tumor necrosis (Table ). All procedures accomplished a wide excision. The overall median necrosis was 70% (range. 0-100%). Ten of the 22 patients had greater than 95% necrosis. There was no difference in necrosis rate between the osteosarcoma (17) and nonosteosarcoma patients (5). The median necrosis for the patients treated by LSP and amputation were 88% and 45%. respectively. The median necrosis or the LSP-to-LSP (9 patients) and the amputation-to-lsp (9 patients) groups was 75% and 95% respectively. All Number Septem nine som 50% grea T rnor nos< ou corr Sign met to Th, TABLE. Criteria for Amputation Versus Resection Relationship to Percent Tumor Necrosis ( ~ 22) Type {!( A o. Surgery Pauents BX :\"1. PQliJ. FX,\led/fll/ Tumor Sf'('~H~is [. Limb Sparing i.s.is A-lS Total lsps Amputation A-A ls-a Total A Overall total 9 9 18 4 5 75'1; 95'''" xw"o 3 2 2 35% 5U% 4..15% 22 5 7 2 to ~Oo-o LS. limb sparing: A. amputation. BX. poor biopsy with extensive contamination: NV. neurovascular inn)i\cment: Path. FX. pathologic fracture: Size. large tumor size. 7 the pre am u.

cuoeer QrthopaeOlcs and Related Research -------~,~-------------- Number 270 September, 1991 Preoperative Chemotherapy 217.nts), ~ 6e tumor with neuro-.achment in which there was rfe margin (7 patients), and :ture ( patient). rgicat decision. The definitive formed following pact were our amputations, representing oaring rate. The types of proceed are presen ted in Table 2. J of pre- verslis post-poct deine of 12 patients (75%) ini-.inresectable were converted to g procedure; the remaining received amputations. Nine of leemed initially resectable re- :he remaining patient had local sion with a secondary fracture, high above-knee amputation. 'J a/type a/surgery to tumor ne- ). All procedures accorn- : excision, The overall median 0% (range, 0-100%). Ten of the ad greater than 95% necrosis. difference in necrosis rate be-,osarr--~ (17) and nonosteo-.nts \ 'he median necrosis ts treatec by LSP and arnputaand 45%, respectively. The meof the LSP-to-LSP (9 patients) putation-to-lsp (9 patients) 5% and 95% respectively. All ction Relationship ~2) Si :e Median Tllmor Necrosis 75% 7 95% 88% 2 35% 50% 45% 10 70% ation: NV. neurovascular involvement: TABLE 2. Types of Surgical Procedures Performed After Two Cycles of pact Limb-sparing surgery (18) Segmental replacement (custom or modular) Expandable replacement Resection alone Saddle (pelvic) prosthesis Proximal femoral prosthesis Neer prosthesis Allograft nine patients in the LSP-to-LSP group had some response; the minimum response was 50%, and in four patients response was greater than 95% necrosis. TUMOR STATUS AND TUMOR NECROSS The median follow-up period is 30 months. Two patients were initially diagnosed with metastatic disease (Stage ). Six of the remaining 20 patients (five osteosarcoma, one malignant schwannoma), without signs of metastases initially, have developed metastatic disease. The disease-free interval to pulmonary metastasis was 16.6 months. The median tumor necrosis of the primary FNAL SURGCAL Types of Surgical Procedures Performed (N = 22) OUTCOME NTAL SURGCAL OECSON V PROCEDURE PERFORMED L-S L~ (N 22, 1985'19681 10 POST CT 4 18 FG.. Schematic showing the relationship or the initial (pre CT) surgical decision versus the procedure performed (post CT) with respect to amputation (AMP) and limb-sparing procedure (L-S) 9 2 3 Amputations (4) Above-knee amputation Hemipelvectomy Forequarter amputation specimen of these patients was 45%. Four of the five osteosarcoma patients had less than 90% necrosis. There has been one relapse among the six good respondents and two relapses among the nine poor respondents with osteosarcoma. Overall local tumor control was 95% (21 patients). One patient developed a local recurrence following a limb-sparing procedure for a Stage lb proximal femoral sarcoma and required a secondary hemipelvectomy. The primary specimen showed 95% tumor necross. COMPLCATONS Deep muscle (myocutaneous) necrosis occurred in two patients, Neither required surgical debridement. Two patients developed arterial thrombosis (one before surgery and one intraoperatively): both required embolectomies. One patient developed an infection secondary to an infected subclavian line following a distal femoral prosthetic replacement. necessitating an above-knee amputation. DSCUSSON The primary purpose of this prospective study was to determine the impact of preoperative chemotherapy on the choice of surgical procedure. as measured by the number of limb-sparing procedures versus amputations performed. 2

218 Malawer et al. Clinical Orthopaedics and Related Research The criteria of patient selection for LSP are based upon careful evaluation of clinical and radiographic parameters. n this study the initial (i.e.. before POCT) surgical decision was LSP for 45% of the 22 patients. This percentage in itself is higher than that reported in studies that have involved postoperative adjuvant therapy alone; it reflects, in part, the authors' positive institutional experience with LSP. Following POCT, the LSP rate rose to 81 %. Nine of 12 patients who initially werejudged to need an amputation were converted to a LSP. Only one patient became ineligible for limb sparing during POCT because of tumor progression. The major preoperative factors necessitating an amputation were a poorly performed biopsy with significant local contamination of the soft tissue, adjacent joint or (both), and proximity of tumor to the neurovascular bundle. Tumor size in itself is usually not a contraindication; it affects the choice of surgery only in the sense that larger tumors tend to displace the major vessels or require a large amount of muscle to be resected and thereby reduce the function of the extremity. The neurovascular structures were displaced in some of our patients; however, they were not involved or infiltrated by tumor. Thus, a good response safely permits the widening of the margin along the vessels, either by shrinking the tumor or by destroying presumed local rnicrometastatic disease in the reactive zone of the tumor. This study showed that the tumor necrosis of LSP patients was almost double that of patients who received amputations (88% versus 45% respectively). nterestingly, the arnputation-to-lsp patients also had a greater necrosis (95% versus 75%) than did the initial LSP patients, although both groups had a higher rate than the patients who required amputation. The same degree of tumor necrosis was observed for both osteosarcoma and tumors other than osteosarcoma. Almost all patients treated by a LSP had a tumor necrosis of greater than 50%, whereas those undergoing amputation had less than 50% response rate. Thus LSP can be performed safely if tumor necrosis is greater than 50%. This is different than the often quoted 90%, which relates to the reported improved likelihood of survival rather than to increased local tumor control. The authors recommend reserving amputation for those patients who do not respond clinically to POCT. Most patients, they noted, with greater than 60% tumor necrosis have a successful LSP. The low local failure rate of 5% observed in the current study indicates that a short course of POCT is effective in eradicating microscopic tumor nodules which, if left untreated, would lead to the development of a local recurrence (Fig. 2). Unfortunately, attempts to ascertain tumor necrosis before surgery have been unsuccessful. Preoperative biopsies are fraught with the danger of unrepresentative sampling errors. Preoperative scanning studies, if reli- Femoral Ven & Arlery Postulated Mechanism of Regional Effect of Neoadjuvant Chemotherapy lnlr..rle".t c.thete, Presumed microscopc ~P turnot nodules (oc metastases) FG. 2. Schematic demonstrating the postulated regional tumor effect or intraarterial chemotherapy on presumed local microscopic disease and pulmonary micrometastases. The increased arterial concentration (large arrow) represents increased drug concentration to the local/regional tumor area. which may eradicate local skip lesions as well as the major tumor. thus permitting a safe local procedure to be performed.

Clmicalormooeeccs ana Related aesearen ~~-----------.ents), _.arge tumor with neurooachment in which there was afe margin (7 patients), and cture ( patient). urgical decision The definitive :rformed following POCT were four amputations, representing ;paring rate. The types of proceled are presented in Table 2. ip of pre- versls post-poct del. Nineof2patients(75%)iniunresectable were converted to ng procedure; the remaining ; received amputations. Nine of deemed initially resectable rethe remaining patient had local -ssion with a secondary fracture a high above-knee amputation. ip of type of surgery 10 tumor ne-! ). All procedures accorn- e excision. The overall median 70%(range, 0-100%). Ten of the lad greater than 95% necrosis. ) difference in necrosis rate beteos?~a (17) and nonosteoients The median necrosis us treated by LSP and arnputaband45%, respectively. The me- ; of the LSP-to-LSP (9 patients) nputation-to-lsp (9 patients) 75% and 95% respectively. All Number 270 September. 1991 Preoperative Chemotherapy 219 ection Relationship 22) Size Median Tumor Necrosis 7 2 1 10 75% 95% 88% 35% 50% 45% 70% \- FGS.3A-3F. A 16-year-old patient with a distal femoral osteosarcoma. (A) Plain anterioposterior (AP) roentgenogram prior to POCT. (B) Lateral plain roentgenogram prior to POCT. (C) Plain AP roentgenogram following POCT. (0) Lateral roentgenogram following POCT. Note the smooth margins of the new bone and the reossification (arrows) of several small tumor nodules along the medial aspect. (E) CT of the distal femur prior to POCT. (F) CT of the distal femur following POCT. FOllowing POCT there is marked rimming (arrows) of the extraosseous component, which indicates a good tumor response, even though the size did not significantly decrease. This patient initially was considered for amputation but was treated with a limb-sparing procedure instead. The gross specimen showed >90% tumor necrosis. nation: NY. neurovascularinvolvement; \ \i

220 Malawer et al. Cumcal Onno, aedics ana Related Af':>earch N,m 5ep' F, turn repl. in a Fros. 4A AND 48. MR and gross specimen of a large distal femoral osteosarcoma. (A) MR folic wing POCT demonstrates a large soft-tissue component with evidence (black line. arrows) of new rimming bone. (8) The gross specimen shows marked tumor necrosis (>90%). This patient was converted to a limb-sparing procedure. abl< this ity t Dat SUgj infc me, as v indi calc tion c1in A due mot the cros COn find corj \!

Clinical Orthopaed;cs and Related Researcn Number 270 September. 1991 Preoperative Chemotherapy 221.., -' r 1 ",.', c,..,.,.\..'...'. ".,,. "~... ~.>...~ FG. 5. Low-power photomicrograph of an osteosarcoma demonstrating a typical pattern of complete tumor necrosis following POCT. There is malignant osteoid with empty lacunae and the stroma has been replaced by a reparative fibrovascular tissue. There is no evidence of viable tumor cells. (Stain, hematoxylin and eosin; original magnification, x50). oral osteosarcoma. (A) MR following (black line. arrows) of new rimming 0%). This patient was converted to a able and reproducible, would help resolve this dilemma; however, the value and reliability of radiographic signs are far from settled. Data from the M.D. Anderson group'":" suggest that angiograms may provide helpful information; however, these authors recommend using other studies and clinical criteria as wel. 7,9,14.17-19 Plain roentgen rays and CT indicators of response include an increase in calcification, a decrease in, and better, definition of the soft-tissue component, and remodeling of the cortex (Figs. 3 and 4). Angiographic signs of necrosis include a reduction in vascular blush and possibly in tumor size.":? Some authors have reported that the correlation of CT and plain film with necrosis found at pathology is poorer than the correlation of angiography and histologic findings with necrosis. Other authors have incorporaieo a combination DC clinical indica- tors such as a reduction of pain, size, temperature, and range of motion into their definition of response.":'? The authors were unable to place full confidence in any single modality; however, in most cases, the combination of clinical and radiologic studies gave a good, although not completely quantifiable, assessment of the expected tumor response. The authors' data suggest that one can probably convert an amputation to a LSP provided that there is no obvious tumor progression and radiographic studies indicate some regression (e.g.. decreased vascularity and increased rimming seen on CT, plain roentgenography, or MR; Figs. 3-5). Such a decision was made for 75% of the patients in this study, who had initially been designated for treatment by amputation. The authors do not recommend exc1uding any patient from a trial of POCT, although an infected biopsy

222 Malawer et al. Climcal OrthopaedCs and Related Researet'l site (rare) or gross displaced fracture (unusual) may require a primary amputation. Pathologic fracture was a contraindication for LSP in this study. n selected cases where healing occurs following pact, it may, however, be possible to attempt resection. This would have to be a subjective decision taken by a surgeon and based on experience. n summary, this study suggests the following: () A shon course (2 cycles) of pact is safe and permits a significant amputation-to- LSP conversion. (2) Tumor necrosis of 50% or greater correlates with successful LSP. (3) No single preoperative indicator can reliably predict tumor response. Both clinical response and combined imaging data are required. (4) Most patients with high-grade bone sarcomas deserve a trial of pact in an effort to replace amputation with LSP. (5) The impact of the route of administration and increase in number and intensity of courses of pact on surgical decision-making requires further investigation. REFERENCES l. Campanacci, M., Bacci, G., Bertoni. F., Picci, P., and Minutillo. A.. Franceschi, c.: The treatment of osteosarcoma of the extremities: Twenty years' experience at the nstituto Orthopedico Rizzoli. Cancer 48: 1569. 1981. 2. Chuang. V. P.. Wallace. S., Benjamin. R. S.. Jaffe... Ayala. A., Murray.J.. Zornoza.L. Pan, Y., Mavligit. G.. Charnsangvej. C. and Soo. c.: The therapy of osteosarcoma by intra-arterial cis platinum and limb preservation. Cardiovasc. ntcrvent. Radiol. 4:229.1981. 3. Eilber. F. R.. Eckhardt. L and Morton. D. L.: Advances in the treatment of sarcomas of the extremity. Current status of limb salvage. Cancer 54:2695. 1984. 4. Eilber. F., Morton. D. L. Eckardt. J., Grant. T.. and Weisenburger, T.: Limb salvage for skeletal and soft tissue sarcomas: M ultidiscioli nary preoperative therapy. Cancer 53:2579,1984. 5. Goorin, A. M.. Abelson, H. T. and Frei. E.: Osteosarcoma: fifteen years later. N. Engl. J. Med. 313: 1637. 1985. 6. Huvos. A. G.. Rosen, G., and Marcovc. R. c.: Primary osteogenic sarcoma. Pathologic aspects in 20 patients after treatment with chemotherapy. en bloc resection. and prosthetic bone replacement. Arch. Pathcl. Lab. Med. 101:14. 1977. 7. Jaffe. N.. Raymond. K.. Ayala, A., Carrasco. C H.. Wallace. S.. Robertson.R.. Griffiths, M.. and Wang. Y.: Effect of cumulative courses of intra-arterial cis Diamminedichloroplatin- on the primary tumor in OSteosarcoma. Cancer 63:63.1989. 8. Jaffe, N., Robertson. R., Ayala, A.. Wallace, S., Chuang. V.. Anzai. T, Cangir. A., Wang, Y.M., and Chen, T: Comparison or intra-arterial cis-diammin. edichloroplatinum with high dose methotrexate and citrovorum rescue factor in the treatment of primary osteosarcoma. J. C1in. Oncol. 3: 101. 1985. 9. Jaffe, N.. Robertson, R., Takaue. Y.. Cangir, A., Wallace, S.. Carrasco, H., Eftakhari. F., Ayala, A" and Wang, A.: Control of primary osteosarcoma with chemotherapy. Cancer 56:461. 1985. 10. Kotz. R.. Winkler. K.. Salzer-Kuntchik. M.. GOtsaunder-Wolf. F.. Kickinger, W.. Schiller, C.. RitschL P.. Heise. U.. and TorgglerS.: Surgical margins influencing oncological results in osteosarcoma. n Yamamuro T. (Ed.): New Developments for Limb Salvage in Musculoskeletal Tumors. Tokyo, Springer-Verlag, 1989, pp. 83-91.. Link. M. P., Goorin. A. M., Miser. A. W., Green, A. A., Pratt, C. B.. Belasco, J. B.. Pritchard. J., Malpas. J. S., Baker, A. R., and Kirkpatrick, J.A.: The effect of adjuvant chemotherapy on relapse-free survival in patients with osteosarcoma of the extremity. N. Engl. J. Med. 314,1600,1986. 12. Marcove. R. C,, and Rosen, G.: En bloc resection for osteogenic sarcoma. Cancer 45:3040. 19&0. 13. Monon, D. L. Eilber. F. R., To w nsend. C. M., Jr.. Grant. T. T. Mirra, L and Weisenburger. T. H.: Limb salvage from a multidisciplinary treatment approach for skeletal and soft tissue sarcomas of the extremity. Ann. Surg. 184:268. 1976. 14. Raymond. K. A., Chawla. S. P.. Carrasco, C. H.. Ayala, A., Fanning, C H., Grice, 8., Annen, T., Plager. c., Papadopoulos, N. E. J.. Edeikcn. J.. Wallace. S.. Jaffe. N.. Murray, L and Benjamin, R. S.: Osteosarcoma chemotherapy effect: a prognostic factor. Semin. Diagn. Patbol. 4(3):212, 1987. 15. Rosen, G., Caparros, 8., Huvos. A. C, Kasloff C, Nirenberg, A., Cacavio, A.. Marcove. R. C. Lane, J. M.. Mehta, B., and Urban, C: Preoperative chemotherapy for osteogenic sarcoma: Selection of postoperative adjuvant chemotherapy based upon the response of the primary tumor to preoperative chemotherapy. Cancer 49:1221. 1982. 16. Rosen. G., Marcove. R. C.. Caparros. B., Nirenberg. A., Kosloff C, and Huvos. A. G.: Primary osteogenic sarcoma. The rationale for preoperative chemotherapy and delayed survey. Cancer 43:2163, 1979. 17. Wallace. S. lnterventional radiology intra-arterial therapy. J. RadioL 65:499,1984. 18. Wallace. S., Carrasco. C H.. Chamsangavej, C, Lee. Y. Y.. Wright. K. C. and Gianturco. C: Percutaneous transcatheter infusion and infarction in the treatment of human cancer. Part. Curr. Probl. Cancer 8:. 1984. 19. Wallace. S.. Chamsangavej. C. Carrasco. H., and Bechtel, W.: nfusion-embolization. Cancer 54:2751. 1984. 20. WalS. H.G.: ntroduction to resection of musculoskeletal sarcomas. Clin. Orthop. 153:31. 1980. 21. Winkler. K.. Beron, G., Katz. R.. Salzer-Kuntschik, M., Beck. L Beck. w.. Brandeis, W.. Ebell. W.. El1tman. R.. and Gobel. U.: Neoadjuvant chemotherapy for osteogenic sarcoma: Results of a cooperative German/Austrian study. J. Clin. Oncol. 2:617, 1984. Lir SAM Limb-spari treatment 4 survival, h. descrepanc of an expa: the probler lem of prr long-term ods of fix! permanent skeletal m either be implant or clinical sit n the matic irnj tients W' chernoth: very effe. mission.' peered to vival bey 80%." Recen surgical ( plication vage sur; logically patients. From tl Medical C Reprint Sinai Med 1188,5 E Receive \