- Email: [email protected]
Intraarterial Chemotherapy with Limb-sparing Resection of Large Soft-Tissue Sarcomas of the Extremities
lnte~entionalRadiology in Cancer
Intraarterial Chemotherapy with Limb-sparing Resection of Large Soft-Tissue Sarcomas of the Extremities1 - -
Michael C. Soulen, M D Jeffrey R. Weissmann, MD2 Kevin L. Sullivan, M D Richard D. Lackman, M D Marcelle J. Shapiro, M D Joseph Bonn, M D Arthur J. Weiss, M D Geoffrey A. Gardiner, Jr, M D
Index terms: Chemotherapeutic infusion, 48.32 Extremities, neoplasms, 48.32 Sarcoma, 48.32
JVIR 1992; 3:659-663
From the Departments of Radiology (M.C.S.,J.R.W., K.L.S.,M.J.S.,J.B., G.A.G.), Orthopedic Surgery (R.D.L.),and Medicine (A.J.W.),Thomas Jefferson University Hospital, Philadelphia. Received January 14, 1992; revision requested March 18; revision received April 27; accepted May 27. Address reprint requests to M.C.S., AngiographyIInterventional Radiology, Hospital of the University of Pennsylvania, 3400 Spruce St, Philadelphia, PA 19104. 2 Current address: Department of Radiology, Forrest General Hospital, Hattiesburg, Miss. 1
c
SCVIR, 1992
Fifteen patients with large (average, 15-cm),high-grade soft-tissue sarcomas of the extremities received prolonged selective intraarterial infusions of chemotherapeutic agents in an attempt to permit limb-sparingresection of these tumors, which would otherwise have required amputation. There were seven malignant fibrous histiocytomas, four liposarcomas, two fibrosarcomas, one leiomyosarcoma, and one rhabdomyosarcoma; 73%were grade 111. Seven patients underwent two catheterizations, for a total of 22 infusions, which averaged 11.3 days each. There were four catheterization-related complications, including catheter occlusion or dislodgment in one patient each and two cases of arterial thromboembolism in patients in whom anticoagulant dose was not adequate. Both of the latter patients required thrombectomy; one developed gangrene, which precluded limb-sparingsurgery. Thirteen of the 15 patients underwent limb-sparingresections, and two underwent amputations. No wound complications occurred. With a median follow-up of 36 months (mean, 34 months), life-table analysis indicates overall and diseasefree survivals of 72%and 59%,respectively, at 2 years and 64%and 59%at 3 years. In comparison to other reported therapies, this technique permits limb salvage in most patients without the high wound complication rate associated with preoperative radiation therapy, with equivalent local disease control and survival. L A R G E , high-grade soft-tissue sarcomas have a local recurrence rate of 50% or more and are associated with poor long-term survival when treated by means of excision alone (1). For this reason, amputation became the surgical therapy of choice for these aggressive tumors. Recently, adjuvant and neoadjuvant radiation therapy and chemotherapy have been employed in an attempt to permit limbsparing resection without sacrificing local and systemic control of the disease (2-10). Local control and survival rates equal or superior to those of historical controls have been achieved using short, single-drug courses of intraarterial chemotherapy followed by radiation therapy prior to limb-sparing resection; however, wound complication rates of up to 40% have occurred (7-10). We report our experience treating large,
high-grade soft-tissue sarcomas with prolonged, multidrug intraarterial chemotherapy and limb-sparing resection, without use of preoperative radiation.
PATIENTSAND METHODS Criteria for entry into the protocol were the presence of a soft-tissue sarcoma of an extremity of sufficiently large size or high grade as to require amputation (or such an extensive resection as to render the limb nonfunctional) in the absence of neoadjuvant therapy, absence of metastatic disease, and absence of any prior chemotherapy or radiation therapy. From 1985 to 1989,90 patients with soft-tissue sarcomas were evaluated a t our hospital, and 15 (16%)met these criteria. Eight were men. Ages
65
660
Journal of Vascular and Interventional Radiology
November 1992
Table 1 Tumor Characteristics Tumor Type Malignant fibrous histiocytoma (n = 7) Liposarcoma (n = 4) Fibrosarcoma (n
=
2)
Leiomyosarcoma (n = 1) Rhabdomyosarcoma (n = 1)
Grade 111
I (n = 2) I11 (n = 2) I1 (n = 1) 111 (n = 1) I1 I11
ranged from 20 to 76 years (mean, 40 years). Tumor characteristics are summarized in Table 1. Tumor grade and stage were determined prior to angiography by means of biopsy and computed tomographic (CT) scanning of the chest and affected extremity. Diagnostic angiography was performed from the common femoral artery approach. Lower extremity tumors were studied from the contralateral femoral artery to permit visualization and, if necessary, perfusion of the deep femoral artery. A 4or 5-F catheter was then placed as selectively as possible into the affected extremity without excluding potential feeding vessels and was secured with sutures. In one patient with tumor confined to the forearm, an ipsilateral brachial approach was used for the infusion after a high brachial bifurcation was excluded. In the later years of the study, heparinbonded catheters (Anthron; Toray Industries, Tokyo) were used whenever feasible. Heparin was infused intravenously to raise the activated partial thromboplastin time to 1.5-2 times the control value. Patients were restricted to bedrest. They were examined daily by members of the medical oncology service and the interventional radiology service and were monitored for signs or symptoms suggesting complications related to the catheter (thromboembolism, bleeding, catheter malposition) or the chemotherapy. Arteriography was performed for suspected cathe-
Size (cmi 9-25 10-18 17,20 8 17
Location Elbow (n = I ) , thigh (n = 2), knee (n = 21, calf (n = 2) Forearm (n = 11, thigh (n = 3) Upper arm, knee Calf Thigh
intravenous chemotherapy (n = 71, or both (n = 6) a t the discretion of the oncologist. One patient declined postoperative therapy. Physical examination was performed, and radiographs and CT scans of the extremity and the chest were obtained every 3 months for 2 years, every 4 months for the next year, then every 6 months thereafter. Complications of the protocol were reviewed. Overall and disease-free survival were assessed by means of life-table analysis.
ter-related problems, and the catheRESULTS ter was removed or repositioned as A total of 22 infusions were perneeded. The protocol called for continuous formed in the 15 patients, with an intraarterial infusions of doxorubicin average duration of catheterization of (Adriamycin; Adria Laboratories, 11.3 days (range, 5-17 days). Nine catheterizations of longer than 10 Dublin, Ohio) (8 mg/m2per day) and floxuridine (0.5 mg/m2 per day) addays resulted because complications ministered for 10 days. Cisplatin (120 required temporary interruption of mg/m2 over 4 hours) was given inthe chemotherapeutic infusion. traarterially on the 4th and the final Three infusions were stopped premaday of the infusion, unless contrainturely. One infusion was stopped afdicated because of renal insufficiency. ter 5 days because the patient was Endpoints for the infusion were sesuspected to have developed appendivere stomatitis, dermatitis, neutrope- citis; findings at an exploratory lapania, or thrombocytopenia. Seven parotomy were negative. Infusions were tients returned for a second, identical terminated after 8 days in two painfusion 1 month later. The other tients because of catheter-related eight patients did not receive a seccomplications. ond infusion because of advanced age, poor tolerance of the first course, Catheterization-related dramatic resolution of the tumor Complications Six patients had complications dimass after the first course, or patient rectly or indirectly related to the refusal. catheterization procedure: two artePatients were observed for several rial occlusions, two cases of deep vein days in the hospital after the infuthrombosis, and one case each of sions were completed until past their blood cell count nadir. Definitive sur- catheter dislodgment and occlusion requiring catheter replacement. Both gery was performed 3-6 weeks after completing the final infusion. Limbarterial occlusions occurred in patients who were not receiving therasparing resection was performed unpeutic doses of heparin, and both reless resection would have resulted in a nonfunctional extremity; in the lat- quired surgical intervention. In one of these patients, who had an anteter situation an amputation was performed. Specimens were sent for rou- grade catheter positioned in the sutine pathologic examination, which perficial femoral artery, the femoral, popliteal, and tibial arteries thromincluded assessment of margins and an estimate of the degree of tumor bosed after a n unusually prolonged necrosis. After 6-8 weeks of healing, 16-day catheterization. Despite patients received a postoperative thrombectomy and infrapopliteal bycourse of radiation therapy (n = l ) , pass, intractable gangrene developed
Soulen et a1
661
Volume 3 Number 4
Table 2 Cumulative Probability of Overall and Disease-free Survival Cic Survival Interval (mo) Overall Disease-free 0-6 100 93 + 24 73 t 50 6-12 87 + 32 67 ? 57 12-18 80 k 37 59 ? 63 18-24 72 ? 40 59 t 63 24-30 64 + 45 59 k 63 30-36 64 + 45 59 + 63 36-42 64 + 45 59 r 63 42-48 64 + 45 59 ? 63 48-54 64 k 45 59 ? 63 54-60 64 ? 45 Note.-Values are presented as mean ? standard error.
in the patient's foot. This precluded limb-sparing surgery, and the patient underwent above-the-knee amputation to treat both the gangrene and her tumor. The second patient had a retrograde catheter infusing the ipsilateral leg. The external iliac and common femoral arteries thrombosed along the course of the catheter on the 8th day, and these were managed successfully with thrombectomy alone. This was the only patient for whom the chemotherapeutic infusion was prematurely terminated due to a catheter-related complication. Both cases of deep vein thrombosis occurred in patients who were receiving therapeutic infusions of heparin. These patients had been catheterized for 10 and 16 days, respectively. One was treated with standard anticoagulation therapy. The other patient was given intravenous streptokinase for an iliofemoral vein thrombosis. This patient hemorrhaged into his tumor while receiving thrombolytic therapy. The one case of catheter dislodgment occurred in a patient who had a catheter placed retrogradely into the external iliac artery for an ipsilateral infusion, so that only the distal several centimeters were in the body. The patient experienced extravasation of doxorubicin and developed a small hematoma, but neither surgery nor transfusion were required. An additional nine arteriograms
were obtained during or at the end of infusions. Pericatheter thrombi. ranging in size from 1to 5 mm, &ere detected in two of six cases in which anticoagulation was therapeutic and in one of three cases in which it was not.
Chemotherapy-related Complications Toxic side-effects from the chemotherapy occurred in association with most of the 22 infusions and included myelosuppression in 19 (86961, nausea and vomiting in 18 (82%),and stomatitis in 16 (73%).More severe complications included neutropenic sepsis in four patients (18%),focal skin necrosis in four (18%),and severe leukopenia (white blood cell count I 1,000 x 106/L)or thrombocytopenia (platelet count I 60,000 x 109/L)in 10 (45%).One patient with thrombocytopenia developed hematemesis secondary to gastroduodenitis and required transfusion. No complications from chemotherapy delayed or precluded surgery. Treatment Outcome All patients underwent surgical resection of their tumor. Limb salvage was achieved in 13 patients (87%).Nine patients had a wide local resection, two had total knee replacement, and two had an allograft placed because of involvement of bone by their tumor. Amputations were performed on the patient with the gangrenous foot and on another patient with a malignant fibrous histiocytoma of the proximal arm that was still too large to resect after intraarterial chemotherapy. There were no major perioperative complications. There were no cases of delayed wound healing, wound infection, or skin necrosis. Pathologic examination demonstrated tumor-free margins in all cases. A good response to the chemotherapy, defined as necrosis of more than 60% of the tumor on gross pathologic examination, was achieved in 12 patients (80%).Eleven of these 12 patients were estimated to have 80%-100% tumor necrosis. Of the
two tumors without a good response, one was a 22-cm malignant fibrous histiocytoma, which had only about 10% necrosis. and the other was the rhabdomyosarcoma, which was approximately 40% necrotic. Both of these patients had completed two courses of intraarterial chemotherapy. At a mean follow-ur, of 34 months (median, 36 months), overall survival is 67%. Nine patients are alive and free of disease. One patient with a liposarcoma underwent resection of a local recurrence 11months after surgery, then died tumor-free of a myocardial infarction a t the age of 81 years, 45 months after her initial surgery. There were no other local recurrences. Six patients developed metastases 3-18 months after surgery; five of these have died. The surviving patient had a solitary liver metastasis, which was perfused for 10 days with intraarterial chemotherapy and then resected. The deaths included both patients with fibrosarcomas, the one patient with a rhabdomyosarcoma, and two patients with malignant fibrous histiocytomas. All of these patients' tumors had responded well to the intraarterial chemotherapy (estimated necrosis, 80%-loo%), except for the rhabdomyosarcoma. Four of these patients received postoperative chemotherapy, and the fifth received both chemotherapy and radiation therapy. Cumulative probability of overall and disease-free survival determined with life-table analysis is shown in Table 2 and the Figure. At 2 years, the overall survival is 72% and disease-free survival is 59%. Overall survival drops to 64% a t 3 years. One patient underwent amputation 4 years after tumor resection because of an infected allograft. He remains tumor-free.
DISCUSSION Preoperative (neoadjuvant) therapy for extremity sarcomas has been attempted in several nonrandomized trials (3,7-10). Theoretical advan-
662
Journal of Vascular and Interventional Radiology
November 1992
tages include perfusion and irradiation of a ~ r i m a r vtumor with an intact blooh supp& and oxygenation, factors critical for the delivery and effectiveness of the therapeutic agents, and also earlier therapy for micrometastases (2). The response of the tumor to preoperative therapy facilitates resection by shrinking the tumor or better defining planes between tumor and normal tissue and permitting limb salvage when it would not otherwise be possible. The preoperative response may also help to predict responsiveness to postoperative adjuvant therapy (3). Intravenous neoadjuvant chemotherapy with cyclophosphamide, adriamycin, dacarbazine, k vincristine was used a t the M.D. Anderson Cancer Center (Houston) for 46 patients with large (5-24 cm; mean, 10.6 cm), extremity soft-tissue sarcomas (3). Only 67% were subsequently able to undergo limb-sparing surgery, and 34% of these patients had a local recurrence. Intraarterial administration of chemotherapy adds the advantage of high drug concentrations delivered to the primary tumor, with the same total svstemic dose available for eradication of micrometastases. Intraarterial chemothera~vhas been used effectively for peripheral osteosarcomas, with limb-salvage rates of 87%89% and local recurrence rates of 4%-6% (11,121. Table 3 summarizes the major reports of use of intraarterial neoadjuvant chemotherapy to salvage limbs with large, high-grade soft-tissue sarcomas (7-10). In these series, a 3-day course of intraarterial doxorubicin followed by preoperative radiation in total doses ranging from 30 to 46 Gy (36 Gy was used in most cases) was used. The notable distinctions between this series and prior reports are that tumor size in our study was much larger (73% were 10 cm or larger, 33% were larger than 15 cm) and that no ~ r e o ~ e r a t i radiation ve therapy was used. Tumor size is a recognized risk factor adversely affecting long-term outcome (2,4),with tumors larger than 10 cm carrying a particularly bad prognosis. Only two
."
A
.
Figure. Cumulative probability of survival (A) and disease-free survival (0) determined with lifetable analysis. Numbers are the number % of patients at the end of each interval.
1 20 100 80 60 40
0~.l.l.l.l.l.,.l.l.l.~ 0 6 12 18 24 30 36 42 48 54 60 Months
of our patients had tumors smaller than 10 cm, so poorer results might be expected in our series. Contrary to this expectation, our limb-salvage rate, local recurrence rate, and overall and disease-free survival rates are similar to those achieved with the combination of preoperative radiation and single-drug intraarterial chemotherapy. Precise comparison of outcomes is difficult because of the nonuniformity of selection criteria and of adjuvant therapies among the series. The tumor mix varied among the series; however, histology is not a n important prognostic factor except for rhabdomyosarcoma and synovial sarcoma, which are automatically considered grade I11 tumors because of their aggressive behavior. Only the series reported by Wanebo et a1 (7) had a sizable minority of these sarcomas (20%).The limb-sparing resection rate was 88%-92% in the other series, with the highest rate including patients with radical compartment resections. We did not consider radical resection as limb salvage, since the limb is often not functional after this ~rocedure.The most striking difference in our series is the absence of wound complications, which occurred in 18%-40% of patients receiving preoperative radiation and led to secondary amputations in some patients (7,9). Catheter-related complications occurred more frequently in our series than with the shorter infusions re-
ported previously (40%vs 1%-11%), but most of these were early in our experience and were related to inadequate anticoagulation or to retrograde catheter placement for ipsilateral infusions; the latter practice was discontinued after the risk of catheter dislodgment was recognized. All catheterization-related complications occurred in the first 2 years of the protocol (1985 and 1986), with the exception of a single case of deep vein thrombosis that occurred d e s ~ i t e therapeutic anticoagulation. Meticulous daily assessment of the patient is necessary to screen for catheter-related problems. Malposition or migration of the catheter can lead to selective perfusion of a small branch artery. High local concentration of the chemotherapeutic agents causes an intense, painful focal myositis or dermatitis, which can progress to skin necrosis (8,111. Suggestive signs or symptoms are an indication for halting the infusion until arteriography can be performed to check the catheter position. If necessary, the catheter can be withdrawn and sutured again a t the groin. Advancing the catheter to correct a malposition should be avoided, since bacteremia can result from the colonized external portion of the catheter or the tract. We have encountered one case of sepsis following advancement of an indwelling catheter in a patient with a chondrosarcoma who was not included in this series. If advancing or
Soulen et a1
663
Volume 3 Number 4
Table 3 Summary of Trials of Neoadjuvant Intraarterial Chemotherapy Variable No. of patients Tumor grade Tumor size (cm) Preoperative chemotherapy Preoperative radiation therapy (Gy) No. of catheter-related complications Wound complications (%I Follow-up (mo) Overall survival (%) Disease-free survival (%I Local recurrence Note.-Cis = cisplatinum, Dox * Mean value. Median value.
=
Hoekstra e t a1 (8)
Goodnight e t a1 (9)
9 "High" ND
17
100
I11 for 76%
I11 for 75%
> 5 in 76%
Eilber e t a1 (10)
5-15
< 5 in 40%,
10-15 in 22% Dox for 3 days
Dox for 3 days
35
35-40
35
30-46
1 22 32* 66 55 1
0 29 32' 82 55 0
2 18 32' 76 ND 1
2 40 56' 58 47 0
floxuridine, ND
=
=
Current Report
60
I11 for 82%
Dox for 3 days
doxorubicin, FUDR
Dox for 3 days
Wanebo e t a1 (7)
15
I11 for 73% L 10 in 73%,
> 1 5 in 33% DoxIFUDR for 10 days and Cis on days 4 and 10 None 6 0 34*, 36' 67 60 1
no data available.
'
exchanging the catheter is necessary, a fresh puncture is recommended. The role of anticoagulation in these procedures is unclear. Both asymptomatic pericatheter thrombus formation and symptomatic deep vein thrombosis occurred in some patients despite therapeutic anticoagulation. The more devastating arterial thrombotic complications occurred in patients who did not receive adequate anticoagulation. Because arterial thrombosis can lead to loss of a limb, and the risk of heparin therapy is so low, we advocate full systemic anticoagulation with daily monitoring of the activated partial thromboplastin time in these patients. Use of heparin-bonded catheters mav be advantageous in these patient< although a decrease in ~ericatheterthrombus formation remains unproved. In summary, prolonged multidrug intraarterial neoadjuvant chemotherapy appears to permit limb-sparing resection of very large, high-grade soft-tissue sarcomas of the extremities, which would have been unresectable in the absence of neoadjuvant therapy. Local control and survival are similar to those obtained in previous trials employing neoadju-
vant radiation and regional chemotherapy. Withholding preoperative radiation therapy appears to eliminate the high rate of wound complications associated with previous trials. This benefit is partially negated by the higher toxicity and catheterization-related complication rate engendered by the prolonged infusion. These complications can be minimized with meticulous attention to patient care. Future research might examine the efficacy of a shorter multidrug infusion protocol to determine if similar tumor control could be obtained with fewer adverse effects. References 1. Yang JC, Rosenberg SA. Surgery for adult patients with soft tissue sarcomas. Semin Oncol 1989; 16:289-296. 2. Eilber FR, Huth JF, Mirra J, Rosen G. Progress in the recognition and treatment of soft tissue sarcomas. Cancer 1990; 65:660-666. 3. Pezzi CM, Pollock RE, Evans HL, et al. Preoperative chemotherapy for softtissue sarcomas of the extremities. Ann Surg 1990; 211:476-481. 4. Brennan MF. Management of extremity soft-tissue sarcomas. Am J Surg 1989; 158:71-78. 5. Elias AD, Antman KH. Adjuvant chemotherapy for soft tissue sarcomas: an approach in search of an effective regi-
men. Semin Oncol 1989; 16:305-311. 6. Tepper JE. Role of radiation therapy in the management of patients with bone and soft tissue sarcomas. Semin Oncol 1989; 16:281-288. 7. Wanebo HJ, Temple WJ, Popp MB, Douvill CE, Yablonski M. Combination regional therapy for extremity sarcoma: a tricenter study. Arch Surg 1990; 125:355-359. 8. Hoekstra HJ, Koops HS, Molenaar WM, et al. A combination of intraarterial chemotherapy, preoperative and postoperative radiotherapy, and surgery as limb-saving treatment of primarily unresectable high-grade soft tissue sarcomas of the extremities. Cancer 1989; 63:59-62. 9. Goodnight JE, Bargar WL, Voegeli T, Blaisdell FW. Limb-sparing surgery for extremity sarcomas after preoperative intraarterial doxorubicin and radiation therapy. Am J Surg 1985; 150: 109-113. 10. Eilber FR, Morton DL, Eckardt J, Grant T, Weisenburger T. Limb salvage for skeletal and soft tissue sarcomas: multidisciplinary preoperative therapy. Cancer 1984; 53:2579-2584. 11. Kashdan BJ, Sullivan KL, Lackman RD, et al. Extremity osteosarcomas: intraarterial chemotherapy and limbsparing resection with 2-year followup. Radiology 1990; 177:95-99. 12. Benjamin RS. Regional chemotherapy for osteosarcoma. Semin Oncol 1989; 16:323-327.
Intraarterial Chemotherapy with Limb-sparing Resection of Large Soft-Tissue Sarcomas of the Extremities1 - -
Michael C. Soulen, M D Jeffrey R. Weissmann, MD2 Kevin L. Sullivan, M D Richard D. Lackman, M D Marcelle J. Shapiro, M D Joseph Bonn, M D Arthur J. Weiss, M D Geoffrey A. Gardiner, Jr, M D
Index terms: Chemotherapeutic infusion, 48.32 Extremities, neoplasms, 48.32 Sarcoma, 48.32
JVIR 1992; 3:659-663
From the Departments of Radiology (M.C.S.,J.R.W., K.L.S.,M.J.S.,J.B., G.A.G.), Orthopedic Surgery (R.D.L.),and Medicine (A.J.W.),Thomas Jefferson University Hospital, Philadelphia. Received January 14, 1992; revision requested March 18; revision received April 27; accepted May 27. Address reprint requests to M.C.S., AngiographyIInterventional Radiology, Hospital of the University of Pennsylvania, 3400 Spruce St, Philadelphia, PA 19104. 2 Current address: Department of Radiology, Forrest General Hospital, Hattiesburg, Miss. 1
c
SCVIR, 1992
Fifteen patients with large (average, 15-cm),high-grade soft-tissue sarcomas of the extremities received prolonged selective intraarterial infusions of chemotherapeutic agents in an attempt to permit limb-sparingresection of these tumors, which would otherwise have required amputation. There were seven malignant fibrous histiocytomas, four liposarcomas, two fibrosarcomas, one leiomyosarcoma, and one rhabdomyosarcoma; 73%were grade 111. Seven patients underwent two catheterizations, for a total of 22 infusions, which averaged 11.3 days each. There were four catheterization-related complications, including catheter occlusion or dislodgment in one patient each and two cases of arterial thromboembolism in patients in whom anticoagulant dose was not adequate. Both of the latter patients required thrombectomy; one developed gangrene, which precluded limb-sparingsurgery. Thirteen of the 15 patients underwent limb-sparingresections, and two underwent amputations. No wound complications occurred. With a median follow-up of 36 months (mean, 34 months), life-table analysis indicates overall and diseasefree survivals of 72%and 59%,respectively, at 2 years and 64%and 59%at 3 years. In comparison to other reported therapies, this technique permits limb salvage in most patients without the high wound complication rate associated with preoperative radiation therapy, with equivalent local disease control and survival. L A R G E , high-grade soft-tissue sarcomas have a local recurrence rate of 50% or more and are associated with poor long-term survival when treated by means of excision alone (1). For this reason, amputation became the surgical therapy of choice for these aggressive tumors. Recently, adjuvant and neoadjuvant radiation therapy and chemotherapy have been employed in an attempt to permit limbsparing resection without sacrificing local and systemic control of the disease (2-10). Local control and survival rates equal or superior to those of historical controls have been achieved using short, single-drug courses of intraarterial chemotherapy followed by radiation therapy prior to limb-sparing resection; however, wound complication rates of up to 40% have occurred (7-10). We report our experience treating large,
high-grade soft-tissue sarcomas with prolonged, multidrug intraarterial chemotherapy and limb-sparing resection, without use of preoperative radiation.
PATIENTSAND METHODS Criteria for entry into the protocol were the presence of a soft-tissue sarcoma of an extremity of sufficiently large size or high grade as to require amputation (or such an extensive resection as to render the limb nonfunctional) in the absence of neoadjuvant therapy, absence of metastatic disease, and absence of any prior chemotherapy or radiation therapy. From 1985 to 1989,90 patients with soft-tissue sarcomas were evaluated a t our hospital, and 15 (16%)met these criteria. Eight were men. Ages
65
660
Journal of Vascular and Interventional Radiology
November 1992
Table 1 Tumor Characteristics Tumor Type Malignant fibrous histiocytoma (n = 7) Liposarcoma (n = 4) Fibrosarcoma (n
=
2)
Leiomyosarcoma (n = 1) Rhabdomyosarcoma (n = 1)
Grade 111
I (n = 2) I11 (n = 2) I1 (n = 1) 111 (n = 1) I1 I11
ranged from 20 to 76 years (mean, 40 years). Tumor characteristics are summarized in Table 1. Tumor grade and stage were determined prior to angiography by means of biopsy and computed tomographic (CT) scanning of the chest and affected extremity. Diagnostic angiography was performed from the common femoral artery approach. Lower extremity tumors were studied from the contralateral femoral artery to permit visualization and, if necessary, perfusion of the deep femoral artery. A 4or 5-F catheter was then placed as selectively as possible into the affected extremity without excluding potential feeding vessels and was secured with sutures. In one patient with tumor confined to the forearm, an ipsilateral brachial approach was used for the infusion after a high brachial bifurcation was excluded. In the later years of the study, heparinbonded catheters (Anthron; Toray Industries, Tokyo) were used whenever feasible. Heparin was infused intravenously to raise the activated partial thromboplastin time to 1.5-2 times the control value. Patients were restricted to bedrest. They were examined daily by members of the medical oncology service and the interventional radiology service and were monitored for signs or symptoms suggesting complications related to the catheter (thromboembolism, bleeding, catheter malposition) or the chemotherapy. Arteriography was performed for suspected cathe-
Size (cmi 9-25 10-18 17,20 8 17
Location Elbow (n = I ) , thigh (n = 2), knee (n = 21, calf (n = 2) Forearm (n = 11, thigh (n = 3) Upper arm, knee Calf Thigh
intravenous chemotherapy (n = 71, or both (n = 6) a t the discretion of the oncologist. One patient declined postoperative therapy. Physical examination was performed, and radiographs and CT scans of the extremity and the chest were obtained every 3 months for 2 years, every 4 months for the next year, then every 6 months thereafter. Complications of the protocol were reviewed. Overall and disease-free survival were assessed by means of life-table analysis.
ter-related problems, and the catheRESULTS ter was removed or repositioned as A total of 22 infusions were perneeded. The protocol called for continuous formed in the 15 patients, with an intraarterial infusions of doxorubicin average duration of catheterization of (Adriamycin; Adria Laboratories, 11.3 days (range, 5-17 days). Nine catheterizations of longer than 10 Dublin, Ohio) (8 mg/m2per day) and floxuridine (0.5 mg/m2 per day) addays resulted because complications ministered for 10 days. Cisplatin (120 required temporary interruption of mg/m2 over 4 hours) was given inthe chemotherapeutic infusion. traarterially on the 4th and the final Three infusions were stopped premaday of the infusion, unless contrainturely. One infusion was stopped afdicated because of renal insufficiency. ter 5 days because the patient was Endpoints for the infusion were sesuspected to have developed appendivere stomatitis, dermatitis, neutrope- citis; findings at an exploratory lapania, or thrombocytopenia. Seven parotomy were negative. Infusions were tients returned for a second, identical terminated after 8 days in two painfusion 1 month later. The other tients because of catheter-related eight patients did not receive a seccomplications. ond infusion because of advanced age, poor tolerance of the first course, Catheterization-related dramatic resolution of the tumor Complications Six patients had complications dimass after the first course, or patient rectly or indirectly related to the refusal. catheterization procedure: two artePatients were observed for several rial occlusions, two cases of deep vein days in the hospital after the infuthrombosis, and one case each of sions were completed until past their blood cell count nadir. Definitive sur- catheter dislodgment and occlusion requiring catheter replacement. Both gery was performed 3-6 weeks after completing the final infusion. Limbarterial occlusions occurred in patients who were not receiving therasparing resection was performed unpeutic doses of heparin, and both reless resection would have resulted in a nonfunctional extremity; in the lat- quired surgical intervention. In one of these patients, who had an anteter situation an amputation was performed. Specimens were sent for rou- grade catheter positioned in the sutine pathologic examination, which perficial femoral artery, the femoral, popliteal, and tibial arteries thromincluded assessment of margins and an estimate of the degree of tumor bosed after a n unusually prolonged necrosis. After 6-8 weeks of healing, 16-day catheterization. Despite patients received a postoperative thrombectomy and infrapopliteal bycourse of radiation therapy (n = l ) , pass, intractable gangrene developed
Soulen et a1
661
Volume 3 Number 4
Table 2 Cumulative Probability of Overall and Disease-free Survival Cic Survival Interval (mo) Overall Disease-free 0-6 100 93 + 24 73 t 50 6-12 87 + 32 67 ? 57 12-18 80 k 37 59 ? 63 18-24 72 ? 40 59 t 63 24-30 64 + 45 59 k 63 30-36 64 + 45 59 + 63 36-42 64 + 45 59 r 63 42-48 64 + 45 59 ? 63 48-54 64 k 45 59 ? 63 54-60 64 ? 45 Note.-Values are presented as mean ? standard error.
in the patient's foot. This precluded limb-sparing surgery, and the patient underwent above-the-knee amputation to treat both the gangrene and her tumor. The second patient had a retrograde catheter infusing the ipsilateral leg. The external iliac and common femoral arteries thrombosed along the course of the catheter on the 8th day, and these were managed successfully with thrombectomy alone. This was the only patient for whom the chemotherapeutic infusion was prematurely terminated due to a catheter-related complication. Both cases of deep vein thrombosis occurred in patients who were receiving therapeutic infusions of heparin. These patients had been catheterized for 10 and 16 days, respectively. One was treated with standard anticoagulation therapy. The other patient was given intravenous streptokinase for an iliofemoral vein thrombosis. This patient hemorrhaged into his tumor while receiving thrombolytic therapy. The one case of catheter dislodgment occurred in a patient who had a catheter placed retrogradely into the external iliac artery for an ipsilateral infusion, so that only the distal several centimeters were in the body. The patient experienced extravasation of doxorubicin and developed a small hematoma, but neither surgery nor transfusion were required. An additional nine arteriograms
were obtained during or at the end of infusions. Pericatheter thrombi. ranging in size from 1to 5 mm, &ere detected in two of six cases in which anticoagulation was therapeutic and in one of three cases in which it was not.
Chemotherapy-related Complications Toxic side-effects from the chemotherapy occurred in association with most of the 22 infusions and included myelosuppression in 19 (86961, nausea and vomiting in 18 (82%),and stomatitis in 16 (73%).More severe complications included neutropenic sepsis in four patients (18%),focal skin necrosis in four (18%),and severe leukopenia (white blood cell count I 1,000 x 106/L)or thrombocytopenia (platelet count I 60,000 x 109/L)in 10 (45%).One patient with thrombocytopenia developed hematemesis secondary to gastroduodenitis and required transfusion. No complications from chemotherapy delayed or precluded surgery. Treatment Outcome All patients underwent surgical resection of their tumor. Limb salvage was achieved in 13 patients (87%).Nine patients had a wide local resection, two had total knee replacement, and two had an allograft placed because of involvement of bone by their tumor. Amputations were performed on the patient with the gangrenous foot and on another patient with a malignant fibrous histiocytoma of the proximal arm that was still too large to resect after intraarterial chemotherapy. There were no major perioperative complications. There were no cases of delayed wound healing, wound infection, or skin necrosis. Pathologic examination demonstrated tumor-free margins in all cases. A good response to the chemotherapy, defined as necrosis of more than 60% of the tumor on gross pathologic examination, was achieved in 12 patients (80%).Eleven of these 12 patients were estimated to have 80%-100% tumor necrosis. Of the
two tumors without a good response, one was a 22-cm malignant fibrous histiocytoma, which had only about 10% necrosis. and the other was the rhabdomyosarcoma, which was approximately 40% necrotic. Both of these patients had completed two courses of intraarterial chemotherapy. At a mean follow-ur, of 34 months (median, 36 months), overall survival is 67%. Nine patients are alive and free of disease. One patient with a liposarcoma underwent resection of a local recurrence 11months after surgery, then died tumor-free of a myocardial infarction a t the age of 81 years, 45 months after her initial surgery. There were no other local recurrences. Six patients developed metastases 3-18 months after surgery; five of these have died. The surviving patient had a solitary liver metastasis, which was perfused for 10 days with intraarterial chemotherapy and then resected. The deaths included both patients with fibrosarcomas, the one patient with a rhabdomyosarcoma, and two patients with malignant fibrous histiocytomas. All of these patients' tumors had responded well to the intraarterial chemotherapy (estimated necrosis, 80%-loo%), except for the rhabdomyosarcoma. Four of these patients received postoperative chemotherapy, and the fifth received both chemotherapy and radiation therapy. Cumulative probability of overall and disease-free survival determined with life-table analysis is shown in Table 2 and the Figure. At 2 years, the overall survival is 72% and disease-free survival is 59%. Overall survival drops to 64% a t 3 years. One patient underwent amputation 4 years after tumor resection because of an infected allograft. He remains tumor-free.
DISCUSSION Preoperative (neoadjuvant) therapy for extremity sarcomas has been attempted in several nonrandomized trials (3,7-10). Theoretical advan-
662
Journal of Vascular and Interventional Radiology
November 1992
tages include perfusion and irradiation of a ~ r i m a r vtumor with an intact blooh supp& and oxygenation, factors critical for the delivery and effectiveness of the therapeutic agents, and also earlier therapy for micrometastases (2). The response of the tumor to preoperative therapy facilitates resection by shrinking the tumor or better defining planes between tumor and normal tissue and permitting limb salvage when it would not otherwise be possible. The preoperative response may also help to predict responsiveness to postoperative adjuvant therapy (3). Intravenous neoadjuvant chemotherapy with cyclophosphamide, adriamycin, dacarbazine, k vincristine was used a t the M.D. Anderson Cancer Center (Houston) for 46 patients with large (5-24 cm; mean, 10.6 cm), extremity soft-tissue sarcomas (3). Only 67% were subsequently able to undergo limb-sparing surgery, and 34% of these patients had a local recurrence. Intraarterial administration of chemotherapy adds the advantage of high drug concentrations delivered to the primary tumor, with the same total svstemic dose available for eradication of micrometastases. Intraarterial chemothera~vhas been used effectively for peripheral osteosarcomas, with limb-salvage rates of 87%89% and local recurrence rates of 4%-6% (11,121. Table 3 summarizes the major reports of use of intraarterial neoadjuvant chemotherapy to salvage limbs with large, high-grade soft-tissue sarcomas (7-10). In these series, a 3-day course of intraarterial doxorubicin followed by preoperative radiation in total doses ranging from 30 to 46 Gy (36 Gy was used in most cases) was used. The notable distinctions between this series and prior reports are that tumor size in our study was much larger (73% were 10 cm or larger, 33% were larger than 15 cm) and that no ~ r e o ~ e r a t i radiation ve therapy was used. Tumor size is a recognized risk factor adversely affecting long-term outcome (2,4),with tumors larger than 10 cm carrying a particularly bad prognosis. Only two
."
A
.
Figure. Cumulative probability of survival (A) and disease-free survival (0) determined with lifetable analysis. Numbers are the number % of patients at the end of each interval.
1 20 100 80 60 40
0~.l.l.l.l.l.,.l.l.l.~ 0 6 12 18 24 30 36 42 48 54 60 Months
of our patients had tumors smaller than 10 cm, so poorer results might be expected in our series. Contrary to this expectation, our limb-salvage rate, local recurrence rate, and overall and disease-free survival rates are similar to those achieved with the combination of preoperative radiation and single-drug intraarterial chemotherapy. Precise comparison of outcomes is difficult because of the nonuniformity of selection criteria and of adjuvant therapies among the series. The tumor mix varied among the series; however, histology is not a n important prognostic factor except for rhabdomyosarcoma and synovial sarcoma, which are automatically considered grade I11 tumors because of their aggressive behavior. Only the series reported by Wanebo et a1 (7) had a sizable minority of these sarcomas (20%).The limb-sparing resection rate was 88%-92% in the other series, with the highest rate including patients with radical compartment resections. We did not consider radical resection as limb salvage, since the limb is often not functional after this ~rocedure.The most striking difference in our series is the absence of wound complications, which occurred in 18%-40% of patients receiving preoperative radiation and led to secondary amputations in some patients (7,9). Catheter-related complications occurred more frequently in our series than with the shorter infusions re-
ported previously (40%vs 1%-11%), but most of these were early in our experience and were related to inadequate anticoagulation or to retrograde catheter placement for ipsilateral infusions; the latter practice was discontinued after the risk of catheter dislodgment was recognized. All catheterization-related complications occurred in the first 2 years of the protocol (1985 and 1986), with the exception of a single case of deep vein thrombosis that occurred d e s ~ i t e therapeutic anticoagulation. Meticulous daily assessment of the patient is necessary to screen for catheter-related problems. Malposition or migration of the catheter can lead to selective perfusion of a small branch artery. High local concentration of the chemotherapeutic agents causes an intense, painful focal myositis or dermatitis, which can progress to skin necrosis (8,111. Suggestive signs or symptoms are an indication for halting the infusion until arteriography can be performed to check the catheter position. If necessary, the catheter can be withdrawn and sutured again a t the groin. Advancing the catheter to correct a malposition should be avoided, since bacteremia can result from the colonized external portion of the catheter or the tract. We have encountered one case of sepsis following advancement of an indwelling catheter in a patient with a chondrosarcoma who was not included in this series. If advancing or
Soulen et a1
663
Volume 3 Number 4
Table 3 Summary of Trials of Neoadjuvant Intraarterial Chemotherapy Variable No. of patients Tumor grade Tumor size (cm) Preoperative chemotherapy Preoperative radiation therapy (Gy) No. of catheter-related complications Wound complications (%I Follow-up (mo) Overall survival (%) Disease-free survival (%I Local recurrence Note.-Cis = cisplatinum, Dox * Mean value. Median value.
=
Hoekstra e t a1 (8)
Goodnight e t a1 (9)
9 "High" ND
17
100
I11 for 76%
I11 for 75%
> 5 in 76%
Eilber e t a1 (10)
5-15
< 5 in 40%,
10-15 in 22% Dox for 3 days
Dox for 3 days
35
35-40
35
30-46
1 22 32* 66 55 1
0 29 32' 82 55 0
2 18 32' 76 ND 1
2 40 56' 58 47 0
floxuridine, ND
=
=
Current Report
60
I11 for 82%
Dox for 3 days
doxorubicin, FUDR
Dox for 3 days
Wanebo e t a1 (7)
15
I11 for 73% L 10 in 73%,
> 1 5 in 33% DoxIFUDR for 10 days and Cis on days 4 and 10 None 6 0 34*, 36' 67 60 1
no data available.
'
exchanging the catheter is necessary, a fresh puncture is recommended. The role of anticoagulation in these procedures is unclear. Both asymptomatic pericatheter thrombus formation and symptomatic deep vein thrombosis occurred in some patients despite therapeutic anticoagulation. The more devastating arterial thrombotic complications occurred in patients who did not receive adequate anticoagulation. Because arterial thrombosis can lead to loss of a limb, and the risk of heparin therapy is so low, we advocate full systemic anticoagulation with daily monitoring of the activated partial thromboplastin time in these patients. Use of heparin-bonded catheters mav be advantageous in these patient< although a decrease in ~ericatheterthrombus formation remains unproved. In summary, prolonged multidrug intraarterial neoadjuvant chemotherapy appears to permit limb-sparing resection of very large, high-grade soft-tissue sarcomas of the extremities, which would have been unresectable in the absence of neoadjuvant therapy. Local control and survival are similar to those obtained in previous trials employing neoadju-
vant radiation and regional chemotherapy. Withholding preoperative radiation therapy appears to eliminate the high rate of wound complications associated with previous trials. This benefit is partially negated by the higher toxicity and catheterization-related complication rate engendered by the prolonged infusion. These complications can be minimized with meticulous attention to patient care. Future research might examine the efficacy of a shorter multidrug infusion protocol to determine if similar tumor control could be obtained with fewer adverse effects. References 1. Yang JC, Rosenberg SA. Surgery for adult patients with soft tissue sarcomas. Semin Oncol 1989; 16:289-296. 2. Eilber FR, Huth JF, Mirra J, Rosen G. Progress in the recognition and treatment of soft tissue sarcomas. Cancer 1990; 65:660-666. 3. Pezzi CM, Pollock RE, Evans HL, et al. Preoperative chemotherapy for softtissue sarcomas of the extremities. Ann Surg 1990; 211:476-481. 4. Brennan MF. Management of extremity soft-tissue sarcomas. Am J Surg 1989; 158:71-78. 5. Elias AD, Antman KH. Adjuvant chemotherapy for soft tissue sarcomas: an approach in search of an effective regi-
men. Semin Oncol 1989; 16:305-311. 6. Tepper JE. Role of radiation therapy in the management of patients with bone and soft tissue sarcomas. Semin Oncol 1989; 16:281-288. 7. Wanebo HJ, Temple WJ, Popp MB, Douvill CE, Yablonski M. Combination regional therapy for extremity sarcoma: a tricenter study. Arch Surg 1990; 125:355-359. 8. Hoekstra HJ, Koops HS, Molenaar WM, et al. A combination of intraarterial chemotherapy, preoperative and postoperative radiotherapy, and surgery as limb-saving treatment of primarily unresectable high-grade soft tissue sarcomas of the extremities. Cancer 1989; 63:59-62. 9. Goodnight JE, Bargar WL, Voegeli T, Blaisdell FW. Limb-sparing surgery for extremity sarcomas after preoperative intraarterial doxorubicin and radiation therapy. Am J Surg 1985; 150: 109-113. 10. Eilber FR, Morton DL, Eckardt J, Grant T, Weisenburger T. Limb salvage for skeletal and soft tissue sarcomas: multidisciplinary preoperative therapy. Cancer 1984; 53:2579-2584. 11. Kashdan BJ, Sullivan KL, Lackman RD, et al. Extremity osteosarcomas: intraarterial chemotherapy and limbsparing resection with 2-year followup. Radiology 1990; 177:95-99. 12. Benjamin RS. Regional chemotherapy for osteosarcoma. Semin Oncol 1989; 16:323-327.