Marginal Excision for Osteosarcoma With Caffeine Assisted Chemotherapy

The treatment of osteosarcoma has evolved remarkably during the past 2 decades, with corresponding dramatic improvements in prognosis and reconstructive surgery. The current standard treatment for osteosarcoma includes neoadjuvant chemotherapy, limb sparing wide surgical resection, and reconstruction.1,10,13,14,24 Recently, demand for minimization of surgical margins with the aim of preservation of better limb function has been on the increase while limb sparing surgery combined with wide tumor resection has come into widespread use. This is because minimal surgical intervention leading to good limb function is ideal for improving the quality of life of patients with osteosarcoma if minimal surgical intervention can be achieved without increasing the risk of local recurrence or jeopardizing the patient's life.

In 1992 the authors published a preliminary report on the possibility of marginal excision for osteosarcoma along with the use of caffeine assisted chemotherapy,23 because caffeine can enhance the cytocydal effects of anticancer drugs through its deoxyribonucleic acid repair inhibiting effect.22 Because caffeine assisted chemotherapy yielded a high rate of complete response, the authors subsequently investigated the effects of marginal excision combined with caffeine assisted chemotherapy on local recurrence, survival, and functional evaluation in osteosarcoma. Intentional marginal excision is different from marginal excision secondary to wide excision, because the former aims at the preservation of healthy tissue, leading to good limb function. This procedure can preserve important structures such as major vessels and nerves, muscles, epiphyses, tendons, and ligaments. The authors report here the results of intentional marginal excision in conjunction with caffeine assisted chemotherapy for osteosarcoma with the emphasis on indications for, and principles of, this procedure and on functional evaluation of the affected limbs. The procedure is contrasted with the usual treatment of osteosarcoma consisting of neoadjuvant chemotherapy and wide or radical resection.

Twenty-seven patients with osteosarcoma (22 with Stage IIB and give with Stage IIIB disease), 18 males and nine females with an average age of 20 years (range, 8-67 years), were included in this study. Tumors were located in the distal femur in 12 patients, the proximal tibia in eight, the proximal fibula in four, the proximal humerus in two, and the proximal femur in one patient.

Intraarterial cisplatin and caffeine with doxorubicin were administered three to five times preoperatively. Postoperative chemotherapy was performed on a neoadjuvant basis. Two chemotherapy protocols were used. In the K1 protocol (Fig 1), patients first were treated with three courses of intraarterial cisplatin (120 mg/m², 1-2 hours) and caffeine (1.5 g/m²/day for 3 days continuously) at 2-week intervals. If there was little or no response to the first course of intraarterial cisplatin and caffeine, doxorubicin (30 mg/m²/day for 2 days continuously) was added from the second course on. High dose methotrexate (8-12 g/m²) along with the citrovorum factor (15 mg 10 times) and vincristine (1.5 mg/m²) then were administered intravenously twice at 2-week intervals. Good responders received intravenously administered cisplatin and caffeine with or without doxorubicin and high dose methotrexate three times each as postoperative chemotherapy. The K2 protocol (Fig 1), which was developed to make limb sparing surgery safer, consisted of five courses of intraarterial cisplatin, caffeine, and doxorubicin at 3-week intervals as preoperative chemotherapy. For postoperative chemotherapy, intravenous cisplatin and caffeine with doxorubicin, and high dose methotrexate combined with citrovorum factor and vincristine were administered three times each. Nonresponders received a combination of ifosfamide, etoposide, and caffeine, and a combination of bleomycin, cyclophosphamide, actinomycin-D, and caffeine three times each. In this study, 15 patients were treated with the K1 protocol and 12 patients were treated with the K2 protocol.

Fig 1. Treatment schedule (K1 and K2 protocols).

The histologic response to the preoperative chemotherapy was evaluated with the grading system of Rosen et al11: Grade I, little or no effect of chemotherapy observed; Grade II, some areas of histologically viable tumor cells, but also areas of acellular tumor osteoid, necrotic, and/or fibrotic material; Grade III, predominant areas of change attributable to chemotherapy, with only scattered foci of viable tumor cells; and Grade IV, no evidence of viable tumor cells seen in extensive sampling (at least a full cross section of the tumor).

After the preoperative chemotherapy, and after a good response was observed with at least two of four radiologic examinations, namely, plain radiographs, angiogram, magnetic resonance (MR) imaging, and T1 201 scan, limb salvage surgery was conducted by means of the intentional marginal procedure. Evidence of good chemotherapeutic response consists of sclerotic changes or good margination of the tumor observed on plain radiographs, marked shrinkage of tumors extending into soft tissue on MR images, the disappearance of tumor stains on angiograms, or the disappearance of abnormal accumulation on T1 201 scintigrams. Wide excision was indicated when poor clinical response to the chemotherapy was detected on radiographic evaluations.

Important structures for good limb function were preserved intentionally by marginal tumor excision. These consisted of the rotator cuff, deltoideus, and short head of the biceps brachii muscle in case of a proximal humeral lesion; the gluteus medius tendon in case of a proximal femoral lesion; the collateral ligaments, quadriceps muscle, and epiphysis in case of distal femoral lesion; the collateral ligaments, patella ligament, and epiphysis in case of a proximal tibial lesion; and common peroneal nerve, lateral collateral ligament, and biceps femoris tendon in case of a proximal fibular lesion (Table 1). Major vessels and nerves around those areas also were preserved, but nonessential muscles surrounding the tumor were excised at less than 1 cm thickness. The margin was decided around structures beneficial for limb function, not around the entire area. For reconstruction, distraction osteogenesis was used to fill the skeletal defects when the epiphysis was preserved. After excision of the entire epiphysis and the tumor, a megaprosthesis was used or an osteochondral allograft, allograft and prosthesis composite, autoclaved bone and prosthesis composite, or autoclaved bone grafting were used when available.

TABLE 1. Structures Preserved by Intentional Marginal Procedure*

The function of the affected limb was assessed at the last followup with the patient by using the evaluation system of Enneking et al4 which consists of six factors each for upper and lower extremities: pain, function, emotional acceptance, hand positioning, dexterity, and lifting ability for upper extremities; and pain, function, emotional acceptance, supports, walking, and gait for lower extremities.

The Kaplan-Meier method was used for survival analysis.6 Event free survival was defined as the time from a patient's entry into this study until death with the disease in remission, local recurrence of disease after definitive surgery, development of distant metastases, development of a second malignancy, or last contact. Survival was defined as the time from study entry until death or last contact.

Intentional marginal excision was performed on 26 (21 with Stage IIB and five with Stage IIIB disease) patients with good response to the preoperative chemotherapy. These 26 patients received intravenous cisplatin and caffeine with or without doxorubicin and high dose methotrexate as postoperative chemotherapy. Only one patient with poor response to the preoperative chemotherapy and treated with wide excision received high dose methotrexate and ifosfamide combined with etoposide, methotrexate, and caffeine. High dose methotrexate was eliminated from the postoperative chemotherapy for a 67-year-old woman treated with the K2 protocol. Side effects induced by caffeine administration were found to be tolerable for most patients as a result of the concomitant use of a major tranquilizer. No symptomatic cardiac or neurologic side effects attributable to caffeine were observed.

The histologic responses of 22 patients with Stage IIB osteosarcomas to the preoperative chemotherapy were classified as Grade IV (no viable cells) in 19 patients, Grade III (only scattered foci of viable cells) in two, and Grade II (some areas of viable cells) in one patient. All five patients with Stage III osteosarcomas showed a Grade III response. The rate of complete response (Grade IV) was 70% (86% in the Stage IIB series), and the overall efficacy rate (Grades IV and III) was 96%. Local tumor recurrence was seen in only one patient with a Stage IIIB proximal fibular lesion. Eighteen patients are still disease free with a mean followup of 61 months. The mean followup of all patients was 48 months. One patient with Stage IIB osteosarcoma died of chemotherapy related side effects and a second patient died of unknown causes. The overall 5-year cumulative survival rate of 21 patients with Stage IIB osteosarcomas was 91%, and the 5-year event free survival rate was 79% (Fig 2).

Fig 2. Survival rates of 21 patients with good response to the preoperative chemotherapy. The overall 5-year survival rate was 91% and the event free 5-year survival rate was 79%.

Intentional marginal excision, which was performed on the 26 patients with a good chemotherapeutic response seen on radiographs, resulted in the preservation of the major nerves and vessels around the lesions in all cases. For lesions around the knee, the common peroneal nerve, lateral collateral ligament, and biceps femoris tendon were preserved in all four patients with a proximal fibular lesion (Fig 3); the patellar ligament was preserved in all eight patients with a proximal tibial lesion; and the epiphysis was preserved in eight of the 19 patients with a distal femoral or proximal tibial lesion. For reconstruction, epiphyseal preservation and distraction osteogenesis were performed on eight patients; allograft and prosthesis composite, autoclaved bone and prosthesis composite, autoclaved bone grafting, and osteochondral allograft on two patients each; megaprosthesis on six patients, and resection alone on four patients (Table 2). Reconstruction with distraction osteogenesis after epiphyseal preservation resulted in restoration of normal knee joint function (Fig 4). The average functional evaluations for various types of reconstruction were 86% for distraction osteogenesis, 93% for allograft and prosthesis composite, 85% for autoclaved bone and prosthesis composite, 75% for autoclaved bone grafting, 95% for osteochondral allograft, 100% for megaprosthesis, and 100% for resection alone (Table 2). The average functional evaluations were 75% for proximal humeral lesions, 94% for distal femoral lesions, 89% for proximal tibial lesions, and 100% for proximal fibular lesions. The mean range of motion obtained was 20° flexion, 20° abduction, and 30° extension for proximal humeral lesions; 120° flexion, 35° abduction, and 0° extension for a proximal femoral lesion; 127° ± 26° flexion and -3°± -6° extension for distal femoral lesions; 123°± 26° flexion and -5°± -8° extension for proximal tibial lesions; and a full range of the knee joint motion for proximal fibular lesions. The overall average functional evaluation of the 26 patients treated with intentional marginal procedure was 91% of normal (Table 3). Complications accompanying this limb sparing surgery consisted of deep infection and fracture in two patients each, and bone resorption and skin necrosis in one patient. No revision surgery was performed in this series.

Fig 3. Preservation of the common peroneal nerve in a 15-year-old boy with a conventional osteosarcoma in the proximal fibula resulting in peroneal nerve palsy. The peroneal nerve palsy was cured completely after effective preoperative K1 chemotherapy. The subsequent intentional marginal procedure resulted in preservation of the common peroneal nerve, lateral collateral ligament, and biceps femoris tendon. The collateral ligament and biceps femoris tendon were reattached to the tibia with a spike washer. The patient has had normal ankle and knee joint function and no evidence of disease for 70 months. (solid arrow: deep branch; open arrow: superficial branch; *:lateral collateral ligament and biceps femoris tendon; T:tumor)

TABLE 2. Reconstructions of 26 Patients Treated With Marginal Procedure

Fig 4. Preservation of epiphysis of the distal femur. (A) Shown here on the T2 weighted MR image is an osteosarcoma in the distal femur of a 9-year-old girl with no tumor extension into the epiphysis. (B) The epiphysis of the distal femur was preserved by intentional marginal excision. (C) The 7 cm bone loss after tumor resection was shortened and (D) The femur distracted by 9 cm at the diaphysis with an intramedullary nail. The affected limb was overextended by 2 cm to compensate for anticipated limb length discrepancy. (E) The patient has had normal knee joint function and no evidence of disease for 24 months.

TABLE 3. Functional Evaluation of 26 Patients Treated With Marginal Procedure*

The current standard treatment for osteosarcoma, which includes neoadjuvant chemotherapy and limb sparing wide surgical resection, has resulted in a 49% to 76% disease free survival during 5- to 8-year followups.10,13,24 The local tumor recurrence rate after limb sparing surgery with a wide procedure combined with chemotherapy ranges from 5.4% to 15.5%.2,3,24

Limb sparing surgery with wide tumor resection and reconstruction provides good limb function in the early stages. However, function tends to deteriorate with time, particularly more than 5 years after surgery, because of complications such as infection, fracture, implant failure, bone absorption, limb length discrepancy, and nerve sacrifice.17 To maintain good limb function for a long time and reduce the incidence of complications, intentional marginal excision for the purpose of preservation of healthy tissue and important structures for limb function was conducted for the present study. Because limb sparing surgery has become increasingly important in musculoskeletal tumor surgery, it now is essential to minimize surgical margins to improve the function of the affected limb and to provide optimal quality of life for the patients. Limb sparing surgery can be expected to result in superior function and appearance compared with amputation or Van Nes rotationplasty. Limb salvage for high grade osteosarcoma of the distal femur has been found to provide little measurable benefit in the quality of life of survivors during long term follow-up compared with amputation. Rotationplasty yields satisfactory limb function with the aid of artificial limbs.7,12,15

Comprehension of marginal or intralesional margins secondary to wide excision usually brings about poor prognosis for patients with osteosarcoma.19 With no adverse effects on survival or local recurrence, however, intentional marginal excision for high grade osteosarcoma was found to be advantageous in conjunction with caffeine assisted chemotherapy, showing a good response on radiologic evaluations and resulting in excellent limb function (91% of normal). Because a good chemotherapeutic response is an important favorable prognostic factor, much attention has been focused on preoperative prediction of response to marginal excision. When a good response is observed with at least two of four methods, namely, plain radiography, angiography, MR imaging, and T1 201 scan, the use of chemotherapy can be expected to result in more than 90% tumor necrosis. Recently, Tc-99m-methoxyisobutylisonitrile is being used to assess tumor response to chemotherapy.16

In this series, a high rate of local control (96% overall efficacy and 70% complete response rate) led to a low incidence of local recurrence (3.7%). A marginal excision was done only around structures important for limb function to reduce the risk of local recurrence as much as possible. If complete response can be detected preoperatively by means of radiologic examination, total marginal tumor excision may be feasible, but the present state of radiologic assessment does not allow for recurrence differentiation between complete and partial response.

Most reconstructions, such as prosthesis replacement, allograft implantation, and allograft and composite, provide excellent results at first, but complications occur with high frequency with time, leading to deterioration of function and reconstruction survival.5 A study indicated that prosthetic survival for large segment replacements for high grade bone sarcomas is 83% at 5 years and 67% at 10 years.8 For a proximal femoral lesion, 10-year survival is 76% for patients treated with a composite graft and 58% for those treated with a megaprosthesis.25 Similarly, allografts are considered to be successful for more than 20 years after implantation in 75% of recipients.9 Although wound necrosis is a significant cause of prosthetic removal and loss of limb,5 maximal preservation of healthy soft tissue by means of intentional marginal excision can help to solve this problem. Patients treated with a hinged endoprosthesis and who have excellent knee function have been found to have more quadriceps muscle mass.18 In addition to muscle mass, the intentional marginal excision described here can preserve structures such as ligaments, tendons, epiphyses, and nerves. This procedure is advantageous for yielding and maintaining excellent long term limb function in musculoskeletal tumor surgery. In particular, epiphyseal preservation and reconstruction using distraction osteogenesis can reproduce normal joints for selected cases.20 Another important advantage is that there is no need to worry about complications with this reconstruction once it has been achieved. Furthermore, distraction osteogenesis also yields a normal knee provided the thin articular surface is preserved at the proximal or distal tibia.21 In this series, epiphyseal preservation and joint reconstruction were achieved for eight of the 19 patients with a distal femoral or proximal tibial lesion showing little or no epiphyseal involvement. Additional efforts are needed to improve limb sparing surgery without jeopardizing patient survival.

Intentional marginal excision of osteosarcoma in conjunction with caffeine assisted chemotherapy was found to be a beneficial alternative to wide tumor excision and amputation. This approach should help to improve the success rate of limb sparing surgery for osteosarcoma, and to preserve better function of the affected limb without any adverse impact on survival or local recurrence.