Background: Negative surgical margins are uncommon for skull base, spinal/paraspinal sarcomas. As such, preoperative and/or postoperative radiation therapy is often recommended to improve local control (LC). Surgical resection achieving negative margins and delivery of adequate radiation therapy dose is challenging because of the neighboring delicate structures in these anatomical locations. Material and Methods: A comprehensive literature search was conducted using PubMed search engine. Relevant peer-reviewed articles written in English were the basis for the findings described in this chapter. Results: Surgery can be morbid for some patients in these locations and negative margins are hard to achieve. Radiation therapy helps with LC, and advances in technology have enabled the safer delivery of higher doses of radiation therapy. Conclusions: Radiation therapy appears to improve LC of chordomas and chondrosarcomas of the base of skull and spine. Prospective randomized data are scarce limiting the ability to make definitive treatment recommendations.
KeywordsChondrosarcomas, Chordomas, Management, Proton therapy, Radiation therapy, Surgery, Systemic therapy, Treatment
General Comments 363
General Considerations 364
Spine Chordomas and Chondrosarcomas 364
Sequencing of Radiation Therapy and Surgery 366
Illustrative Patient Treatment Plans 366
Further Reading 368
© 2018 Elsevier Inc. All rights reserved. Please note that the copyright for the original figures submitted by the contributors is owned by Contributors.
Chordomas and chondrosarcomas are quite uncommon, with an estimated annual incidence of ∼1320 chondrosarcomas and ∼300 chordomas in the United States. Accordingly, the clinical outcome data for different management strategies are quite modest and data from randomized clinical trials are virtually nonexistent.
A multidisciplinary team with expertise in managing these rare tumors is favored. The primary treatment strategy for chondrosarcoma and chordoma of the spine has been and remains surgery, provided tumor-negative margins are achieved, although recent evidence suggests a role for definitive radiation for selected patients who decline surgery in anatomic locations such as the upper sacrum where the morbidity associated with surgery is unacceptable to some patients. Obtaining pathologically clear margins for sarcomas that have extended into the extraosseous paravertebral tissue/structures is difficult in a high proportion of patients, even for experienced oncologic surgeons, provided reasonable levels of function are preserved. This is a particularly grievous problem when there has been extension into the vertebral canal with displacement of the dura and/or the cord. This latter situation inevitably means a positive or extraordinarily close margin and eradication of the sarcoma cells on the dura is not easy. Because of these issues with tumor eradication, the natural inclination is to combine surgery and radiation therapy to treat these patients.
The potential morbidity associated with surgery in the skull base, spine and paraspinal regions can be significant. En bloc resection resulting in negative margins is challenging because of the tumor proximity to vital and delicate structures such as spinal cord, nerve roots, brainstem, and cranial nerves. The ultimate surgical goal is to achieve the best possible eradication of the tumor while preserving neurological function. The same normal structures that pose difficulty for resection are also a problem for the radiation oncologist, because the dose limitations offered by the delicate structures in close proximity to the target region make it hard to deliver the high dose of radiation therapy necessary to control these tumors. Important technological advances in radiation oncology have enabled conformal delivery of high-dose radiation therapy, translating into decreased normal tissue toxicity and at the same time enabling local disease control.
For medically or technically inoperable patients, or patients who want to forgo surgery, high-dose definitive radiation therapy using modern technology achieves a useful tumor control probability (TCP), with decreasing efficacy in patients with tumors larger than 500 cc.
Type of resection and margin status are important prognostic factors for the local control (LC) of chordomas and chondrosarcomas. Unfortunately, the difficult and intricate anatomical territory makes it hard to achieve tumor-negative margins and the intraoperative and postoperative complication rates are as high as 46%.
En bloc resection is favored in the spine because this type of surgical resection leads to superior LC. Boriani et al. found that en bloc resection played an important role in decreasing local recurrence rates and also improved survival.
Margin status is thought to influence the risk of local recurrence; however, in the presence of radiation therapy, patients with contaminated margins may have similar rates of LC and survival as patients with negative margins. Schwab et al. focused on sacral chordomas treated with wide negative margins versus contaminated margins ; in the presence of radiation therapy, there was no difference in LC or survival rates for patients with sacral chordomas treated with wide versus contaminated margins. Delivering sufficient dose is essential to providing adequate LC—the dose of radiation therapy should be higher than 60 Gy; the effective dose for microscopically positive margins is generally 66–70 Gy and for gross residual disease >70 Gy.
Intralesional resections of spine chordomas are associated with inferior LC. Rotondo et al. reported rates of local and locoregional control for intralesional resection versus en bloc of 55% versus 72% ( P = .016) and 53% versus 70% ( P = .033), respectively. Ozaki et al. presented the outcome of a series of 26 chondrosarcomas managed by intralesional surgery. Out of these, 18 were of the axial skeleton and 8 were located in the extremities. The rate of relapse-free survival at 20 years was 7%, in contrast to 64% for patients whose margins were negative. Local regrowth was observed in 7 of 7 patients treated by curettage and in 14 of 19 patients following attempted radical resection. Radiation was employed in 4 of the 26 intralesional cases; the doses were 15, 30, 60, and 60 Gy. Of these, there was LC in one patient with sacral chondrosarcoma following excision and 60 Gy. Shives et al. described the outcome of 19 patients treated surgically at Mayo Clinic for chondrosarcoma of the spine; the procedure was described as intralesional or extralesional with contaminated margins in all patients. Five patients also received radiation, only one of whom received ≥50 Gy. No patient received chemotherapy. Local failure developed in 14 of these 19 patients. That is, inadequate surgery for low-grade chondrosarcomas was not inevitably followed by local regrowth. In fact, five patients are listed as NED at more than 60 months after treatment. Although, this was true in their experience, even for tumors categorized as chondrosarcoma grade 1, ideal management is negative margin resection, combined with radiation if wide or negative margins are not achieved.
Spine Chordomas and Chondrosarcomas
Treatment of mobile spine and sacral chordomas with surgery followed by low doses of radiation therapy (<60 Gy) results in suboptimal outcomes. For such low doses, the reported LC rates range from 0% to 50% for chordomas located in the mobile spine and sacrum.
Since the rate of local recurrence of spine chondrosarcomas removed by en bloc resections with tumor-free margins is as low as 8%, postoperative radiation therapy is usually omitted in such cases. However, because such R0 resections can be difficult to achieve in the spine, the rate of local recurrence for all spinal chondrosarcomas is higher. Fortunately, LC can be improved by adjuvant or neoadjuvant radiation therapy (high-dose photon/proton therapy ± intraoperative brachytherapy).
Goda et al. reported a crude LC rate of 90% for high-risk extracranial chondrosarcomas, of which 48% were pelvic/lower extremity and 17% were spine/sacral tumors, using preoperative RT (median, 50 Gy) and postoperative RT (median, 60 Gy) in 40% and 60% patients, respectively. Patients with R0, R1, and R2 resections had LC rates of 100%, 94%, and 42%, respectively.
Schoenfeld et al. treated 21 chondrosarcomas of the mobile spine (between 1984 and 2006) with surgery and, in all but one, radiation therapy, usually a combination of photons and protons to an average dose of 71 Gy (range: 53–83 Gy). Low tumor grade, tumor-free margins, and absence of metastasis were associated with superior overall survival (OS). En bloc resection was associated with significantly superior disease-specific survival ( P = .05). About 24% of patients experienced local recurrence and 43% developed metastatic disease. The rate of 5-year OS was 61%.
Bhatt et al. treated 44 patients with high doses of radiation therapy, using a combination of photons and protons (median dose: 70.2 GyRBE, range: 19.8–77.4 Gy). Intraoperative radiation therapy with electrons (10–12 Gy total dose) was used in 5 cases, and in 3 cases tumor-involved dura was treated with dural plaque brachytherapy. Rates of LC were 76% at 2 years and 57% at 5 years. The rates of OS were 90% and 68% at 2 and 5 years, respectively (CTOS Annual Meeting, Berlin, 2014).
Hug et al. treated 6 chondrosarcomas with high-dose proton/photon-beam radiation therapy (54–82 CGE with 1.8–2.0 CGE per fraction) alone or combined with surgery. The rates of 5-year OS and LC were both 100% ; for spine chordomas treated similarly, rates of 5-year OS and LC were only 53% and 50%, respectively. Holliday et al. also found better 5-year LC with proton therapy in chondrosarcomas than in chordoma, 77% versus 33%, respectively.
The ideal of en bloc resection with tumor-free margins is difficult to achieve for most spinal chordomas. Often, maintaining good neurological function requires leaving some tumor. In such cases, adjuvant high-dose radiation therapy improves the LC of these tumors. Schwab et al. treated 42 sacral chordomas with moderate-dose radiation therapy (5400–6480 cGy). The duration of median survival was 84 months and the rates of 5-year disease-free survival and disease-specific survival were 56% and 77%, respectively. Local recurrence occurred in 40% of patients. Rates of LC ( P = .5) and survival ( P = .9) were similar for resections having wide tumor-free (noncontaminated) margins and those with tumor-contaminated margins. Metastasis affected 31% of patients. Cryosurgery was used to treat five patients who had tumor-positive margins on frozen section analysis of a biopsy specimen.
Bjornsson et al. treated 40 chordomas between 1933 and 1990. Among these, 29 had incomplete surgery followed by radiation therapy. Death resulted from tumor in 25 patients (63%); the mean duration of survival was 56 months.
Hug et al. treated 14 spine tumors with high-dose proton/photon-beam radiation therapy (54–82 CGE and 1.8–2.0 CGE per fraction) alone or combined with surgery. Rates of 5-year OS and LC for chordomas were 53% and 50%, respectively, outcomes less favorable than those for chondrosarcomas.
The prospective Phase II study at Massachusetts General Hospital (MGH) of high-dose proton/photon treatment of 50 thoracic/lumbar/sacral spine tumors included 29 chordomas, 14 chondrosarcomas, and 7 other tumors. Twenty-five patients underwent gross total resection, 12 had subtotal resection, and 13 had only biopsy. Patients received 50.4 GyRBE for subclinical disease, 70.2 GyRBE for residual microscopic disease in the tumor bed, and 77.4 GyRBE for gross residual disease. Rates of LC at 8-years were 85% for primary tumors and 74% for the entire group. Some patients received a boost to the external dura mater surface via the dural plaque technique with 90 Y to doses ranging from 7.5 to 10 Gy. About 50% of the patients treated for a recurrent chordoma experienced another recurrence. This contrasts with the high LC rate achieved when treating primary chordomas, of which only 11% recurred locally ( P = .002). Long-term results of this phase II study with a median follow-up of 7.3 years confirmed the earlier high LC rates for patients with primary chordomas treated with high-dose radiation therapy and surgery.
Rutz et al. treated 26 chordomas (9 cervical, 2 thoracic, 8 lumbar, and 7 sacrococcygeal) with surgery (18/26) had gross total resection and scanning beam proton therapy to a median dose of 72 CGE. Rates of 3-year local progression-free, distant disease-free survival, and actuarial OS were 86%, 86%, and 84%, respectively, with a median follow-up of 35 months. Patients with implanted spine hardware were more prone to develop local failure ( P = .034).
Rotondo et al. treated 127 chordomas in 126 patients with surgery and high-dose proton-based radiation therapy. After a median follow-up of 41 months, the rates of 5-year OS, and local, locoregional, and distant controls were 81%, 62%, 60%, and 77%, respectively. The 5-year LC rate for primary chordoma was superior to that for recurrent tumors, 68% versus 49% ( P = .058), respectively. Local control was improved by preoperative RT (5 year) rate of 72% versus 54% ( P = .113). Among primary tumors, rates of 5 year LC and LRC were higher with preoperative RT, 85% ( P = .019) and 79% ( P = .034), respectively, than without preoperative RT, 56% and 56%, respectively. None of the 28 primary chordomas treated with preoperative RT followed by en bloc resection and postoperative proton boost radiation recurred.
MGH currently is accruing patients to a study of high-dose scanning beam protons ± surgery for spine and sacral chordomas. For patients undergoing surgery, a component of radiation is delivered preoperatively whenever possible. Because so few primary chordomas recur locally when preoperative radiation therapy is given, we have reduced the total radiation dose for patients who have undergone en bloc resection with tumor-free margins after a needle core biopsy (as opposed to an open biopsy, which has higher risk of tumor spillage) from a total dose of 70.2 GyRBE to 68.4 GyRBE and, more recently, to 64.8 GyRBE, without diminution of tumor control thus far ( www.clinicaltrials.gov ).
Side effects of high-dose proton/photon-based radiation therapy include temporary epilation, skin erythema, mucositis, and GI distress. DeLaney et al. reported sacral neuropathy in 3 of 25 patients treated with doses of 76.6–77.4 GyRBE (1.8 GyRBE per fraction), and no neurological complication for patients treated with doses ≤ 70.2 GyRBE (1.8 GyRBE per fraction).
Proton-based high-dose definitive radiation therapy to 77.4 Gy without surgery to 24 chordomas (most of which were upper sacral) yielded high LC (79.8% at 5 years). With a median follow-up of 56 months, rates of local progression-free survival at 3 and 5 years were 90.4% and 79.8%, respectively. Complications of such high-dose definitive radiation therapy include insufficiency fracture (8 in 24 patients), and neurologic deficits—foot drop (1), worsening of urinary and fecal incontinence (2), erectile dysfunction (1), and perineal numbness (1). These neurologic toxicities, however, must be viewed in the context of the nearly 100% risk of neurological injury with resection of chordomas involving S1–S3 sacral segments.
Carbon ions are an alternative particle therapy for unresected chordomas. In 95 unresected sacral chordomas receiving carbon ion treatment between 1996 and 2007, rates of 5 years LC and OS were 88% and 86%, respectively. Side effects included severe skin erosion requiring skin grafting in two patients and sciatic nerve irritation requiring medication in 15 patients.
Currently a randomized controlled clinical trial is comparing proton and carbon ion boost for patients with sacral chordomas ( clinicaltrials.gov/ct2/show/NCT01811394?term=carbon+ions&cntry1=EU%3ADE&rank=2 ).