Stereotactic radiosurgery for spinal metastases: a literature review

ABSTRACT Objective: The spine is the most common location for bone metastases. Since cure is not possible, local control and relief of symptoms is the basis for treatment, which is grounded on the use of conventional radiotherapy. Recently, spinal radiosurgery has been proposed for the local control of spinal metastases, whether as primary or salvage treatment. Consequently, we carried out a literature review in order to analyze the indications, efficacy, and safety of radiosurgery in the treatment of spinal metastases. Methods: We have reviewed the literature using the PubMed gateway with data from the MEDLINE library on studies related to the use of radiosurgery in treatment of bone metastases in spine. The studies were reviewed by all the authors and classified as to level of evidence, using the criterion defined by Wright. Results: The indications found for radiosurgery were primary control of epidural metastases (evidence level II), myeloma (level III), and metastases known to be poor responders to conventional radiotherapy – melanoma and renal cell carcinoma (level III). Spinal radiosurgery was also proposed for salvage treatment after conventional radiotherapy (level II). There is also some evidence as to the safety and efficacy of radiosurgery in cases of extramedullar and intramedullar intradural metastatic tumors (level III) and after spinal decompression and stabilization surgery. Conclusion: Radiosurgery can be used in primary or salvage treatment of spinal metastases, improving local disease control and patient symptoms. It should also be considered as initial treatment for radioresistant tumors, such as melanoma and renal cell carcinoma.


INTRODUCTION
The spine is the most common site for bone metastases (1) . Spinal involvement occurs in up to 40% of patients with cancer during progression of the disease, with 5 to 10% of these patients developing epidural compression at some point of their progression (2) . More than 90% of the spinal metastases are located extradurally, whereas about 5% are intradural and less than 1% are intramedullar (3) . Symptomatic compression occurs more frequently in the thoracic spine (50 to 70%), followed by the lumbar spine (20 to 30%) and the cervical spine (10 to 30%) (4) . This probably occurs because the thoracic canal has the smallest diameter and the largest number of vertebrae in the spine. The posterior half of the vertebral body is the most commonly involved (4) , spreading latterly to other regions (anterior body, lamina, and pedicles). Also important is the fact that by the time of diagnosis, multiple lesions can be found at noncontiguous levels in up to 40% of the patients. Regarding histology, up to 50% of the metastases come from one of the three following cancers: breast, lung, or prostate (1) .
Since spinal metastases suggest a non-controlled tumor, treatment is mainly based on palliation and local disease control. Although in selected cases chemotherapy can be used, radiotherapy, with or without concomitant surgery, remains the main treatment modality (5,6) . Surgical techniques have progressed substantially in the last years, allowing circumferential decompression of the spinal cord and spinal fixation for early stabilization (6) . The main goals of surgery are spinal stabilization, restoration, and maintenance of the global alignment and decompression of neural structures, especially in radioresistant tumors (5) . Surgery can also be an option in unknown primary tumors (when a tissue sample is necessary) or in cases where local pain persists after other treatment modalities have been used. However, the surgical option should only be considered for patients with good performance scores, since it may entail serious morbidity and possible medical complications. Prognostic scoring systems, like the one proposed by Tokuhashi et al. and Tomita et al., are useful in surgical decision-making (7,8) .

Role of radiation therapy in spinal metastases
Since surgery is only considered in selected cases, conventional external beam radiotherapy (CER) remains the mainstay treatment modality for spinal metastases. The total dose of radiation ranges from 25 to 40Gy, fractionated daily into 8 to 20 doses (6) . A broad margin is used in the radiation field, typically one or two vertebral segments above and two segments below the affected level in order to compensate for internal vertebral motion. This technique exposes healthy tissue to radiation, including the sensitive spinal cord. The dose of radiation is fractioned to allow recovery of the normal tissue, thus improving tolerance (6) .
Spinal stereotactic radiosurgery (SRS) is a more recently developed type of radiation therapy that delivers a high dose of radiation to the tumor, whilst minimizing the amount delivered to the healthy neighboring tissues. The target is defined by high-resolution stereotactic imaging, and requires rigid spine immobilization to offer a precise conformal dose, using frame or frameless techniques, based on external immobilization devices with fiducial markers. Treatment is fractionated into fewer sessions (one to five), with total doses ranging from 8 to 30Gy. When one session is used, the term "radiosurgery" must be preferentially used, whereas "fractionated stereotactic radiosurgery" is better used when multiple sessions are performed. Early versions of SRS in the spine required surgical fixation of frames in the spinous processes to avoid movement artifacts. The novel systems are frameless, based on internal skeletal anatomy, and implanted fiducial points, or infrared surface markers, with near real-time images for motion correction. Radiation intensity modulation increases the conformity of radiation to the tumor, minimizing the dose to normal tissue (6) . However, the effectiveness, safety, precise indications, and long-term evaluation regarding disease control of the effects of SRS remain unclear.

OBJECTIVE
To perform a literature review of the indications, safety, and efficacy of spinal radiosurgery when compared to conventional external radiotherapy for the treatment of spinal metastases.

Study selection
We included clinical papers of patients who received radiosurgery for the treatment of spinal metastases.
We excluded non-English language papers, literature reviews, case reports, and papers not focused on radiosurgical treatment of spinal metastases.

Data extraction
One hundred and twenty-four studies were reviewed and scrutinized according to study design (retrospective case series, prospective studies, and randomized control trials), patient characteristics, and primary findings by four independent researchers (AFJ, EG, HT, LG). After analysis of the four authors, reselection was made and a total of 31 studies were selected and included in our review.

Measurements and outcome evaluation
The results of the studies were grouped according to their main purpose: pain control, radiological control, overall survival, comparison of treatment modalities (SRS versus CER versus surgery), indications for spinal radiosurgery, and methods of evaluating disease control and response to treatment.

Evaluation of the level of evidence
The studies were classified according to the level of evidence for therapeutic studies, defined by Wright et al. (9) . 1. Higher quality randomized controlled trial. 2. Lesser quality randomized controlled trial; prospective comparative study. 3. Case-control study; retrospective comparative study. 4. Case series. 5. Expert opinion.

RESULTS
Out of 31 selected studies, a total of 2,241 patients were treated for spinal metastases. All papers are summarized on chart 1. Median follow-up of 11.1 months 59% of patients experienced no acute side effects from treatment; there were 3 cases of acute grade 3 toxicity Local control and overall survival were 91 and 65%, respectively Pain improvement was seen in 75% of symptomatic metastases at 6 months post-treatment Patel et al. (20) Case
The use of spinal radiosurgery for treatment of intradural extramedullary and intramedullary spinal metastases and for epidural compression of myeloma was reported in one study each (level III) (34,35) .

Dose intensity and number of fractions
The dose obtained in the studies for epidural metastases ranged from 8 to 30Gy (different low doses were used for intradural tumors) (12) , in one to five fractions. Moulding et al. considered that when lower doses were used (18 to 21Gy), worse local control was obtained compared to higher doses . Some authors administered the same doses for all patients (18) , whereas others reported that factors such as prior conventional radiotherapy, distance of the epidural metastases from the cord (29) , ...continuation

Study type Radiosurgery status/indication n Dose Results
Massicote et al. (36) Case series Primary irradiation after minimally invasive spinal surgery for unstable spinal metastases 10 patients 8 patients were symptomatic at baseline The median follow-up was 13 months (range: 3-18) Following surgery, the median time to SBRT treatment planning was 6.5 days and subsequent median time to treatment was 7 days Local control was observed in 7 of the 10 patients. Improvements in VAS, ODI, and QOL were observed post-SBRT Moulding et al. (37) Case series Surgical decompression and stabilization for epidural compression followed by spinal radiosurgery 21 patients (20 tumors; 95% were considered highly radioresistant to conventional external beam radiotherapy) 18-24Gy (median 24) single dose Planned target volume received a high dose (24Gy) in 16 patients (76.2%), and a low dose (18 or 21Gy) in 5 patients (23.8%) 15 (72%) of 21 patients died, and in all cases death was due to systemic progression as opposed to local failure The median overall survival after radiosurgery was 310 days One patient (4.8%) underwent repeat surgery for local failure and 2 patients (9.5%) underwent spine surgery for other reasons Local control was maintained after radiosurgery in 17 (81%) of 21 patients until death or most recent follow-up, with an estimated 1-year local failure risk of 9.5% Of the failures, 3 of 4 were noted in patients receiving lowdose radiosurgery Patients receiving adjuvant stereotactic radiosurgery with a high dose had a 93.8% overall local control rate (15 of 16 patients), with a 1-year estimated failure risk of 6.3% Garg et al. (38) Prospective case series EBRT cost 29-71% of the SRS treatment, had more acute toxicity (but self-limited and with low grade), and more of them need further intervention (surgery/kyphoplasty) No late complication in either groups Similar pain relief and tumor histology (23) , among others, were used to individualize the dose. It also should be mentioned that lower doses were used for intradural and intramedullar metastases as well as for salvage treatment after prior conventional radiotherapy, in order to avoid late radiation myelopathy.

Pain control
Although the lack of standardized outcome measurements in the reviewed studies precluded deep or statistical analyses, pain was substantially controlled (up to 90%) in many series, suggesting good efficacy in decreasing pain symptoms (10)(11)(12)14,16,19,23,24,29,33,35,38,39) . A potential advantage of SRS is the time required to obtain some pain relief, as early as 24 hours, probably quicker than CER (21) . However, it is important to emphasize that the methods utilized for pain evaluation in different papers were very heterogeneous, which precluded a uniform analysis (verbal analysis scale, and subjective measurements of pain relief, among others). In some tumors considered radioresistant to CER, pain control was achieved in 96% of the patients with melanomas and in 89% of the patients with renal cell carcinoma (23,24) . However, we also observed that low rates of pain control (43%) were obtained in a series using a lower dose of SRS (11) .

Toxicity
Radiation-related toxicity was mild in almost all series, with rare reports of severe toxicity (10,(12)(13)29,31) . Transient radiculitis was reported in 2 of 31 patients of the Benzil et al. series (12) . Three patients of 41 in the cases published by Martin et al. presented with toxicity grade 3 (19) . These rarely reported cases of moderate toxicity in all the cases reviewed suggested that SRS is a safe procedure.

Local disease control
The efficacy of spinal radiosurgery in local disease control was assessed clinically and radiologically in some studies. Chang et al. reported radiological control of spinal metastases in 90% of the patients at 6 months and in 80% at 12 months, similar to Garg et al., who obtained an imaging control rate of 88% of the patients after 18 months of follow-up (15,17) . Sheehan et al. (33) also reported decreased or stable tumor volume in 82% of the patients treated (mean of 12.7 months of follow-up). Some papers also reported improvement in neurological deficits after treatment with SRS (16,25,39) .
Regarding tumor recurrence, some interesting information was obtained: Koyfman et al. reported that the presence of paraspinal disease at the time of CER (prior to SRS) was associated with high rates of local recurrence (p=0.02) (28) . A dose lower than 16Gy was also associated with local recurrence (p=0.14). Regarding the radiation target, Klish et al. evaluated a prospective series of 58 patients with good disease control and no damage to healthy tissues, and concluded that irradiation one or two levels above or below the involved segment is not necessary when SRS is performed (27) .

Cost-effectiveness of CER versus SRS
In one prospective case series, Haley et al. (40) compared the efficacy/cost-effectiveness of CER versus SRS. At 1-month follow-up, they reported that both methods have similar outcomes regarding pain relief, but CER costs about 29 to 71% less than the SRS treatment. However, CER has a high rate of acute toxicity (although it was self-limited and with low intensity) and required more interventional procedures (surgery, kyphoplasty etc.) at the level of the treated spinal. Neither method had late complications.

DISCUSSION
The ideal treatment of spinal metastases would be one that obtains cure without adding excessive morbidity. However, since no treatment has been proven to increase life expectancy, the main objectives of treatment remain local disease control, pain relief, reestablishment of a normal neurological status, and avoidance of spinal deformity/instability. Surgery is reserved for an unstable spine, to obtain tissue samples, for pain relief in cases of instability, or for urgent decompression. Regardless of the indication for initial surgical treatment, almost all patients should be referred to radiation therapy treatment.
As shown in our data, SRS was efficient in controlling pain in many series (10)(11)(12)14,16,19,23,24,29,33,34,38,39) . Most of the studies showed pain control of up to 90%, even for radioresistant tumors such as melanomas and renal cell carcinoma. Since a cure is not possible, the palliative effect of SRS in pain control has been one of the main reasons for radiation treatment in these patients.
Clinical stability and radiological disease control were obtained in up to 95% of the series (15,17,33) . The efficacy of local control decreased with time, but most of the patients died from progression of the systemic disease before local recurrence (37) . The presence of paraspinal disease increases local recurrence probably because it is associated with late stages of cancer and a more systemic disease. This radiological control obtained with SRS can decrease or stabilize tumor volume, improving the neurological status in some series (16,35,39) . All these features confirm SRS as an efficient method for treating spinal metastases regarding local control.
Ideally, an efficient method that improves the patient's symptoms should not carry high morbidity rates. In this context, the studies assessed in our review with more than 2,000 patients did not demonstrate severe toxicity related to SRS. Moreover, when reported, toxicity was mild and self-limited, attesting to the safety of SRS in treatment of spinal cancer (10,12,13,29,32) .
Since safety and efficacy are warranted, the oncology group needs to individualize the patient's treatment and decide between SRS and CER. Although both methods require certainty of the diagnosis, SRS has its best indications for patients with more restricted spinal/ paraspinal metastases (involving one to two spinal segments); on the other hand, CER is more efficient for multiple level spinal involvement (34) . As stated before, although costs favor the indication of CER (40) , this assertion does not really reflect the truth, especially for patients who live far from radiotherapy centers, as SRS requires less fractions (1 to 5) compared to CER (8 to 20 daily fractions).
Although many questions are not yet solved, SRS has proven to be an efficient and safe alternative to CER for treating patients with restricted spinal metastases. The best indication for using SRS as a primary treatment of spinal metastases is likely for disease restricted to one or two levels of a known poor responder to CER, such as renal cell carcinoma and melanomas. Another potentially good clinical scenario for SRS is in cases in which CER has failed to achieve local disease control.

CONCLUSION
SRS can be used for the treatment of epidural spinal metastases (level II of evidence) and spinal myeloma (level III), especially in tumors known to be poor responders to CER (such as melanoma or renal cell carcinoma) (level III). SRS is also indicated for salvage treatment when CER fails to control disease (level II), as reirradiation of patients without any other potential treatment available. There is some evidence (level III) of the efficacy and safety of SRS in the treatment of intradural extramedullary and intramedullary spinal metastases. The same is true for its use after surgical procedures for canal decompression and stabilization (level III). Further prospective studies are necessary to compare the potential benefits of SRS over CER.