More than 1.6 million new cases of cancer are diagnosed annually in the United States.^2,24 Symptomatic spinal metastases are estimated to occur in up to 10% of all cancer patients.²⁸ The majority occur in the extradural compartment, most frequently presenting within the vertebral bodies.¹² Most of the metastatic burden (70%) is within the thoracic spine, followed by the lumbar spine (20%), cervical spine, and sacrum.²⁰ As survival rates for primary tumors improve, it is expected that the prevalence of spinal metastases will continue to increase.⁶

The treatment paradigm for management of spinal metastatic disease has shifted dramatically over the last 2 decades.^5–7 Conventional external beam radiation therapy (cEBRT) has been the principal method of delivering radiation to the metastatic spine.¹⁶ However, cEBRT lacks the precision to deliver large radiation doses to the spine without also exposing nearby radiosensitive structures such as the spinal cord. This limits the ability of cEBRT to deliver a cytotoxic dose to a tumor.^{9,11,14,15,22–25}

Stereotactic radiosurgery (SRS), on the other hand, delivers high-dose radiation to spinal metastases while limiting irradiation of the spinal cord.^10,25,27 SRS has been shown through multiple studies to be accurate, safe, and effective in providing local tumor control and alleviating pain.^{1,8,10,21,23,30} SRS therapy has also been shown to improve outcomes independent of tumor histology, which previously had been a limitation with cEBRT as described above.¹⁸

Surgical treatments have undergone a similar evolution,¹³ a nd s urgical s trategy h as b een i nfluenced by t he success of radiotherapy. In 2005, Patchell et al.¹⁸ demonstrated the superiority of surgical decompression and instrumentation followed by conventional radiation therapy versus radiation therapy alone.

Radiosurgery alone has recently been shown to be an effective noninvasive treatment for spinal metastases with limited epidural compression.^10,27 In cases of higher-grade epidural compression, however, radiosurgery may not offer a safe or effective strategy for tumor control and decompression of neural elements. For these cases, operative decompression can be used as a tool to allow for appropriate radiosurgical treatment. This method has been shown to be a safe and effective method for local disease control.^5,6,8,14

The current paradigm at our institution, as described by Bilsky and colleagues,^5–8,14 is to consider each patient in terms of neurological function, oncological histology and prognosis, the mechanical stability of the spine, and systemic disease and medical comorbidities (NOMS). A patient's neurological function is usually directly related to the degree of epidural spinal cord compression (ESCC) caused by tumor growth (Fig. 1). Those with high-grade ESCC and/or mechanical instability undergo debulking and instrumentation, or “separation surgery,” to create a tumor-free margin around the thecal sac and to stabilize the spine.^6,18 High-dose SRS may be used as the primary therapy if a safe margin exists between the spinal cord and tumor on initial imaging, eliminating the need for surgery.

Grading of ESCC. Left: Grade 0 is disease confined to bone, Grade 1a disease involves the epidural space but does not compress the dura, Grade 1b disease compresses the dura but does not abut the spinal cord, and Grade 1c disease abuts but does not compress or adjust the course of the spinal cord. Center: Grade 2 disease compresses the spinal cord, but CSF is still visible. Right: Grade 3 disease compresses the spinal cord, obliterating all visible CSF at that level. (Reproduced with permission from Moulding et al: J Neurosurg Spine 13:87–93, 2010). Figure is available in color online only.

This paradigm has shown greater efficacy for tumor control, neurological outcome, and duration of response when compared with cEBRT, regardless of tumor histology.¹⁶ In this study, we evaluated the efficacy of radiosurgery with or without separation surgery for treatment of oligometastatic disease at our institution in a series of consecutive patients.

Methods

Results

A total of 63 patients treated from January 1, 2007, through December 31, 2011, were identified, but 6 of these patients were lost to follow-up. Therefore, a total of 57 patients with 69 lesions were included in this analysis. Twenty-one lesions were treated with surgical decompression followed by SRS, and 48 lesions were treated with SRS alone. At patient death or latest follow-up, 63 of 69 cases (91.3%) had local disease control, and the remaining 6 cases (8.7%) demonstrated radiographic progression. Tables 1 and 2 show patient demographics and locations of the tumors treated. The majority of cases (65%) involved single-level disease and 22% had 2-level disease. Table 3 shows the histology and rates of local failure of the tumors treated. The 6 patients with local recurrence are described in Table 4; 3 were from the SRS+surgery group and 3 were from the SRS-only group. The mean time to recurrence was 8.1 months (range 2.5–16.6 months). There were recurrences within each of the fractionation schemes in the study. Three instances were found on surveillance imaging and were asymptomatic, and one of these actually occurred outside the SRS field in the paraspinous musculature. Two patients presented with radicular symptoms and worsening pain. One patient had imaging studies as part of a workup for a new foot drop, which revealed recurrence, but the foot drop was ultimately found to be caused by peroneal nerve compression in the popliteal fossa. Further treatment was determined on a case-by-case basis and consisted of observation, resection, or re-resection as needed.

A univariate regression analysis was performed to assess which factors might predict recurrence. The following variables were analyzed: treatment type (SRS alone vs SRS+surgery), age (< 60 vs ≥ 60 years), histology, region of spine involvement, number of levels treated, prior conventional radiation, fractionation (hypofractionated vs single), and dose (hypofractionated 20, 27, or 30 Gy and single fraction 16 Gy vs 20–22 Gy). None of the variables were found to be statistically significant factors in predicting recurrence.

The median length of follow-up was 9.8 months in the SRS-only group, 13.7 months in the SRS+surgery group, and 10 months overall. Of the 48 cases in the SRS-only group, 43 (90%) had a stable Frankel grade, and the remaining 10% had an improved Frankel grade. Of the 21 cases in the SRS+surgery group, 17 (81%) had a stable Frankel grade, and 3 (14%) showed an improved grade. The remaining patient, who had a decline in Frankel grade, developed new cervical radiculopathy that was not due to a compressive lesion.

There were 5 fractures (7.2%) overall; hypofractionated therapy was performed in 4 of the 5 cases, and single-fraction therapy was performed in the remaining case. None of the patients who received SRS and surgery sustained a pathological fracture. The most common histological findings were renal cell, breast, and lung carcinomas, which accounted for 66% of the cases (Table 3). There was 1 recurrence in the breast/lung/renal cell group. There was 1 case of mild esophagitis that resolved within 3 months. There were no cases of radiation-induced myelopathy. There were 2 cases of durotomy in the surgical group, both of which were closed primarily with no further sequelae. Data on pain were available for 45 of 69 cases (65%), with an overall average visual analog scale score decrease of 3.4 ± 2.6 (median 3) following treatment.

The median overall time to recurrence was 9.6 months (range 2.5–16.6 months). A total of 32 patients (56%, median follow-up 6.6 months) died; of these, 3 had local progression. Twenty-five patients (44%) were alive with a median follow-up of 15.4 months. Of the patients still living, 3 (12%) had local progression of their disease. The remaining 22 patients were alive and free of local failure at a median follow-up of 15.6 months. A Kaplan-Meier survival analysis was performed to demonstrate the time to local failure (Fig. 2).

Cumulative incidence of local progression as a function of time from SRS completion (Kaplan-Meier method). The dashed line represents the group receiving hypofractionated therapy and the solid line represents the group receiving single-fraction SRS.

Overall, there was a 5.8% risk of local failure at 1 year. Tables 5 and 6 show the percentage risk of local failure at 1 year for the SRS-only and SRS+surgery groups and subdivide them by fractionation scheme. The 1-year risk of failure for SRS alone was 4.2% and for SRS+surgery was 9.5%.

Discussion

The goals of treatment for metastatic spinal disease are palliative in nature and consist of alleviating pain, providing mechanical stability, maintaining or improving neurological function, and achieving durable local tumor control. Patchell et al.¹⁸ and Bilsky⁵ demonstrated that surgery combined with cEBRT was superior to radiation therapy alone in achieving these goals. More recently, high-dose, single-fraction SRS has also been shown to be a safe and effective palliative treatment independent of tumor histology.^{1,10,22,27,29} SRS following separation surgery is also safe and effective.^17,19 A study in 2010 demonstrated better results for those receiving higher doses of SRS following surgery.¹⁷

Recently, Laufer et al.¹⁴ reported a series of 186 patients with ESCC treated with SRS and surgery. This is the largest series to date showing the efficacy of separation surgery followed by adjuvant SRS. Importantly, the authors demonstrated that high-dose hypofractionated (24–30 Gy in 3 fractions) therapy provided a benefit in local tumor control compared with low-dose (18–36 Gy in 5 to 6 fractions) hypofractionated adjuvant therapy. The rate of recurrence in the high-dose hypofractionated group was lower than that in the single-fraction group, although it was not statistically significant.

The objective of this study was to evaluate the efficacy of our approach to treatment of oligometastatic disease to the spine. This includes SRS alone for patients with minimal ESCC and separation surgery followed by SRS for patients with high-grade ESCC. Those receiving SRS alone had a 1-year risk of local failure of 4.2% (95% CI 1.6%–10.1%). This result is similar to those found in previous studies by Gerstzen et al.¹⁰ and Yamada et al.²⁹

Gerstzen et al. presented 500 consecutive lesions treated with moderate-to high-dose (15–22.5 Gy) singlefraction SRS with a median follow-up of 21 months.¹⁰ The authors reported long-term tumor control of 88% for all cases, and 100% in breast, lung, and renal cell metastases when SRS was delivered as the primary therapy. Our series demonstrates a 95.8% (46 of 48) local control rate when using SRS alone. Furthermore, after subdividing our patients with breast, lung, and renal cell cancer there was only a single recurrence in a patient with renal cell carcinoma (Table 4). The rates of recurrence in this subgroup are consistent with the findings presented by Gerstzen et al.¹⁰

In our series there were 21 cases of surgery followed by adjuvant SRS with a 1-year rate of local failure of 9.5% (Table 6). This is the same rate of failure at 1 year as reported by Moulding et al. in 2010.¹⁷ Recently, Laufer et al.¹⁴ reported a series of 186 patients with ESCC treated with SRS and surgery, which is the largest series to date showing the efficacy of separation surgery followed by adjuvant SRS. Importantly, the authors demonstrated that highdose (24–30 Gy in 3 fractions) hypofractionated therapy provided a benefit in local tumor control compared with low-dose (18–36 Gy in 5–6 fractions) hypofractionated adjuvant therapy. Additionally, the rate of recurrence in the high-dose hypofractionated group was lower than the single- fraction group, although not statistically significantly.

Our study shows a difference in local progression between single-fraction and hypofractionated therapy in patients who also underwent surgery (0% vs 28.6% 1-year risk of local failure); however, this difference was not statistically significant. Any conclusion about the greater efficacy of single-fraction over hypofractionated radiotherapy may, therefore, be premature based on our results and those from the recent study by Laufer et al.¹⁴ Additionally, we did not identify any factors that could predict which patients might experience a recurrence.

The limitations of this study include the retrospective nature of data collection, small sample size, and single institution analysis, thus limiting generalizability. Another clear limitation of this study is the limited duration of follow-up. This study population has a limited lifespan because of the nature of metastatic disease. Nonetheless, this should not detract from the demonstrated efficacy of such a palliative approach until the time of death. Further investigations can use these data to develop studies that are more adequately powered to investigate SRS with or without surgery as an effective treatment in preventing local progression and preserving neurological function.

Conclusions

In this retrospective case series, SRS represents a safe, viable, and effective method in managing spinal metastases independent of tumor histology. It was especially effective for the most common histological entities of breast, lung, and renal cell carcinoma. We identified no predictors of treatment failure in this series. SRS alone was a reliable and effective option in treating patients with low-grade ESCC. In patients with high-grade ESCC requiring surgical decompression, SRS was an effective adjuvant therapy in preventing local progression and preserving or improving neurological function.