Stereotactic radiosurgery for high-grade metastatic epidural cord compression

Abstract

Purpose

The standard of care of patients with high-grademetastatic epidural compression is open decompression with or without stabilization. However, many patients are unwilling or unable to undergo open surgical decompression. This study investigated the outcomes of treating patients with high-grade (Ryu/Rock radiographic grade IV and V, Spine Oncology Study Group Grade II and III) metastatic epidural spinal cord compression with spinal radiosurgery as first-line therapy in lieu of surgical decompression.

Methods

Utilizing the Henry Ford Spinal Tumor Database, patients with metastatic lesions causing advanced radiographical grade (IV or V) epidural spinal cord compression who received stereotactic radiosurgery (SRS)with adequate clinical and radiological follow-up were identified from 2007-2011. These patients were retrospectively reviewed for clinical and radiological response to radiosurgery.

Results

33 patients with 35 metastatic lesions causing Ryu/Rock radiographical grade IV or V compression were identified with a median follow-up of 435 days. Of the 34 lesions in 32 patients who were ambulatory pre-SRS, 23 (67%) were ambulatory at last follow-up. 6/33 progressed early (less than 2 months) neurologically and an additional 5 patients developed late progressive neurologic deficit. The one patient who was initially non-ambulatory was able to regain ambulatory status. Radiologically, there was a significant epidural tumor response rate of 74%. Ultimately, 9 patients (27%) eventually required surgery for neurologic compromise or mechanical instability. There was one patient who received EBRT previously who experienced radiation myelopathy as a complication of SRS.

Conclusion

Radiosurgery as an initial therapy for high-grade metastatic epidural compression appears to be a viable treatment paradigm for selected patients with close clinical and radiological follow-up. However, a significant minority will progress necessitating the need for rigorous monitoring. Further study is needed prospectively analyze the effectiveness of SRS with or without open surgical decompression.

1. INTRODUCTION

Metastatic epidural spinal cord compression (MESCC) occurs when secondary tumors that grow in the spine or epidural space cause compression of the spinal cord. It is a frequent complication of cancer with an incidence between 2.5 and 14% of cancer patients [1-3] and an overall survival of 3-6 months [2]. Prognosis varies depending on primary tumor histology and location [4], degree of metastasis, comorbid conditions, earliness of diagnosis [5], and ability to ambulate before treatment [6, 7]. Patients who are ambulatory at the time of treatment are more likely to have better outcomes than those who are not [1, 8-12].

The widely accepted standard of treatment of patients with high-grade metastatic epidural canal compromise is open surgical decompression [12, 13] and/or with radiotherapy [7, 14]. Surgery is suitable for patients with a reasonable life expectancy, rapidly evolving neurologic deficit, symptomatic spinal instability, and tumor types that are more aggressive and relatively radioresistant. In select patients, surgery combined with external beam radiotherapy (EBRT) is superior to EBRT alone [13]. Patients with multiple lesions in disperse spine locations, expected survival of less than four months, advanced systemic disease, or patients who are unwilling or unable to undergo open surgical decompression are poor candidates for surgical intervention but might still benefit from radiotherapy. The potential complications of surgery include post-operative pain, wound-healing issues (observed in 11% of patients), and spinal instability [15]. After treatment with EBRT, about 50% of patients maintained ability to ambulate and some nonambulatory patients regained the ability to walk [16]. The rates of tumor control (ie; defined as a lack of tumor progression) by EBRT with various fractionation regimens are variable [17, 18].

In recent years, stereotactic radiosurgery (SRS) has gained popularity as a primary treatment modality for MESCC.We have shown that rapid pain relief was achieved in a median 14 days after radiosurgery and the pain control was durable [19, 20].In a phase II study, 62 patients with previously untreated MESCC were treated up-front with SRS. Radiosurgery doses ranged from 14-20 Gy according to a dose escalation scheme. Overall, epidural tumor response rate was 80% (complete response 27%, partial response 30%, and minimal response 23%), with progression seen in 6%. At the level of the most significant spinal cord compression, thecal sac patency increased from 55±4% to 76±3% (p<0.01). Neurological outcome was also documented in this trial. Overall, 85% remained intact or ambulatory after radiosurgery. Notably, 94% of the patients who were intact before radiosurgery remained so. Of the 27 patients who presented with neurological deficit, 52% (14/27) had complete recovery to normal, 11% (3/27) improved, and 11% (3/27) remained stable [21]. Thus, in this group of patients who were not selected specifically for degree of epidural compression, spinal radiosurgery was found to be effective.

Despite these encouraging findings, there still exists a concern regarding treating patients with high-grade compression with SRS. Because of the proximity to the spinal cord the epidural portion of the tumor is somewhat underdosed, leading to fear that this portion of the tumor may be treated inadequately. As such, the current treatment paradigm at most centers is that high-grade MESCC is considered a surgical disease. However, using SRS as the primary treatment modality for this group of patients is attractive, if they are unwilling or unable to undergo open surgical decompression and stabilization. This study focuses on the neurological outcomes of treating patients with high-grade metastatic MESCC (Ryu/Rock grade IV and V, Spine Oncology Study Group Grade 2 and 3) [22] with spinal SRS as first-line therapy in lieu of surgical decompression. We hypothesize that, even in this group of patients with high-grade MESCC, SRS can be utilized safely and effectively.

2. METHODS

The treatment algorithm for MESCC treated through the Henry Ford Spine Tumor Board is as follows. Patients were offered radiosurgery if they had relatively intact neurologic function (grade a or b). If patients had significant weakness (grade c or d), they were offered surgery as first-line therapy. If patients refused surgery or were determined not to be a surgical candidate due to medical comorbidities, they were then treated with radiosurgery upfront. Patients undergoing SRS at our institution are entered prospectively into a Spinal Tumor Database. Over time, the radiosurgery dose to the epidural tumor has been escalated. Radioresistant tumors such as melanoma, renal cell carcinoma, and soft tissue sarcomas were treated using the same method. All patients included in this analysis were discussed at the weekly multidisciplinary spine tumor board and had their radiographical images reviewed pre-SRS and during follow-up visits at 2, 4 and 6 months after radiosurgery. All patients in this study had at least one follow-up imaging study 2 months after SRS, with the exception of one patient who required surgery 30 days post-SRS, thus imaging follow-up was less than two months. Nonetheless, this patient was included in our study since it represented a radiosurgical failure.

Spinal metastasis was diagnosed by computed tomography (CT) and magnetic resonance imaging (MRI). Patients were given steroids and received radiosurgery within 24-48 hours after diagnosis. Chemotherapy was not halted due to radiosurgery. Steroids were tapered after radiosurgery. For radiosurgery, the entire vertebral body of the involved spine and the epidural tumor were included within the target volume of the radiosurgery treatment. The target mass and spinal cord were delineated by combining contrast-enhanced simulation CT images fused with T1-weighted MRIs with and without gadolinium contrast and T2-weighted MRIs. The paraspinal soft tissue component of the tumor, if present, was also included in the treatment volume. The radiosurgery dose ranged from 16-18 Gy prescribed to the 90% isodose line. For purposes of calculating an organ-at-risk dose-volume histogram, spinal cord volume extended from 6 mm above to 6 mm below the radiosurgery target. The spinal cord dose constraint was 10 Gy to 10% partial volume of the cord [20]. The radiosurgery treatment was delivered by intensity-modulated radiation with a precise beam-shaping technique and noninvasive positioning as previously described [19-25].

The primary endpoints were improvement of neurological function and improvement of thecal sac patency. The extent of epidural compression was measured using T2-weighted and T1-weighted MRI and radiographic grades were assigned using the Ryu/Rock scale [22]. The Ryu/Rock grading system combines radiographic and neurological data to rank epidural compression since the radiographic extent of cord compression does not necessarily correlate with the neurological examination. Radiographic grade 0 (SOSG grade 0) has no spinal canal compromise, grade I (SOSG grade 1a) has epidural tumor with no thecal sac compression, grade II (SOSG grade 1b) has mild compression of the thecal sac, grade III (SOSG grade 1c) has impingement of the spinal cord, grade IV (SOSG grade 2) has cord displacement with cerebrospinal fluid (CSF) still visible, and grade V (SOSG grade 3) is spinal cord compression with no visible CSF between the tumor and the spinal cord. For neurological grading, grade a is symptom free, grade b has pain or radiculopathy, grade c is ambulatory paresis with muscle strength >3 of 5, grade d is a non-ambulatory paresis with muscle strength ≤3 of 5, and grade e is plegic muscle strength, with or without urinary or rectal incontinence.

In this study, a positive outcome is when a patient with a pre-SRS Ryu/Rock neurological grade of a, b or c had an improved or stable neurological examination at the end of the study period (ie, post SRS neurological grade a, b, or c). If the patient declined neurologically, though remained ambulatory, this would be considered a negative outcome. If a patient began with grades d or e and improved to a, b or c, this was also considered a positive outcome. A negative outcome is defined as any situation in which the post-SRS grade is either d or e, regardless of the pre-SRS grade. In addition if the patient required open surgical decompression at any point, this was regarded as a radiosurgical failure or negative outcome. Patients were excluded from this study if they were lost to follow-up or were unable to obtain neurological or radiological follow-up.

Further analyses using McNemar’s test were done to compare pain-free rates (ie, grade a on neurological examination) between pre- and post-surgical measurements. Chi-squared tests were performed to compare the rates of additional surgery for the three different locations (ie, lumbar, thoracic and cervical). Kaplan-Meier methods were used to perform time-to-event analyses for death, neurological progression and radiological progression. Time was measured as the number of months between SRS and the event of interest or last follow-up. Patients were censored if they did not have the event of interest at their last follow-up. Logrank tests were used to compare the event rates among the imaging and neurology examination groups. All testing was done at the 0.05 level. Analyses were performed using SAS version 9.2.

3. RESULTS

We identified 44 patients with a total of 46 distinct, advanced radiographical grade (Ryu/ Rock grade IV or V) MESCC lesions at cervical, thoracic or lumbar levels who received SRS from 2007-2011, and had not had prior neurosurgical decompression. Of these 46 total lesions, 33 patients with 35 lesions had adequate radiological (range 30-2001 days, median 336 days) and clinical follow-up (range 34-2070 days, median 435 days) to be included in the study. These patients were reviewed retrospectively for clinical and radiological response to SRS . 22 (66%) were male and the mean age at SRS was 64 (s.d.=11.7) with a range from 41 to 87 years. The distribution of primary tumors is given in Table 1. Of the 35 procedures performed 20 (60%) were for grade IV lesions and 15 (40%) were for grade V lesions. The distribution of locations was 28 (80%) thoracic, four (11%) cervical and three (9%) lumbar. The majority of the procedures used a dose of 18Gy (n=33, 94%) with two (6%) patients prescribed doses of 16Gy (adjacent level of previously radiosurgically treated spine). The mean GTV was 57.0 cc +/- 34.1 cc. No margin was added for treatment planning. See Figure 1 for a sample treatment plan.

Figure 1.

MRI showing an 83-year-old male with prostate cancer that metastasized to T11/12 (A= Pre SRS, B=Post-SRS). The patient had significant pain improvement, remained neurologically stable, and was followed for 55 months. The treatment plan is also shown.

At the time of analysis with a median follow-up of 435 days (range 34-2070 days), 6 patients were still alive (18%). Of 32 patients who were ambulatory pre-SRS, 23 (67%) continued to be ambulatory with a stable or improved neurological examination at last follow-up (25 of 35 lesions). Of the 33 total patients, 6 (18%) worsened neurologically, i.e., changed Ryu/Rock neurological grade, early (less than 2 months) despite radiosurgical treatment and an additional 5 patients developed late progressive neurologic deficit due to radiosurgical failure. The one patient who was initially non-ambulatory did regain ambulatory status after radiosurgery.

Of the 33 patients who obtained the SRS procedure, overall outcome was positive for 21 (64%) and negative for 12 (36%).

Of the 21 Ryu/Rock grade IV (SOSG Grade 2) lesions, 14 had stable or improved radiographic compression at last follow-up, 5 lesions worsened, and another 3 initially responded to therapy and worsened subsequently. Of 14 Ryu/Rock grade V lesions (SOSG grade 3), 12 were improved or stable after spinal radiosurgery whereas the other two progressed. Overall, radiologically there was a favorable response rate of 71%.

The overall 6-, 12-, and 24-month survival rates were 82%, 61%, and 23%, respectively (Figure 2). The time variable for this graph is months between SRS and death or last follow-up. The difference in survival rates for the imaging outcomes was significant (p=0.018, Figure 2), with those patients whose condition improved post radiosurgery having the highest survival rates, followed by those whose condition remained the same. Patients whose condition worsened following radiosurgery had the poorest survival rates. The difference in survival rates for the neurology examination outcomes was also significant (p=0.002), and the patterns were similar to those seen in the imaging groups. Histology did not appear to be a significant predictive factor for response to therapy, which may be in part due to the small number of patients and diverse histologies. Furthermore, when histologies were grouped as “radiation sensitive” (breast, prostate, merkel cell, myeloma) or “radiation insensitive” (lung, renal, gastrointestinal, Ewing’s sarcoma, melanoma, thyroid), again, no significant difference was found (p=0.36) (Figure 3). Pre-treatment radiological examination grade was also not found to be a significant predictor of neurologic or radiologic outcome (p=0.20). Likewise, neurologic exam pre-treatment grade was also not found to be a significant prognostic factor (Figure 4).

Figure 2.

Kaplan-Meier survival curves showing overall survival and stratified by radiological outcome (p=0.018). B) Better = 16 (46%), Same=10 (29%), Worse=10 (29%).

Figure 3.

Kaplan-Meier curves for freedom from neurologic (a, p=0.44) and radiologic (b, p=0.39) progression stratified for the most common histologies. Also shown are Kaplan-Meier curves for freedom from neurologic (c, p=0.36) and radiologic (d, p=0.51) progression stratified for radiation-sensitive and insensitive histologies. A and B) Breast = 5 (14%), lung = 8 (23%), prostate = 6 (16%), and C and D) radiation sensitive y=14 (40%), n=21 (60%).

Figure 4.

Kaplan-Meier curves for freedom from neurologic (a, p=0.75) and radiologic (b, p=0.20) progression when stratified by radiologic pre-treatment grade. Also shown are Kaplan-Meier curves for freedom from neurologic (c, p=0.64) and radiologic (d, p=0.19) progression when stratified by neurologic pre-treatment grade. A and B: Grade 4=21 (60%), grade 5=14 (40%); C and D: a=3 (9%), b=25 (71%), c=6 (17%) , d=1 (3%).

Radiosurgical failures and complications

Of the grade IV patients, five patients eventually required surgery. One patient required early surgery (within 60 days) for mechanical instability. Two patients underwent open decompression and fusion due to early neurologic decompensation. Both of these patients unfortunately did not have good return of function. One patient did undergo corpectomy and fusion for recurrent pain 5 months after radiosurgery. An additional patient underwent lumbar decompression 1 year after SRS. His compressive pathology was due to epidural lipomatosis rather than metastatic tumor, however. For the grade V patients, two patients did require early surgical decompression due to neurologic decline. One other patient underwent decompression and fusion due to intractable pain 3 months post-SRS. In addition, one patient had a surgical decompression 1 year after SRS due to imaging progression although the patient’s neurological status was unchanged. Overall, nine patients required additional neurosurgical decompression following the SRS procedure for spinal fractures and for neurological compression. The only statistically significant factor associated with additional surgery was location: four (14%) of the 28 thoracic MESCCs, two(50%) of the cervical MESCCs and three(100%) of the lumbar MESCCs required additional neurosurgical decompression (p=0.004, Fisher’s Exact test).

There was one patient with grade IV compression who experienced radiation toxicity after SRS to a metastatic lesion at C1-2. She had previously undergone external beam therapy. She eventually underwent surgical decompression and fusion but unfortunately her neurologic function declined significantly to near complete quadriplegia.

4. DISCUSSION

This retrospective study evaluates the use of SRS in lieu of surgery for treatment of patients with advanced MESCC. A limitation to this study is due to the fact that most of these patients had significant advanced systemic disease hence some patients were unable to obtain follow-up MRI images. Although data suggests improved neurological outcomes in patients with advanced MESCC, the small sample size of this study limits its statistical significance. Because of the small number of total patients and heterogeneous tumor histologies, this study is underpowered to make general conclusions. Patients were excluded if they did not have a repeat radiological scan at least 1 month after SRS; the demographics of these patients was not significantly different from that of the patients included in the study except that patients that died early were more likely to have advanced systemic disease. Of the 11 patients who were not included in the study due to inadequate follow-up or imaging, two were lost completely to follow-up soon after treatment due to moving outside of the health system. The other nine patients did not receive any post-treatment imaging follow-up. These patients, however, continued to be followed by their medical oncologist (median 97 days, range 24-295). At last follow-up all patients were deceased and because of the lack of formal neurological testing, it is difficult to determine their exact neurologic grading. There did appear to be one patient who had worsening neurologic status which could be ascribable to MESCC. The remainder seemed to maintain their neurologic status.

Because of the intimate relationship of the epidural tumor to the spinal cord, surgical decompression is generally the treatment of choice in patients with high-grade spinal canal compromise. Indeed, if radiosurgery is performed in this patient group, it can be argued that underdosing of the tumor area can cause recurrence of both the lesion and epidural compression. However, even partial reduction of the epidural lesion may provide sufficient decompression of the spinal cord and thus preserve neurologic function. In this series, we did observe a lower rate of tumor control in this study than in our other larger series, which could have been due to insufficient dosing to the epidural portion of the tumor [23].

Because radiosurgical management of these patients can lead to rapid, permanent loss of neurologic function in some, it further highlights the need for rigorous follow-up and also close collaboration between the radiation oncologist and surgeon. This is not a paradigm which can be followed without immediate availability of surgical support, should the urgent need arise. Certainly, open surgical decompression is a more definitive therapy as immediate and rapid decompression of the neural elements is achieved. This may represent a better management algorithm for a certain group of patients, especially since those most likely to fail cannot always be identified.

Nonetheless, a majority of patients in this study were able to avoid invasive surgery as a result of treatment with radiosurgery resulting in stable or improved neurologic examinations. By current standard of care, all of these patients would have otherwise undergone open surgical decompression. Radiosurgery is more convenient for these patients since it is a noninvasive outpatient procedure thus presenting an attractive option if clinically appropriate. Of the patients with radiographic grade IV and V, 27% needed surgery, including 17% who needed surgery for neurologic deficit. The patients who progressed after radiosurgery may have had better outcomes if they had had early surgical decompression, especially those with rapid neurological deterioration, spinal instability, and/or retropulsed compression fracture.

A low incidence of spinal toxicity was reported in this study, despite the close proximity of the spinal cord to the radiation dose. One patient who received radiosurgery after external beam therapy experienced radiation toxicity in the form of intense skin pain around the radiosurgery site. Of the patients who initially had improved neurological outcomes and then subsequently deteriorated, it is likely that the deterioration was from progression of their disease as opposed to the side effects of long-term radiation.

5. CONCLUSION

Surgical decompression for high-grade metastatic epidural compression remains the gold standard. However, radiosurgery as an initial therapy for high-grade metastatic epidural compression appears to be a viable treatment paradigm in certain patients. In this prospectively recorded case series a majority of patients who would traditionally be regarded as ‘surgical’ were managed effectively and safely with SRS. A significant minority of patients, however, did require surgical intervention or declined neurologically highlighting the need for close clinical and radiological follow-up. In addition, the patients who had neurologic progression early after radiosurgery had uniformly poor outcomes. This study was unable to identify those patients most likely to fail after radiosurgery. Given these findings, a prospective trial comparing the efficacy of stereotactic radiosurgery with or without open surgical decompression is needed. In addition histology may prove to be an important prognostic factor, but more patients would need to be studied to determine this possible effect.

6. Acknowledgements

The authors report no conflicts of interest or sources of funding. The authors would like to thank Michelle Jankowski, MAS for her additional statistical assistance. In addition, the authors sincerely thank Susan Macphee-Gray for her invaluable editorial input.