Introduction

The spinal column represents the most common skeletal site of cancer metastasis, affecting 40% of all cancer patients, and 20,000–30,000 cases per year in the USA [1,2,3]. Up to 20% of patients with spinal metastasis will develop symptomatic epidural spinal cord compression (ESCC) requiring treatment [4]. Some form of therapeutic intervention, tailored to the spinal metastatic disease, is required in about half of the patients with spinal metastasis regardless of the presence of ESCC. Approximately 5–10% of these patients require surgical intervention [5, 6].

Assessing which patients with spinal metastasis will benefit from surgical intervention is challenging. Potential indications for surgery in this patient population include evidence of ESCC, neurological compromise, evidence of mechanical spinal instability, favorable prognosis (e.g., dependent on primary tumor type and physical status of the patient), and tumors known to be resistant to radiation treatment [5].

The definition of spinal instability has been controversial with multiple stability classification systems published in the trauma setting [7,8,9,10,11]. Among these trauma classification systems are the White-Panjabi stability checklist, the three-column system of Denis, the subaxial injury classification and the thoraco-lumbar classification system [7, 9, 11]. Following these publications, SINS was introduced in 2010 to address spinal stability unique to epidural bony metastases in order to facilitate communication of factors pertaining to spinal instability in the oncology setting. SINS uses clinical and radiographic data to provide clinicians with a heuristic assessment tool to assess stability in patients with spinal metastatic disease. SINS was formulated by the Spine Oncology Study Group on the basis of a systematic review and modified Delphi criteria evaluating factors crucial for the assessment of spinal stability [12]. It stratifies patients into three categories; those with a stable spine, potentially unstable spine, and unstable spine.

Since SINS is used to define important hallmarks of spinal instability in metastatic spinal disease, we sought to investigate its clinical utility in relation to surgical practice and clinical factors important in surgical decision-making. These factors include the severity of ESCC, the patient neurologic status at presentation and patient prognosis. This study also aimed to study the rate of progression of instability (as defined by SINS) and the extent of spread of spinal metastasis (as defined by Tomita et al. [13]) in patients treated with radiotherapy. This may provide insights in improving follow-up mechanisms and timing of surgical interventions in this patient population.

Methods

Study sample

Ethics approval was obtained by the local research ethics board (REB). Patients with spinal epidural metastatic disease were identified via a centrally administered oncology database compiled by the provincial Cancer Centre which manages ~50–70 referrals concerning spinal metastasis annually. Patients were allocated during the period between January 2010 and January 2015. The inclusion criteria used were (1) patients older than 18 years, (2) documented diagnosis of primary tumor type (confirmed through tissue biopsy), and (3) radiographic evidence of spinal epidural metastatic disease on computed tomography (CT) scan and/or magnetic resonance imaging (MRI). The presence of pain and whether it related to mechanical characteristics was documented as part of the pain component in SINS. Cases with intradural or leptomeningeal metastatic disease were excluded. Treatment decisions in relation to surgical interventions and radiotherapy were determined by the treating spine surgeon and radiation oncologist.

Radiographic scoring

Scoring of SINS components was completed by a neurosurgery spine fellow. SINS has been previously validated and has been found to have near-perfect inter- and intra-observer reliability [14]. CT imaging was used in order to determine: (1) the location and morphology of the bone lesions, (2) spinal alignment, (3) the degree of vertebral body collapse, and (4) and posterolateral element involvement. In cases of multiple metastatic lesions, the lesion with the highest SINS score was used in the analysis.

Based on SINS, patients were divided into three categories: those with stable (0–6), potentially unstable (7–12), and unstable spines (13–18). An assessment of ESCC was conducted utilizing a scale consisting of grades 0–3 on the basis of MRI [15, 16]. Grade 0 or 1 represents disease confined to bone or that with epidural space involvement but no spinal cord or cauda equina compression, respectively; grade 2: epidural impingement with spinal cord or cauda equina compression but CSF signal is still present; grade 3: spinal cord or cauda equina compression with complete obliteration of CSF signal. The anatomical classification suggested by Tomita et al. [13] was applied to categorize spinal metastasis as intra-compartmental, extra-compartmental or multiple based on MRI findings.

Data collection

Patient demographics and clinical features were obtained via retrospective review, including tissue-proven tumor histology, timing of diagnosis of the primary malignancy, timing of spinal epidural metastases, patient prognosis, neurologic status and treatment modalities. In addition, the Karnofsky performance status (KPS) was utilized to to assess the functional level of patients on the basis of an 11-point activity assessment ranging from 0 (death) to 100% (no symptoms, no evidence of disease) [17]. KPS categorizes patients into those with good (≥80%), intermediate (50–70%) and poor physical condition (≤40%). The neurologic status was assessed by means of the Frankel grading system and categorized into three groups including those with Frankel grades A–C (useless or absent motor function below the level of the lesion), Frankel grade D (useful motor function below the level of the lesion), and Frankel grade E (normal motor and sensory function) [18]. Ambulation was also assessed at presentation and designated three categories; ambulates independently, ambulates with assistance (cane or walker) and non-ambulatory (wheelchair- or bed-bound).

Statistical analysis

Data analysis was performed using SPSS version 24.0 (Armonk, NY: IBM Corp.). Results were summarized as single proportions and percentages or means with standard deviations. Differences between proportions of cases in study comparisons were assessed with the Pearson chi-squared test. For normally distributed data, two-sample Student’s t-test for independent samples was used for non-parametric data. Alternatively, the Mann–Whitney U test was used. All levels of significance were two-tailed and set at p < 0.05.

Results

Our database identified 300 patients with spinal epidural metastases; of these, 15 cases were excluded due to concomitant presence of intradural or leptomeningeal metastases (n = 12) and unavailable CT imaging (n = 3). We included a total of 285 patients who met the inclusion criteria. The SINS classification identified 35.1%, 52.3% and 12.6% of patients as stable, potentially unstable and unstable criteria, respectively (Table 1). The majority of bony lesions were lytic (67.4%) and the overall rate of kyphotic or scoliotic deformities associated with spinal metastases was 9%, with subluxation or translation affecting 3% of cases (Table 2). The most common histological diagnosis was non-small cell lung cancer (25%), followed by breast (18%) and prostate (14%).

Table 1 Breakdown of SINS in 285 patients with spinal metastases.
Full size table
Table 2 Clinical and pathological characteristics of patients with spinal metastases as classified by SINS.
Full size table

SINS and patient prognosis

Survival data was available for 225 patients. Of these, 83, 117, and 25 patients were stable, potentially unstable and unstable, respectively. The mean survival in the stable, potentially unstable and unstable categories was 270 ± 323, 280 ± 488, and 248 ± 411 days post-diagnosis of spinal metastases, respectively. Kaplan–Meier analysis showed no difference in survival between the three SINS groups (p = 0.98) (Fig. 1).

Fig. 1
figure 1

Kaplan–Meier plot showing survival data of different SINS groups.

Full size image

SINS and the extent of ESCC

Spinal MRIs were available in 185 patients. There was a statistically significant relationship between grading of SINS and the severity of ESCC (X2 = 38, p < 0.001). A statistically higher percentage of patients with unstable SINS (69%) was found to have grade 3 ESCC, whereas 86% of patients with stable SINS had grade 0 or 1 ESCC (Table 2).

SINS and neurological status

Pre-treatment Frankel grading and ambulatory status were available for all patients. Among patients with unstable SINS, 39% of patients were wheelchair-bound, compared with 19 and 18% of potentially unstable and stable SINS groups, respectively (Table 2). This difference was not statistically significant (X2 = 7.9, p = 0.09). When comparing whether any specific SINS component was related to worse ambulatory status, the presence of bilateral posterolateral neoplastic involvement was significantly associated with non-ambulatory status at presentation; about 47% of non-ambulatory patients demonstrated bilateral posterolateral disease, as compared with 40% of patients with unilateral disease and 9% of patients with no posterolateral disease (X2 = 12.6, p = 0.014). Other SINS components were not significantly associated with worse ambulatory status including the location of metastatic lesion (p = 0.27), presence and quality of pain symptoms (p = 0.69), type of bone lesion (p = 0.91), spinal deformity (p = 0.39), and the extent of vertebral body collapse (p = 0.86).

Among patients with unstable SINS, about 28% had Frankel grades A–C, compared with 14.1 and 16% of potentially unstable and stable SINS groups, respectively (X2 = 5.5, p = 0.23) (Table 2). SINS was not significantly associated with differences in the Frankel grade of patients at presentation. There was no significant association between the Frankel grade of patients and the location of metastatic lesion (p = 0.28), presence and quality of pain (p = 0.86), type of bone lesion (p = 0.56), spinal deformity (p = 0.068), and the extent of vertebral body collapse (p = 0.73). However, 47% of patients with Frankel grades A–C at presentation demonstrated bilateral posterolateral neoplastic involvement, as opposed to 40% with unilateral involvement and 13% with no posterolateral neoplastic spread (p = 0.028).

SINS and treatment interventions

There were 59 patients treated surgically with post-operative radiotherapy, 198 treated with radiotherapy, and 28 treated medically with chemotherapy. There were 58% of patients with unstable SINS treated surgically, compared with 12% of patients with stable SINS (X2 = 14.3, p < 0.001) (Table 2). The average overall SINS score was significantly higher in the surgical group in which spinal instrumentation was used (10.7 ± 3.2) as compared with stand-alone decompression (7.1 ± 2.3, p = 0.002) or percutaneous vertebral augmentation (6.6 ± 3.5, p = 0.03). Furthermore, 93% of patients with unstable SINS received spinal instrumentation, compared with 81 and 33% in the potentially unstable and stable groups (X2 = 15.5, Fisher exact p = 0.002); none of the cases classified as unstable SINS underwent stand-alone decompression (Fig. 2).

Fig. 2
figure 2

The relationship between SINS and the use of spinal instrumentation.

Full size image

Changes in metastatic spinal instability after radiotherapy

A total of 95 out of 198 patients were treated with radiotherapy alone and had available follow-up CT imaging of the spine 364 ± 244 days after the initial diagnosis of spinal metastasis. Of these patients, a deterioration in SINS from potentially unstable to unstable category was observed in 30% and from stable to potentially unstable in 8.7% (X2 = 39, p < 0.001) (Table 3). In the unstable category, 83% of patients remained unstable whereas 17% improved to potentially instability post-radiotherapy alone. Only 2 out of 95 patients demonstrated an improvement. One patient improved from potentially unstable to a stable disease and the other from unstable into potentially unstable disease. The improvement in SINS after radiotherapy was attributed to alleviation of pain and changes from lytic or mixed bony morphology into blastic.

Table 3 Ordinal changes in SINS in the radiotherapy group (n = 95) at mean 364 ± 244 days (columns), as compared with SINS at initial presentation (rows), X2 = 39, p < 0.001.
Full size table

Radiographic evidence of multiple spinal metastasis was found in 68% of patients. This group was divided into extra-compartmental (epidural and/or paravertebral extension) and compartmental solitary lesions which were vertebral body (7%), involving the pedicle (5%), or lamina (1%) (Table 4). In the radiotherapy group, a significantly higher proportion of cases with extra-compartmental solitary spinal metastases progressed into multiple metastases (64%), compared with 40% of cases with solitary intra-compartmental lesions (X2 = 42.4, p < 0.01; Fisher’s exact p < 0.01).

Table 4 Progression of spinal metastases in the radiotherapy group at mean 364 ± 244 days (columns), compared with the initial presentation (rows), X2 = 46, p < 0.01.
Full size table

Discussion

In this single institution retrospective review, we described multiple patient factors and their association with SINS. Spinal instability, as defined by SINS, was associated with surgical interventions and the use of spinal instrumentation. Although SINS was not significantly associated with the degree of neurologic disability, we observed that patients classified in the unstable SINS category had evidence of severe epidural spinal cord compression on MRI. In addition, SINS is not associated with survival post-metastatic epidural disease.

SINS is viewed as a surrogate of spinal instability in the context of spinal epidural metastasis [12]. It only applies to patients with spinal metastatic disease and aids physicians in identifying the presence of instability or impending instability that might require surgical intervention. SINS also assists in quantifying the severity of spinal metastatic instability which facilitates clinical communication for both academic and decision-making purposes. In this study, we found that a greater proportion (than expected) of patients with unstable SINS required surgical stabilization procedures (93%) or percutaneous vertebral augmentation (7%). In addition, patients requiring spinal instrumentation demonstrated significantly higher SINS scores as opposed to patients treated with stand-alone decompression. This finding suggests that SINS might facilitate in determining the choice of surgical modality. Surgical options for spinal metastasis range from minimally invasive procedures such as stand-alone percutaneous pedicle-screw instrumentation for isolated instability (e.g., in patients with unstable SINS and grade 0–1 ESCC) to open surgical decompression with spinal instrumentation and anterior column reconstruction (in patients with unstable SINS and grade 3 ESCC) [19, 20]. Other surgical techniques such as percutaneous radiofrequency ablation has demonstrated a therapeutic role for maintaining stability and treating focal pain in patients with isolated focal disease and grade 0–1 ESCC in the absence of instability [21]. There is an emerging pattern of surgical interventions according to SINS and the degree of ESCC.

Assessment of a patient’s prognosis influences surgical decision-making and the choice of treatment modality. For example, there is no absolute agreement regarding life expectancy required to justify radical surgical intervention with some authors suggesting that an estimated survival <3 months is considered contraindication to radical resection [22]. Such patients might benefit from a less invasive surgical technique that addresses a focal unstable lesion and alleviates pain. Whether the presence of spinal instability (as defined by SINS) impacts patient survival has been a question of interest in the literature [23,24,25,26,27,28]. This study examined whether spinal metastatic instability was an important predictor of prognosis, and whether it should be taken into account when considering the expected survival of patients. Unlike other publications [23, 24], we did not find an association between SINS and survival which is in agreement with other investigators [27, 28]. Indeed, the assessment of prognosis continues to evolve as more recent studies identified that prognostic scoring systems such as those of Tukahashi and Tomita were less accurate than historically thought [29, 30].

The randomized controlled trial by Patchell and colleagues et al. [31] examined the effect of surgical decompression followed by radiotherapy and demonstrated a significant improvement and maintenance of ambulation post-operatively as opposed to radiotherapy alone in patients with symptomatic ESCC [31]. We included a select patient population excluding patients with intradural metastasis, which is a confounding etiology for neurological deterioration other than severe ESCC and instability. A significantly higher proportion (than expected) of patients with unstable SINS had grade 3 ESCC (69%) with all patients treated surgically utilizing an open surgical technique with instrumentation. This suggests that SINS may assist physicians identify patients with severe symptomatic ESCC on the basis of CT when assessing the components of SINS.

Neurological deterioration may occur in patients with metastatic spinal epidural disease due to spinal instability resulting in mechanical pain and/or severe symptomatic ESCC [32]. This is the first study to examine the relationship between SINS and the functional or neurological status of patients with spinal metastatic disease. Although we observed an increased proportion of patients being wheelchair-bound in the unstable SINS category, this difference did not reach statistical significance possibly due to the lower number of patients. There was no difference in the proportions of patients with severe Frankel grades A–C (useless or absent motor function below the level of the lesion) between the different SINS categories. Only a significantly higher proportion (than expected) of patients with bilateral posterolateral metastatic disease was found to be wheelchair-bound and had Frankel A–C grades. According to Denis’s three-column theory underlying spinal stability, the posterior column is supported by the pedicle, lamina, facets and their capsules as well as the ligamentum flavum, the interspinous and supraspinous ligaments [33]. This finding suggests that metastatic bilateral disruption of these posterior column elements might be a significant contributory factor to disabling pain and/or severe ESCC associated with functional and neurological compromise.

Since SINS was introduced there has been an increase in the reporting of metastatic spinal instability as defined by SINS in the literature [34, 35]. Sahgal and colleagues et al. [35] conducted a multi-institutional study of 410 patients with spinal epidural metastasis showing that SINS predicts high risk features of vertebral compression fractures (VCF) in patients post spine stereotactic body radiotherapy (SBRT). The overall VCF was 14% in their cohort occurring at a median of 2.5 months. Lytic tumor and the presence of spinal deformity were the most predictive of VCF among SINS features and not mechanical pain. The authors argued that their findings “do not reflect the utility of SINS as a tool to communicate spinal instability”, but to predict the risk for VCF post-radiotherapy [35]. Similarly, Lee et al. examined a cohort of 79 patients with spinal metastasis and found a higher rate of 40% of VCF post SBRT [36]. In this study, we observed progression of metastatic spinal instability into unstable SINS in about 9% and 30% of patients with stable and potentially unstable SINS categories, respectively. This observation was noted in patients treated with radiotherapy at an average of 1-year follow-up. Improvement of SINS after radiotherapy occurred due to alleviation of mechanical pain and changes in the osseous morphology of the lesion from lytic into blastic characteristics. Both intra- and extra-compartmental spinal metastases progress into multiple spinal metastases at varying rates of 40% and 64%, respectively. This demonstrates the utility of SINS as a surveillance tool to monitor for progression of spinal neoplastic instability especially in patients treated with radiotherapy. These findings suggest the need for closer follow-up of patients at increased risk of progression of instability in order to facilitate an intervention at a timely fashion.

The use of SINS in assessment of spinal metastasis could influence the pattern of clinical referrals and treatment decisions and help identify patients early in need for surgical or radiation treatments. Indeed, the Spine Study Oncology Group (SOSG) recommends consultation with a spine surgical expert in patients with SINS > 7 (i.e., potentially unstable disease) [12]. In one retrospective cohort of 38 patients matched to 76 control patients, high SINS scores were associated with increased risk for radiotherapy failure defined as retreatment for clinical symptoms (commonly pain) resulting from the index lesion; the odds ratio of radiotherapy failure for potentially unstable and unstable lesions were 5.9 and 12.8, respectively [37]. Interestingly, Gallizia et al. found a significant correlation between pain response and higher SINS scores in 121 patients with spinal metastases (p = 0.007) [38]. The authors suggested that SINS might be used to identify patients who can benefit the most from radiotherapy. Furthermore, Verteeg et al. [39] studied the impact of the initial SINS publication [12] on SINS median score difference between patients treated surgically as compared with those treated with radiotherapy. There was a significant difference in median SINS scores between the two groups before and after SINS was published; median scores decreased significantly from 11.2 to 10.3 and from 8.4 to 7.2 in the surgical and radiotherapy groups, respectively. The authors suggested that awareness of spinal instability has been enhanced after the introduction of SINS, and that might result in earlier identification of spinal metastatic instability with timely surgical referrals.

Limitations

We report the results of a retrospective single-center study in which we observed a deficiency in clinical data for some patients. Survival data was not available in 60 patients, which were subsequently excluded from the respective analysis. In addition, the total number of patients treated surgically was 59 out of 285 patients, which necessitates caution when interpreting surgical results. Our data collection period overlapped with the publication of SINS. Indeed, in our institution the routine application of SINS assessment did not commence until after the study period, which attempts to highlight its value in clinical practice. Furthermore, all imaging studies were evaluated by a single author (A.D.) without inter- or intra-observer reliability assessments; however, SINS has been shown to have near-perfect inter- and intra-observation differences among spine surgeons [14, 40, 41]. Lastly, our neurological grading system was based on the Frankel grading system which, although similar, has largely been replaced by the American Spinal Injury Association (ASIA) Impairment Scale.

Conclusions

This study highlights the clinical role of SINS in the identification and monitoring of spinal metastatic epidural disease as well as its interaction with other clinical factors important in treatment decision-making. We observed a significantly higher proportion (than expected) of patients with unstable SINS requiring spinal instrumentation. SINS interacted with ESCC, and while a higher proportion of non-ambulatory patients were observed to have an unstable SINS category, this relationship was not significant. Although spinal instability causes significant disability, it was not associated with prognosis in this cohort. Assessment of SINS, ESCC, the neurological and functional status as well as prognosis might aid in allocating the appropriate treatment modality to patients with spinal metastatic epidural disease. In addition, SINS can be used as an important tool for surveillance of spinal metastatic instability, especially in cases initially deemed non-surgical.