The incidence of spinal metastases is increasing due to early detection and advances in the treatment of the primary tumour [1]. Spinal metastases can result in pathological fracture leading to instability back pain or even paralysis due to metastatic epidural spinal cord compression (MESCC). Instability back pain and/or neurological deficit can significantly affect quality of life [2]. Conventionally, patients with advanced disease are treated with nonoperative palliative care [34]. However, recent studies have demonstrated better quality of life, longer survival and improved neurological function in patients with MESCC who are treated with open decompressive surgery followed by radiotherapy, compared to radiotherapy alone [56]. Open surgical procedures are associated with significant risks and morbidities. Open surgery requires extensive muscle dissection with a higher risk of infection, increased blood loss, significant postoperative pain, longer hospitalization and time for functional recovery [789]. With the development of minimally invasive stabilization (MISt) using percutaneous pedicle screws, the morbidity associated with open surgery can be avoided. The case illustration in Fig. 1 is a good example illustrating the advantages using MISt in advanced spinal metastatic patients. The purpose of this paper is to report the outcome of 50 patients with spinal metastases who were treated with MISt using fluoroscopic-guided percutaneous pedicle screws with/without minimally invasive decompression.

This was a prospective study carried out between January 2008 and December 2013. Eighty one consecutive spinal metastatic patients who presented with pathological fracture(s) associated with severe instability pain with/without neurological deficit. The selection criterion for surgery included severe instability pain with presence of pathological fracture that did not improve with external orthosis support with/without neurological deficit secondary to MESCC. Instability pain was defined as mechanical pain in the presence of radiological evidence of pathological fracture(s) resulting in difficulty/limitation in turning the trunk in supine position, propping up in bed and/or sitting up from a lying position. The exclusion criteria were presence of neurological deficit without instability pain, poor candidates for general anaesthesia secondary to poor premorbid condition after assessment by an anaesthesiologist or extensive metastatic disease resulting in end organ failure (i.e., deranged liver functions, renal failure or brain with altered sensorium), pathological fracture(s) involving the cervical or sacral spine, isolated single level lesion when excision surgery was possible, poor visualization of pedicles in fluoroscopic images, sclerotic metastatic lesions and patient refusal for surgery. After exclusion, 52 patients were recruited. They underwent MISt with/without minimally invasive decompression surgery. Two patients were lost to follow up. The remaining 50 patients comprised 25 males and 25 females, with a mean age of 56.3 years (range, 16–82 years). The primary tumors were breast (34.0%, 17 cases), lung (22.0%, 11 cases), kidney (10.0%, 5 cases), prostate (10.0%, 5 cases), colorectal (8.0%, 4 cases) and others (16.0%, 8 cases). The demographic data and cases included are presented in Tables 1 and 2.

The mean Tomita score was 6.3±2.4. The mean number of vertebra levels with pathological fractures was 1.8±0.8. Of the 50 patients, 19 (38.0%) involved single level pathological fracture, 22 (44.0%) involved two levels and 9 (18.0%) involved three levels or more. Most (80.0%) presented with various grade of neurological deficits. Lesions at thoracic spine (52.5%) was the main region causing the neurological deficit followed by thoracolumbar spine (40.0%) (Table 1, Fig. 2).

Thirty seven patients (74.0%) required decompression surgery in addition to MISt. The average operation time was 3.1 hours (range, 1–7 hours). When MISt without decompression was carried out, the mean operating time was 2.3 hours (range, 1–3 hours) while the mean operating time was 3.4 hours (range, 1.5–7 hours) when decompression was needed. The average blood loss was 1.4±1.0 litres. Blood loss for MISt was 0.4±0.2 L and was 1.7±0.9 L for MISt plus decompression. The average number of instrumented vertebrae was 7.8 with the longest instrumentation spanning across 15 levels (Fig. 1, Table 3). The average number of screws inserted was 8 per patient. Twelve cases (24.0%) displayed significant anterior column deficiency and required anterior column augmentation with either cement augmentation or cage (Table 3).

There was no postoperative infection, wound dehiscence, pulmonary complications, screw related injury or early death (<30 days) noted in this study (Table 3). However, there was one case of implant failure (case 31) diagnosed with renal cell carcinoma (Table 2). This case had a breakage of both rods at the thoracolumbar junction 27 months after the index surgery. The broken implants were revised using domino connectors. The patient has survived 51 months at the time of manuscript submission. None of patients had tumour recurrence at decompression site leading to recurrence or worsening of neurology. However, 2 patients had recurrence at a different site from the index level, which was decompressed. One patient required repeated decompression while the other received radiotherapy.

Overall, the mean preoperative, 2-week and 3-month postoperative pain score (VAS) was 7.9±1.4, 2.5±1.2 and 2.8±1.3, respectively. There was significant pain reduction 2 weeks and 3 months after surgery compared to preoperatively (both p<0.001). Three months after surgery, the VAS for pain was no different from the 2-week postoperative value (p=0.249). The 40 patients who presented with neurological deficit had an average of one Frankel grade recovery at final review; 70% had improvement of 1 grade and 5% had an improvement of 2 grades. None of the patients had worsening of neurology after surgery (Table 3). Of the 10 patients who did not have neurological improvement, 50% presented with complete paralysis (Frankel A) preoperatively (Fig. 3). None of the patients were bed-ridden after the surgery. The average time to ambulation was 3.4±1.8 days.

The overall mean survival was 11.3 months (range, 2–51 months). During the final review, 39 patients had died, with an average survival of 8.5 months (range, 2–36 months), and 11 patients were still alive with an average survival of 21.5 months (range, 8–51 months). Based on the Kaplan-Meier survival analysis, the probability of survival was 0.54 at 6 month and 0.39 at 1 year (Fig. 4). When compared the patients with Tomita score <8 and those with Tomita score ≥ 8, the pre- and postoperation pain score was higher for the latter group (Table 4). The Tomita <8 group had slightly shorter time to ambulation as compared to the Tomita ≥8 group, but they were not significantly different. The group with the Tomita score <8 had a significant longer survival as compared to the group with the Tomita score ≥8, with a mean survival of 14.1±12.5 months and 6.8±4.9 months, respectively (p=0.019) (Table 4).

The incidence of skeletal metastases is as high as 70% in autopsy of cancer patients and spine is the most common site of bony metastases [1112]. Most patients with spinal metastases present with instability back pain due to pathological fracture(s) [1314]. Approximately 5-10% present with spinal cord compression resulting in neurological deficit [1516]. Both instability pain and neurological deficit can significantly affect quality of life [2]. The patients are likely to be bed-bound because of intractable pain and neurological deficits. Immobilization in bed can lead to complications including orthostatic pneumonia, pressure sore, urinary tract infection, thromboembolism and joint contractures [1718].

The treatment for advanced spinal metastases (patients with Tomita score ≥8) often involves nonoperative palliative care due to their grave prognosis. Even patients with spinal metastases who present at an early stage with isolated spinal metastasis can rarely be treated with radical resection [3]. This is because any spread of the primary tumour to the spine is often considered the end stage of the disease. The decision for surgery is governed by three main factors: life expectancy, surgical indication and surgical risk. Life expectancy depends on the nature of the primary tumour, the extent of the metastases and the patients' general condition [34]. The indications for surgery are presence of spinal instability with/without neurological deficit [1920]. However, the benefits of the surgery should outweigh the surgical risks when deciding for surgical intervention in this group of patients.

Life expectancy plays a major role when deciding surgery for spinal metastatic patients. Most surgeons recommend that surgery should be offered when life expectancy exceeds 3 months [31921]. Tokuhashi et al. [4] suggested that surgery should be reserved for patients with expected survival of more than 6 months. These timing decisions are decided mainly because open procedures carry significant morbidity and mortality in patients with advanced spinal metastasis. However, in the absence of any surgical intervention, they suffer from intractable instability pain with an ever-ongoing poor quality of life. They will be more likely to bed-bound and to develop many complications as a result of immobilization in bed due to severe instability pain. In this study, there were 7 patients (14.0%) that survived only 2 months after the index surgery. The main indication of surgery for these patients was to palliate and relieve their instability pain. MISt surgery had achieved significant reduction of pain in our study as evidenced from the pre- and postoperation pain score of VAS 7.9 and 2.5, respectively (p<0.001). Despite having 2 months of survival, these patients could sit up and had better ambulation after stabilization of their pathological fractures. The average duration to ambulate after surgery in this study was 3.4 days. This rapid recovery from surgery is attributed to the minimal muscle disruption in MISt [822]. Despite the short life expectancy in this group of patients, we believed that MISt can prevent them from being bed ridden and thus avoiding complications associated with immobilization. Even a short duration of survival with preservation of their ability to ambulate is very meaningful for these dying patients.

In this study, patients with Tomita score ≥8 had mean survival of 6.8 months. This result compared favorably to a prior study, which reported a mean survival of 5.9 months for patients with a score ≥8 [3]. The finding of longer survival in our patients that had undergone MISt with/without minimally decompression is consistent with prior results [56]. The longer survival could be attributed to the preservation of the ability to ambulate following stabilization of their pathological fracture through MISt and, therefore, reduction in complications of immobilization. Presently, we also noted that patients with Tomita score ≥8 had significant more pain than those patients with Tomita score <8. Therefore, MISt can benefit these patients with Tomita score ≥8 despite their poorer prognosis.

Open surgery in patients with spinal metastases is associated with higher complication rate [2123]. The patients often display a poorer general condition as a consequence of their primary disease process as well as the adverse effects of chemotherapy and radiotherapy. The postoperative wound infection rate can be up to 12% for this group of patients [524]. If the patient has underwent radiotherapy prior to the surgery, the risk of dehiscence and wound infection is as high as 32% [24]. Minimal invasive surgery reduces the infection rate because it reduces the physiological insult, decreases the recovery time and minimizes morbidity associated with more conventional open approach. O'Toole et al. concluded in their study that minimally invasive spinal surgery may reduce postoperative wound infections as much as 10-fold compared to open surgery [25]. In our study, there was no wound breakdown or infection in all 50 patients. In the case illustration (Fig. 1), we also demonstrated the versatility of MISt in avoiding an irradiated skin that is at higher risk of dehiscence. With the advancements of this surgical technique and instrumentation, MISt offers a new treatment option in patients with spinal metastases.

Patients with a low Tomita score (<6) are expected to have 2 or more years of life expectancy [3]. The surgical strategy recommended by Tomita et al. [3] for this group of patient is excision. In our series, there were 20 patients with a Tomita score <6. The low score was due to their slow to moderate growth of primary tumours such as breast, prostate and kidney cancers, despite having multiple bony metastases. We excluded those patients with isolated or single level lesions when excision surgery was possible. Although they had a protracted prognosis, palliative surgery (i.e., MISt) was performed for them due to multiple lesions. All patients were given postoperative radiotherapy for local control.

The MISt technique allows the diseased spine to be effectively stabilized. However, this technique does not allow fusion of the diseased segment. The durability of the implants in patients with low Tomita score (<6) is questionable. Of the 20 patients with low score, 5 survived longer than 24 months. One patient with a Tomita score 7 survived longer than 24 months (case 36). Only one implant failure happened in our series (case 31, Tomita score 4), giving rise to 2.0% failure rate (overall) and 16.7% (1/6) failure in those with protracted survival. In this case, the implant failure occurred at the mobile junctional segment of the spine (i.e., the thoracolumbar junction). The reason for the failure was attributed to the long survival (51 months) of the patient. Therefore, patients with potentially long survival i.e., Tomita score<6 will have to be informed that there is a risk of implants failure in the event they survive beyond 24 months. In such cases minimally invasive fusion in addition to MISt via minimally invasive facet joint fusion or minimally invasive percutaneous BMP application could be safely performed [2627].

In our series of patients who survived 3 months or less, three had Tomita score <6. None of them died because of surgery-related complications. The first patient (case 30) had renal cell carcinoma with Tomita score of 4. He had Frankel A paralysis at presentation and his neurology did not improved despite decompression surgery. He passed away due to nosocomial pneumonia. The second patient (case 47) had multiple myeloma and a Tomita score of 4. This patient died of renal failure. The last patient (case 49) was diagnosed with nasopharyngeal carcinoma. He had a Tomita score of 3. His cause of death was respiratory failure secondary to uncontrolled locally advanced lymph nodes involvement.

There were a few limitations in this study. The primary tumours in these 50 patients with spinal metastasis were heterogenous. Different primary tumour would have different biological behavior and prognosis. The Spine Instability Neoplastic Score is the consensus from the Spine Oncology Study Group published in 2010 for assessment of instability in spinal metastasis [28]. This score was not used in our study because about half of the patients in this study were recruited from 2008-2010. There was also lack of assessment for quality of life in this study.

Percutaneous pedicle screw stabilization offers a treatment option for patients with spinal metastasis with spinal instability with or without neurological deficit secondary to MESCC. It provides significant relieve of pain with low risk of complications even in patients with advanced metastatic disease.