As advances in chemotherapy prolong the life expectancy of patients with solid tumors, the frequency of spinal metastases is likely to increase. Survival is a key factor physicians consider when selecting a treatment for spinal metastases. Various assessment systems can predict the length of survival and facilitate the selection of the most appropriate treatment. The assessment systems, however, differ with respect to the parameters assessed and the significance assigned to each parameter in the total score. Spinal palsy is one of the prognostic parameters in the Tokuhashi scoring system [1, 2] but is not included in the Tomita scoring system [3]. A limitation of these scoring systems is that studies examining them have included different tumor types.

The aim of this study was to evaluate the usefulness of preoperative neurological status as a prognostic factor in non-small-cell lung cancer patients with spinal metastases who underwent surgical treatment.

The mean patient age was 61.6 years (range, 20–87 years), and there were 34 males and 16 females. The tumor sites included the thoracic spine (n = 28), lumbar spine (n = 12), thoracolumbar spine (n = 6), cervical spine (n = 3), and sacrum (n = 1). Adenocarcinoma (32 patients) was the most common histological type, followed by squamous cell carcinoma (9 patients). The mean intraoperative blood loss volume was 975 mL (range, 350–6500 mL), and the mean surgical time was 4.8 h.

Neurologic improvement by ≥1 Frankel grade was noted in 37 of the 50 cases (74 %). Twelve patients showed no improvement, and 1 patient showed deterioration from Frankel grade B to grade A. Overall, 68 % of patients (34/50) were ambulatory after surgery. Twenty-two of 40 (55 %) nonambulatory (Frankel B/C) patients became ambulatory (Frankel D/E).

One patient developed symptomatic tumor recurrence at the level of previous decompression, and 3 patients developed new symptomatic spinal cord compressions because of noncontiguous metastases. These patients underwent repeat decompressive surgeries.

Of 12 observed complications (Table 1), 11 were surgery related. There was no intraoperative mortality; however, 3 patients died during the postoperative period. One patient died from respiratory failure 14 days after surgery, 1 died from hepatic and respiratory failure 23 days after surgery, and 1 died from nonsurgery-related sigmoid colon perforation and sepsis 37 days after surgery. The median postoperative survival time was 7.5 months (95 % CI: 4.2–10.9 months). The Kaplan-Meier curve (Fig. 1) showed that 58 % (29/50) of the patients survived >6 months. The Frankel A + B group (palsy score 0 in Tokuhashi system) contained 8 patients, and the Frankel C + D group (palsy score 1 in Tokuhashi system) contained 42 patients. The median survival was 2.5 months (95 % CI: 1.22–16.3 months) in the Frankel A + B group and 8.0 months (95 % CI: 5.52–9.89 months) in the Frankel C + D group (Fig. 2; p = 0.87).

Complication	Number of patients
Neurologic progression	1
Wound dehiscence	1
Wound infection	3
Respiratory failure	2
CSF leakage	2
Sigmoid colon perforation	1
30-day mortality	2

CSF cerebrospinal fluid

Tokuhashi et al. [2] stated that the average survival period was longer (10.4 ± 13.6 months) in patients without neurologic deficits than in patients with complete palsy (3.7 ± 3.9 months). Therefore, they included “spinal cord palsy” as a prognostic parameter in their study. Several authors have reported that patients with paralysis at presentation or posttreatment have a considerably shorter life expectancy than ambulatory patients [6–10]. Prasad and Schiff [11] reported that other than the nature of the primary tumor, the presence of paraparesis prior to surgery had the most detrimental effect on survival.

In 2001, Tomita et al. [3] developed a scoring system that does not use neurologic status as a prognostic factor for survival in patients with spinal metastases. The authors described that a long survival period can be possible with appropriate treatment, even in cases with paraplegia, and suggested that spinal cord decompression can improve paralytic conditions. Patients with paralysis tend to have shorter survival because of cancer progression and not due to the paralysis itself [3]. Spiegel et al. [12] reported that neurologic deficits did not significantly influence survival of melanoma patients. North et al. [13] observed that the preoperative ambulatory status predicted the duration of postoperative ambulation but was only marginally associated with survival. Yamashita et al. [14] used the revised Tokuhashi scoring system to predict survival in patients with spinal metastases and found that Frankel grade was not significantly associated with survival. Kumar et al. [15] studied 87 patients with spinal metastases from nasopharyngeal cancer and found the modified Tokuhashi score was the best to predict prognosis; however, neurological status had no significant association with survival. Quraishi et al. [16] studied the effect of surgical timing on neurological outcome and survival in spinal metastases patients and found that earlier surgical treatment resulted in significantly better neurological outcomes. However, the timing of surgery did not influence survival.

Leithner et al. [17] compared 7 preoperative prognostic scoring systems for spinal metastases, including the Bauer [18], modified Bauer [19], Tokuhashi [2], revised Tokuhashi [1], Tomita [3], van der Linden [20], and Sioutos [21] scoring systems. In their analyses, primary tumor and visceral metastases were the only parameters significantly associated with survival. Their results did not show pretreatment neurological status as a prognostic factor; therefore, the authors did not consider paralysis as a predictive of survival [17, 22]. Wibmer et al. [23] evaluated the same preoperative scoring systems and found that primary tumor, status of visceral metastases, and systemic therapy were significantly associated with survival. Leithner et al. [17] and Wibmer et al. [23] further concluded that the Bauer and modified Bauer scoring systems are the most reliable systems for prediction of survival. The modified Bauer scoring system includes 4 positive prognostic factors: absence of visceral metastases, solitary skeletal metastasis, non-primary lung cancer, and primary tumor of the breast or kidney, lymphoma, or myeloma. It does not include preoperative neurological palsy as a parameter.

The parameters of the revised Tokuhashi scoring system include the patient’s general condition, number of extraspinal bone metastases, number of metastases in the vertebral bodies, presence of metastases in the major internal organs, the primary site of the cancer, and the presence of palsy. The first 5 parameters are all associated with disease severity, but palsy score is not. In our study, the survivals of the Frankel A + B and Frankel C + D groups were not statistically different. One patient with preoperative Frankel B status improved to Frankel D postoperatively and survived for 16 months. One patient with preoperative Frankel B status improved to Frankel C postoperatively and survived for 36 months. Patients with preoperative Frankel B neurological status can still survive for a long duration.

These results may be because paralytic condition is not associated with disease severity, and paralytic condition can be improved with adequate spinal cord decompression. In a patient with multiple spine metastases, palsy might be absent. However, palsy might be noted in a patient with only 1 vertebral metastasis. A patient with multiple spinal metastases generally has higher disease severity than a patient with a single vertebral metastasis. Also, the number of vertebrae involved and the pattern of spinal cord compression might be the same in patients with different palsy score (Figs. 3 and 4). Deterioration of neurological status might occur in 1 week, and the survival time should not differ too much in such situation. After adequate surgical decompression and stabilization, palsy should be reversed in all patients except those with a neurological status of Frankel A, or those with a poor overall medical status [24]. Thus, palsy should not be a major prognostic factor in lung cancer patients with spinal metastasis who underwent spinal surgery. Patients with paralysis tend to have shorter survival because of cancer progression, not due to the paralysis itself [3]. The duration of survival largely depends on the disease severity and the ability of other modalities (such as chemotherapy, targeted therapy, or radiotherapy) to control the tumor [25–28].

This study has some limitations that should be considered. First, as a retrospective study from a single center, several potential biases may exist, including referral bias and patient characteristics. Second, the sample size was small. Third, we did not include all parameters in the analyses. However, this is the first study that focused only on non-small-cell lung cancer patients with spinal metastases who underwent spinal surgery. The result of this study should be valuable in the decision of treatment of non-small-cell lung cancer patients with spinal metastases.

Preoperative palsy score had no statistically significant association with survival in non-small-cell lung cancer patients with spinal metastases who underwent spinal surgery.