INTRODUCTION
Each year, over half of all cancer cases progress to metastasis, with the skeletal system involved in 50% of these cases [2,10]. Among skeletal metastases, approximately 60% affect the spine [18]. In such cases, up to 50% of patients require treatment, with 5-10% necessitating surgical intervention [16]. As cancer treatments continue to advance, extending the survival of patients with metastatic disease, improved patient selection and earlier surgical interventions have made spinal metastases more common [5,18,20].
Most patients with spinal metastases experience pain, spinal deformities, and neurological deficits, further complicated by systemic issues such as chemotherapy, immunosuppression, and malnutrition [5,8,17]. Consequently, the primary goal in managing metastatic spinal tumors is to alleviate symptoms and restore neurological function as part of palliative care [4,7,10]. Despite significant advances in radiology and oncology, surgery remains a cornerstone of treatment for spinal metastases [1,17]. Intralesional resection for decompression to alleviate neurologic deficits is essential for the effective management of spinal metastatic disease [1,4,7]. Although less invasive techniques, such as separation surgery and minimally invasive surgery, have emerged, the evidence supporting their effectiveness remains limited due to incomplete tumor removal, which may result in tumor regrowth before initiating radiotherapy [1,17].
Over 80% of spinal metastases occur in the thoracic (50-60%) and lumbar (30-35%) spine. However, very few evidencebased studies exist that address the optimal surgical approach and instrumentation levels for metastatic spinal tumors [19]. Therefore, this study aims to compare short-segment fixation with traditional long-segment fixation in patients with metastatic tumors in the thoracic and lumbar spine to assess the relative efficacy of these approaches.
MATERIALS AND METHODS
This study was approved by Institutional Review Board (IRB) of Asan Medical Center (IRB No. 2023-1333), which waived the requirement for informed consent because of the study’s retrospective design.
Materials
We conducted a retrospective review of all patients who underwent spinal fusion for symptomatic metastatic spinal tumors at a single tertiary referral hospital. A total of 91 consecutive patients with metastatic thoracic and lumbar spinal tumors who underwent either short-segment fixation with circumferential tumor debulking and anterior support or long-segment fixation without anterior support, followed by posterolateral fusion, were included in the study. The surgeries were performed between April 2014 and December 2022 by two surgeons employing distinct surgical approaches. All patients presented with neurological deficits and intractable pain. Surgical intervention was considered for patients expected to survive more than 3 months, based on their overall health status and the malignancy of their primary cancers. We excluded patients with follow-up radiographs of less than 2 weeks, those who underwent spondylectomy or corpectomy, and those with cervical spine metastases.
Surgical techniques
Short-segment fixation
Under general anesthesia, patients were positioned prone on a Jackson spinal table. A midline incision was made, extending one level above and below the index level. Using the Solera® system (Medtronic Sofamor-Danek, Memphis, TN, USA) guided by a navigation system, pedicle screws (5.5 mm to 7.5 mm in diameter, depending on pedicle size) were inserted at one level above and below the index vertebra. At the index level, a unilateral pedicle screw was placed on the contralateral side, determined based on pedicle osteolysis observed on computed tomography (CT) and tumor lateralization identified on magnetic resonance imaging (Fig. 1A). Temporary rods were then placed on one side.
A unilateral laminectomy and pediculectomy were performed to decompress the spinal cord (Fig. 1B). The pediculectomy provided sufficient access to the ventral aspect of the spinal canal and vertebral body. The osteolytic portion of the vertebral body was excised using an osteotome and reverse-angled curette (Fig. 1C). Following tumor resection and confirmation of ventral spinal cord decompression, the anterior column was supported by inserting 15 cm3 of allograft bone chips and a titanium cage (Fig. 1D). Posterolateral fusion was performed on the remaining unilateral lamina using 15 cm3 of allograft bone chips (Fig. 1E). Finally, intraoperative plain radiographs and CT scans confirmed correct screw and cage placement (Fig. 1F and Supplementary Video 1).
Long-segment fixation
Under general anesthesia, patients were positioned prone on a Jackson spinal table. A midline incision was made, extending two levels above and two levels below the site of tumor invasion. Subperiosteal dissection was performed to expose the spinous processes and laminae. Following laminectomy and/or facetectomy, the epidural tumor compressing the spinal cord was circumferentially excised to achieve separation surgery [12]. Pedicle screws were placed two levels above and two levels below the index vertebra, avoiding the index vertebra. Bilateral posterolateral fusion was performed on the remaining facet joints and/or transverse processes. Intraoperative radiographs were obtained to confirm correct screw placement.
Clinical outcomes
Demographic data, including age, sex, underlying comorbidities (e.g., diabetes, hypertension, cardiovascular disease), and primary tumor origin, were compared between groups. Neurological function was evaluated preoperatively and at the last follow-up using the Frankel grading system (A-E). Neurological outcomes were categorized based on changes in Frankel grades : improvement (an increase of at least one grade), maintenance (no change), or deterioration (a decrease of at least one grade).
Radiologic outcomes
Segmental Cobb angles were measured on plain radiographs immediately after surgery and at the last follow-up, with separate analyses for thoracic and lumbar segments. For thoracic spine lesions, thoracic kyphosis (TK) was measured, while for lumbar lesions, lumbar lordosis (LL) was assessed to evaluate the global curvature. The segmental Cobb angle was defined as the angle between the upper endplate of the uppermost instrumented vertebra and the lower endplate of the lowermost instrumented vertebra in both the short- and long-segment groups (Fig. 2). Loss of correction was assessed by comparing immediate postoperative segmental Cobb angles with those at the last follow-up.
Statistical analysis
Statistical analyses were performed using SPSS software (version 28; IBM SPSS Statistics, Armonk, NY, USA). All tests were two-sided, with a p-value of <0.05 considered statistically significant. Continuous variables were presented as mean±standard deviation (SD) or median (range) and were compared using Student’s t-test or the Wilcoxon rank-sum test. Categorical variables were expressed as counts and percentages and were analyzed using the chi-squared test or Fisher’s exact test.
RESULTS
Demographics
A total of 91 patients were included in this study, with 44 undergoing short-segment fixation and 47 undergoing long-segment fixation (Table 1). No significant differences were observed between the groups in terms of follow-up duration (thoracic short-segment : 8.79±12.86 months vs. thoracic long-segment : 7.31±10.71 months, p=0.620; lumbar short-segment : 6.00±6.63 months vs. lumbar long-segment : 14.80±18.40 months, p=0.350), age (short-segment : 60.3±12.7 years vs. long-segment : 56.5±14.1 years, p=0.181), sex (short-segment : 56.8% male vs. long-segment : 61.7% male, p=0.636), or preoperative comorbidities such as diabetes mellitus (short-segment : 22.7% vs. long-segment : 10.6%, p=0.120), hypertension (short-segment : 22.7% vs. long-segment : 36.2%, p=0.161), cardiovascular disease (short-segment : 0.0% vs. long-segment : 4.3%, p=0.495) and chronic kidney disease (short-segment : 4.6% vs. long-segment : 4.3%, p=0.495). The five most frequent primary tumor sites were the lung (25 patients, 27.5%), liver (15 patients, 16.5%), kidney (nine patients, 9.9%), breast (eight patients, 8.8%), and prostate (seven patients, 7.7%) (Table 1). Other primary cancers included cervix (six patients, 6.6%), lymphoma (six patients, 6.6%), ampulla of Vater (five patients, 5.5%), stomach (four patients, 4.4%), and thyroid (two patients, 4.4%).
Operative characteristics
Long-segment surgery at the thoracic level was performed more frequently (41 patients, 87.2%) than short-segment surgery (28 patients, 63.6%), although this difference was not statistically significant (p=0.942). However, the short-segment procedure had a significantly longer operating time (204.6±44.8 vs. 165.3±49.0 minutes, p<0.001).
Hospital stay and complications
Patients in the short-segment group had a significantly shorter hospital stay (15.5±17.9 days) compared to the long-segment group (27.6±25.2 days, p<0.001). However, no significant differences were found between the groups regarding postoperative complications (short-segment : five patients, 11.4% vs. long-segment : 10 patients, 21.3%; p=0.318) or revision surgery rates (short-segment : two patients, 4.6% vs. long-segment : three patients, 6.4%; p>0.99) (Table 2). Postoperative complications included surgical site infection (two cases in each group), wound issues (one case in the short-segment group, two cases in the long-segment group), pneumonia (no cases in the short-segment group, one case in the long-segment group), epidural hematoma (one case in each group), and recurrence within 3 months (0 cases in the short-segment group, three cases in the long-segment group).
Neurological outcomes
Neurological evaluations (categorized by Frankel grade) from preoperative to final follow-up are shown in Table 3. In the thoracic region, the short-segment group demonstrated a significantly higher rate of neurological improvement compared to the long-segment group (17.9% vs 4.8%, p=0.048). Among the short-segment group, five patients showed improvement (two from grade C to D, and three from grade D to E), whereas only two patients in the long-segment group showed improvement (one from grade C to D, one from grade D to E). One patient in the short-segment group experienced neurological deterioration (from grade C to B), while no deterioration was observed in the long-segment group. In the lumbar region, neither the short-segment (n=16) nor the long-segment group (n=5) showed any neurological changes during the follow-up period, with no significant differences between the groups (p=1.000). All patients maintained their initial neurological status from the preoperative assessment to the final follow-up.
Radiological outcomes
The mean and SD of the Cobb angle were analyzed as continuous variables. Surgical correction was defined as the difference between postoperative and preoperative angles, while loss of correction was defined as the difference between the final follow-up and postoperative angles. Thoracic and lumbar spine measurements were analyzed separately (Table 4). There were no significant differences in TK or LL changes between the two groups. However, for thoracic lesions, long-segment fixation resulted in significantly greater surgical correction compared to short-segment fixation. Nevertheless, the loss of correction was also significantly greater in the long-segment group, resulting in no significant difference between the two groups at the final follow-up. The results are presented in Table 4 and Fig. 3. Although not statistically significant, LL in the short-segment group exhibited greater surgical correction and was maintained throughout the follow-up period.
DISCUSSION
The results revealed no significant radiological differences between the two groups. Patients who underwent long-segment fixation without anterior support showed significant postoperative changes in sagittal alignment, indicating surgical correction. However, considerable loss of correction was observed at the final follow-up, resulting in no significant difference in the final outcome (Table 4 and Fig. 3).
Regarding neurological outcomes, significant improvements were noted in the thoracic spine region. The group that underwent short-segment fixation with anterior support showed significant neurological improvement (Table 3). This is likely due to the anterior support procedure, which inherently involves performing a ventral lesion resection, enabling 360-degree circumferential decompression of the spinal cord.
No significant differences were observed in the lumbar region corresponding to the cauda equina. Additionally, no differences in complications were noted between the two groups. Although the short-segment group had longer operation times, the length of hospital stay was significantly shorter.
These findings suggest that short-segment fixation is neither superior nor inferior to traditional long-segment fixation. While it may seem counterintuitive that short-segment surgery involved longer operating times than long-segment fixation, given that one advantage of short-segment surgery is its shorter operating time [6,11], this discrepancy can be attributed to intraoperative delays related to screw fixation using a navigation system that required two CT scans, as well as additional time for corpectomy procedures [3,9,15]. Despite the longer operating time, the short-segment group had a significantly shorter hospital stay, likely due to the smaller surgical wounds associated with this procedure.
As patients undergoing surgery for metastatic spinal disease are often older and physically frail, their ability to tolerate surgery must be carefully considered [13]. Failure to account for patient vulnerability may lead to substantial surgical challenges and increased morbidity. Therefore, reducing complications, operating time, and recovery duration is crucial to achieving better outcomes in these patients. Traditional long-segment spinal fusion, which typically involves tumor excision with multi-level pedicle screw fixation, requires a long incision, significant blood loss, and extended operating time, potentially increasing morbidity and complication rates, particularly in patients with multiple tumor-related comorbidities [14]. In contrast, short-segment fixation involves fewer motion segments, a smaller incision, and a shorter recovery time while maintaining sagittal alignment [6]. Consequently, patients with spinal metastases may benefit from short-segment fixation due to its less invasive nature and expedited recovery.
Moreover, advancements in metastatic tumor treatments have prolonged patients’ lifespans, increasing the likelihood of requiring multiple surgeries across different spinal levels. For example, one patient in our short-segment group had two metastatic tumors located at the thoracic and lumbar levels, necessitating two separate spinal fusions (Fig. 4). Since the initial short-segment fixation involved only two or three levels, the subsequent surgical procedure was simpler and shorter, as additional vertebrae were available for further instrumentation.
Finally, short-segment fixation may be as effective as separation surgery in achieving complete tumor resection. Although separation surgery is a well-established technique for treating spinal metastases [1], it does not prevent tumor regrowth following spinal decompression, as the tumor is not fully excised. In our study, all patients who underwent long-segment fixation also had separation surgery, and in some cases, tumor regrowth was observed in as little as 3 months after surgery, which is considered short-term recurrence. Conversely, short-segment fixation cases achieved maximal tumor resection, preventing postoperative short-term recurrence. Furthermore, incompletely excised tumors from separation surgery may regrow before radiotherapy can be started, meaning that separation surgery may not be the optimal choice for treating metastatic spinal cancers.
Limitations
This study had several limitations, including potential selection biases inherent in its retrospective design. The small sample size of lumbar long-segment cases limited the statistical power to detect significant outcomes. Additionally, the involvement of multiple surgeons across different periods introduces potential bias. The heterogeneity of tumor types across groups, as well as the evolution of cancer treatments, particularly the introduction of targeted therapies, could have substantially influenced outcomes over the study’s duration. These factors may have affected the results and warrant detailed discussion. Moreover, the authors should acknowledge the selection bias associated with the choice of surgical approaches and consider how advances in chemotherapy and targeted therapies during the study period might have affected outcomes. Lastly, long-term follow-up data were unavailable for most patients, as many succumbed shortly after surgery due to the severity of their disease.
CONCLUSION
Our study demonstrated that short-segment pedicle screw fixation combined with anterior support significantly improves neurological outcomes and reduces hospital stay length for patients with metastatic spinal tumors. Radiological outcomes also indicated that short-segment fixation is not inferior to long-segment fixation. Furthermore, short-segment spinal fusion, when combined with maximal tumor resection, may facilitate rapid recovery, reduce the risk of tumor regrowth, and preserve more operable levels in cases of adjacent tumor invasion. Due to the selection biases inherent in this retrospective study, prospective studies are needed to further elucidate the advantages of short-segment fixation.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation 
Supplement
Print 
