Impact of Surgical Resection Without Spinal Fusion for Thoracic Dumbbell Tumors on Postoperative Global Spinal Sagittal Alignment and Clinical Outcomes

Abstract

Study Design

A retrospective comparative study.

Objectives

This study investigated radiographical changes in global spinal sagittal alignment (GSSA) and clinical outcomes after tumor resection without spinal fusion in patients with thoracic dumbbell tumors.

Methods

Thirty patients with thoracic dumbbell tumors who were followed up for at least 3 years were included in this study. Variations in the outcome variables were analyzed using individual GSSA parameters measured on radiography. Clinical outcomes were assessed using the modified McCormick scale (MMCS), Japan Orthopaedic Association (JOA) score, and visual analog scale (VAS). To assess the impact of the affected levels on these outcomes, we divided the patients into three groups according to the location of the tumor (upper [T1-4], middle [T5-8], or lower [T9-12] thoracic spine).

Results

The GSSA parameters (cervical lordosis, T1 slope, thoracic kyphosis [global, upper, middle, and lower], thoracolumbar kyphosis, lumbar lordosis, sacral slope, pelvic incidence, and pelvic tilt) of all the patients did not change significantly after surgery. Eleven of thirty patients had preoperative gait disturbances but they could walk without support (MMCS grade I or II) at the final follow-up. The JOA score and VAS showed significant postoperative improvements. No statistically significant differences were observed in each postoperative sagittal profile or clinical outcome between the upper, middle, and lower groups.

Conclusions

Tumor resection without spinal fusion did not affect the various GSSA parameters and resulted in satisfactory clinical outcomes, indicating that spinal fusion may not always be necessary when resecting thoracic dumbbell tumors.

Keywords

spinal cord tumor thoracic dumbbell tumor spinal sagittal alignment thoracic kyphosis modified McCormick scale (MMCS)Japanese orthopaedic association score

Introduction

Dumbbell-shaped tumors are relatively rare neoplasms, with an incidence rate of 18% of all spinal cord tumors.¹ The tumors most commonly occur in the cervical spine (44%), followed by the thoracic spine.¹ To date, several surgical approaches, including anterolateral one-way, posterior one-way, and combined approaches, have been considered for resecting such tumors.^2-8 In particular, a posterior one-way approach, such as laminectomy with unilateral facetectomy, is a standard technique that allows the safe removal of the intra- or extraspinal and foraminal components of the tumors with sufficient space after direct decompression at the level of tumor localization.^3,5,7 However, this approach might cause postoperative spinal deformities, such as local instability or subsequent kyphosis. This is because it causes more extensive damage to the posterior elements of the spinal column than that caused by the usual resection of thoracic spinal cord tumors localized to the spinal canal.^3,4,6 Although concomitant spinal fusion is occasionally required to prevent post-laminectomy deformity and stabilize the spine after tumor resection, it may cause postoperative complications related to spinal instrumentation.^9-11 Thus, the appropriate surgical procedure for resecting thoracic dumbbell tumors is still controversial.

Postoperative global spinal sagittal alignment (GSSA) has recently been regarded as an increasingly important subject associated with clinical and functional outcomes in patients with spinal cord tumors.^12-15 Some reports have shown that spinal deformity occurs at a rate of approximately 10% following surgery for spinal cord tumors.^1,3,12 Postoperative spinal sagittal malalignment can lead to neurological deterioration, pain, and posture problems.^12-16 However, several studies have focused on assessing only the changes in sagittal alignment around the levels of tumor resection and have not examined the postoperative changes in overall GSSA.^13,14,17 As no studies have examined radiographical changes in GSSA after surgical resection without spinal fusion in patients with thoracic dumbbell tumors, investigating the clinical association of the procedure with the GSSA parameters is essential to identify the benefits of the surgical procedure.

This study aimed to examine radiographical changes in various GSSA profiles and clinical outcomes following surgical resection without spinal fusion in patients with thoracic dumbbell tumors and evaluate the possible factors that influence postoperative GSSA.

Materials and Methods

Study Design, Patient Demographics, and Characteristics

Thirty patients with thoracic dumbbell tumors who required primary resection via the posterior approach without spinal instrumentation at our institution between 2007 and 2017 were included and consecutive in this study. For tumor resection, all patients underwent spinous process-splitting laminectomy (SPSL) or hemilaminectomy (HL) with unilateral facetectomy (approximately one-half to two-thirds of the inside facet joint was removed) at the affected level, depending on the extent of the tumor. The inclusion criteria were as follows: (1) diagnosis of thoracic dumbbell tumor (Eden type I, intradural and extradural; II, intradural, extradural, and paravertebral; III, extradural and paravertebral; or IV, foraminal and paravertebral¹⁸) at 18 years of age or above; (2) absence of congenital vertebral deformity, trauma, neuromuscular disorders, or other pathological conditions; (3) no history of spinal surgery; and (4) a minimum of 3-year follow-up after surgery. The exclusion criteria were as follows: history of postoperative radiotherapy and/or chemotherapy and presence of cervical or lumbar spinal cord tumors with the center of the tumor located above C7 or below L1, intramedullary spinal cord tumors, non-dumbbell tumors localized to only the spinal canal, and extradural spinal cord tumors, such as metastatic tumors. To analyze the impact of the affected levels on postoperative GSSA parameters and clinical outcomes, the 30 patients who met the criteria for the study were divided into three groups based on tumor location (in which the central region of the tumor was located at the level of upper, T1-4; middle, T5-8; or lower, T9-12) as previously reported for assessing ossification of the posterior longitudinal ligament and thoracic spinal cord tumors.^14,19 All information regarding demographics (age at surgery, sex, Eden classification, tumor location, and histopathology of tumor), imaging, and surgical data (surgical time, estimated blood loss [EBL], surgical procedure, rate of tumor resection, postoperative complications, and postoperative follow-up period) were retrospectively obtained by reviewing the patient’s records and radiological images.

This study was approved by the Ethics and Institutional Review Board (Keio university school of medicine, approval number: #20110142), and all participants provided informed consent to be included in the study before tumor resection.

Radiographical and Magnetic Resonance Imaging Data

All patients in this study underwent whole-spine radiological assessment at the preoperative and final follow-up stages. Radiographic findings included standing erect whole-spine posteroanterior and lateral radiographs. To obtain lateral radiographs, the patients stood looking straight ahead with their knees locked, feet shoulder-width apart, elbows bent, and knuckles in the supraclavicular fossa bilaterally.²⁰

We examined the following GSSA profiles: C2-7 cervical lordosis (CL), T1 slope, T1-12 global thoracic kyphosis (GTK), T10–L2 thoracolumbar kyphosis (TLK), L1-5 lumbar lordosis (LL), sacral slope (SS), pelvic incidence (PI), and pelvic tilt (PT). GTK was subdivided into three segments, including T1-4 upper thoracic kyphosis (UTK), T5-8 middle thoracic kyphosis (MTK), and T9-12 lower thoracic kyphosis (LTK). Three of the study authors independently analyzed the radiographs. An experienced spine surgeon reviewed the medical records and radiographs of all the patients. The intraobserver reliability tested using the intraclass correlation coefficient (ICC; 1, 1) and ICC (1, 3) formula was .83 and .92, respectively. To assess the interobserver error, 30 randomly chosen radiographs were assessed by three spine surgeons who were not directly involved in the surgeries independently. The interobserver reliability tested using the ICC (2, 1) and ICC (2, 3) formula was .85 and .93, respectively.

MRI was performed postoperatively to confirm the extent of tumor resection, which was classified into two types according to the operative records: gross total resection (GTR) and partial resection (PR). The standard definition of GTR was used in this study (i.e., 100% removal of the tumor based on the absence of a remnant tumor documented microscopically and no evidence of neoplasm at the site of tumor origin). The procedure was considered PR when a small tumor fragment or an extraforaminal or paravertebral portion was not resected, based on a documented removal of 50%–99% of the tumor on intraoperative ultrasonography or postoperative MRI findings.²¹

Clinical Outcomes and Functional Evaluation

Clinical outcomes and functional effectiveness were assessed at the preoperative and final follow-up stages using the Japanese Orthopaedic Association (JOA) scoring system (11 possible points), and the degree of chest-to-toe pain or numbness was assessed using the visual analog scale (VAS; the score ranges from 0 to 100, with lower scores indicating a comfortable condition without any pain or numbness). The JOA score recovery rate (%) was calculated as follows: (final follow-up JOA score − preoperative JOA score)/(11 − preoperative JOA score) × 100. We also evaluated the degree of gait disturbance preoperatively and at the final follow-up using the modified McCormick scale (MMCS; grade I, normal gait; grade II, mild gait disturbance not requiring support; grade III, gait disturbance requiring support; grade IV, requiring assistance; and grade V, wheelchair needed).^22,23

Statistical Analysis

All continuous data are expressed as means ± standard error of the means, and categorical data are presented as percentages. Radiographical findings and clinical outcomes were compared between the preoperative and final follow-up stages using the Wilcoxon signed-rank test, between the upper, middle, and lower groups using the Kruskal–Wallis test, followed by Dunn’s multiple comparisons post-hoc test, and between SPSL and HL groups using the Mann–Whitney U test, respectively. P values <.05, <.01, and <.001 were considered statistically significant. All statistical analyses were performed using SPSS version 26.0 (IBM Corp., Armonk, NY, USA).

Results

Patients’ Demographics, Clinical Characteristics, and Surgical Data

The current study included 17 men (56.7%) and 13 women (43.3%) with ages at surgery ranging from 20 to 86 years (mean age, 50.1 ± 17.2 years). According to the Eden classification, 2 cases (6.7%) were type I, 9 (30.0%) were type II, 12 (40.0%) were type III, and 7 (23.3%) were type IV. The most frequent tumor locations were the T8-9 and T9-10 levels (5 cases [16.7%]).

The histopathologic findings of the thoracic dumbbell tumors were schwannoma in 27 cases (90.0%), meningioma in two (6.7%), and neurofibroma in one (3.3%). The mean number of resected laminae was 2.2 ± .3. The mean surgical time was 190.2 ± 45.0 (range, 88-279) min, and the EBL was 132.9 ± 151.5 (range, 5-725) mL. Regarding the surgical procedure, 15 patients (50.0%) underwent SPSL with facetectomy, and the remaining 15 (50.0%) underwent HL with facetectomy. Although GTR was attempted for all the study participants, it was achieved in 23 patients (76.7%), whereas seven patients (23.3%) resulted in PR. For one patient, the paravertebral component of the tumor was extremely significant; therefore, surgical resection and thoracic surgery were performed concomitantly. Four patients experienced postoperative complications. Two patients (6.7%) developed a wound infection. Tumor recurrence and regrowth after surgery, which required additional tumor resection, occurred in two patients (6.7%). The mean postoperative follow-up period was 48.8 ± 29.9 (range, 24-150) months. The JOA score (preoperation, 9.3 ± 2.1 vs. final, 10.7 ± .7; P < .001) and VAS (preoperation, 53.8 ± 23.1 vs. final, 1.7 ± 4.6; P < .001) significantly improved after tumor removal, and the JOA score recovery rate was 74.5 ± 45.3% at the final follow-up (Table 1).

Table 1.

Demographics, Clinical Characteristics, and Surgical Data of all Patients.

Patients with Thoracic Dumbbell Tumors	n = 30
Age at surgery (y/o)	50.1 ± 17.2
Sex (n, %)
Men	17 (56.7%)
Women	13 (43.3%)
Eden classification (n, %)
Type I	2 (6.7%)
Type II	9 (30.0%)
Type III	12 (40.0%)
Type IV	7 (23.3%)
Tumor location (n, %)
Upper (n = 6)
T1-2	3 (10.0%)
T2-3	2 (6.7%)
T3-4	1 (3.3%)
T4-5	1 (3.3%)
Middle (n = 14)
T5-6	3 (10.0%)
T6-7	2 (6.7%)
T7-8	3 (10.0%)
T8-9	5 (16.7%)
Lower (n = 10)
T9-10	5 (16.7%)
T10-11	3 (10.0%)
T11-12	2 (6.7%)
Histopathology of tumor (n, %)
Schwannoma	27 (90.0%)
Meningioma	2 (6.7%)
Neurofibroma	1 (3.3%)
Number of resected laminae	2.2 ± 0.3
Surgical time (min)	190.2 ± 45.0
Estimated blood loss (mL)	133.0 ± 151.5
Surgical procedure (n, %)
SPSL + facetectomy	15 (50.0%)
HL + facetectomy	15 (50.0%)
Resection (n, %)
GTR	23 (76.7%)
PR	7 (23.3%)
Postoperative complication (n, %)
Wound infection	2 (6.7%)
Tumor recurrence or regrowth	2 (6.7%)
Postoperative follow-up period (months)	48.8 ± 29.9
JOA score (/11)
Preoperation	9.3 ± 2.1
Final	10.7 ± 0.7
P-value (Preoperation vs Final)	<.001^***
JOA score recovery rate (%)	74.5 ± 45.3
VAS
Preoperation	53.8 ± 23.1
Final	1.7 ± 4.6
P-value (Preoperation vs Final)	<.001^***

Values indicate mean ± standard error; ^***P < .001, statistically significant difference. SPSL, spinous process-splitting laminectomy; HL, Hemilaminectomy; GTR, gross total resection; PR, partial resection; JOA, Japan Orthopaedics Association; VAS, visual analog scale.

Postoperative Changes in Sagittal Spinal Alignment

The mean values of the cervical (CL, P = .917; T1 slope, P = .363), thoracic (GTK, P = .087; UTK, P = .363; MTK, P = .108; LTK, P = .328; TLK, P = .133), and lumbosacral sagittal (LL, P = .660; SS, P = .116; PI, P = .917; PT, P = .748) parameters did not change in all the patients after tumor resection, indicating no statistically significant difference between the preoperative and final follow-up results (Table 2). At the final follow-up, no evidence of new spondylolisthesis or thoracic scoliosis development at the level of tumor removal was observed.

Table 2.

Comparison of Preoperative and Final Follow-Up Spinal Sagittal Parameters of the Patients.

Parameter	Preopearion	Final	P-value
CL (C2-7, °)	8.5 ± 10.2	10.1 ± 9.7	.917
T1 slope (°)	21.3 ± 11.0	25.0 ± 9.3	.363
GTK (T1-12, °)	41.4 ± 10.8	43.8 ± 9.6	.087
UTK (T1-4, °)	15.9 ± 6.2	15.5 ± 7.2	.363
MTK (T5-8, °)	21.7 ± 7.1	24.1 ± 7.5	.108
LTK (T9-12, °)	7.3 ± 4.3	8.3 ± 4.5	.328
TLK (T10-L2, °)	5.3 ± 6.5	8.0 ± 7.1	.133
LL (L1-5, °)	36.0 ± 9.6	36.1 ± 11.3	.660
SS (°)	30.7 ± 6.5	30.8 ± 6.5	.116
PI (°)	25.3 ± 24.6	24.5 ± 24.5	.917
PT (°)	16.6 ± 5.0	16.9 ± 7.0	.748

Values indicate mean ± standard error. CL, cervical lordosis; GTK, global thoracic kyphosis; UTK, upper thoracic kyphosis; MTK, middle thoracic kyphosis; LTK, lower thoracic kyphosis; TLK, thoracolumbar kyphosis; LL, lumbar lordosis; SS, sacral slope; PI, pelvic incidence; PT, pelvic tilt.

We investigated whether age at surgery or individual preoperative thoracic kyphosis parameters influenced postoperative changes in GTK, UTK, MTK, or LTK (Table 3). We found no statistically significant correlation between postoperative changes in individual thoracic kyphosis parameters and the age at surgery (GTK: r = −.356, P = .233; UTK: r = −.172, P = .573; MTK: r = −.358, P = .236; and LTK: r = −.310, P = .294) or individual preoperative thoracic kyphosis profiles (GTK: r = −.289, P = .338; UTK: r = .171, P = .577; MTK: r = −.364, P = .247; LTK: r = −.275, P = .364).

Table 3.

Correlation Analysis Between the Postoperative Changes in Individual Thoracic Kyphosis Parameters and Age at Surgery or Preoperative UTK, MTK, LTK, or GTK.

Parameter	Item	r	P-value
Changes for GTK	Age	−.356	.233
Changes for GTK	Preoperative GTK	−.289	.338
Changes for UTK	Age	−.172	.573
Changes for UTK	Preoperative UTK	.171	.577
Changes for MTK	Age	−.358	.236
Changes for MTK	Preoperative MTK	−.364	.247
Changes for LTK	Age	−.310	.294
Changes for LTK	Preoperative LTK	−.275	.364

GTK, global thoracic kyphosis; UTK, upper thoracic kyphosis; MTK, middle thoracic kyphosis; LTK, lower thoracic kyphosis.

Changes in Each Modified McCormick Scale Grade After Thoracic Dumbbell Tumor Resection Without Spinal Fusion

Preoperatively, 19 patients (63.3%) presented with MMCS grade I, 7 (23.3%) with grade II, 3 (10.0%) with grade III, and 1 (3.4%) with grade IV. At the final follow-up, 28 patients (93.3%) presented with MMCS grade I and 2 (6.7%) with grade II. Regarding changes in postoperative gait status according to each preoperative MMCS grade, 6 of 7 patients (85.7%) with preoperative MMCS grade II, 2 of 3 patients (66.7%) with preoperative MMCS grade III, and the only patient (100.0%) with preoperative MMCS grade IV presented with no gait disturbance (grade I) at the final follow-up. None of the patients in this study had worse postoperative gait disturbance (Table 4).

Table 4.

Summary of Changes in the MMCS Grade of Patients at the Final Follow-Up.

MMCS grade (Preoperation)	MMCS grade (Final)					Total
MMCS grade (Preoperation)	I	II	III	IV	V	Total
I	19	0	0	0	0	19
II	6	1	0	0	0	7
III	2	1	0	0	0	3
IV	1	0	0	0	0	1
V	0	0	0	0	0	0
Total	28	2	0	0	0	30

MMCS, modified McCormick scale.

Comparison of Each Sagittal Spinal Profile and Clinical Outcomes Between the Upper, Middle, and Lower Groups

According to the tumor localization level, six patients were in the upper group, fourteen in the middle group, and ten in the lower group. The individual spinal sagittal parameters did not change significantly after tumor resection in the three groups. In addition, no statistically significant group differences were observed in the preoperative and postoperative sagittal spinal parameters (Table 5).

Table 5.

Comparison of Each Sagittal Spinal Parameter Between the Upper, Middle, and Lower Groups.

	Upper (n = 6)	Middle (n = 14)	Lower (n = 10)	P-value (Upper vs Middle)	P-value (Upper vs Lower)	P-value (Middle vs Lower)
CL (C2-7, °)
Preoperation	6.2 ± 3.9	10.9 ± 9.2	7.8 ± 12.2	.508	.940	.632
Final	10.0 ± 10.2	11.9 ± 8.6	9.8 ± 11.6	.904	.998	.887
P-value (Preoperation vs Final)	.180	.463	.686
T1 Slope (°)
Preoperation	21.2 ± 2.8	21.3 ± 11.9	21.2 ± 12.7	.872	.967	.955
Final	26.0 ± 1.8	22.6 ± 6.9	25.6 ± 11.6	.963	.980	.997
P-value (Preoperation vs Final)	.180	.917	.686
GTK (T1-12, °)
Preoperation	40.2 ± 3.8	45.4 ± 4.4	37.0 ± 16.5	.769	.909	.413
Final	49.5 ± 0.4	43.8 ± 5.9	41.6 ± 14.5	.688	.522	.927
P-value (Preoperation vs Final)	.109	.753	.345
UTK (T1-4, °)
Preoperation	20.3 ± 4.0	15.0 ± 6.1	15.3 ± 7.3	.463	.524	.996
Final	24.8 ± 0.7	13.1 ± 5.7	14.7 ± 7.9	.061	.105	.910
P-value (Preoperation vs Final)	.109	.093	.500
MTK (T5-8, °)
Preoperation	24.7 ± 4.0	24.5 ± 3.8	17.1 ± 9.4	.998	.270	.177
Final	24.6 ± 1.6	27.7 ± 7.8	19.5 ± 6.7	.782	.567	.147
P-value (Preoperation vs Final)	.593	.293	.225
LTK (T9-12, °)
Preoperation	8.0 ± 4.3	8.1 ± 4.1	6.0 ± 5.1	.998	.820	.722
Final	6.0 ± 6.2	7.7 ± 4.0	10.1 ± 4.8	.849	.439	.656
P-value (Preoperation vs Final)	.109	.917	.080
TLK (T10–L2, °)
Preoperation	1.2 ± 0.4	6.8 ± 8.5	4.7 ± 4.1	.187	.645	.529
Final	3.2 ± 0.8	10.0 ± 10.1	7.3 ± 3.5	.500	.716	.916
P-value (Preoperation vs Final)	.183	.753	.138
LL (L1-5, °)
Preoperation	42.6 ± 5.4	34.9 ± 12.0	35.4 ± 7.4	.638	.653	.998
Final	43.4 ± 6.4	33.8 ± 14.6	36.3 ± 8.8	.556	.716	.956
P-value (Preoperation vs Final)	.191	.701	.600
SS (°)
Preoperation	27.7 ± 0.1	30.6 ± 7.4	32.1 ± 3.4	.736	.473	.837
Final	29.8 ± 0.9	29.9 ± 9.2	31.2 ± 3.5	.917	.955	.703
P-value (Preoperation vs Final)	.204	.075	.345
PI (°)
Preoperation	43.7 ± 0.8	46.2 ± 8.5	45.6 ± 13.7	.990	.743	.558
Final	48.5 ± 1.1	43.8 ± 9.8	43.8 ± 13.5	.838	.970	.622
P-value (Preoperation vs Final)	.331	.345	.686
PT (°)
Preoperation	16.0 ± 0.8	15.6 ± 6.0	18.3 ± 4.5	.759	.380	.069
Final	18.7 ± 0.1	14.8 ± 9.1	18.5 ± 5.6	.694	.902	.331
P-value (Preoperation vs Final)	.217	.173	.893

The mean JOA scores improved significantly after tumor resection in the middle (P = .024) and lower (P = .018) groups. In contrast, the mean preoperative JOA score was high in the upper group preoperatively and remained unchanged after surgery (P = .317). Moreover, the JOA scores (postoperative and at final follow-up) and JOA score recovery rates were comparable among the three groups. Analyses of sensory disturbances indicated that VAS significantly improved in each of the three groups postoperatively. No significant preoperative or postoperative differences were observed among the three groups (Table 6).

Table 6.

Comparison of Neurological Status and Clinical Outcomes Between the Upper, Middle, and Lower Groups.

	Upper (n = 6)	Middle (n = 14)	Lower (n = 10)	P-value (Upper vs Middle)	P-value (Upper vs Lower)	P-value (Middle vs Lower)
JOA score
Preoperation	10.8 ± 0.4	8.8 ± 2.5	9.0 ± 1.7	.115	.204	.965
Final	10.9 ± 0.2	10.5 ± 1.0	10.8 ± 0.4	.428	.949	.519
P-value (Preoperation vs Final)	.317	.024^*	.018^*
JOA score recovery rate (%)	50.0 ± 0.0	67.2 ± 57.0	85.7 ± 37.8	.335	.128	.696
VAS
Preoperation	55.0 ± 28.8	47.9 ± 27.2	61.5 ± 8.3	.803	.850	.343
Final	1.7 ± 4.1	1.4 ± 5.3	2.0 ± 4.2	.994	.990	.995
P-value (Preoperation vs Final)	.039^*	.002^**	.005^**

JOA, Japan Orthopaedics Association; VAS, visual analog scale. Values indicate mean ± standard error. *P < .05, ** <.01, statistically significant difference.

Comparison of Each Sagittal Spinal Parameter by Surgical Approach

When the patients were grouped by the SPSL or HL surgical approach, none of the various postoperative spinal sagittal parameters changed significantly in both groups. Furthermore, we did not find any significant group differences between preoperative and postoperative spinal sagittal parameters (Table 7).

Table 7.

Comparison of Each Sagittal Spinal Parameter Between the SPSL and HL Groups.

	SPSL (n = 15)	HL (n = 15)	P-value (SPSL vs HL)
CL (C2-7, °)
Preoperation	12.0 ± 10.9	7.4 ± 9.3	.459
Final	11.8 ± 9.5	10.7 ± 14.4	.802
P-value (Preoperation vs Final)	.843	.576
T1 Slope (°)
Preoperation	24.6 ± 11.0	21.0 ± 10.4	.833
Final	23.8 ± 8.3	27.7 ± 12.0	.601
P-value (Preoperation vs Final)	.881	.330
GTK (T1-12, °)
Preoperation	43.3 ± 11.2	34.9 ± 7.4	.225
Final	44.3 ± 3.2	42.4 ± 9.6	.746
P-value (Preoperation vs Final)	.857	.288
UTK (T1-4, °)
Preoperation	17.0 ± 4.6	12.5 ± 10.6	.458
Final	16.7 ± 6.1	12.0 ± 10.9	.392
P-value (Preoperation vs Final)	.803	.851
MTK (T5-8, °)
Preoperation	22.8 ± 7.4	24.3 ± 6.2	.794
Final	24.3 ± 6.2	23.3 ± 12.6	.711
P-value (Preoperation vs Final)	.656	.728
LTK (T9-12, °)
Preoperation	7.4 ± 4.1	6.8 ± 5.7	.771
Final	7.8 ± 4.6	10.2 ± 4.2	.589
P-value (Preoperation vs Final)	.813	.304
TLK (T10–L2, °)
Preoperation	5.3 ± 4.5	8.7 ± 11.2	.222
Final	5.6 ± 3.3	8.8 ± 10.1	.205
P-value (Preoperation vs Final)	.793	.831
LL (L1-5, °)
Preoperation	39.7 ± 8.1	35.9 ± 10.4	.556
Final	39.7 ± 9.7	34.7 ± 16.0	.508
P-value (Preoperation vs Final)	.894	.811
SS (°)
Preoperation	31.7 ± 5.8	28.8 ± 6.4	.766
Final	29.4 ± 5.9	29.6 ± 7.6	.884
P-value (Preoperation vs Final)	.492	.833
PI (°)
Preoperation	48.1 ± 7.6	47.5 ± 5.0	.791
Final	46.6 ± 8.6	49.5 ± 9.2	.667
P-value (Preoperation vs Final)	.740	.779
PT (°)
Preoperation	16.3 ± 5.8	18.7 ± 4.9	.734
Final	17.2 ± 8.4	19.9 ± 2.7	.565
P-value (Preoperation vs Final)	.733	.749

Values indicate mean ± standard error. SPSL, spinous process-splitting laminectomy; HL, Hemilaminectomy; CL, cervical lordosis; GTK, global thoracic kyphosis; UTK, upper thoracic kyphosis; MTK, middle thoracic kyphosis; LTK, lower thoracic kyphosis; TLK, thoracolumbar kyphosis; LL, lumbar lordosis; SS, sacral slope; PI, pelvic incidence; PT, pelvic tilt.

Representative Case

A 31-year-old woman patient complained of lower extremities numbness and gait disturbance. Her MRI showed a dumbbell-shaped tumor (Eden type II) on the left side at the T10/11 spinal level (Figure 1A). The posterior SPSL approach with facetectomy was selected for tumorectomy, which resulted in GTR. The histopathologic finding of the tumor was schwannoma. Four years postoperatively, her MRI did not indicate any tumor recurrence (Figure 1B), the patient was in good condition without any symptoms. Global spinal sagittal alignment did not show any significant progression of instability or degeneration on 4-year postoperative whole spine standing sagittal radiographs for this patient (Figure 1C–D).

Figure 1.

Representative MR and radiographic images of a patient with thoracic dumbbell tumor. (A–B) The arrow indicated the area where the resected tumor existed at the T10/11 level in the preoperative and postoperative T2-weighted MR images. (C–D) No occurrence of spinal instability or deformity was observed at the lesion level. Whole spine standing sagittal radiographs did not demonstrated significant changes from preoperation to 4-year postoperation (CL, 8.6° to 5.3°; T1 slope, 12.8° to 11.5°; GTK, 28.7° to 33.2°; UTK, 9.1° to 8.0°, MTK, 12.1° to 16.9°; LTK, 7.5° to 8.5°; TLK, 2.3° to 8.1°; LL, 41.3° to 41.2°; SS, 32.2° to 29.1°; PI, 47.3° to 40.1°; PT, 15.1° to 11.0°). CL, cervical lordosis; GTK, global thoracic kyphosis; UTK, upper thoracic kyphosis; MTK, middle thoracic kyphosis; LTK, lower thoracic kyphosis; TLK, thoracolumbar kyphosis; LL, lumbar lordosis; SS, sacral slope; PI, pelvic incidence; PT, pelvic tilt.

Discussion

This study is the first to compare various GSSA parameters before and after surgical resection without spinal fusion in patients with thoracic dumbbell tumors. We found that individual GSSA parameters were unaffected by tumor removal, and clinical outcomes significantly improved. In addition, when the participants were grouped on the basis of tumor distribution (upper, middle, and lower thoracic spine), no significant group differences were observed in individual preoperative and postoperative GSSA profiles. Our findings have important implications for studies that aim to evaluate the utility of tumor resection without spinal fusion in patients with thoracic dumbbell tumors as no studies have examined changes in postoperative GSSA parameters.

During the resection of thoracic dumbbell tumors, the facet joint must be occasionally sacrificed to remove the intra- or extra-foraminal and paravertebral portion of the tumor, which may cause postoperative spinal sagittal malalignment.^2-5 Furthermore, several reports have shown that removing more than two laminae or performing facetectomy during an initial surgery would increase the incidence of postoperative spinal deformity.^4,8-10 In contrast, Aizawa et al. reported that 10 patients with thoracic dumbbell tumors who underwent hemilaminectomy and facetectomy without spinal fusion had limited postoperative changes in local kyphosis and scoliosis angles on radiographs.⁵ Rong et al. reported that 14 patients with thoracic dumbbell tumors had no aggravating deformity or instability after surgery.²⁴ Moreover, Ishikawa et al. examined the mid-to-long-term clinical and radiological outcomes of patients who underwent unilateral total facetectomy without instrumented fusion for the resection of thoracic dumbbell tumors. They indicated that no patient demonstrated postoperative instability or spinal deformity that required additional surgery.⁷ However, these studies focused on assessing only the radiographical changes in local sagittal parameters around the levels of tumor resection and did not examine changes in the overall GSSA. Our study showed that various GSSA parameters did not change significantly after surgery, and the surgical outcomes were satisfactory, suggesting that resection without spinal fusion is valuable for improving the outcomes of patients with thoracic dumbbell tumors.

Previous reports have demonstrated that the development of postoperative spinal deformity would depend on the tumor localization level.^13,14,16,17 Particularly, concerning thoracic spinal cord tumors, Kobayashi et al. reported that sagittal spinal alignment did not change postoperatively in patients who underwent middle and lower thoracic intraspinal tumor resection, even without spinal fusion. In contrast, postoperative kyphosis may increase when the tumor is located in the upper thoracic spine.¹⁴ Ishikawa et al. reported that postoperative radiographic measurements and clinical outcomes were comparable between patients with thoracolumbar dumbbell tumor and those with middle thoracic dumbbell tumors after tumor resection without spinal fusion.⁷ Furthermore, in this study, no significant differences were observed in both preoperative and postoperative GSSA parameters of the upper, middle, and lower groups. These discrepancies among the studies may be due to differences in the number of resected laminae. In particular, the upper thoracic levels (T1-4) are included in the cervical–thoracic junction (CTJ), which is in the C7-T3/4 region, and the CTJ is strongly correlated with postoperative instability owing to its anatomical features. Moreover, the transition from the mobile articulated cervical spine, which is lordotic, to the rigid thoracic spine, which is kyphotic in the CTJ, creates a mechanical vulnerability that may increase postoperatively.²⁵ In intraspinal canal tumors, the number of resected laminae is necessarily significant because the tumor grows and invades the spinal canal craniocaudally. In contrast, dumbbell-shaped tumors often require the resection of a relatively small number of laminae because the tumor invasion into the spinal canal is less and into the intervertebral foramen and paravertebral bodies is extensive. Thus, it is likely that the number of resected laminae due to these developmental tumor patterns could influence the surgical invasion of the posterior supporting tissues and contribute to progressive kyphosis.

Concomitant spinal fusion with instrumentation is occasionally required to prevent post-laminectomy deformities or instability in patients with dumbbell-shaped spinal cord tumors. Moreover, if postoperative spinal sagittal alignment and clinical outcomes deteriorate, fixation could be considered as a second-stage surgery for the thoracic spine after careful observation.^9,10 This procedure has several disadvantages, including blurred postoperative MRI findings due to metallic artifacts and biofilm formation around the spinal instrumentation.¹¹ In addition, more than three vertebrae must be fused using bilateral fixation in some cases, resulting in the loss of mobility in the additional mobile segment.³ Unfortunately, no patient had undergone tumor resection with spinal fixation for thoracic dumbbell tumors at our institution. Thus, we could not compare the postoperative surgical results and the effects on GSSA between patients treated with and without spinal fusion. Therefore, comparative studies should be conducted in future.

In patients with thoracic dumbbell tumors, the association between changes in GSSA on radiography after surgical resection and related clinical symptoms remains unclear.^2-5 Ishikawa et al. showed in their study that VAS for pain and the Oswestry disability index (ODI) for patient-reported outcomes were acceptable in 80% of the patients with thoracic dumbbell tumors who underwent tumor resection without spinal fusion. Furthermore, the SF-36 scores were comparable or better than the reference values of the standard Japanese population. Moreover, disc and facet joint degeneration, observed on MRI, did not progress after surgery.⁷ In the present study, no patient showed changes in global sagittal balance or complained of low back pain from the period after surgery until the final follow-up. Unfortunately, we could not evaluate whether there was a significant difference in health-related quality of life (HRQOL) before and after tumor removal. Furthermore, we did not use preoperative or postoperative patient-based questionnaires. Therefore, future investigations on the clinical correlates of HRQOL, ODI, and physiologic GSSA preservation would be helpful to validate our clinical findings and identify the actual benefits of surgical resection without fusion in patients with thoracic dumbbell tumors. Moreover, we intend to include this investigation in a future prospective study.

Our study had several significant limitations. First, it used a retrospective approach, which inevitably lowered the level of evidence. Second, the sample size was relatively small, and the statistical power was insufficient to draw conclusions about the precise clinical outcomes of patients with thoracic dumbbell tumors. Moreover, thoracic sagittal alignment, such as GTK, worsened postoperatively in only three of 30 patients, which is too low to allow a statistical comparison between the worsening and non-worsening groups. Third, we did not include a comparison group that underwent fixation; thus, we could not show the superiority of surgical resection without spinal fusion. Finally, the postoperative follow-up period in this study was relatively short to investigate the development of postoperative spinal malalignment. Therefore, future evaluations with a long-term observation period are necessary to validate our proposed surgical procedure. Despite these limitations, this study showed actual results concerning the absence of notable changes in various GSSA parameters in patients with thoracic dumbbell tumors after tumor resection without spinal fusion. The significance of the present study is highlighted by the unprecedented analyses that focused on the presence or absence of variations in each postoperative GSSA profile in patients with thoracic dumbbell tumors who underwent tumor resection without spinal fusion.

Conclusion

The present study evaluated the changes in various GSSA parameters and clinical outcomes after tumor resection without spinal fusion in patients with thoracic dumbbell tumors. This surgical procedure did not affect the individual postoperative GSSA profiles investigated in this study, and it maintained acceptable sagittal alignment with satisfactory functional and clinical outcomes at the final follow-up. However, as the number of patients in this study was relatively small, future studies with larger sample sizes are necessary to provide more precise conclusions.

Footnotes

Acknowledgments

The authors would like to thank Editage () for the English language review.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Toshiki Okubo

Narihito Nagoshi

References

Ozawa

Kokubun

Aizawa

Hoshikawa

Kawahara

. Spinal dumbbell tumors: An analysis of a series of 118 cases. J Neurosurg Spine. 2007;7(6):587-593. doi:10.3171/SPI-07/12/587

Thorat

Rajendra

Thirugnanam

IHB

. Single-stage posterior midline approach for dumbbell tumors of the thoracic spine, with intraoperative CT guidance. Surg Neurol Int. 2011;2:31. doi:10.4103/2152-7806.77272

Ando

Imagama

Ito

, et al. Removal of thoracic dumbbell tumors through a single-stage posterior approach: Its usefulness and limitations. J Orthop Sci. 2013;18:380-387. doi:10.1007/s00776-013-0370-9

Dong

. Minimally invasive resection of extradural dumbbell tumors of thoracic spine: Surgical techniques and literature review. Eur Spine J. 2016;25:4108-4115. doi:10.1007/s00586-016-4677-z

Aizawa

Ozawa

Hashimoto

Kanno

Kokubun

Itoi

. Benign thoracic dumbbell tumor can be removed through hemilaminectomy and facetectomy without spinal reconstruction. J Neurosurg Sci. 2016;60:145-148.

Liu

Zhang

Zhu

. Surgical strategy for spinal dumbbell tumors: A new classification and surgical outcomes. Spine (Phila Pa 1976). 2017;42(12):E748-E754. doi:10.1097/BRS.0000000000001945

Ishikawa

Ohashi

Hirano

, et al. Mid- to long-term outcomes after resection of thoracic dumbbell tumors managed by laminectomy and unilateral total facetectomy without instrumented fusion. Global Spine J. 2021;13(3):771-780. doi:10.1177/21925682211008836

McCormick

. Surgical management of dumbbell tumors of the cervical spine. Neurosurgery. 1996;38(2):294-300. doi:10.1097/00006123-199602000-00012

Ando

Imagama

Ito

, et al. Unilateral instrumented fixation for cervical dumbbell tumors. J Orthop Surg Res. 2014;9(1):2. doi:10.1186/1749-799X-9-2

10.

Simon

Auerbach

Garg

Sutton

Telfeian

Dormans

. Efficacy of spinal instrumentation and fusion in the prevention of post-laminectomy spinal deformity in children with intramedullary spinal cord tumors. J Pediatr Orthop. 2008;28(2):244-249. doi:10.1097/BPO.0b013e3181623819

11.

Cummings

. Recurrent meningitis secondary to infection after spinal arthrodesis with instrumentation. A case report. J Bone Jt Surg Am. 1998;80(5):722-724. doi:10.2106/00004623-199805000-00013

12.

Montano

Trevisi

Cioni

, et al. The role of laminoplasty in preventing spinal deformity in adult patients submitted to resection of an intradural spinal tumor. Case series and literature review. Clin Neurol Neurosurg. 2014;125:69-74. doi:10.1016/j.clineuro.2014.07.024

13.

Nori

Iwanami

Yasuda

, et al. Risk factor analysis of kyphotic malalignment after cervical intramedullary tumor resection in adults. J Neurosurg Spine. 2017;27(5):518-527. doi:10.3171/2017.4.SPINE16956

14.

Kobayashi

Kawabata

Nishiyama

, et al. Changes in sagittal alignment after surgical excision of thoracic spinal cord tumors in adults. Spinal Cord. 2019;57(5):380-387. doi:10.1038/s41393-018-0235-3

15.

Okubo

Nagoshi

Tsuji

, et al. Spinous process-splitting laminectomy approach for tumor excision at conus medullaris or cauda equina level results in satisfactory clinical outcomes without affecting global spinal sagittal alignment. Global Spine J. 2021;13:1745. doi:10.1177/21925682211047460

16.

McGirt

Garc'es-Ambrossi

Parker

, et al. Short-term progressive spinal deformity following laminoplasty versus laminectomy for resection of intradural spinal tumors: Analysis of 238 patients. Neurosurgery. 2010;66(5):1005-1012. doi:10.1227/01.NEU.0000367721.73220.C9

17.

Papagelopoulos

Peterson

Ebersold

Emmanuel

Choudhury

Quast

. Spinal column deformity and instability after lumbar or thoracolumbar laminectomy for intraspinal tumors in children and young adults. Spine (Phila Pa 1976). 1997;15(22):442-451. doi:10.1097/00007632-199702150-00019

18.

Eden

. The dumb-bell tumours of the spine. British J Surg. 1941;28:549-570.

19.

Aizawa

Sato

Ozawa

, et al. Sagittal alignment changes after thoracic laminectomy in adults. J Neurosurg Spine. 2008;8(6):510-516. doi:10.3171/SPI/2008/8/6/510

20.

O’Brien

Kuklo

Blanke

, et al. Radiographic Measurement Manual. USA: Spinal Deformity Study Group; 2018.

21.

Okubo

Nagoshi

Tsuji

, et al. Imaging characteristics and surgical outcomes in patients with intraspinal solitary fibrous tumor/hemangiopericytoma: A retrospective cohort study. Global Spine J. 2021;13:276. doi:10.1177/2192568221994799

22.

Manzano

Green

Vanni

Levi

. Contemporary management of adult intramedullary spinal tumors-pathology and neurological outcomes related to surgical resection. Spinal Cord. 2008;46(8):540. doi:10.1038/sc.2008.51

23.

Matsuyama

Sakai

Katayama

, et al. Surgical results of intramedullary spinal cord tumor with spinal cord monitoring to guide extent of resection. J Neurosurg Spine. 2009;10:404-413. doi:10.3171/2009.2.SPINE08698

24.

Rong

Fan

, et al. Management of dumbbell and paraspinal tumors of the thoracic spine using a single-stage posterolateral approach: Case series. Orthop Surg. 2018;10:343-349. doi:10.1111/os.12405

25.

Vaccaro

Cotler

Lin

. Spinal disorders at the cervicothoracic junction. Spine (Phila Pa 1976). 1994;19:2557-2564. doi: 10.1097/00007632-199411001-00011