The radiological outcomes of one-stage posterior-only hemivertebra resection and short segmental fusion for lumbosacral hemivertebra: a minimum of 5 years of follow-up

After general anesthesia, the patient was placed in a prone position on a Jackson table, and a standard midline incision was performed. The posterior elements of the LSHV and the adjacent normal vertebrae that needed to be fused were carefully exposed. The pedicle screws were inserted into the adjacent normal vertebrae using the freehand technique. The posterior elements of the LSHV, including the facet joints, laminae, transverse processes, and posterior parts of the pedicle, were subsequently excised. After that, the lateral cortex of the LSHV was carefully exposed by blunt dissection. To stabilize the spine, a precontoured rod was provisionally screwed on the concave side and left unlocked. Then, the body of the hemivertebra was removed completely, followed by excision of the adjacent discs. Similarly, the contralateral facet and disc were also removed completely to obtain circumferential release. Cancellous bone from the hemivertebra was used for interbody fusion. Then, the convex side was gradually compressed to close the gap completely. In cases where a large gap was left due to the excision of a large hemivertebra, interbody fusion with a cage was used. During the whole procedure, the dural sac and the nerve roots were cautiously protected. The coronal balance and lumbosacral horizontality were checked under fluoroscopy before wound closure. All of the surgeries were performed under the neuromonitoring of sensory evoked potential (SEP) and motor evoked potential (MEP).

All radiographs were analyzed by two authors and the average values were calculated. Standing erect posteroanterior and lateral radiographs of the whole spine were evaluated before surgery, immediately after surgery, 2 years after surgery, and at the last follow-up. Three-dimensional computed tomography (CT) reconstruction was reviewed to identify the location and segmentation of the hemivertebrae preoperatively. Preoperative magnetic resonance imaging (MRI) of the whole spine was used to record the associated intraspinal malformations.

The parameters measured in the coronal plane included the Cobb angles of the lumbosacral and compensatory curves, the upper instrumented vertebra (UIV) tilt and trunk shift. Lumbosacral lordosis, lumbar lordosis, thoracic kyphosis, and the sagittal vertical axis (SVA) were measured in the sagittal plane. Lumbosacral scoliosis and lordosis were defined as the angle between the upper endplate of the vertebra above the hemivertebra and that of the sacrum [7]. The UIV tilt was measured as the angle between the superior endplate of the UIV and the horizontal line. Trunk shift was defined as the horizontal distance between the plumb line drawn from the middle of the C7 body and the central sacral vertical line (CSVL). Lumbar lordosis was measured as the angle between the superior endplate of L1 and the superior endplate of the sacrum. Thoracic kyphosis was defined as the angle between the superior endplate of T5 and the inferior endplate of T12 (Fig. 1).

All parameters were analyzed with standardized statistical software (SPSS; version 22). Continuous values were described as the mean ± standard deviation. Parameters at different time periods were compared using the paired Student’s t test. Correlation analysis was used to identify the relationships between the decrement of UIV tilt and the improvement in trunk shift preoperatively to postoperatively immediately. A P value < 0.05 was regarded as statistically significant.

There were 14 boys and 9 girls with an average age of 7.8 ± 3.5 years (range, 2.5–13.0 years) at surgery. The hemivertebrae were located at L5 in 10 patients, L5–S1 in 11 patients, and L6–S1 in 2 patients. Among all of them, twenty-one (91.3%) patients were identified with a grade 0 Risser sign and the other 2 (8.7%) had grade 1 at presentation. The mean follow-up time was 88.6 ± 28.5 months (range, 60–156 months). At the last follow-up, 4 (17.4%) patients showed a grade 0 Risser, 2 (8.7%) were grade 1, 3 (13.0%) were grade 4, and 14 (60.9%) were grade 5.

Of the 23 patients, 11 patients were identified to have fully segmented hemivertebrae and 12 with semi-segmented hemivertebrae. The fusion span averaged 3.0 ± 0.7 levels (range, 2–4 levels), including 6 patients with 2 levels, 12 with 3 levels, and 5 with 4 levels. Interbody cage fusion was performed in 2 patients. The mean operation time was 196.7 ± 13.2 min (range, 175–230 min), and the average amount of blood loss was 271.7 ± 32.3 ml (range, 200–320 ml). One patient was identified with tethered cord and syringomyelia (Table 1).

The lumbosacral curve averaged 30.7 ± 10.4° before surgery, 6.7 ± 7.1° (79.1% correction, P < 0.001) immediately after surgery, 7.3 ± 6.1° (76.9% correction) 2 years after surgery, and 8.1 ± 7.0° (73.3% correction) at the last follow-up. The UIV tilt significantly improved from 15.3 ± 6.4° before surgery to 3.8 ± 4.3° (P < 0.001) immediately after surgery, 4.0 ± 5.5° 2 years after surgery, and 4.6 ± 6.7° at the last follow-up. Accordingly, trunk shift was significantly improved from 27.3 ± 8.6 mm preoperatively to 11.7 ± 9.4 mm immediately postoperatively, 10.8 ± 8.2 mm 2 years postoperation, and 10.4 ± 8.8 mm at the last follow-up (Table 2, Fig. 2). Interestingly, the decrease in trunk shift was significantly correlated with the change in the UIV tilt (r = 0.615, P = 0.002).

Transient neurologic complications were observed in one patient. After 3 months of conservative treatment, this patient recovered completely. No infection, instrumentation breakage, or pseudoarthrosis was observed during the long-term follow-up. Another patient encountered S1 screw loosening on the right side at the 1-year follow-up and eventually received revision surgery.