INTRODUCTION
Endoscopic posterior thoracic approaches have been employed to treat thoracic myelopathy caused by ossification of the ligamentum flavum (OLF) [
1-
8]. However, endoscopic methods have limitations in addressing severely progressed thoracic OLF, particularly in cases with extensive dural ossification. In such instances, microscopic surgery is preferred, often requiring a wide laminectomy and additional instrumentation [
3,
4,
9-
11].
To minimize injury to surrounding structures, we are exploring the use of the unilateral biportal endoscopic (UBE) system for treating advanced thoracic OLF with extensive dural ossification. UBE employs independent 2-handed techniques like those in microscopic surgery, while significantly enhancing surgical visibility [
8,
12-
15]. By leveraging the advantages of UBE surgery, we successfully manage the various types of dural ossification [
16] associated with thoracic OLF according to preplanned surgical strategies based on preoperative imaging.
This study aims to present surgical strategies informed by preoperative computed tomography (CT) findings during UBE thoracic OLF surgery, while also evaluating the concordance between preoperative CT images and intraoperative video observations.
RESULTS
We included 34 patients (14 men and 20 women) who underwent biportal endoscopic posterior thoracic laminectomy. The mean age of the patients was 68.1±9.1 years (
Table 1). The mean follow-up period was 12.7±3.9 months, and the average length of hospital stay was 5.9±1.6 days (
Table 1). A total of 41 segments were involved, with the operating spinal levels indicated in
Table 1. The mean operative time per segment was 64.5±17.3 minutes.
Three patients demonstrated an epidural hematoma on postoperative MRI. One patient was successfully managed with conservative treatment, while 2 patients required microscopic hematoma removal due to severe pain at the surgical site and progression of neurological deficits in both legs. Following revision surgery, the acute symptoms were alleviated, and the patients recovered without any additional neurological deficits. One patient experienced a wrong-level surgery, which was identified on postoperative MRI, necessitating an additional procedure to address the targeted lesion. No serious intraoperative neurological injuries, such as those resulting from direct spinal cord contusion or high-pressure irrigation, were observed. Fur-thermore, segmental instability did not occur, despite most surgeries being performed in the lower thoracic and thoracolumbar junction areas.
VAS scores for back and leg pain showed significant improvement following surgery (
Table 1). Neurological function was also restored, as evidenced by an increase in the JOA score (preoperative: 12.3±1.0; final follow-up: 15.4±0.7;
Table 1).
Among the 41 operative levels, the OLF types were classified according to the Sato classification [
19]: 14 levels were categorized as type A, 18 as type B, 5 as type C, 2 as type D, and 2 as type E (
Table 1). Nine levels exhibited the ‘tram-track sign’ of dural ossification on preoperative CT images, indicating dural ossification (
Table 2,
Fig. 1). During the review of intraoperative videos for cases exhibiting the tram-track sign, focal dural ossification was observed to be fused to the OLF mass in 5 levels (1 classified as Sato type B and 4 as Sato type C); In 3 cases, the dural ossification located in the outer dural layer was removed together with the OLF mass by peeling it away from the inner layer of the dura mater (
Supplementary Video Clip 1). Two levels demonstrated ossification of the entire dural layer, resulting in the ossified lesion being left intact and floated focally to facilitate dural expansion (
Supplementary Video Clip 2). However, Dural ossification was not clearly visible in the surgical videos during the removal of the OLF mass in 4 levels exhibited the ‘tramtrack sign,’ all classified as Sato type B (
Table 2); instead, only adhesive tissue was observed between the dura and the OLF mass. In these cases, it was suspected that the dural ossification was not fused to the OLF mass, allowing for clear separation of the OLF mass from the ossified dura. As a result, in all cases with dural ossifications exhibiting the tram-track sign, the OLF mass was successfully removed without resulting in any significant dural defects.
One OLF case exhibiting the ‘bridge sign’ of dural ossification was classified as Sato type D (
Table 2,
Fig. 2). Intraoperative videos revealed that the dural ossification was located in the dorsal portion of the dural sac and was completely fused to the OLF mass; however, the ossified dura did not extend to the lateral dura mater (
Supplementary Video Clips 3). Consequently, the circumferential floating technique was feasible for managing this type of dural ossification.
Two levels exhibited the ‘comma sign’ of dural ossification, classified as Sato type E (
Table 2,
Fig. 3). Surgical videos revealed that the broad ossified dura extended to both the lateral and ventral portions of the dura mater and was completely fused to the OLF mass (
Supplementary Video Clip 4). The ossified dura was circumferentially excised to achieve neural decompression, followed by dural reconstruction. Both patients recovered neurological function without experiencing any instances of CSF leakage postoperatively.
DISCUSSION
In this study, the tram-track sign, indicative of focal dural ossification, was most frequently observed in preoperative CT scans. Focal dural ossification that is not fused to the OLF mass is visualized as the tram-track sign on preoperative CT images; however, completely fused ossified dura could not be confirmed through preoperative imaging. The detection of focal dural ossification on axial CT slices may be inaccurate, and the actual prevalence of unconfirmed focal dural ossification may exceed the findings noted in the preoperative scans.
Furthermore, upon reviewing intraoperative videos, only half of the levels displaying the tram-track sign were confirmed as distinct dural ossification, while 5 out of 9 levels exhibiting the sign did not show distinguishable dural ossification. Nonfused ossified dura could be smoothly separated from the OLF mass, but remaining focal dural ossification was not visualized during surgery. Conversely, fused ossified dura was identified through endoscopic visualization while elevating the OLF mass, irrespective of whether the outer dura was ossified or the entire dural layer was affected (
Supplementary Video Clips 1 and
2). When operating on these cases, it is essential to anticipate the potential for dural involvement, and the management of dural ossification should be adapted to the intraoperative findings, either through removal of the outer dural layer or by employing a focal floating technique. The en bloc removal with a bisected flap technique is considered safer than a single-piece removal pattern, as it allows for direct confirmation of the space between the OLF and the dura mater.
All instances of focal dural ossification exhibiting the tramtrack sign were successfully managed without any dural defects or neurological injuries, irrespective of whether the ossified dura was fused to the OLF. Focal dural ossification may not indicate a significant risk for dural defects during thoracic OLF removal, particularly when associated with OLF classified as Sato type B. However, focal dural ossification in enlarged thoracic OLF (Sato type C) displayed invasive characteristics in the surgical videos (
Supplementary Video Clip 2).
Extensive and aggressive dural ossification, indicated by the ‘comma sign’ and ‘bridge sign,’ is typically associated with aggressive thoracic OLF classified as Sato types D or E. A comprehensive understanding of the signs of dural ossification—particularly regarding their characteristics and the extent of dural involvement—enables us to effectively manage these challenging cases. The benefits of UBE surgery, such as enhanced visualization of the lesion and the use of a free 2-handed technique, can be leveraged in these situations [
1].
The ‘bridge sign’ indicates ossification of the midline portion of the dura mater, which forms a connection between the bilateral OLF. This type of dural ossification typically did not extend to the lateral dura mater (
Supplementary Video Clip 3). The ossified dura is completely fused to the OLF mass, making en bloc removal of the OLF mass infeasible without resulting in a significant dural defect. Therefore, floating the ossified dura attached to the thinned OLF—similar to an island in a vast lake—is a safer and more efficient approach than complete removal.
Most aggressive forms of dural ossification are indicated by the ‘comma sign,’ where the lateral and ventral dura mater is ossified as it crosses the midline. This type of dural ossification is commonly observed in OLF classified as type E, signifying anterior growth of the aggressive OLF. The comma sign encompasses the bridge sign of midline ossification; however, floating the dural ossification is not feasible due to the absence of expandable dura mater at the lateral dural borders. To achieve adequate neural decompression, it is essential to remove the dorsolateral portion of the dural ossification. A critical consideration when removing dural ossification associated with thoracic OLF is that no barrier exists between the OLF and the ossified dura. The spinal cord may be unexpectedly exposed during drilling of the OLF, and once widely exposed, further drilling becomes challenging due to the risk of direct spinal cord contusion or injury from the water pressure of infused saline [
20,
21]. Therefore, it is imperative to assess the extent of the ossified dura bilaterally and cranio-caudally using preoperative CT scans. Important anatomical landmarks include the bilateral facet joints. If the facet joint cartilage is not visualized during drilling, there is an increased risk of encountering the spinal cord. Once the facet joints are displayed, the lateral edge of the dural ossification can be secured along the medial pedicular line [
4] (
Supplementary Video Clip 4). The tip of the SAP extends to the neuroforamen, and the cranial end of the OLF is typically positioned at or slightly cranially to the neuroforamen. The extent of cranial drilling should be guided by the location of the SAP tip.
One case classified as Sato type D did not present clear evidence of dural ossification on preoperative CT; however, unilateral broad dural ossification was discovered during the surgical procedure. The unexpected presence of dural ossification resulted in a dural tear during the
en bloc removal of the OLF mass [
22]. The remaining dural ossification and attached OLF were floated, and the dural defect was subsequently repaired using a sealant patch. When dural ossification is completely fused to the bulky OLF mass, the signs of dural ossification may not be evident on the preoperative CT scan, particularly when the ossified dura is extensively located in the lateral dural area. Therefore, during the removal of OLF in cases classified as Sato types D and E, it is crucial to anticipate the potential presence of advanced dural ossification in all instances and to plan the surgical approach accordingly [
7].
These surgical methods can assist endoscopic spine surgeons in safely performing thoracic OLF removal in cases with dural ossifications. Prior to the treatment of thoracic OLF, a preoperative CT scan should be conducted to confirm the presence of dural ossification and assess its characteristics. Making customized surgical plans based on the findings from the CT scan facilitates successful removal of the thoracic OLF and effective management of the dural ossification.
During endoscopic thoracic decompression surgery, inexperienced surgeons face an increased risk of insufficient decompression when confronting severe conditions [
2]. Inadequate decompression of the spinal canal can result in delayed progression of myelopathy, potentially necessitating revision surgery [
3]. Prioritizing adequate decompression and minimizing complications is more critical than achieving minimal invasiveness, regardless of whether microscopy or endoscopy is utilized. Therefore, I recommend that inexperienced endoscopic spine surgeons consider microscopic surgery when faced with aggressive OLF associated with the comma sign, given the elevated risk of spinal cord injury and the potential for complications due to inadequate decompression. Furthermore, intraoperative neuromonitoring can reduce the risk of spinal cord injury in these cases.
The study has limitations. We focused on the management methods of dural ossification during the removal of thoracic OLF, which means that details of pre- and postoperative symptoms were not thoroughly described. Furthermore, the cases of aggressive dural ossification are limited, and the shapes of dural ossifications vary significantly; thus, the application of the recommended techniques may have limitations. A study with a larger sample size, incorporating a more diverse range of dural ossification shapes, is needed to establish the efficacy of the techniques demonstrated in this study.