The short-term effectiveness of precise safety decompression via double percutaneous lumbar foraminoplasty and percutaneous endoscopic lumbar decompression for lumbar lateral spinal canal (LLSC) stenosis

Purpose: This prospective study reports on a new technique, namely precise safety decompression via double percutaneous lumbar foraminoplasty (DPLF) and percutaneous endoscopic lumbar decompression (PELD) for lateral lumbar spinal canal (LLSC) stenosis, and its short-term clinical outcomes. Methods: The study analyzed 69 patients with single-level LLSC stenosis simultaneously occurring in both zones 1 and 2 (dened as retrodiscal space and upper bony lateral recess respectively by new LLSC classication) who underwent DPLF–PELD from November 2018 to April 2019. Clinical outcomes were evaluated according to preoperative, three months postoperatively, and last follow-up, via leg pain/low back pain (LBP) visual analog scale (VAS) scores, Oswestry disability index (ODI) scores, and the Macnab criteria. The postoperative MRI and CT were used to conrm the complete decompression, and exion-extension x-rays at the last follow-up were used to observe lumbar stability. Results: All patients successfully underwent DPLF–PELD, and the stenosis was completely decompressed, conrmed by postoperative MRI and CT. The mean follow-up duration was 13 months (range: 8–17 months). The mean preoperative leg pain VAS score is 7.05 ± 1.04 (range 5–9), which decreased to 1.03 ± 0.79(range: 0–3) at three months postoperatively and to 0.75 ± 0.63 (range: 0–2) by the last follow-up visit(p < 0.05). The mean preoperative ODI was 69.8 ± 9.05 (range: 52–85), which decreased to 20.3 ± 5.52 (range: 10–35) at the third month postoperatively and to 19.6 ± 5.21 (range: 10–34) by the nal follow-up visit(p < 0.05). The satisfactory (excellent or good) results were 94.2%. There was one patient with aggravated symptoms, which were relieved after an open surgery. Two patients had a dural tear, and two patients suffered postoperative LBP. No recurrence or segmental instability was observed at the nal follow-up. Conclusion: DPLF–PELD could be a good alternative for the treatment of LLSC stenosis patients whose stenosis occurred in both zones 1 and 2.


Introduction
Along with the aging of society, the incidence of degenerative lumbar disorders has increased, becoming one of the main reasons for lumbar surgery in elderly patients [1,2]{Inoue, 2016 #73}. Owing to the remarkable evolution of percutaneous endoscopic lumbar decompression (PELD), the application of spinal endoscopy is shifting from the treatment of soft disk herniation to complex lumbar spinal stenosis. Satisfactory results of PELD in treating lumbar spinal stenosis (LSS) are reported at 82%-92% [3].
It has been widely accepted that LSS anatomically involved the central canal, lateral recess, foramen, or any combination of these locations [4,5]. However, the concept of a lateral recess still has no universal de nition, and was frequently represented by other ambiguous terms, such as radicular canal, lateral recess zone or nerve root canal [5][6][7]. After carefully analyzing the spinal anatomy and clinical facts, Yu Wang et al. recently rede ned the concept "lateral lumbar spinal canal" (LLSC) and creatively provided a new classi cation of LLSC with ve different zones [8]. We found that the retrodiscal space (zone 1) and upper bony lateral recess (zone 2) were the two most common regions for occurrence of lumbar degeneration. In clinical practice, stenosis simultaneously occurring in both zones 1 and 2 were most common (43.4%) [8]. Unfortunately, endoscopic decompression for patients was di cult, even for experienced endoscopic spine surgeons, due to the complicated compressive situation.
Percutaneous endoscopic lumbar foraminoplasty was initially used to enlarge the foramen by trephine and/or high-speed drill. Thereafter, foraminoplasty was used as an e cient decompressive method for treating LSS [3,9]. The procedure of foraminoplasty was facilitated by changing the speci c location of the needle tip and trajectory of trephine to decompress different compressive pathology. Foraminoplasty was performed in order to resect the upper-ventral aspect of the superior articular process (SAP) in a classical transforaminal endoscopic spine system (TESSYS) technique [10]. However, the removed scale was not enough for stenosis patients with both zones 1 and 2 involvement; additional endoscopic dorsal decompression was required, using a high-speed drill during the operation. The disadvantages were obvious: increased surgery time, additional risk of iatrogenic nerve root/dural sac injury, and particularly postoperative low back pain (LBP) and potential spinal instability caused by excessive removal of SAP.
To realize more accurate decompression and minimize the destruction of the facet joints, we creatively applied the accurate double percutaneous lumbar foraminoplasty and PELD (DPLF-PELD) by separately removing the upper-medial-ventral portion of the facet joint and lower medial-ventral part of SAP for stenosis patients with involvement of both zones 1 and 2. In our previous retrospective study in 2016,29 patients achieved satisfactory clinical outcomes, with an excellent rate of 93.3% by using this innovative technique [11]. The present study was prospectively designed to re-evaluate the clinical outcomes of recent similar patients by using DPLF-PELD with the help of our specially designed depth-limited instruments. Technique notes and short-term outcomes are included in this report.

Materials And Methods
This prospective study was approved by the Ethics Committee of West China Hospital, Sichuan University and was registered with the Chinese Clinical Trial Registry (ChiCTR1800019551). The study was conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants prior to surgery.
Here, 69 patients with single-level LLSC stenosis, simultaneously occurring in both zones 1 and 2 from November 2018 to April 2019, were enrolled. All of the patients underwent DPLF-PELD performed by one endoscopic spine surgeon (KQQ). Table 1 presents the characteristics of 69 patients. LBP, muscle weakness of the lower limbs, extremity radiating pain with/without gluteal pain, and neurogenic intermittent claudication were observed in 5 (7.2%), 1 (1.4%), 62 (89.8%), and 53 (76.8%) patients, respectively. No patients had a history of surgeries.
The study included the following patients: (1) Those who manifested a single nerve root symptom, such as single-side extremity pain, numbness, or weakness with or without LBP. (2) Those who possessed full preoperative radiological information. The method of distinguishing the stenotic zone has been described in a previous study ( gure 1) [8]. Stenosis in zone 1 was diagnosed by sagittal T2-weighted MRI scans through the paracentral region: the anteroposterior distance measured less than 1 mm. Stenosis in zone 2 was diagnosed by axial bone window CT scans, which showed that the anteroposterior distance in the lateral recess region was less than 3 mm. The radiological diagnosis should be related to clinical symptomatology. The preoperative blocking of the nerve root could be applied in some intractable cases.
(3) Those who presented with obvious symptoms (preoperative leg pain visual analog score [VAS] score over 6) after over three months of ineffective conservative treatment. (4) Those who provided informed consent for our study and agreed to attend all required follow-up visits.
The study excluded the following patients: (1) those with lumbar segmental instability indicated by preoperative lumbar exion-extension x-rays; (2) those who were diagnosed with lumbar central canal stenosis; (3) those who were diagnosed as having a pure lumbar disk herniation; (4) those with a highgrade lumbar spondylolisthesis with multilevel spinal stenosis or other deformity; (5) those exhibiting a high iliac crest, with the peak of the iliac crest surpassing the lower quarter of the L4 vertebral body, hindering puncture at L5/S1; and (6) those with any type of surgical contraindication.

Special surgical tools
Specially designed depth-limited trephine for foraminoplasty (ZL 201621149959.2) [12]: consisted of a trephine, handle, and stopper ( gure 2). The saw tooth of the trephine was no different from the traditional trephine. The diameter of the trephine had various sizes: 6.5 mm, 7.5 mm, and 10 mm.
However, the total length of the trephine is identical at 243 mm. At the proximal end of the trephine, 17 circular grooves are designed. The rst groove is 165 mm away from the distal end of the trephine, and there is an interval of two mm in the next adjacent grooves. The depth of foraminoplasty was accurately controlled and limited by the stopper located in the trailing end of the trephine. The stopper is locked in one of the grooves, thus preventing the advancement of the trephine. The exact foraminoplasty depth required can be adjusted based on the stopper location. The handle can be easily assembled and disassembled from the trephine.

Surgical techniques
All DPLF-PELD procedures employed by the author were essentially a classic THESSYS technique popularized by Hoogland [9]. The entire procedures were performed under local anesthesia, with the patient in the prone position on a radiolucent table, using C-arm uoroscopy. A 1.6 mm Spinal Guiding Cannulas (SPINENDOS, Germany) was inserted into the safe zone of Kambin's triangle. The puncture point was 12 ± 2 cm from the midline, according to the size of the body and surgical level, and 2-5 cm from the horizontal line of the target intervertebral disk. After in ltrating 15-20 ml of 0.5% lidocaine into the subcutaneous soft tissue, around the SAP, and intervertebral foramen, the needle was replaced with a 0.7-mm-diameter guiding wire. A dilator 2-channel (7.0-, 6.3-, or 5.3 mm-diameter) was passed over the guiding wire under uoroscopic control. A trephine protection tube (6.5-, 7.5-, or 8.5-mm-diameter) was introduced over the obturator until it was situated in the proper position. The depth-limited trephine designed by us (6.5-, 7.5-, or 8.5-mm-diameter, selected based on pathologic conditions) was used to perform two-time foraminoplasty, which was facilitated by changing the trajectory of the trephine, to aim for different compressive portions. The details of the two foraminoplasty procedures are shown in table 2 and gure 3.
In the rst foraminoplasty, the scale of the resection could be slightly adjusted, based on different pathologic conditions. After the rst foraminoplasty, a radiofrequency probe was endoscopically used (working tube with an elevator tip, ID 7.2 mm, OD 8.0 mm, and L178 mm; spinal endoscope, 30° direction of view, WC 3.75 mm, OD 6.3 mm, and WL 181 mm) to control bleeding and to adequately expose bony structures by resecting any adherent soft tissue. The margin of exposure should run from the upperventral surface of the SAP to the lower-ventral surface of the SAP and upper surface of the pedicle. Next, a 1.5-mm Kirschner wire was inserted into the aiming site. After removing the spinal endoscope, the second foraminoplasty was then performed, after positioning the trephine protection tube over the Kirschner wire and adjusting its tip to embrace the ventral-basal aspect of the SAP. In some severe stenosis cases, to prevent injury to the nerve roots, we only inserted the trephine into the three quarters of the SAP, thus breaking the involved SAP, instead of performing a complete resection by the trephine. For the two foraminoplasty procedures, the trephine needed to be underdraught, thus aiming to resect more of the SAP. The order of the two steps can be adjusted according to different situations.
In the following step, the trephine protection tube was replaced with the working tube with an elevator tip. High-speed drilling was then used to resect the remaining hypertrophied SAP or IAP as needed. The working tube was adjusted to completely remove decompressive factors: the hypertrophied ligamentum avum, facet joints, and anterior herniated disk. To reduce the recurrence rate of lumbar disk herniation (LDH), we did not perform discectomy (only decompress dorsal compressive factors) for patients whose annulus was not damaged. The compressed nerve root was decompressed and explored from the distal end to near-end, especially at the attachment point of the annulus. The surgeon could see and mobilize both the traversing nerve root and the exiting nerve root under endoscopic visualization. Free movement of the dural sac and nerve root could be a sign of complete decompression. Epidural bleeding was controlled with a radiofrequency probe under saline irrigation.
For each operation duration, times of intraoperative C-arm uoroscopy use and any complications were recorded. Every patient was asked to wear lumbar protection devices for two to four weeks after the operation, and to take muscle function exercise the initial two weeks following the surgery.

Outcome assessment
Outcomes were evaluated via follow-up interviews (WY) at three months and a nal follow-up postsurgery. We used LBP, leg pain VAS, and Oswestry disability index (ODI) to evaluate the outcomes of surgery. Function outcomes were assessed using the modi ed Macnab criteria [13]. All patients routinely underwent 3D-reconstructive CT scans two days after the operation as well as MRI and CT scans after three months to con rm complete decompression. In the nal follow-up, patients underwent CT to con rm no recurrence of LLSC stenosis, and exion-extension x-rays to observe for lumbar stability. All patients' postoperative radiological exams are permitted to be discharged.

Statistical analysis
Statistical analysis was performed with SPSS 23 software (SPSS Inc., Chicago, IL). Preoperative and postoperative (three-month and nal follow-up) VAS and ODI scores were analyzed with ANOVA. Here, p < 0.05 was considered the threshold for signi cance.

Clinical outcomes
All patients successfully underwent DPLF-PELD without hematoma formation, a change to open surgery, or any nerve root injuries. Leg pain was immediately eased after the operation. The mean follow-up period was 13 months (range, 8-17 months). All clinical outcome results are shown in table 3. Two patients had dural tear complications. They were cured after a conservative treatments. There was one patient whose preoperative symptoms were not relieved by the surgery. The postoperative CT scan illustrated a small separated bony segment that had moved into the spinal canal. After two months of conservative treatment, the symptoms became aggravated, and we performed open surgery. The symptoms completely disappeared immediately. Two patients complained of moderate postoperative LBP without lumbar muscle weakness, which disappeared after conservative treatment. All three-month postoperative CT and MRI exams con rmed that the compressive factors were completely removed, by showing complete removal of the dorsal hypertrophied SAP on CT, and CSF lling around the compressed nerve root on MRI. The exion-extension x-rays and CT at nal follow-up indicated no recurrence or lumbar segmental instability. A case presentation is shown in gure 4.

Discussion
The term "lateral lumbar spinal canal" was rst introduced by Lee et al. in 1988, and was divided into entrance, middle, and exit zones [5]. However, problems were noted with the label, such as ambiguous borders of each zone and improper names [14]. Prior to our systemically de ning LLSC, and providing a classi cation, there was no universally accepted de nition of LLSC, including the lateral recess region [8].
We found that the retrodiscal space (zone 1) and upper bony lateral recess (zone 2) are the two most common regions in which lumbar degenerative changes occur. As we analyzed [15], zone 1 is surrounded by soft tissue whose dorsal compressive element was the ligamentum avum and joint capsules; zone 2 was formed by tricortical bony structures whose compressive element was hypertrophied SAP. Consequently, regardless of LSS occurring in zone 1 and/or 2, to ensure the effectiveness of surgery, accurate and complete surgical decompression is important. However, for LLSC stenosis patients with the involvement of both zones 1 and 2, complete surgical decompression necessitates higher requirements. Conventional open surgery can adequately treat LLSC stenosis patients via a resecting laminectomy and medial arthrotomy, however, the drawback is obvious: a longer operation time, lengthier recovery period, and more complications [16].
In recent years, PELD has greatly developed, making revolutionary progress [17][18][19]. The application of foraminoplasty greatly expanded the indications for PELD in treating LSS [3,20,21]. Various foraminoplasty methods are available by adjusting the location of needle tip and the trajectory of trephines, achieving the purpose of decompression on speci c targets. For example, the classical TESSYS technique popularized by Hooglan et al. rst introduced the foraminoplasty procedure by resecting the upper-ventral aspect of the SAP [22]. Afterwards, variations of the TESSYS technique were created, aiming to remove the lower-ventral portion of an hypertrophied SAP in lateral recess stenosis patients [9,23]; recently, our team created trans-articular and trans-isthmus approaches as foraminoplasty methods to treat central/paracentral and high upper-migrated LDH patients, respectively [12,24].
However, endoscopic decompression toward zones 1 and 2 required different foraminoplasty targets ( gure 3). It was very di cult to realize full-course decompression via single foraminoplasty with a classical TESSYS technique, without the use of an endoscopic high-speed drill. The frequent usage of a high-speed drill is bound to cause additional length of surgery, a higher risk of iatrogenic nerve root/dural sac injury and excessive resection of the SAP, which can cause potential postoperative LBP and lumbar segmental instability [3]. Therefore, after carefully analyzing the anatomical, pathological, and biomechanical features of LLSC mentioned in the published study [8], and combined with extensive endoscopic surgical practice, we creatively designed precise safety decompression via DPLF-PELD, which performed programed and accurate foraminoplasty toward zones 1 and 2 separately. The advantages were obvious. On the one hand, the programed operation greatly improved decompression e ciency and accuracy. This guaranteed the full-course and complete decompression of the two regions with a shorter timeframe. On the other hand, more of the normal SAP can be retained because most principle compression in zones 1 and 2 would be accurately resected in the two-time programed foraminoplasty procedures. This can largely prevent the occurrence of the postoperative LBP and potential lumbar segmental instability.
Of the 69 patients in our study, the preoperative leg pain VAS score was 7.05 ± 1.04, which decreased to 1.03 ± 0.79 postoperatively (p < 0.05). Additionally, we did not nd the increasing postoperative LBP VAS score in our group (p > 0.05). This indicates that our new technique did not necessarily increase iatrogenic postoperative LBP, which was an additional bene t as compared to conventional open surgery.
We owing this to the advantages of minimally facet joint damage. The nal follow-up results revealed an excellent rates (94.2%), which was similar to conventional microsurgical techniques [25],and was higher than other endoscopic techniques: 82% in Kambin [26], 85% in Lewandrowski [23] and 89.2% in Yeung [27]. No incomplete decompression, nerve root injury, or other complication appeared. No recurrence and segmental instability was observed in any patient during the follow-up period. We propose that the good clinical outcome was a result of using the classi cation, con rming compressive factor preoperatively, using TPLF-PELD realizing full-course complete decompression, and avoiding unnecessary resection of SAP intraoperatively. Furthermore, our specially designed the depth-limited trephine effectively guaranteed the safety of the procedures. For those severe stenosis patients, the nerve root was tightly compressed by hypertrophied facet joint. The nerve root may be easily injured by excessive advance of the trephine without depth limitation.
Among all 69 patients, only one suffered a severe postoperative complication, consisting of small separated bony material remaining in the spinal canal, which was subsequently removed by open surgery two months later. This was caused by insu cient experience in the early period. Additionally, there were two cases of intraoperative dural tear, cured by conservative treatment. We attributed these occurrences to the severe adhesion between the dural sac and surrounding bony structures caused by long-term stenotic changes. The complication rate was 4.3%, which was apparently lower than in other studies ranging from 5.5%-13.2% [9,26,27]. The above-mentioned results proved the effectiveness and rationality of the DPLF-PELD in treating LLSC stenosis in both zones 1 and 2.
We were also concerned about the effect on postoperative lumbar segmental stability, following removal of a portion of the facet joint. Although it has been demonstrated that facetectomy decreases the stiffness and increases the mobility of the spinal motion segment in all modes of loading [28,29], there is still no evidence that injured or damaged facet joints consequently induce the mechanical instability of the spine [30]. Furthermore, our technique only resects a small segment of the facet joint, comprising less than 10%-20% of the entire facet joint. Osman studied the pathoanatomic and exibility changes after posterior and transforaminal decompression in a cadaver biomechanical study [31], which even much more destruction than ours. No exibility change or instability was noted, identical results as obtained in our previous reports [11,12,15,24]. In our study, we designed lumbar dynamic position x-rays at each patient's nal follow-up. No postoperative iatrogenic segmental instability was observed.
The limitations of this study should be noted, particularly the small sample size and a short follow-up period. However, the aim of the study was to introduce an alternative approach for treating LLSC stenosis in both zones 1 and 2, rather than to compare it with other methods. In addition, TPLF-PELD has a steep learning curve and relative narrow indication: the surgery is only suited to simple single-level LLSC stenosis patients who have both zones 1 and 2 involvement. The number of patients who quali ed for the study was therefore quite limited, and further studies are needed. Conclusion DPLF-PELD is a minimally-invasive, effective, and safe surgical method that can well treat LLSC stenosis patients whose stenotic region has occurred in both zones 1 and 2, with the advantages including less damage to lumbar spine anatomy, a lower complication rate, and good short-term clinical outcomes.

Declarations
-Ethics approval and consent to participate This prospective study was approved by the Ethics Committee of West China Hospital, Sichuan University and was registered with the Chinese Clinical Trial Registry (ChiCTR1800019551). Informed consent was obtained from all individual participants included in the study.
-Consent to publish A copy of the written consent is available for review by the Editor-in-Chief of this journal.
-Availability of data and materials       The specially designed depth-limited trephine.  A 55-year-old male patient complained of severe left radicular pain for 12 months. He could not walk for three months due to severe left buttock and leg pain. Left L4/5 LLSC stenosis in both zones 1 and 2 was con rmed. We con rmed the totally decompression by postoperative CT and MRI. The leg pain was relieved immediately after the operation. No lumbar instability was indicated in the nal follow-up. a, b Preoperative X-ray in AP and lateral view. c Preoperative sagittal CT scans indicated stenosis of the retrodiscal space (zone 1) on the left at L4/5 (red circle). d Preoperative axial CT scans indicated upper bony lateral recess (zone 2) stenosis on the left at L4/5 (red circle). e, f Preoperative sagittal and axial T2weighted MRI scans showing the L4/5 left zone 1 stenosis caused by a lumbar disk bulge anteriorly, and hypertrophied, curled ligamentum avum posteriorly (red circle). g, h Fluoroscopy during surgery demonstrates the trajectory of the trephine in the rst foraminoplasty in the AP and lateral views. i, j Fluoroscopy during the operation shows the trajectory of the trephine in the second foraminoplasty in the AP and lateral views. k, l Three months postoperative sagittal and axial CT bony-window scans clearly demonstrate complete decompression of zones 1 and 2. m, n Three months postoperative sagittal and axial T2-weighted MRI scans indicate that the nerve root was decompressed without recurrence. o, p Postoperative exion-extension x-rays at the nal follow-up con rm that no lumbar instability occurred.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.