Endoscopic surgery in the lumbar spine: experience of the ABC Medical Center

Introduction

The history of percutaneous full-endoscopic spine surgery dated to 1989 when Schreiber, Suezawa, and Leu added a “discoscopy” to the percutaneous discectomy described first by Hijikata in 1975 and later popularized by Kambin, who added the description of the security triangle that holds his name, to approach the intervertebral dfisk on 1986¹. Later Mayer adopted this technique and was the first to compare endoscopic discectomy to microdiscectomy by 1993, observing similar overall outcomes with more patients in the endoscopy group returning to their previous occupational activities after surgery, he concluded that endoscopy was best suited for patients with contained and small subligamentary disk herniations². Later in 1999, Yeung popularized Kambin’s approach by describing a surgical technique in which we start discectomy from inside the disk and finish searching the outside herniation; this “inside-outside” technique carries increased risks of residual and recurrent herniation because of an indirect vision and additional annular damage³. Therefore, in 2005, Rothen improved the surgical technique by describing the trans and extraforaminal technique that avoids entering the disk and focuses on the disk herniation directly, thus avoiding residual and recurrence risks⁴. Rothen also described the interlaminar approach; both techniques are the current state-of-the-art for endoscopic discectomy^4,5. Currently, endoscopy is indicated for a variety of spinal diseases, including full-spine Intervertebral disk herniation (central, paramedian, foraminal, extraforaminal, and migrated disk), whole spine spinal stenosis (lateral recess, central canal, and foraminal stenosis as by ossification of the ligamentum flavum), Infective spondylodiscitis (pyogenic, epidural abscess), spondylolisthesis grade 2 or less, and even revision surgery (recurrent disk herniation, cage displacement, bone cement leakage into the canal or foramen)⁶. In addition, evolving applications and indications are coming in the field of tumor surgery, fusion, and pain-related surgery, either as complementary or hybrid fusion⁶.

In Mexico, endoscopy is still an evolving field, with very few surgeons trained in full-endoscopic techniques and even less research being performed^7,8. However, contrary to the few surgeons trained adequately in full-endoscopic techniques, many surgeons incorporate endoscopy into their daily practice through hands-on workshop training or attending endoscopy conferences^9,10. The lack of proper training in developing endoscopic surgery is demonstrated by higher rates of complications during critical aspects of the surgical technique, including instrument insertion, docking, or removal-related injuries; complications related to motorized or cutting instruments with higher risks of aggressive or deep penetration outside of the intervertebral disk with subsequent vascular or visceral injury, or even inadequate planning related complications such as wrong level surgery¹¹.

Provided this background and the precarious research productivity in the field in Mexico, we developed the present work to show our experience with the first one-hundred cases performed in our institution in the hands of an expert spine surgeon who was adequately trained in full-endoscopic techniques, intending to describe our results, compare them to international series from top class countries in terms of outcomes and complications to demonstrate that we are in the top benchmark of endoscopic spine surgery.

Objective

Methods

We performed a retrospective review of all medical files since the first endoscopy in our hospital early in 2017 to gather the first 100 surgical procedures. We included the following information: age, gender, body mass index, diagnosis, preoperative pain score as by VAS for lumbar and leg pain by the side, clinical status (motor or sensory deficit), preoperative Oswestry score, a surgical procedure performed, surgical time and bleeding, intra and postoperative complications, postoperative VAS for lumbar and leg pain as well as postoperative Oswestry score at 3, 6, and 12 months. We excluded those records with incomplete medical records on the surgical technique employed. As shown below, we described the surgical technique and operative setting for all the cases. Later, we used central dispersion statistics for demographical data and preoperative data. We employed a student t-test to compare VAS and Oswestry scores’ evolution at different intervals. Finally, we compared our results to the current series for discussion.

Surgical technique

On an individual basis, we chose and performed either of the three following surgical techniques that compound a full-endoscopic armamentarium for lumbar discectomy:

Transforaminal Endoscopic Lumbar Discectomy (TELD)^4–12

We use prone positioning on a pro-axis or Jackson table for all the cases. We use a posterolateral or lateral approach depending on the surgical goals desired; often, we prefer the most lateral entry point possible to be able to access the extraforaminal, foraminal, and central canal areas. Important anatomical considerations should be emphasized, avoiding the iliac crest and the renal fossa; therefore, TELD is ideal for an L4-L5 segment. We use a fluoroscopic-assisted technique where we search for a true anteroposterior view of the vertebral segment, aiming to have parallel platforms at the intervertebral disk to treat, we mark the midline, and then we trace a lateral projection of the intervertebral disk space (Fig. 1A), Later we proceed on Lateral view to identification the merging of the posterior vertebral wall (Fig. 1B), and we intersect both lines to delineate the entry point for the incision (Fig. 1C). We insert the spine needle on anteroposterior view up to the foramen in the medial pedicular line, searching for a shift in resistance indication of the intervertebral disc, then we proceed to lateral view to confirm the proper spinal needle positioning in the intervertebral disk space at the level of the posterior wall of the respective vertebras. Then we introduce the spinal needle to the disk up to the midline on an anteroposterior view. Later we introduce the inner needle and introduce the guide wire removing the outer sheath of the needle. We proceed to make a skin incision 7 mm wide over the wire to insert the dilator up to the final guide-wire position and confirm its proper situation in the caudal third of the foramen (Fig. 1D) over the caudal pedicle over the intervertebral disk at the level of the posterior wall of the respective vertebrae (Fig. 1E) before guide wire removal (Fig. 1F). Later on, the beveled working sheath is carefully inserted (Fig. 1G) and docked over the caudal pedicle in the lower third of the foramen (Fig. 1H), and the dilator removed (Fig. 1I). Finally, the endoscope is introduced over the working cannula. Once we introduce the endoscope, we proceed to blood vessel coagulation in case of bleeding. Later we use the radiofrequency tip as palpatory to localize the disk space or caudal pedicle, and the visual field is freed from muscle or ligament structures interposing to the intervertebral disk with the aid of coagulation radiofrequency or disk clamps. Once the disk space is visualized, we look for the disk herniation as in preoperative imaging studies identification on surgical planning. Most of the time, the herniated disk is the first structure to visualize even before any blood vessel bleeding; in those situations, we proceed directly to fragmentectomy of the herniated disk with the aid of grasps or rongeurs and later continue to hemostasis by bipolar coagulation if needed. Other times, we need to go inside the annulus with the aid of a Penfield dissector, an endoscope scissor, or the proper dissector for disk incision. Once inside the annulus defect, the herniated disk is removed uneventfully using disk grasps as previously described, and any bleeding is controlled by bipolar coagulation or radiofrequency before endoscope removal. We close the skin incision with one knot of 3-0 nylon suture.

Figure 1. Transforaminal lumbar endoscopic discectomy. A: level localization by lateral and B: AP fluoroscopic view, C: skin Intersection point, D: dilator positioning on lateral, and E: AP fluoroscopic view, F: dilator position on surgical view, G: cannula docking on lateral, and H: AP fluoroscopic view, I: annula docking surgical view.

Special considerations

Cranial levels over L4-L5 require additional preoperative planning for lateral entry point identification to avoid the renal fossa. On preoperative magnetic resonance imaging, we trace a midline over the entire vertebral body alienated to the spinous process; later, we trace an oblique line starting at the center of the vertebral body and passing outside the inner and posterior wall of the ipsilateral kidney up to the skin to mark the most lateral entry point, care should be taken to work medial to this projection line during instrument insertion in order to avoid complications derived from coursing inside the renal fossa.

L5-S1 level requires similar planning to avoid the iliac crest; this last structure will be demarked on the skin surface by palpation. This line determines the most lateral entry point by searching its intersection with the intervertebral disk space’s projection line and the endoscopy’s suitability for an L5-S1 approach.

Extraforaminal Endoscopic Lumbar Discectomy⁴

The surgical technique follows the same principles for TELD until needle insertion. We use prone positioning on a pro-axis or Jackson table for all the cases. We use a posterolateral or lateral approach depending on the surgical goals desired; often, we prefer the most lateral entry point possible to be able to access the extraforaminal, foraminal, and central canal areas. Important anatomical considerations should be emphasized, avoiding the iliac crest and the renal fossa; therefore, TELD is ideal for an L4-L5 segment. We use a fluoroscopic-assisted technique where we search for a true anteroposterior view of the vertebral segment, aiming to have parallel platforms at the intervertebral disk to treat, we mark the midline, and then we trace a lateral projection of the intervertebral disk space (Fig. 1A), Later we proceed on lateral view to identification the merging of the posterior vertebral wall (Fig. 1B), and we intersect both lines to delineate the entry point for the incision (Fig. 1C). The main difference concerning TELD is that needle insertion is directed to the caudal vertebral pedicle (Fig. 2A) instead of the lower third of the foramen. Rotation movements are performed to introduce the needle into the cortical bone (Fig. 2B). Afterward, the inner needle is removed, and the guide wire is introduced into the cortical bone before the outer sheath of the needle is removed (Fig. 2C). A 7 mm incision is performed over the guide wire for the insertion of the dilator. The dilator position is confirmed before cannula placement (Fig. 2D). Following dilator insertion, the working cannula for the endoscope is inserted, and the dilator is removed after fluoroscopic position confirmation (Fig. 2E). Once the endoscope is inserted, we must realize that increased endoscope control is required for this approach, provided that it is inserted into the foramen, which stabilizes the endoscope. Once inserted, the endoscope shows the pedicle and must be redirected to the intervertebral disc, avoiding nerve root injuries provided the disk herniation can displace it. Once the intervertebral disk is visualized, we proceed to annulus defect finding or incision to proceed with the removal of the herniated disk with the use of disk grasps (Fig. 2F) as previously described and any bleeding controlled by bipolar coagulation or radiofrequency before endoscope removal. We close the skin incision with 1 knot of 3-0 nylon suture.

Figure 2. Extraformaminal lumbar endoscopic discectomy. A: level localization by AP, and B: lateral fluoroscopic view, C: guide-wire insertion, D: dilator positioning on AP, and E: lateral fluoroscopic view, F: endoscope placing and disk insicion.

Interlaminar Endoscopic Lumbar Discectomy^5,12

We use prone positioning on a pro-axis or Jackson table for all cases with lumbar flexion to rectify lumbar lordosis and open interlaminar space. Fluoroscope is positioned at a straight 90°degree to the floor, irrespective of a true anteroposterior view of the intervertebral disc. The desired intervertebral disk is localized in this fluoroscopic view, and a vertical line of 7 mm at 2 mm from the midline is marked for skin incision. The skin incision is performed at the marked site, and a dilator is inserted up to the ligamentum flavum, and the position is confirmed by fluoroscopy on the anteroposterior (Fig. 3A), and lateral view (Fig. 3B). Following dilator insertion, the working cannula for the endoscope is inserted with the beveled side in a lateral direction, and the dilator is removed after fluoroscopic position confirmation for endoscope insertion. Muscle tissue is dissected to visualize the ligamentum flavum (Fig. 4A). We maintain tension on the ligamentum flavum with the working cannula to ease opening while scissors (Fig. 4B) resect the ligamentum layer by layer in a posterior to anterior direction. The last layer is darker (Fig. 4C); at this point, we verify that the non-cutting side of the scissor is directed to the spinal canal. This layer can be opened by the water pressure of the endoscope, providing a space between the ligamentum and the dural sac. Lastly, we observe the epidural fat (Fig. 4D), and the flavum ligamentum is resected from the medial to the lateral up to the articular facet and the lateral recess. Once the dural sac is visible, we localize the traversing nerve root, and we locate a 2 mm curved Penfield between the lateral recess and the nerve root up to feeling the intervertebral disk in the deep, taking care to hold the tip on the lateral side. Later the dissector is turned 180 degrees with the tip to the neural structures displacing the traversing nerve root to the medial side, making space for looking at the intervertebral disk and the herniation defect in the deep with fluoroscopic confirmation on anteroposterior (Fig. 3C) and lateral view (Fig. 3D). We insert the beveled cannula over the rotated Penfield dissector up to the intervertebral disk. Later we perform axial rotation to the sides to discover which side permits cannula rotation so that the bezel protects the neural structures serving as nerve root separator. In this situation, if we consider the view of the endoscope that of a clock, the cannula bezel points to the 12 (medially), cephalic structures are positioned at 3 when on a right-side approach, whereas for a left-side approach points to 9 and the lateral recess points to 6, and ideally the herniated disk is in the center of the view. Finally, discectomy is performed conventionally with a Penfield dissector and disk grasps as in other techniques. We perform bleeding control by radiofrequency or bipolar coagulation if needed before endoscope removal. We close the skin incision with one knot of 3-0 nylon suture.

Figure 3. Interlaminar lumbar endoscopic discectomy. A: level localization by AP and B: lateral fluoroscopic view. C: endoscope placing and disk incision on AP and lateral view.

Figure 4. Interlaminar lumbar endoscopic discectomy. A: ligamentum flavum identification, B: ligamentum flavum incision with the aid of scissors, C: dark deep layer of ligamentum flavum signaling peridural fat below, D: peridural fat below ligament flavum. Erative outcomes (VAS for lumbar and radicular pain by side and Oswestry score. Statistically significant differences between preoperative and postoperative periods (p < 0.01) and nonsignificant differences among postoperative periods (p > 0.05).

Neuromonitoring

All patients underwent intraoperative neurophysiological monitoring by recording somatosensory evoked potentials of the median and ulnar nerves in the thoracic limbs and in the peroneal or tibial nerves for the pelvic limbs to monitor the integrity of the posterior cords and spinothalamic sections (proprioceptive pathways), employing surface electrodes and skull recording according to the 10-20 International technique in Fz-Cz in lower limbs and C3-C4, C3-Fz, C4-C3, C4-Fz in upper limbs as well as Obtaining dermatomal evoked potentials in the corresponding myotomes for the corresponding level at cervical, thoracic, and lumbosacral levels. Furthermore, in addition to stimulating the scalp with subdermal electrodes in the C1-C2 region and recording distally, motor potentials are also obtained to monitor the corticospinal pathway and electromyographic activity is continuously recorded for real-time control of the corresponding myotomes, thus constituting multimodal intraoperative neurophysiological monitoring¹³.

Results

Lumbar endoscopy was performed on 100 patients in the period between May 2017 and January 2023; historically, the first case of lumbar endoscopy was performed on 9^th May 2017. A total of 55 patients were male, and 45 patients were female. The mean age was 54 (95% confidence interval (CI), 95% CI: 51.24-57.66, standard deviation (SD): 16.16). The most frequent diagnosis was disk herniation in 92% of the cases, Lumbar stenosis in 6%, foraminal stenosis, and combined Lumbar stenosis and sic herniation in 1%, respectively. Of the 93% of cases with disk herniation, 73% were disk extrusions, and 20% were disk protrusions. Surgery was performed in a single level in 92% of the cases and multilevel in 8% of the cases (two-level surgery in seven cases and three-level surgery in one case). The most frequent segment was L5-S1 in 49% of the cases, followed by L4-L5 in 42% of the cases, L3-L4 in 11%, and L2-L3 in 7% of the cases. Only 57 cases were left-sided, 42% were right-sided, and 1% were bilateral. The surgical approach most frequently used was the interlaminar approach in 55% of the cases, Transforaminal in 38%, and combined (extraforaminal/interlaminar) in 1%. Thirteen patients were simultaneously treated with an additional lumbar block (nonspecified). A preoperative motor deficit was registered in 36% of the cases, 30% had no motor deficit, and the rest was not specified in medical records; 26 had evidence of motor improvement in the postoperative period, one did not improve, and the rest was not recorded in medical files. We had a 16% rate of complications; the most frequent was the need for microsurgery conversion in 7% of the cases (one by durotomy, another by endoscopy failure, and the rest by difficulty or complexity of the case as in the case of dural adhesions), followed by recurrence in 6% of the cases, radiculitis in 2%, and cauda equina in 1%. Seven patients required a second surgery in their follow-up, three nonspecified, one required anterior lumbar interbody fusion in L4-L5 and L5-S1, disk prosthesis in one, and discectomy by symptomatic reherniation in one case. About 72 patients had a complete medical file to determine lumbar and radicular pain, Oswestry in the preoperative and surgical time, and bleeding. Mean axial pain as determined by the VAS was at 5.85 (95% CI: 2.96-4.85, SD 3.33), mean left and right radicular pain was 4.76 (95% CI: 3.74-5.79, SD 4.372) and 3.9 (95% CI: 2.96-4.85, SD 4.018) respectively; mean preoperative Oswestry Score was at 36.93 (95% CI: 33.25-40.61, SD 15.66). Mean surgical time was 103.13 min (95% CI: 54.79-151.46, SD 205.70), while surgical bleeding was at 9.78 ml (95% CI: 7.44-12.12, SD 9.95). In 70% of three cases had follow-up information at one month, where axial pain improved at a mean 1.8 (95% CI:1.40-2.20, SD 1.66), left and right radicular pain at 1.27 (95% CI: 0.87-1.68, SD 1.69) and 1.26 (95% CI: 1.73-1.03, SD 1.97) respectively; while mean Oswestry was registered at 11.39 (95% CI: 8.58-14.19, SD 11.77). Only 55% of the cases had follow-up information at 3 months, where axial pain improved at a mean of 1.69 (95% CI:1.17-2.21, SD1.93), left and right radicular pain at 0.95 (95% CI: 0.52-1.37, SD 1.56) and 0.73 (95% CI: 0.33-1.13, SD 1.48) respectively; while mean, Oswestry was registered at 8.73 (95% CI: 5.87-11.59, SD 10.58). Only 41% of the cases had follow-up information at 6 months, where axial pain registered a mean of 1.95 (95% CI:1.14-2.76, SD 2.56), left and right radicular pain at 1.00 (95% CI: 0.41-1.59, SD 1.86) and 0.71 (95% CI: 0.22-1.20, SD 1.55) respectively; while mean Oswestry was registered at 10.98 (95% CI: 5.81-16.14, SD 16.35). Finally, only 25% had follow-up information at 12 months, where axial pain registered a mean 1.64 (95% CI: 0.92-2.36, SD 1.75), left and right radicular pain at 0.52 (95% CI: 0.16-0.88, SD 0.87) and 0.92 (95% CI: 0.34-0.77, SD 1.41) respectively; while mean, Oswestry was registered at 8.88 (95% CI: 4.85-12.91, SD 95.36). All the variables comparing axial and radicular pain and Oswestry score in the preoperative and postoperative periods were statistically significant (p < 0.01), while changes comparing 1-month records to 3, 6, and 12 months were not statistically significant. Preoperative versus postoperative outcomes (VAS for lumbar and radicular pain by the side and Oswestry score) are shown in Fig. 5.

Figure 5. Preoperative versus postoperative outcomes (VAS for lumbar and radicular pain by side and Oswestry score. Statistically significant differences between preoperative and postoperative periods (p < 0.01) and nonsignificant differences among postoperative periods (p > 0.05).

Discussion

Full-endoscopy lumbar spine surgery is a challenging, minimally invasive technique gaining acceptance among spine surgeons and represents the edge of the state-of-the-art in lumbar surgery^14,15. Full-endoscopy of the lumbar spine is best suited for decompressive techniques. In contrast, other indications such as fusion, tumor, and spine deformity surgery are still evolving field that requires tubular or mini-open approaches, and others can only be treated by open surgery. Clinical outcomes such as VAS and Oswestry score show comparable results between endoscopies and other minimally invasive techniques and open surgery¹⁴. Endoscopy demonstrates improved outcomes in terms of intraoperative bleeding and surgical time for simple cases^16,17, while complex cases have an increased risk of complications that prolong surgical time and carry a risk of surgical conversion to a microscopic approach because of bleeding, dural injury, technical nuances or difficulty in reaching surgical goals; a taught patient selection is mandatory to avoid such complications^15,18. Our series compare to those reported by authors from countries where endoscopy is the preferred surgical approach for most of the described diagnosis, with revision surgery in the range of 4-10% and dural tear between 3 and 8%, while nerve injury and infection rates below 3%^18–21. We had very few complications, all minor complications such as surgical conversion to microsurgery, as we had no one case of a neural injury²¹. We also had a shallow recurrence rate below that reported for the international series. Our infection rates up to the follow-up period were null, and the rate of radiculitis was also below that reported (below 10%); we do consider this comes about the careful preoperative planning and strict surgical protocol we follow during surgical procedures, avoiding distractions or unplanned maneuvers²¹. A formal fellowship training by the senior surgeon and his mastery of the surgical technique during this process of the first 100 cases should also have played a role in having a low complication rate in our series. It will be the object of future research^{11,18,19,21,22}. In the present study, we have provided detailed operative planning and surgical technique so that future generations can benefit from our acquired experience. In our institution, we have recently developed a formal spine surgery fellowship program that integrates open, mini-open, tubular, and endoscopy training so that future generations benefit from a complete and robust surgical training program by a joint effort of the different spine surgeons that constitute the Neurological Center of the American-British Cowdray Medical Center. We consider that our experience will provide a background to project improved outcomes and the highest quality of care for the benefit of our patients. This series is not without limitations; our series describes decompressive techniques treated by full-endoscopy lumbar spine surgery; nevertheless, complex cases, fusion, and tumor surgery are expected to follow, as well as cervical and thoracic procedures are being developed as mastering the technique takes place and more surgeons star involving in endoscopy techniques in their practice, as discussion of cases between pairs and cooperative efforts can improve our performance.

Lastly, we have demonstrated that endoscopy is a safe, reproducible, and successful surgical technique that has comparable results with the current gold standard and should be strongly considered in cases of neural decompression and disk extrusions because of shortened surgical time and decreased bleeding and an almost nil risk of surgical site infection while preserving improved outcomes compared to current gold-standard.

Conclusion

Endoscopy is a safe, reproducible, and successful surgical technique that has comparable results with the current gold standard. Our results compare to those in the international series.

Funding

The authors declare that they have not received funding for this study.

Conflicts of interest

The authors declare that they have no conflicts of interest.