Korean Journal of
Veterinary Service

Intervertebral disc herniation is a common neurological disease that has been widely studied in previous reports (Macias et al, 2002; Bergknut et al, 2012). In chondrodystrophic dogs, especially dachshunds, the incidence of intervertebral disc disease is relatively high (Bergknut et al, 2012). The clinical signs of intervertebral disc herniation include spinal pain and neurological dysfunction, depending on the site of compression . Among these disc sites, the most affected sites in the thoracolumbar region are located between T12-T13 and L2-L3 (Macias et al, 2002). Conventional hemilaminectomy is usually the preferred surgical treatment for intervertebral disc herniation in the thoracolumbar region (Brisson, 2010; Moore et al, 2016)

Nevertheless, conventional neurosurgery has lost ground to minimally invasive spine surgery (MISS) in human medicine (Postacchini and Postacchini, 2011; Truumees, 2015). MISS offers several advantages such as decreased tissue trauma, minimized scar formation, decreased bleeding, increased paraspinal stability, reduced pain, and a faster return to normal life (Hettlich, 2018). In recent decades, despite the availability of several MISS methods and systems, percutaneous endoscopic spinal surgery has been emerging as one of the common surgical methods for disc herniation. According to various previous studies, the percutaneous endoscopic approach shows better results than other approaches in some respects (Qin et al, 2018; Shi et al, 2019; Muthu et al, 2021).

In veterinary medicine, several studies on the percutaneous endoscopic approach have been reported in recent years. To the best of our knowledge, Hwang et al. were the first to use a percutaneous endoscopic approach for spinal surgery in dogs and performed percutaneous endoscopic thoracolumbar pediculectomy in small-breed dogs (Hwang et al, 2016). Moon et al. performed percutaneous endoscopic thoracolumbar mini-hemilaminectomy in small-breed dogs and compared the surgical outcomes between the thoracic and lumbar regions (Moon et al, 2017). Yang et al. presented the first trial of a surgical approach for lumbosacral disc disease via a percutaneous endoscopic approach and performed percutaneous endoscopic limited-lumbosacral-dorsal laminectomy in small-breed dogs (Yang et al, 2021). However, in these studies, the surgical techniques were only performed on small-breed dogs and were not compared to other techniques. Moreover, although minimally invasive foraminotomy in large-breed dogs has been performed using a tubular retractor system, no research has been conducted to utilize the percutaneous endoscopic approach for minimally invasive foraminotomy in large-breed dogs (Lockwood et al, 2014).

This study aimed to evaluate the feasibility of a minimally invasive surgical technique via the percutaneous endoscopic approach in comparison with the conventional open approach technique in large-breed dogs. We hypothesized that percutaneous endoscopic thoracolumbar foraminotomy would allow clear exposure of compression of the spinal cord in the spinal canal and sufficient decompression of the spinal cord, with the advantage of less tissue morbidity in comparison with conventional hemilaminectomy.

Six fresh cadavers of mixed-breed dogs (weight, 27.15±2.26 kg) were used in the present study. The dogs were euthanized for reasons unrelated to this study, and preoperative computed tomography (CT) confirmed that the dogs had a normal vertebral structure with no risk of spinal disease.

Experimental model of intervertebral disc herniation

An experimental model of intervertebral disc herniation was established using the protocol described by Lockwood et al. (2014) with a few modifications. In a 100-mL flask, powdered barium sulfate (0.4 g; Daejung, Seoul, South Korea) and agarose (0.15 g; Duchefa, Haarlem, Netherlands) were dissolved in tap water (10 mL), and the flask was sealed with cotton to minimize vaporization. The mixture was heated until boiling by using a microwave, aspirated into a 3-mL syringe, and allowed to cool to room temperature for solidification before injection.

To determine the optimal volume of BA-gel, a preliminary experiment was conducted. In this experiment, two canine cadavers positioned in sternal recumbency were prepared, and the hair of the dorsal area from T9 to L5 was clipped. Under fluoroscopic guidance (KMC-950; COMED, Gwangju, South Korea), four different intervertebral disc spaces (T10-T11, T12-T13, L1-L2, and L3-L4) were selected and, through the interarcuate space, an 18-gauge spinal needle was inserted into the ventral side of the spinal cord at each site, which allowed injection of various volumes of BA-gel into the epidural space. Subsequently, the optimal injection volume of BA-gel was determined to be 1 mL by comparing the experimental findings with clinical cases on CT images, and the optimal volume of BA-gel was injected into two operative spaces (T12-T13, L2-L3) in each cadaver.

Preoperative assessment (CT)

CT scanning was performed before and after surgery using a helical scanner (Alexion^TM, Toshiba, Otawara, Japan) with our imaging protocol for clinical cases of intervertebral disc disease. CT images were obtained from T11 to L4 and analyzed using the ViewRex software (Techheim, Seoul, South Korea). The vertebral structure, spinal cord, and simulated disc material were confirmed through preoperative scans. The volume of simulated disc material was calculated on the basis of a previous study (Roach et al, 2012). The simulated disc area on the first transverse CT image was multiplied by the slice thickness, and the disc areas on subsequent images were multiplied by the slice interval. The total volume of the simulated disc material was calculated by adding these values.

Surgical instruments

Surgical instruments were provided by Richard Wolf GmbH (Knittlingen, Germany). The instruments included the rod lens optics of a discoscope (5.9 mm diameter, 122 mm length, 25° direction of view), a dilator (5.9 mm diameter, 2 channels), working sleeve (7.0 mm diameter, 80 mm length, 30° beveled window), micro-bone punches (2.5, 3.0 mm diameter, 290 mm length), micro-punches (2.0, 2.5 mm diameter, 290 mm length), micro-rongeurs (2.0, 2.5 mm diameter, 290 mm length), reamer (3.0 mm diameter, 350 mm length), an exploring hook (2.0 mm diameter, 290 mm), a dissector (2.5 mm diameter, 350 mm), and an exploring probe with flexible tip (2.5 mm diameter, 290 mm) (Fig. 1). The optics contained an eccentric working channel (3.1 mm diameter), inlets for rinsing fluid and light, and intra-endoscopic inlets. Various burrs (oval, round, and diamond) equipped with a PowerDrive ART1 (Richard Wolf GmbH, Knittlingen, Germany) were used for drilling (Fig. 1). The fluid management system (Continuous Wave 3; Arthrex, Naples, FL, USA) was utilized for irrigation.

Figure 1. Surgical instruments for performing percutaneous endoscopic foraminotomy (Richard Wolf GmbH, Knittlingen, Germany). (A) Reamer. (B) Diamond burr. (C) Round burr. (D) Oval burr. (E) Exploring hook. (F) Dissector. (G) Micro-bone punch (2.5 mm). (H) Micro-bone punch (3.0 mm). (I) Micro-punch (2.0 mm). (J) Micro-punch (2.5 mm). (K) Micro-rongeur (2.0 mm). (L) Micro-rongeur (2.5 mm). (M) Dilator. (N) Discoscope. (O) Working sleeve. (P) Exploring probe with flexible tip.

Surgical technique

The two surgical techniques were allocated to the two surgical sites in each cadaver by using block randomization to avoid bias toward a single side. Considering the distribution of the BA-gel on preoperative CT, the lateral aspect where surgery was performed (left, right) was chosen. Before surgery, each dog was clipped and positioned in a sternal recumbency position. Under fluoroscopic guidance, an 18-gauge spinal needle was inserted into the dorsal muscle over the injection site to indicate each surgical space. All the surgical procedures were planned and performed by the same surgeon (S.H.L.).

Percutaneous endoscopic foraminotomy (PEF)

At the skin lateral to the spinous process, a 17-gauge spinal needle was advanced toward the lateral aspect of the articular process over the targeted intervertebral space and palpated under fluoroscopic guidance. The spinal needle was then walked ventrally toward the cranial end of the caudal vertebra. Fluoroscopic images were used to confirm that the tip of the spinal needle landed correctly on the cranial end of the caudal vertebra. Once the landing point was correctly determined, an 8 mm K-wire was inserted into the spinal needle, which was then removed. Then, a small stab incision (<7 mm) was made on the site of K-wire insertion into the skin, a dilator was advanced over the K-wire, and a working sleeve was placed over the dilator in order. Finally, the K-wire and dilator were pulled out and the endoscope was introduced into the working sleeve (Fig. 2).

Figure 2. Minimally invasive approach to the vertebra under fluoroscopic guidance. (A) The lateral aspect of the articular process over the targeted intervertebral space was palpated using a 17-gauge spinal needle under fluoroscopic guidance, and the spinal needle walked ventrally toward the cranial end of the caudal vertebra. (B) The position of the spinal needle was confirmed under fluoroscopic guidance in the dorsoventral view. (C) A 0.8-mm K-wire was inserted into the spinal needle. (D) The spinal needle was extracted from the cadaver. (E) The dilator was advanced over the K-wire. (F) The working sleeve was placed over the dilator. (G) The K-wire and dilator inside the working sleeve were removed. (H) The endoscope was passed through the working sleeve. (I) The position of the endoscope inside the working sleeve was confirmed on a dorsoventral fluoroscopic image.

All PEF procedures were performed under continuous lavage through an endoscope. The epaxial musculature was elevated using a dissector, micro-punches, and micro-rongeurs to expose the intervertebral foramen and pedicle of the caudal vertebra. When the pedicle was revealed, foraminotomy was performed with the burrs until the layer of the inner cortical bone was exposed. Once the inner cortical bone layer was adequately visualized, it was removed cautiously with the burrs, micro-rongeurs and a reamer to expose the BA-gel and spinal cord. Then, the BA-gel was revealed, and after the creation of the bone window was completed using the burrs, micro-rongeurs, and micro-bone punches, the BA-gel was removed gently with micro-rongeurs, exploring hook, and exploring probe. Finally, the endoscope and working sleeve were retracted and the skin and fascia were sutured in a routine fashion (Fig. 3).

Figure 3. Endoscopic view of the spinal canal after percutaneous approach to the intervertebral foramen. (A) The pedicle of the caudal vertebra was exposed after elevating the epaxial muscle. (B) The outer cortical bone and cancellous bone were sequentially removed during foraminotomy and the layer of inner cortical bone was revealed. (C) The inner cortical bone was cautiously removed, and the BA-gel injected earlier was observed. (D) After removing the BA-gel, the spinal cord was decompressed.

Open hemilaminectomy (OH, Control)

Hemilaminectomy was performed as described in previous studies (Huska et al, 2014; Shores, 2017). First, a dorsal midline skin incision was made, and the length was extended cranially and caudally to the simulated disc space. The epaxial muscle was dissected and reflected from the lateral aspect of the vertebra. The tendinous attachments were also severed, and Gelpi retractors were placed to increase exposure. The articular processes over the simulated disc space were removed using bone rongeurs, and hemilaminectomy was performed using a pneumatic drill and Kerrison rongeurs. The BA-gel over the designated disc space was carefully removed with probes, and the incision was closed in layers.

Operative and postoperative assessments

During the operation, the procedure time was recorded from incision to closure. Furthermore, we evaluated whether the target site could be clearly exposed and whether decompression could be performed properly.

After the operation, the length of the incision was measured using a surgical ruler. Then, the size of the vertebral defect and the amount of remaining BA-gel were calculated. The amount of BA-gel removed was determined by comparing pre- and postoperative BA-gel volumes.

Statistical analysis

A statistical analysis software package (GraphPad Prism ver. 5.01; GraphPad Software, San Diego, CA, USA) was used to perform all analyses. Differences between the surgical techniques were evaluated using the Mann–Whitney U test. All data are presented as mean± standard error of the mean (SEM).

In the preliminary experiment, various volumes of BA-gel were injected into each epidural space, because of which the BA-gel occasionally spread far from the target intervertebral space or was compacted too much in the spinal canal. Moreover, the amount of BA-gel was too small to cause compression of the spinal cord in other spaces. Consequently, the optimal volume of injected BA-gel was determined to be 1 mL, which was validated by comparison of CT images with those obtained in clinical cases in our hospital.

In both techniques, reaching the desired intervertebral foramen, exposing the spinal cord, nerve root, disc, and BA-gel appropriately, and removing the BA-gel in the spinal cord were achievable without intraoperative complications, such as nerve root injury (Fig. 3, 4). In PEF, the operating site was illuminated and magnified through the endoscope and irrigated constantly until the end of the surgery. However, while removing the injected BA-gel, the surgeon’s view on the screen was blocked by debris from the BA-gel, which made its removal difficult. By extracting debris using micro-rongeurs or increasing the rate of lavage, the view through the endoscope became clear.

Figure 4. Endoscopic view of the spinal canal with spinal cord, nerve root, and epidural fat visible.

The two techniques showed no significant difference in the amount of BA-gel removed (P=0.309, Fig. 5). Sufficient decompression of the spinal cord was achieved after both procedures (Fig. 6). The average volume of BA-gel removed was 684.27±96.38 mm³ by PEF and 747.39±73.48 mm³ by OH (Table 1). On postoperative CT scans, residual BA-gel was detected at all surgical sites, but it could not cause compression of the spinal cord.

Table 1 . Mean±SEM results for outcome measures following percutaneous endoscopic foraminotomy (PEF) or open hemilaminectomy (OH).

	PEF (n=6)	OH (n=6)	P valuea
Incision size (mm)	7.9±0.05	101.6±0.71	0.004
Window size (mm2)	37.04±4.02	119.81±9.30	0.002
Procedure time (min)	55.58±9.11	35.87±2.78	0.002
Amount of BA-gel removed (mm3)	684.27±96.38	747.39±73.48	0.309

^aBased on Mann-Whitney U tests..

Figure 5. (A) Length of skin incision, (B) area of bone defect, (C) procedure time, and (D) amount of barium/agarose gel removed for percutaneous endoscopic foraminotomy (PEF) and open hemilaminectomy (OH). Variables with a significant difference (P<0.05) between the techniques are indicated by asterisk. Values are given as mean±SEM.

Figure 6. Preoperative and postoperative transverse computed tomography images. (A) Preoperative image. The spinal cord is displaced to the right by the injected BA-gel (asterisk). (B) Postoperative image. A small amount of BA-gel remains visible, but the residual BA-gel does not appear to produce compression of the spinal cord. Arrowheads indicate the bone defect created during PEF.

The incision size in the PEF group was much smaller than that in the OH group (P<0.01, Fig. 5). The mean value of incision length was 7.9±0.05 mm for PEF and 101.6±0.71 mm for OH (Table 1). Likewise, the window dimensions created during OH were larger than those created during PEF (P<0.01, Fig. 5). The mean window size in PEF was 37.04±4.02 mm², while that in OH was 119.81±9.31 mm² (Table 1). The three-dimensional images reconstructed using CT are presented in Fig. 7. The surgical duration of PEF was longer than that of OH (P<0.01, Fig. 5). On average, PEF took 55.58±9.11 min and OH required 35.87±4.43 min (Table 1).

Figure 7. Preoperative and postoperative 3D-reconstruction images. (A) Preoperative image. The injected BA-gel is confirmed inside the spinal canal (asterisk). (B) Postoperative image. The dotted line delineates the foraminotomy. The bone defect created after PEF is indicated by arrowheads.

The minimally invasive surgical technique we attempted in this study originated from surgical procedures used in human medicine. Since the first description of percutaneous nucleotomy (Hijikata, 1975), percutaneous endoscopic spine surgery has developed rapidly. Minimally invasive spinal surgery via the percutaneous endoscopic approach offers several advantages over the open approach, such as minimal skin incision, a percutaneous approach, minimal muscle retraction and bone resection, lower complication rate, and earlier return to normal life (Ahn, 2021). However, few studies have addressed minimally invasive spinal surgery via a percutaneous endoscopic approach in veterinary medicine. The present study was conducted to evaluate the feasibility of decompression by minimally invasive foraminotomy via a percutaneous endoscopic approach in the thoracolumbar region of canine cadavers.

An OH was selected as the control procedure for this experiment. This was because OH is the most commonly used surgical method in dogs with intervertebral disc herniation. According to one study, 89% of surgeons performed this technique to facilitate spinal cord decompression (Moore et al, 2016). In addition, various studies associated with conventional hemilaminectomy, such as biomechanical evaluations or postoperative assessments, have been well documented in veterinary literature (Viguier et at, 2002; Vicente et al, 2013; Woelfel et al, 2021)

A clear view of the operating site is important to provide surgeons with improved surgical conditions. In the percutaneous endoscopic approach, the operating site was magnified and illuminated using an endoscope, which allowed more effective observation of the simulated disc material and spinal cord in comparison with bare sight.

The volume of BA-gel removed through the minimally invasive approach was similar to that removed in the open approach. Thus, PEF can cause as much decompression of the spinal cord as OH. In the study by Lockwood et al. two minimally invasive approaches using a tubular retractor system also allowed similar removal of simulated disc material compared to the open approach (Lockwood et al, 2014). Therefore, sufficient decompression of the spinal cord is possible in large-breed dogs through minimally invasive surgery via a percutaneous endoscopic approach or tubular retractor system.

Incision size was evaluated to compare the degree of soft tissue damage between the PEF and OH groups. Since the epaxial musculature was stabbed along a corridor of the incision, the length of the incision could be used to represent soft tissue morbidity. The skin incision in PEF was less than 1 cm and much smaller than that in OH. This ensured various advantages, such as decreased tissue trauma, minimized scar formation, reduced pain, and faster return to normal life, in veterinary medicine, similar to human medicine (Hettlich, 2018). Lockwood et al. applied two minimally invasive approaches using a tubular retractor system; the incision length in large-breed dogs was 25±0.04 mm in foraminotomy via the illuminated port (FP) and 24±0.1 mm in endoscopic foraminotomy (EF). The difference in incision size between the MISS procedures seems to result from the different approach systems. Since the working sleeve for the percutaneous endoscopic approach is not altered according to the size of the patient, we believe that this approach is more advantageous for large-breed dogs than small-breed dogs in terms of soft tissue morbidity.

The vertebral window size was measured and compared after the procedure. On average, the window size in OH was larger than that in PEF. OH was performed in accordance with the standardized anatomical limits. However, the margin of burring in PEF is limited to the extent required to properly remove the artificial disc material through the endoscope. This difference appears to create a large gap in the size between the PEF and OH. According to Lockwood et al, the window size in large-breed dogs was 38.2±10.4 mm² in FP and 32.0±5.0 mm² in EF. Thus, the window size in PEF was similar to that in FP and EF. This may be attributed to differences in the surgical approach, although the foraminotomy is the same. Generally, a smaller window offers several advantages related to soft tissue damage, surgical time, biomechanical stability, and postoperative morbidity (Huska et al, 2014). Conventional hemilaminectomy has the potential to cause spinal instability and increased morbidity owing to its large defect size (Sharp and Wheeler, 2005). Although hemilaminectomy extending up to three adjacent sites or bilaterally up to two adjacent sites cannot increase the risk of spinal instability, it can increase the risk in large-breed dogs (Corse et al, 2003). Moreover, removal of articular processes decreases intervertebral joint stiffness, possibly because spinal rotational stability is maintained by both the articular facet and annulus fibrosus (Viguier et al, 2002; Vicente et al, 2013). The higher morbidity and lower surgical success rate of conventional hemilaminectomy in large-breed dogs in comparison with small breed dogs could be related to these factors (Cudia and Duval, 1997; Olby et al, 2003). Therefore, PEF with a relatively small bony defect and preservation of the articular facet would contribute to spinal stability and prognosis.

The mean duration of PEF was longer than that of OH. In PEF, removal of bone and artificial material is the most time-consuming part of the surgical procedure since the drill used in the PEF is less aggressive than the pneumatic drill used in OH; this instrumental difference seems to contribute to the gap in duration between the procedures. Moreover, as a minimally invasive spinal surgery, PEF has a substantial learning curve (Lee et al, 2008), which can be another reason for the gap in operating time between the procedures. Although the preliminary experiment and some cadaveric training sessions were conducted before the experiment, it is difficult to conclude that we overcame the learning curve. According to one study (Lee and Lee, 2008), the correlation between the learning curve and the operating time of the percutaneous endoscopic approach has been confirmed in human medicine. Therefore, if mastery of PEF is achieved, the gap can be narrowed.

The extruded disc material is not always confined to the intervertebral disc space (Besalti et al, 2006). If a portion of the extruded disc extends beyond the bounds of the intervertebral disc space, it is referred to as a migrated or dispersed disc (Mateo et al, 2011). In one previous study, the percentage of intervertebral extrusion confined to the intervertebral space was only 32.7% in large-breed dogs (Gomes et al, 2016). Because of the instruments used and the relatively smaller window size of PEF, the extent to which the extruded disc material could be removed through PEF would be limited to some degree. Thus, if the extruded disc material is highly migrated from the intervertebral disc space, widening the vertebral window or conversion to an open surgical approach may be required for full decompression. Migration of disc material can also occur in human medicine, and various studies related to migration, e.g., studies examining migration patterns of herniated disc fragments, have been conducted (Daghighi et al, 2014). Several studies have been conducted to overcome the issue of far-migrated disc material in human medicine (Wu et al, 2016; Nakamura and Taguchi, 2019).

This study had some limitations. First, the sample size was smaller than that of other studies. Second, this was a cadaveric study, which precluded the evaluation of several postoperative factors. Guevar et al. used successive neurological evaluations, such as measurements using the modified Frankel scale (MFS), to quantify the degree of neurological deterioration. Moreover, mechanical sensory thresholds, creatine kinase levels, and histopathological findings of the spinal cord have also been evaluated for assessing postoperative pain, soft tissue damage, and spinal cord integrity, respectively (Guevar et al, 2020). If this experiment is executed in live dogs, the abovementioned postoperative factors should be evaluated. In addition, intraoperative bleeding could not be assessed in this cadaveric study. Future studies are warranted to evaluate intraoperative bleeding and the abovementioned postoperative factors in live dogs.

Our results suggest that percutaneous endoscopic thoracolumbar foraminotomy is feasible for treating intervertebral disc herniation in large-breed dogs. This technique allowed observation of the simulated disc material and spinal cord clearly and removal of as much simulated disc material as with OH. In comparison with OH, PEF can be performed with a much smaller incision and bony defect. Although the surgical duration of PEF was longer than that of OH, the gap between the two procedures could be narrowed after the learning curve is overcome. In conclusion, PEF could be an effective surgical option to treat intervertebral disc herniation of the thoracolumbar region in large-breed dogs.

The authors would like to thank Richard Wolf GmbH for providing their instrumental support.

No potential conflict of interest relevant to this article was reported.