INTRODUCTION
In the late 1980s, Kambin and Hijikata introduced percutaneous decompression for herniated lumbar discs using a posterolateral approach [1,2]. Since then, percutaneous endoscopic surgery has evolved significantly and is now widely recognized as full-endoscopic spine surgery (FESS).
Over the past 3 decades, FESS has become a accepted technique for treating various lumbar conditions, including lumbar foraminal stenosis (LFS) with or without stable spondylolisthesis, disc herniation, lumbar spinal stenosis, and L5 isthmic spondylolisthesis [1-5]. Among those in particular, the surgical technique applied to LFS is named full-endoscopic lumbar foraminotomy (FELF) or transforaminal endoscopic lumbar foraminotomy [6,7].
FESS can be performed via either a transforaminal or interlaminar approach. The transforaminal approach accesses the intervertebral disc space and spinal canal through Kambin triangle [8], while the interlaminar approach utilizes the interlaminar space [9]. A key advantage of transforaminal FESS over the interlaminar approach is its feasibility under local anesthesia, which enables a direct route to the foramen. This characteristic makes the procedure suitable for patients with significant comorbidities who are not candidates for general anesthesia [10]. Additionally, transforaminal FESS reduces the risk of epidural scarring [11].
The success rate of FELF for lumbar foraminotomy is relatively high, approximately 85%, compared to 80% for microscopic foraminotomy [7]. Given its high success rate and superior accessibility to the foramen, FELF has become the gold standard for managing LFS across a wide age range, from patients in their 20s to those in their 90s, including those with multiple comorbidities [12].
However, some studies have reported short- and midterm complications following FELF, including dural tears, postoperative dysesthesia, and revision surgeries due to recurrent leg pain or inadequate decompression [4,11,13,14]. One contributing factor to recurrent LFS and associated leg pain is postoperative progression of scoliosis. At our facility, we have observed cases where scoliosis progression following FELF for L5–S1 foraminal stenosis necessitated revision surgeries or lumbar interbody fusion. Nevertheless, not all cases exhibit scoliosis progression. One notable factor that appears to mitigate postoperative scoliosis progression is the anatomical condition in which the L5 transverse process contacts the sacral alar, which we have termed the “lateral kissing spine” (Figure 1).
In this study, the postoperative course after FELF on L5–S1 of 2 cases with and without lateral kissing spine are presented. Secondly, we discuss the risks of scoliosis progression after FELF and propose potential indications for selecting FELF or lumbar fusion as the preferred treatment strategy.
CASE PRESENTATION
An 82-year-old female presented with left leg pain due to left L5–S1 lumbar stenosis. The foramen was narrowed by osteophytes and facet joint hypertrophy. She underwent L5–S1 left FELF under local anesthesia. Postoperatively, the left leg pain resolved by day 1 and did not recur during the 1-year follow-up period.
Preoperative x-rays revealed a coronal lumbar Cobb angle of 36°. At the 1-year follow-up, the Cobb angle remained at 36°. Preoperative computed tomography (CT) scans showed a foraminal space of 91 mm², which expanded to 198 mm² postoperatively. Additionally, the width of the central part of the foramen at the midheight of the disc expanded from 9 mm to 16 mm (Figures 2 and 3).
We have termed the condition where the L5 transverse process touches the sacral area as "lateral kissing spine" (Figure 1). This condition was present on the left side of the lumbar spine in this patient.
The following case is a 78-year-old male patient without lateral kissing spine, in whom postoperative scoliosis progression was observed. The patient presented with right leg pain caused by right L5–S1 LFS. This patient had also foraminal stenosis due to osteophytes and facet joint hypertrophy. He underwent L5–S1 right FELF under local anesthesia. Postoperatively, the right leg pain resolved by day 1 and remained absent for 6 months. However, at the 1-year follow-up, the patient experienced a recurrence of right L5–S1 lumbar stenosis, resulting in a return of right leg pain.
Preoperative x-rays showed a coronal lumbar Cobb angle of 10°, which increased to 15° at the 1-year follow-up. Preoperative CT scans demonstrated a foraminal space of 75 mm², which expanded to 164 mm² on CT imaging performed 1 week postoperatively. Additionally, the width at the central part of the foramen at the middle height of the disc increased from 6.5 mm to 16 mm.
DISCUSSION
1. Clinical Outcomes and Complication After FELF
One of the cases illustrated above showed the failure case of FELF. However, the clinical outcomes of full-endoscopic lumbar foraminotomy (FELF) are relatively good according to the past studies. Ahn et al. [15] reviewed 33 patients (mean age, 64.2; range 27–81 years) treated for LFS, showing significant improvements in pain and functional status over a 2-year follow-up period. The mean visual analogue scale (VAS) score for leg pain decreased from 8.36 preoperatively to 3.36 at 6 weeks postoperatively and further to 1.97 at 2 years (p<0.01). Similarly, the Oswestry Disability Index (ODI) improved from 65.8% to 19.3% (p<0.001), with 81.8% of patients achieving excellent or good outcomes based on the modified MacNab criteria.
Giordan et al. [13] conducted a systematic review and meta-analysis evaluating FELF proceeded with transforaminal approach under local anesthesia. Fourteen studies were included in the study. Mean operative time ranged to 48 to 73 minutes. Leg pain decreased from a starting value of 7.4 (standard deviation [SD], ±1.1) to an average of 5.2 points (SD, ±0.9). The average ODI scores preoperatively were 58.4% (SD, ±13.7%), and improved to an average of 41.2% (SD, ±14.8%). They had concluded that the success rate of the conventional open, microscopic, or mini-invasive foraminotomy demonstrated success rates was 76.9%–83% and it aligns with FELF surgery success rates which were 78%–90%.
Vande Kerckhove et al. [16] compared clinical outcomes of endoscopic foraminotomy versus spinal fusion for LFS through a systematic meta-analysis. Not all endoscopic procedures in their study were performed using the transforaminal FESS technique. The endoscopic group included 10 transforaminal FESS procedures, 2 interlaminar FESS procedures, 3 biportal endoscopic procedures, and 1 unspecified technique. They reported that endoscopic foraminotomy significantly reduced operative time, averaging 69 minutes compared to 119 minutes for fusion. Short-term outcomes (less than 2 years) and midterm outcomes (2–6 years) showed that the proportion of patients with good or excellent MacNab criteria ranged from 81%–100% for endoscopic foraminotomy, with similar midterm results (82%–91%).
Postoperative ODI results demonstrate comparable midterm outcomes between the 2 techniques, with mean ODIs of 19.25 for endoscopic foraminotomy and 20.49 for fusion. In the short term, endoscopic foraminotomy exhibits a slightly better ODI (18.46) compared to fusion (25.40). Statistical analysis using a random-effects model revealed no significant differences between the procedures (p>0.05), with heterogeneity reported as I² = 98% (short term) and 0% (midterm).
VAS/numerical rating scale improvement for postoperative leg pain also showed comparable midterm outcomes, with scores of 2.09 for endoscopic foraminotomy and 1.90 for fusion, with no significant differences (p>0.05). In the short term, endoscopic foraminotomy demonstrated a slight advantage, with a mean score of 1.38 compared to 2.26 for fusion, though this difference was not statistically significant (p>0.05). Heterogeneity was minimal for midterm analysis (I²=0%) and moderate for short-term analysis (I²=60%).
Overall, statistical analysis using a random-effects model indicates that while endoscopic foraminotomy may offer slight short-term advantages in pain reduction, no statistically significant differences were observed in midterm outcomes between the 2 surgical techniques.
These studies suggest that FELF for LFS achieves relatively high success rates comparable to traditional techniques and lumbar interfusion, with similar clinical outcomes.
2. Complications
This paper focuses primarily on the progression of scoliosis following FELF. However, other postoperative complications associated with FELF include dural tears, dysesthesia, hematoma, and the need for revision surgery due to restenosis or inadequate decompression [4,11,13,14].
According to Vande Kerckhove et al. [16], the total complication rate for endoscopic foraminotomy surgery was reported as 2% in the short term and 10% in the midterm.
Leg dysesthesia is the most frequent complication in transforaminal endoscopic surgeries. After FELF, the complication rate for leg dysesthesia averaged 14%, ranging from 0.5% to 28% in previous studies [17-29]. Ju and Lee [14] and Silav et al. [30] commented that this is primarily caused by direct irritation of the exiting nerve root, which lies at the border of the safety zone, by surgical instruments. However, most postoperative leg root disturbances resolve spontaneously or with short-term management.
Dural tears are among the most common complications in endoscopic spinal surgery. The overall incidence of dural tears in endoscopic procedures is 2.7%, ranging from 0% to 8.6% [31]. Giordan et al. [13] reported an intraoperative dural tear rate of 0.9% (95% confidence interval [CI], 0%–1.0%) in their meta-analysis, which is lower compared to the average rate of 2.7%. Ju and Lee [14] reported that the dural tear rate was higher in interlaminar decompression at 2.19%, compared to 1.46% in transforaminal decompression, where it is less common. The reduced risk of dural tears with the transforaminal approach may be attributed to less direct manipulation within the epidural space. All intraoperative dural tears in these cases were repaired immediately during surgery, and no additional interventions were required.
Giordan et al. [13] also reported the overall rate of same-level recurrent stenosis was 1.4% (95% CI, 0%–4.3%), while the revision surgery rate was 1.2% (95% CI, 0%–3.7%).
3. Advance of Scoliosis After FELF
The progression of scoliosis after FELF remains a significant concern, particularly due to restenosis, which has been identified as one of the major complications of foraminotomy with this minimally invasive technique, as reported by Ju and Lee [14] In the context of microscopic foraminotomy, Chang et al. [32] observed that patients with a preoperative lumbar Cobb angle greater than 10° demonstrated reduced improvements in average physical scores (decreased by 9.6 points, p=0.07) and bodily pain scores (decreased by 12.1 points, p=0.02) compared to those with a Cobb angle of 10° or less. However, regarding scoliosis progression, their study found no significant advancement of scoliosis during the 5-year follow-up period in patients with preoperative Cobb angles greater than 7.5°. The mean Cobb angle slightly decreased from 10.52° to 9.8°, which was not statistically significant (p=0.37). The authors concluded that scoliosis progression was effectively mitigated due to the facet-preserving nature of their microsurgical foraminotomy approach.
In contrast, a study by Seo et al. [6] reported that a larger preoperative Cobb angle (mean 8.8° in the restenosis group vs. 4.7° in the control group, p=0.004) was significantly associated with an increased likelihood of restenosis, which frequently necessitated revision surgery. This finding highlight that a higher preoperative Cobb angle is a potential risk factor for poorer postoperative outcomes and diminished improvement.
Together, these findings suggest that while facet-preserving techniques may prevent scoliosis progression, higher preoperative Cobb angles present challenges in achieving optimal long-term outcomes after foraminotomy procedures.
4. Optimal Bone Resection in FELF
In both of the cases, the leg symptoms due to the foraminal stenosis vanished postoperatively. To avoid inadequate decompression or postoperative dysesthesia, adequate resection of the foramen is essential. The goal of FELF is to remove the osteophyte on the superior articular process (SAP) that compresses the exiting nerve root (Figure 1). However, resection of the facet joint raises concerns about potential postoperative lumbar segmental instability. As discussed above, facet-preserving techniques are critical to Ahn et al. [4] reported favorable one-year outcomes in patients who underwent transforaminal full-endoscopic foraminotomy, demonstrating significant improvements in both the VAS and ODI scores. Despite the removal of bony structures, including the tip of the SAP, the upper pedicle, and lower vertebral endplates, patients experienced a reduction of over 50% in VAS scores and a 20%–30% improvement in ODI scores within 6 weeks after surgery. Importantly, no clinical or radiological evidence of segmental instability was observed during 1-year follow-up.
Henmi et al. [33] conducted a study assessing the amount of bone resected during foraminotomy using transforaminal FESS. They measured the distance between the posterior edge of the disc and the ventral aspect of the facet joint at the center of the discs before and after surgery. The preoperative average distance was 6.6 mm, which increased significantly to 9.7 mm postoperatively (p<0.05). The study concluded that to ensure safe cannula insertion and avoid exiting nerve root irritation, bone resection of at least 8 mm, corresponding to the usual cannula diameter, is necessary.
Nishidono et al. [34] evaluated the foraminal space on CT scans of 9 patients who underwent FELF, reporting no surgical complications, including postoperative dysesthesia or inadequate symptom improvement (1 at L3–4, 3 at L4–5, and 5 at L5–S1). The mean foraminal space increased significantly from 52.0 mm² preoperatively to 152.8 mm² postoperatively (p<0.05), representing a nearly threefold enlargement.
Another study from Ahn et al. [35] measured the foraminal space on sagittal T2-weighted magnetic resonance imaging (MRI) in 35 patients who underwent FELF. Ninety-seven percent of these patients reported excellent or good symptomatic improvement. The foraminal space increased significantly from 50.05±5.56 mm² preoperatively to 92.03±9.96 mm² postoperatively (p<0.001). Additionally, they measured the middle foraminal width (MFW), defined as the width of the central part of the foramen at the middle height of the disc. This MFW was comparable to the measurements reported by Henmi et al. [33]. In the study of Ahn et al. [35], the MFW increased from an average of 1.47 mm to 4.78 mm.
Both studies observed significant enlargement of the foraminal space after surgery, although the increase was more pronounced in the studies by Henmi et al. [33] and Nishidono et al. [34] compared to Ahn et al. [35]. This discrepancy may be due to the addition of undercutting laminectomy to SAP resection and the removal of soft tissue in the former studies. Another possible explanation is the difference in imaging modalities used (CT vs. MRI).
According to these studies, the amount of bone resection typically ranged from 5–8 mm in width, and the foraminal area increased approximately 50% to 200% increase of the area compared to preoperative values.
Osman et al. [36] made the first cadaveric study to explore the pathological anatomy, intervertebral foraminal area, and flexibility changes between posterior and transforaminal decompression. They had concluded in their study that after transforaminal foraminotomy the flexibility did not change and the surgery induced instability was not observed. However, in their study the bone resection was only 45.5% increase in the intervertebral foraminal area. This is significantly smaller resection compare to the clinical FELF resection area measured in the above 2 study.
While these studies have provided valuable insights into the extent of foraminal enlargement and bone resection, the relationship between the amount of facet joint resection and postoperative instability or the progression of scoliosis remains unexplored.
5. Preventing Scoliosis Progression After FELF at L5–S1: The Role of the Lateral Kissing Spine
In the cases illustrated above, patients without a lateral kissing spine are at a higher risk of postoperative progression of scoliosis. We had reviewed 11 patients who underwent FELF for foraminal stenosis at L5–S1 and had a follow-up period of at least 1 year were reviewed postoperatively. The patients’ average lumbar Cobb angle prior to surgery was 14.9° (range, 1°–38°). The lumbar Cobb angle was measured on coronal lumbar spine x-rays taken before surgery and at the 1-year postoperative follow-up. Changes in the Cobb angle was recorded as positive if tilting occurred towards the side of surgery and as negative if the tilting decreased.
We used the term "lateral kissing spine" to describe the radiological phenomenon where the L5 transverse process makes contact with the sacral ala (Figure 1). Based on this categorization, the patients were divided into 2 groups: those with a lateral kissing spine (7 patients) and those without (4 patients). The lateral kissing spine group had an average lumbar Cobb angle change of only +0.29°, whereas the non-lateral kissing spine group showed a significantly greater average Cobb angle change of +2.1°. Statistical analysis using the t-test revealed a significant difference (p=0.01), suggesting that the presence of a lateral kissing spine mitigates the progression of scoliosis towards the side of surgery. Notably, the lateral kissing spine group included patients with severe preoperative scoliosis (up to 35°), yet the progression of scoliosis was effectively prevented. Of note, patients with severe preoperative Cobb angle (lumbar Cobb angle 35°) did not advance postoperative if when the lateral kissing spine exist.
We had reviewed only patients who underwent FELF on L5–S1 because FELF at the L5–S1 level offers a more extensive bone resection resulting in instability of the facet. Excess bone resection is necessary on L5–S1 due to anatomical constraints, such as the iliac crest, the L5 transverse process, and a hypertrophic L5–S1 facet joint [13,37,38]. This is also important because for sage cannula insertion to avoid postoperative dysthesia which is one of the complications of FELF. Inokuchi et al. [39] reported a case of full-endoscopic discectomy at L5–S1 that required partial resection of the L5 inferior articular process and the S1 pedicle to create adequate space for safe cannula insertion.
Therefore, we suggest that to avoid the need for additional surgery following FELF, patients should be evaluated for the presence of the lateral kissing spine at L5–S1, especially if degenerative scoliosis is present.
6. Indications for FELF: Criteria for Selection and When to Prefer Lumbar Interfusion
Yamashita [40] has proposed that the most suitable indications for FELF include lumbar disc herniation within the foramen and osteophyte formation on the SAP. He also highlighted that progressive spondylolisthesis and scoliosis are not optimal indications for FELF due to the heightened risk of recurrent lumbar stenosis.
This paper further refines the criteria for patient selection by analyzing preoperative Cobb angles and radiographic characteristics. Patients with a postoperative Cobb angle exceeding 10° are at a higher risk of recurrence. Likewise, those with a preoperative Cobb angle greater than 7.5° who require extensive bone resection, such as in L5–S1 FELF procedures, warrant careful evaluation due to the potential for scoliosis progression. Additionally, the presence of lateral kissing spines serves as an important index for surgical decision-making.
While FELF has shown favorable surgical outcomes for foraminal stenosis, particularly in cases with stable lumbar spondylolisthesis, fusion surgery remains the gold standard for foraminal stenosis associated with instability [4]. For patients with a higher risk of recurrence, such as those with a Cobb angle exceeding 10 degrees or undergoing treatment for L5–S1 stenosis without lateral kissing spines, lumbar interfusion should be considered as the preferred surgical approach to minimize the likelihood of complications and disease progression.
CONCLUSION
Foraminotomy via transforaminal FESS is an effective treatment for foraminal stenosis and can be performed under local anesthesia, allowing surgical intervention in patients with severe comorbidities. However, it is important to monitor patients for potential development of degenerative scoliosis post-surgery as a result of instability caused by resection of bone surrounding the foramen. Notably, cases requiring L5–S1 foraminotomy should be evaluated for the presence of the lateral kissing spine. Patients at higher risk of instability or progression of degenerative scoliosis should be considered for lumbar interbody fusion as the initial surgery.





