Ultrasound-Guided Lumbar Nerve Root Hydrodissection: Fluoroscopy-Assisted Evaluation of Injectate Distribution and Feasibility

Abstract

Introduction

Selective nerve root block is commonly performed under fluoroscopy, but reports of ultrasound (US) guidance remain limited. Although hydrodissection has shown favorable outcomes in peripheral nerves, few studies have evaluated the lumbar nerve root. Therefore, we aimed to evaluate the injectate distribution, reproducibility, safety, and clinical efficacy of US-guided hydrodissection of the lumbar nerve root (US-HNR) using fluoroscopic assistance.

Methods

US-HNR was performed in 46 patients with lumbar radiculopathy between January and September 2025. A standardized injectate (contrast agent 3 mL, saline 6 mL, and lidocaine 1% 1 mL; total 10 mL) was administered. Distribution was classified into four types under fluoroscopy. The Numerical Rating Scale (NRS), Oswestry Disability Index (ODI), and patient satisfaction were evaluated pre- and postprocedurally.

Results

Contrast evaluation was possible in 40 patients. Distribution patterns were localized within the periradicular region of the target nerve root in 30%, ipsilateral epidural in 38%, contralateral epidural in 25%, and contralateral nerve root in 8%, suggesting localized spread in one-third of cases. NRS improved significantly from 7.1 to 5.0 at 1 week, 4.2 at 1 month, and 3.8 at 3 months (all p<0.001), with responder rates of 52%, 60%, and 58%. ODI improved from 42 to 32 at 1 month and 28 at 3 months (p<0.01), with responder rates of 54% and 62%. No serious complications were observed. Two experienced transient radiating pain during injection.

Conclusions

US-HNR was safe and feasible, avoiding radiation and direct nerve puncture. This simple technique showed favorable clinical outcomes and may represent a promising treatment for lumbar radiculopathy.

Introduction

Lumbar selective nerve root block (SNRB) is widely used for diagnosing and treating radicular pain associated with lumbar disc herniation and spinal canal stenosis. Traditionally, the procedure is performed under fluoroscopic or computed tomography (CT) guidance, enabling direct confirmation of needle tip position and contrast medium spread. However, it has limitations, including radiation exposure, procedure-related pain, risk of needle injury, and additional time and equipment¹⁾. Recent advances in ultrasonography have highlighted ultrasound (US)-SNRB as a less invasive alternative²⁾. Nevertheless, visualizing deep structures in the lumbar region remains challenging, and technical hurdles limit its widespread adoption^3,4).

Regarding injectate volume, most studies on cervical and lumbar SNRB have employed small amounts (approximately 1-5 mL) of lidocaine-steroid mixtures, as larger volumes are avoided to maintain selectivity^5-10). Recent reports on US-cervical SNRB have supported its efficacy and safety^5,9,11,12). In contrast, data on lumbar US-SNRB remain limited, and the level of evidence is considerably lower than that for procedures in the cervical region.

Furthermore, favorable therapeutic outcomes of hydrodissection in the cervical spine and peripheral nerve regions have been reported^12-15). However, the clinical efficacy and safety of hydrodissection targeting the lumbar nerve root (HNR) have not been sufficiently evaluated. Additionally, hydrodissection typically requires a large injectate volume compared to SNRB; however, its distribution characteristics and clinical significance remain unclear. Nerve root selectivity is crucial; therefore, large-volume hydrodissection should be carefully evaluated, as high volumes may compromise selectivity.

We used a standardized 10 mL mixture comprising 3 mL nonionic contrast agent, 6 mL normal saline, and 1 mL of 0.5% lidocaine^11-15). This study employed fluoroscopic recording of injectate distribution, enabling a hybrid evaluation that combined US and fluoroscopy. This approach integrated the minimally invasive nature of US with the objective, quantitative assessment of contrast distribution provided by fluoroscopy.

We hypothesized that a 10 mL injectate volume would not result in a nonselective distribution and that, depending on anatomical conditions, a proportion of cases would remain localized around the target nerve root. Conversely, in cases with minimal stenosis, we anticipated wider spread patterns through the epidural space. We evaluated changes in the Numerical Rating Scale (NRS) and Oswestry Disability Index (ODI) scores and patient satisfaction to verify the safety and feasibility of this US-HNR technique.

Methods

Ethics

The institutional ethics committee approved this study. In all cases, the procedure's purpose, details, anticipated benefits, and potential risks were thoroughly explained, and written informed consent was obtained.

Patients

Patients were eligible for inclusion if they presented with lumbar radicular symptoms (radiating pain from the lumbar region or buttock to the lower extremity), had failed to achieve sufficient pain relief after more than 1 month of conservative treatment (analgesics or physical therapy)^12,16), showed concordance between magnetic resonance imaging (MRI) findings and clinical symptoms, and personally requested the block. Medication type, disease duration, and baseline NRS scores were not exclusion criteria. The exclusion criteria included cases in which US-guided procedures were technically challenging because of postoperative changes or anatomical deformities that precluded the identification of normal structure (procedures were still performed at postoperative levels if anatomy could be identified), as well as patients with severe renal dysfunction or a history of contrast agent allergy. In patients with severe degenerative changes, postoperative alterations, or marked foraminal narrowing that made essential bony landmarks indistinct, safe US-guided needle placement could not be reliably achieved. These cases were excluded solely for technical and safety reasons, rather than clinical characteristics, to avoid inaccurate needle placement. In these cases, fluoroscopic guidance was used without a contrast medium, with the needle advanced toward the target nerve root until radiating pain was elicited, followed by injection. In fluoroscopy-only cases, contrast medium was not used, and the procedure was performed as a fluoroscopy-guided nerve root hydrodissection rather than US-HNR. Although the injectate did not contain contrast medium, the total volume (10 mL) was unchanged. Since these cases did not receive the contrast-containing injectate and the guidance technique differed from the US-HNR used in the main analysis, including them could have confounded the evaluation of injectate distribution. Therefore, they were excluded from the contrast pattern analysis but were included in the clinical outcome analyses (NRS, ODI, and satisfaction).

Baseline evaluation

Clinical evaluations included an NRS (0-10) measured at baseline, 1 week, 1 month, and 3 months, and the ODI (nine-item version excluding the sexual activity item) assessed at baseline, 1 month, and 3 months. Additionally, patient satisfaction (five-point Likert scale) was recorded at 1 and 3 months¹⁶⁾.

Procedure

The procedures were performed in a fluoroscopy room, with the patient in a prone position, with a pillow placed under the iliac crest to reduce lumbar lordosis. In some cases, at the L5 level, an additional towel was placed beneath the pillow to further adjust the height. A Konica Minolta SNIBLE 2 US system (C5-2 convex probe, 1-6 MHz) was used, and fluoroscopy was performed using a C-arm. The skin was disinfected with alcohol and thoroughly dried before probe placement to determine the puncture site, after which povidone-iodine was applied.

MRI was performed on all patients before the procedure, and clinical findings were correlated with the imaging findings. Using parasagittal images, the depth from the body surface to the target nerve root was measured (mean 5.3±0.6 cm, range 4.5-6.7 cm). When visualizing the nerve root with US, this pre-measured value was referenced as an auxiliary aid to avoid misidentification or positional errors.

Fluoroscopy use

Fluoroscopy was used only for minimal confirmation. In most cases, a single shot was taken before injection to confirm the level and probe alignment, and another single shot was taken after completion of the injection to document contrast spread. Thus, fluoroscopy was typically performed 1-2 times per procedure, resulting in negligible radiation exposure.

Technique

The injection technique was based on the interfacet out-of-plane approach described by Wan et al.¹⁷⁾ and the L5 nerve root block using the transverse process as a landmark with concomitant nerve stimulation, as described by Sato et al¹⁸⁾. In conventional methods that use the transverse process as a landmark, bony references decrease toward the lateral side, making positioning unstable. Katayama et al.¹⁹⁾ recently described the continuous contour formed by the superior articular process (SAP) and accessory process (AP) as the “Half-Mountain View,” providing a stable primary landmark.

The probe was placed over the posterior superior iliac spine and slid medially-cranially to identify the sacrum-L5 lamina slope. A slight lateral slide revealed the Half-Mountain View¹⁹⁾, beneath which the nerve root appeared as a hyperechoic circular structure and served as the needle target (Figures 1-3).

Figure 1.

Probe positioning and Half-Mountain View in a lumbar spine model.

A: Posterior view of a lumbar spine model showing the probe position (red line, L4 nerve root).

B: Sagittal view of Half-Mountain View (white dotted line) and the L4 nerve root.

Figure 3.

Ultrasound image showing the Half-Mountain View (dotted line) and the target L5 nerve root. The Half-Mountain View can appear less distinct at the L5 level due to a smaller accessory process.

Figure 2.

Ultrasound image showing the Half-Mountain View (dotted line) and the target L4 nerve root.

At the L5 level, the AP was small, and the Half-Mountain View¹⁹⁾ was sometimes indistinct. In cases where the intertransverse ligament could be observed superficial to the nerve root, the “pop” sensation upon passing through this ligament was utilized as tactile feedback⁴⁾. When nerve root visualization was difficult, the injection was performed by avoiding vessels delineated with color Doppler, which accompanied the nerve root, assuming that the nerve root was located immediately beneath them (Figure 4).

Figure 4.

Ultrasound image with color Doppler showing periradicular vessels (red signal) passing through the intertransverse ligament and the target nerve root.

During injection, many patients experience transient heaviness or discomfort in the lower extremities; however, no procedure should be interrupted due to pain. Sato et al.¹⁸⁾ reported clinical efficacy, even in the paraneural type (as defined in their classification), suggesting that direct nerve root puncture is not required if the needle tip is in proximity.

Needle placement

The puncture site and needle trajectory were determined based on the probe position, and a 70-mm needle was advanced under US guidance to a point immediately above the nerve root (Figure 5). A mixed injectate comprising 3 mL contrast medium, 6 mL normal saline, and 1 mL 0.5% lidocaine (total 10 mL) was then administered. After confirming a small amount of spread along the nerve root, a full volume was injected. In cases where contrast medium was not used, a contrast-free injectate of the same total volume (10 mL) was administered instead.

Figure 5.

Schematic illustration showing the probe position, needle trajectory, and insertion angle relative to the skin surface.

Post-procedural observation and safety management

Since a relatively large volume (10 mL) containing 3 mL of contrast agent was used, all patients were observed at rest for one hour after the procedure.

Evaluation of contrast patterns

The spread of the contrast medium was classified into the following four types, with modifications from previous reports²⁰⁾: (1) localized within the periradicular region of the target nerve root (Figure 6A), (2) spread to the ipsilateral epidural space (not crossing the midline), (3) spread to the contralateral epidural space (crossing the midline), and (4) spread to the periradicular region of the contralateral nerve root (including other-level nerve roots) (Figure 6B). When a slight spread of the contrast medium was observed toward the dural side, the case was categorized as type 2 or type 3.

Figure 6.

Fluoroscopic contrast distribution patterns after ultrasound-guided lumbar nerve root hydrodissection.

A: Fluoroscopic image after ultrasound-guided injection at L4, showing type 1 distribution localized to the periradicular region of the target nerve root, without epidural spread.

B: Fluoroscopic image after injection at L5, showing type 4 distribution with contrast crossing the midline into the contralateral epidural space and extending to the contralateral S1 and S2 nerve roots.

Statistical analysis

Statistical analyses were performed using paired t-tests or Wilcoxon signed-rank tests, depending on data distribution. Changes in NRS and ODI were evaluated using these paired comparisons. A ≥2-point improvement in NRS was adopted as the responder threshold based on the minimally clinically important difference reported for lumbar radicular pain. Analyses were conducted using a per-protocol approach. To assess potential attrition bias, baseline characteristics (NRS and ODI) were compared between completers and dropouts. A sensitivity analysis using last observation carried forward was also performed for both NRS and ODI.

Results

Overall, 46 patients (age range, 30-82 years; mean age, 64.8; 20 males [43.5%]) were enrolled between January and September 2025. The right and left sides were affected in 24 (52.2%) and 22 (47.8%) cases. The affected levels were L5, S1, L4, and L3 in 33 (71.7%), 7 (15.2%), 4 (8.7%), and 2 (4.3%) patients. The mean depth from the body surface to the nerve root, as measured on MRI, was 5.3±0.6 cm (range, 4.5-6.7 cm). The underlying diagnoses were lumbar spinal canal stenosis (LSCS) and lumbar disc herniation in 30 (65.2%) and 16 (34.8%) patients (Table 1).

Table 1.

Patient Characteristics (n=46).

Variable	Value
Age, years (range, mean±SD)	30-82, 64.8±11.2
Sex, n (%)	Male 20 (43.5), Female 26 (56.5)
Side, n (%)	Right 24 (52.2), Left 22 (47.8)
Affected level, n (%)	L5: 33 (71.7), S1: 7 (15.2), L4: 4 (8.7), L3: 2 (4.3)
MRI-measured depth of target root (cm)	5.3±0.6 (range 4.5-6.7)
Diagnosis, n (%)	Lumbar spinal stenosis 30 (65.2), Disc herniation 16 (34.8)

MRI: magnetic resonance imaging; SD: standard deviation

US-HNR were attempted in all patients, and fluoroscopy-assisted evaluation of contrast patterns was possible in 40 patients (87.0%). In six (13.0%) cases, evaluation was impossible: in four, due to severe degeneration that precluded visualization of US landmarks, and in two, because contrast medium was avoided due to impaired renal function. These cases were treated using fluoroscopy alone and were excluded from the distribution classification.

Contrast patterns

Contrast evaluation was performed in 40 patients. The distribution patterns were as follows (Figure 6): type 1 (localized within the periradicular region of the target nerve root), 12 (30.0%) cases; type 2 (ipsilateral epidural), 15 (37.5%) cases; type 3 (contralateral epidural), 10 (25.0%) cases; and type 4 (spread to the periradicular region of the contralateral nerve root), 3 (7.5%) cases. Even after administering a relatively large volume (10 mL), approximately 30% remained localized within the periradicular region of the target nerve root. The frequency of type 1 was significantly higher in patients with LSCS (p=0.02; Table 2).

Table 2.

Contrast Distribution Patterns (n=40 Evaluable Cases).

Type	n (%)
One Target periradicular	12 (30.0)
Two Ipsilateral epidurals	15 (37.5)
Three Contralateral epidurals	10 (25.0)
Four Contralateral periradicular	3 (7.5)

Data show fluoroscopically confirmed spread classifications following ultrasound-guided hydrodissection of the lumbar nerve root (US-HNR).

NRS

The NRS improved significantly from 7.1±1.4 (n=46) before the procedure to 5.0±1.6 at 1 week (n=42, p<0.001), 4.2±1.8 at 1 month (n=36, p<0.001), and 3.8±2.0 at 3 months (n=31, p<0.001). The mean improvement at 3 months was −3.3 points (46.5% reduction). The responder rates were 52%, 60%, and 58% at 1 week, 1 month, and 3 months (Table 3A).

Table 3A.

Pain Intensity (Numerical Rating Scale, Mean±SD).

Time point	N	Score	p vs. baseline	Responder rate
Baseline	46	7.1±1.4	–	–
One week	42	5.0±1.6	<0.001	52%
One month	36	4.2±1.8	<0.001	60%
Three months	31	3.8±2.0	<0.001	58%

Responder is defined as ≥2-point improvement from baseline.

SD: standard deviation

ODI

The ODI improved from 42±9.1 before the procedure to 32±10.3 at 1 month (n=36, p=0.004) and 28±11.5 at 3 months (n=31, p=0.002). The mean improvement at 3 months was −14 points (33.3% reduction). The responder rates were 54% and 62% at 1 and 3 months (Table 3B).

Table 3B.

Disability (Oswestry Disability Index 9-Item, Mean±SD).

Time point	N	Score	p vs. baseline	Responder rate
Baseline	46	42±9.1	–	–
One month	36	32±10.3	0.004	54%
Three months	31	28±11.5	0.002	62%

SD: standard deviation

Baseline NRS and ODI did not differ significantly between patients who completed follow-up and those who were lost to follow-up (NRS: 7.5 vs 7.8, p=0.80; ODI: 19.7 vs 18.6, p=0.81), suggesting minimal attrition bias.

Patient satisfaction

Satisfaction scores (five-point Likert scale) were 3.7 and 3.6 at 1 and 3 months. The proportions of patients who were satisfied or very satisfied were 61% and 58% at 1 and 3 months.

Complications (Table 4)

Table 4.

Safety and Adverse Events.

Event	n (%)	Clinical course
Severe complications	0	–
Intravascular injection	0	–
Contrast allergy	0	–
Transient radicular pain	2 (4.3)	Resolved spontaneously after injection

No serious complications were observed in any case. Contrast allergy or intravascular injection did not occur in any case. Transient radiating pain during injection occurred in two cases and resolved spontaneously.

Dropout and transition to surgery

Nine (19.6%) patients did not complete the final follow-up, comprising seven who underwent surgery and two who discontinued follow-up. Surgical cases were included in the analysis of contrast findings and short-term scores at the time of the block. The follow-up completion rates were 37 (80.4%; n=37) and 31 (67.4%; n=31) at 1 and 3 months.

Discussion

This study is the first to visualize and evaluate, with fluoroscopic assistance, the spread, safety, and procedural reproducibility of a 10 mL injectate in US-HNR. Although previous lumbar SNRB studies have used small volumes, the behavior of larger hydrodissection volumes remains unclear.

In the cervical spine, studies have shown that 1 mL does not produce epidural spread, whereas 4-5 mL results in spread into the foraminal epidural space in approximately one-quarter of cases^5,9). In the lumbar region, 3 mL has been associated with semicircumferential to circumferential spread around the nerve root, whereas 5 mL often results in circumferential and epidural spread^6-8). However, even 1 mL can spread to adjacent nerve roots¹⁰⁾, raising concerns about diagnostic selectivity with volumes >0.5 mL⁷⁾. Therefore, minimal volumes are generally recommended for lumbar SNRB.

However, our findings showed that even with 10 mL, approximately 30% of cases demonstrated localized spread within the periradicular region of the target nerve root, likely due to anatomical structures (bony structures, ligaments, and epidural fat). This aligns with Hogan²¹⁾, who noted that distribution can remain localized even with 10 mL. Variability in foraminal anatomy (fat, vessels, and ligaments) also influences injectate distribution²²⁾. Additionally, our modified proximal foraminal injection technique, guided by Half-Mountain View¹⁹⁾, may have contributed to localization, possibly due to adjacent bony structures restricting spread. The predominance of older patients with severe stenosis may have reinforced this effect.

In other words, because the needle was positioned just above the nerve root without contacting it, the injectate did not enter the intraneural sheath but instead spread within a confined bony-ligamentous periradicular compartment bordered by the SAP, pedicle, and foraminal ligaments. These foraminal structures can segment the space and substantially limit the available volume²²⁾, particularly in stenotic foramina, thereby reducing the potential for wide dispersion. This anatomical confinement likely contributed to the relatively high rate (30%) of target-localized spread observed in this study.

US-SNRB is commonly used to treat patients with radiculopathy^5,9). However, because myofascial entrapment neuropathy can involve multiple sites, blocking a specific site may only partially relieve pain. Therefore, separating the compressed nerve from the surrounding fascia and soft tissue is important to prevent adhesions and obstruction. In this regard, using a 10 mL volume for hydrodissection offers both mechanical dissection and dilution/washout of inflammatory mediators.

Similar hydrodissection volumes have shown circumferential spread and pain relief in cervical and peripheral nerves, supporting the concept that mechanical dissection contributes to symptom improvement^13-15).

For cervical radiculopathy blocks, only minimal differences in short-term outcomes have been reported between procedures with and without steroids, suggesting that local anesthetics or block techniques contribute to pain relief¹¹⁾. A retrospective study of US-guided hydrodissection for cervical stenosis demonstrated favorable outcomes regardless of stenosis severity¹²⁾, indicating that the procedure itself plays a major role in clinical efficacy. These findings are consistent with those of this study, indicating that treatment outcomes may be influenced more by the mechanical dissection effect and injectate volume than by the specific type of medication used. Furthermore, basic research has demonstrated that even slight nerve compression can induce neuropathic pain, emphasizing the importance of circumferential dissection and the release of the nerve environment¹⁴⁾.

Contrast media concentrations of 180-240 mg I/mL are typically recommended for myelography, as noted in the package inserts of Omnipaque and Isovue-M. However, such high concentrations are unnecessary for nerve root injections, where only basic visualization is required. In this study, contrast medium was necessary for fluoroscopic confirmation, but we also aimed to preserve the low-viscosity characteristics of hydrodissection. Therefore, Omnipaque 240 was first diluted at approximately a 1:1 ratio (4 mL contrast, 5 mL saline, and 1 mL lidocaine), yielding about 120 mg I/mL. Although at least 4 mL of contrast was initially considered necessary, subsequent cases showed that a mixture containing 3 mL of contrast, 6 mL of saline, and 1 mL of lidocaine (approximately 80 mg I/mL) still provided adequate visualization of the nerve root and foramen. Lidocaine (1 mL) was added to reduce discomfort during injection¹³⁾.

Lumbar SNRB has traditionally been performed under fluoroscopic or CT guidance; however, with recent advances in US equipment, reports of US-guided procedures have increased. However, its use in the lumbar region is less commonly adopted than in the cervical region or peripheral extremities, and developing an appropriate technique tailored for this area is necessary.

Using the Half-Mountain View¹⁹⁾ as a landmark, the needle was advanced to a position immediately above the nerve, with the aim of minimizing the risk of direct nerve puncture. Although the needle-nerve distance was not objectively measured, this landmark-based technique provides a stable reference for consistent needle orientation.

The bony contour that creates this view, formed by the SAP and AP, provides a more stable and reproducible landmark than transverse-process-based lateral approaches, where bony references diminish toward the lateral side. By maintaining a consistent osseous outline directly above the nerve root, this landmark facilitates reliable needle orientation even in deeper lumbar levels and helps prevent unintended lateral deviation during advancement. This stability supports consistent needle positioning and represents a key technical advantage of this approach, enhancing its procedural reproducibility.

Although the Half-Mountain View was identifiable in most cases, its clarity varied across levels; the contour was often less distinct at the L5 level and occasionally at other segments as well.

The vascular and ligamentous structures accompanying the nerve root are not consistently visualized under US, and a clear “pop” sensation during needle advancement is not observed in all cases. Accordingly, the Half-Mountain View, together with preprocedural MRI-based estimation of the depth to the periradicular region, serves as the primary landmark for needle guidance in this technique.

When present, such findings support accurate needle placement. For example, visualizing a vessel may indicate that the nerve root lies just beneath; if tactile feedback is perceived when passing through a ligament, the needle can be considered to have reached the compartment immediately above the nerve root. Although not always observed, these findings provide useful evidence for enhancing operator confidence. In this study, using the Half-Mountain View¹⁹⁾ as the basic landmark and supplementing it with vascular and ligament visualization and tactile feedback, the reproducibility and safety of non-contact injection into the nerve root could be further improved.

Important safety findings were also obtained. In fluoroscopy-guided lumbar transforaminal epidural steroid injection, Smuck et al.²³⁾ reported that intravascular injection occurred in 21% of cases, and simultaneous injection into the epidural space and vessels occurred in 8%. Conversely, with the US-HNR used in this study, the vessels could be delineated in advance using color Doppler, which is considered advantageous for reducing inadvertent intravascular injections. No major complications, including intravascular injection, were observed in this study.

Limitations

This study had some limitations. First, because a mixture containing a contrast medium was used, the spread of the injectate might differ from that in actual hydrodissection. Second, this was a single-arm observational study rather than a comparative trial with conventional low-volume SNRB; therefore, it cannot be concluded that this technique is superior. Third, pain scores (NRS), functional scores (ODI), and satisfaction were secondary outcome measures, and interpretation requires caution because patient backgrounds―including medication use, disease duration, and baseline severity―were not uniform. Furthermore, patients in whom severe degenerative change, postoperative alteration, or marked foraminal narrowing obscured essential bony landmarks were excluded for technical and safety reasons. These factors may have contributed to selection bias and heterogeneity in treatment response. Additionally, fluoroscopy use for evaluation differs from US-only procedures in daily practice. Furthermore, the anatomical diversity was insufficiently stratified for analysis, and the observation period was limited to 3 months. In addition, the clarity of the Half-Mountain View varied across levels and was often less distinct at L5. Such variability, combined with the need to interpret subtle ultrasonographic cues in real time, means that reproducibility may partly depend on operator experience, especially when primary landmarks are indistinct. Future multicenter randomized controlled trials with long-term follow-up are warranted.

Conclusions

US-HNR is feasible and showed no serious complications, suggesting a favorable safety profile. Beyond documenting contrast spread, this study demonstrates that the procedure can be performed easily in an outpatient setting without radiation exposure or fluoroscopic equipment and represents the first report worldwide to visualize its dynamics by combining fluoroscopy and contrast. This technique exhibits features that were not previously available for treating lumbar radiculopathy and that have clinical significance.

Author Contributions: Y.A. designed the study, performed the procedures, analyzed the data, and drafted the manuscript. Y.O. and K.K. supervised the study and critically reviewed the manuscript. M.T., R.A., S.T., and Y.K. assisted with patient recruitment and clinical coordination.

Conflicts of Interest: The authors declare that there are no relevant conflicts of interest.

Ethical Approval: Approved by the Institutional Review Board of Yawata Medical Center.

Informed Consent: Informed consent for publication was obtained from all participants in this study.