Article Text
PDF
Abstract
Introduction Recent studies have proposed revised anatomical targets to improve accuracy of genicular nerve (GN) radiofrequency ablation (RFA). This study aims to compare the accuracy of classical and revised techniques for fluoroscopic-guided GN-RFA in cadaveric models.
Materials and methods Fourteen knees from seven fresh frozen human cadavers were included in this study. For each cadaver, RF cannulas were placed to capture the GN according to the current targets in one knee, and the revised targets in the other knee, randomly. The stylet was removed from the cannula, plunged into non-diffusible black paint, and reintroduced entirely in the cannula, to create a limited black spot on the tissues at the top of the active tip. Anatomical dissection was performed, and the accuracy of both techniques was compared.
Results The mean distance from the top of the active tip to the nerve was significantly lower with revised than current targets for the superior-medial GN (0.7 mm vs 17.8 mm, p=0.01) and the descending branch of the superior-lateral GN (3.7 mm vs 24.4 mm, p=0.02). In both superior-medial GN and superior-lateral GN, the accuracy rate was higher with revised than current targets: 100% vs 0% and 64% vs 35%, respectively. In addition, the accuracy of revised targets for the recurrent fibular nerve and the infrapatellar branch of saphenous nerve was 100%.
Conclusion This study demonstrates that the revised targets are more accurate than the current targets for GN-RFA.
- radiofrequency ablation
- pain medicine
- anatomy
Statistics from Altmetric.com
Introduction
First described by Choi et al 1 in 2011, genicular nerve (GN) radiofrequency ablation (RFA) is a promising technique for the treatment of chronic knee pain. It disrupts articular sensory afferent pathways from the knee joint to the central nervous system, decreasing the nociceptive input.2 Because of the limited size of RF lesions, its success relies on the precise placement of the cannulas, as close as possible to the nerves supplying the knee joint capsule. However, the complexity of the knee innervation has resulted in a disparity in procedural techniques among the available controlled and observational studies.3 Two recent reviews have concluded that there is a lack of clarity about anatomical targets of this procedure and that more research is needed to better identify neural targets.3 4 Newer, more robust cadaveric dissection studies5–7 have called into question the anatomical foundations of the targets originally proposed by Choi et al.1 These have resulted in several papers proposing new protocols with revised anatomical targets to capture GN with more accuracy.8–11 Two very recent studies combined a comprehensive literature review with anatomical dissection of 21 cadavers to identify reliable targets,6 and validated their accuracy in a cadaveric model for GN blockade.9
Since the current anatomical targets for GN-RFA used clinically for a decade have allowed chronic knee pain reduction to a certain degree, it is necessary to compare, with a robust methodology, first, the accuracy of the revised targets to the current targets in a cadaveric model, and, second, their effectiveness in patients suffering from chronic knee pain. The current study aims to compare the accuracy of the revised anatomical targets9 to that of the current targets for fluoroscopic-guided GN-RFA in cadaveric knees. Moreover, as the precise location and the target point of the superior medial genicular nerve (SMGN) and superior lateral genicular nerve (SLGN) as well as their relation to the genicular arteries are still being debated,12 13 we hope to clarify these controversies with new anatomical material.
Methods
This study was conducted at the Laboratories of Anatomy and Experimental Surgery, UCLouvain. All the bodies were donated to the institution in accordance with Belgium law and regulations. Approval from the University’s Research Ethics Committee was obtained. The study was conducted in two parts. In the first part, we used four cadaveric knees (two fresh frozen and two embalmed with zinc chloride solution) for pilot trials in order to design the best way to assess the accuracy of anatomical landmarks for fluoroscopic guided GN-RFA. In the second part (main study), we compared the accuracy of current vs revised targets in 14 fresh frozen knees.
Pilot study
In the knees of one fresh frozen cadaver, we tried to reproduce the clinical conditions allowing the creation of RF lesions on the GNs that could be appreciated. After thawing at room temperature for 24 hours prior to the experimentation, both knees were protected in a waterproof bag and heated by immersion in a water bath at 56°C for 30 min, to reach a nearly normal body temperature. A 18-gauge 10 mm active tip, 145 mm long monopolar RF introducer cannula was placed at each anatomic site under fluoroscopic guidance, according to current targets1 2 10 14 15 for one knee and revised targets9 for the contralateral knee. The initial temperature of the limbs was 34°C, as measured by the thermocouple built in each electrode. Each site was lesioned with a RF generator (Baylis Pain Management Generator V3.0) at tissue temperature of 80°C for 90 s. Both limbs were dissected by an anatomist with 9 years of experience, who had performed dissection in several previous studies addressing the innervation of the knee joint. Unfortunately, even with the use of 3.5× magnification lens, no lesion could be identified in the tissues of both limbs. The limbs were closed and we performed several successive RF tests at different sites and different tissues in order to highlight the lesioned tissue so that it could be visually appreciated. We raised the temperature to 90°C for 120 s, injected 2 mL of 0.9% saline or 1% lidocaine prior to lesioning16 and also tested the injection of 1 mL of egg white prior to lesioning.17 Lesions were clearly visualized when RF was applied on muscle and skin, but not on the nerves, tendons, knee capsule, fat, or deep tissues.
In one knee of the second cadaver, 22-gauge, 10 mm active tip, 100 mm long RF cannulas were placed under fluoroscopic guidance at each site, according to current targets for one knee and revised for the other one. Through each cannula, we injected a volume of 0.1 mL of colored lidocaine (1 mg of methylene blue powder per 1 mL of 1% lidocaine solution). After completion of all the injections, the limbs were dissected to check for accuracy. Since our study aimed to assess the accuracy of anatomical landmarks with regards to RF lesion size, we first measured the diffusion surface of blue lidocaine with a sliding digital caliper. The mean dimensions were 14.6 mm (SD 4.9)×21.1 mm (SD 6.1). We observed that the blue solution spread beyond the limits of all types of RF lesion and, therefore, would not reliably assess the accuracy for RFA.
In the next knee, the same RF cannulas were placed, but instead of inserting a RF electrode or injecting a very small volume, the stylet was removed from the cannula and plunged into black paint, then reintroduced entirely into the cannula. The procedure was repeated twice to ensure that the black paint stained the tissue at the top of the active tip like a fountain pen stains a paper. After dissection, a limited black spot was consistently observed, without diffusion to surrounding tissues. We concluded that this would be a more reliable method to assess the accuracy of RF lesions regarding the anatomical targets.
Main study
Fourteen knees from seven fresh frozen cadavers, four men and three women, aged 78.4±10.0 years, without evidence of prior knee surgery or trauma, were used. For each cadaver, one knee was randomly assigned to the RF simulation procedure with current targets,2 10 14 15 while the other one underwent the same procedure with revised targets (figure 1) as described below.
Revised anatomical targets for fluoroscopic guided RF ablation of genicular nerves. (A) Anterior-posterior X-ray view. (B) Lateral X-ray view. (C) Landmarks for IPBSN targeting. (D) Landmarks for RFN targeting. F, fibula head; GT Gerdy’s tubercle; IMGN, inferior medial genicular nerve; IPBSN, infrapatellar branch of saphenous nerve; P, patella; RFN, recurrent fibular nerve; SLGN, superior lateral genicular nerve; SMGN, superior medial genicular nerve; TT, tibial tuberosity.
For the SMGN, we started with an anterior-posterior (A-P) X-ray view. The RF cannula was inserted 1 to 2 cm medial to the femur, at the junction of the medial femoral cortex and the medial condyle, and advanced to the superior edge of the medial condyle until the tip touched the bone. The C-arm was then rotated to have a true lateral view in which both femoral condyles were superimposed. The tip of the cannula was then adjusted to fit above or just in front the adductor tubercle (AT), (figure 1A,B).
The inferior medial genicular nerve (IMGN) was targeted as is currently done in clinical practice1 2 (figure 1A,B).
For the superior lateral genicular nerve (SLGN), on the A-P view, the RF cannula was inserted 1 to 2 cm lateral to the femur and advanced obliquely toward the superior edge of the lateral femoral condyle until the tip touched the bone. Then, in a true lateral view, the needle tip was adjusted to fit the target area located at the junction between the superior edge of the lateral condyle and the posterior femoral cortex (figure 1A,B).
For the infrapatellar branch of saphenous nerve (IPBSN), we drew a transversal line passing by the apex patellae and another one passing by the top of the tibial tuberosity. Four centimeters medially to the apex patellae, we drew a longitudinal line connecting both transversal lines (figure 1C), which corresponds to the targeted treatment line. The RF cannula was inserted perpendicularly through the skin wheal placed at two extremes of this line to mark the limits of the treatment zone.
For the recurrent fibular nerve (RFN), a longitudinal line was drawn below the Gerdy’s tubercle and the target point was located on this line, 1 cm below the inferior edge of the tubercle (figure 1D). The RF cannula was inserted just below the Gerdy’s tubercle and advanced longitudinally under the muscles until the tip touched the bone. The length of the cannula inserted above the 10 mm active tip was marked so as not to exceed 1 cm.
All procedures were performed under fluoroscopic guidance with 22-gauge, 10 mm active tip, 100 mm long RF cannulas (Cosman Medical, Inc, Burlington, Massachusetts, USA). For all the targeted sites, after verification of adequate cannula placement, the stylet was removed, soaked entirely in black paint and reintroduced entirely in the cannula until the distal opening of the active tip. This operation was repeated twice for each site to ensure that the tissue at the top of the active tip was dyed. Fluoroscopic images were obtained before and after the marking to verify the needle position. The stylet was then removed, after which the cannula was removed too.
After the completion of all RF simulation procedures, the limbs were dissected by the same experienced anatomist to assess for accuracy, based on location of the black mark relative to the targeted GN. The shortest distance from the midpoint of the black mark to the targeted nerve was measured with a sliding digital caliper. Then, considering the midpoint of the black mark as the top of the active tip, and considering the direction of cannula placement and its final position during the procedure, we traced the theoretical location of the 10 mm active tip. We measured the shortest distance from the 10 mm active tip to the targeted nerve. To assess the accuracy of each cannula placement, we traced the hypothetical largest RF lesion volume (transversal distance 7.4 mm, longitudinal distance 12.9 mm, distance beyond the tip 2.4 mm)18 19 around the 10 mm active tip to investigate if the targeted nerve would have been captured if a RF lesion had been applied. For the IPBSN, we considered the target accurate if the nerve was found within the limits of the treatment line.
Additionally, during the dissection of the 14 specimens, we checked for the origin of the SMGN and the SLGN, and also investigated if these nerves (or their branches) were running with the SM or SL genicular vessels. Finally, at the end of the dissection of each knee, we fixed a thumbtack through the trunk of SMGN and SLGN into the underlying bone, before their distribution to the knee capsule, and performed true A-P and lateral radiographs (figure 2). These locations, for the 14 specimens, were consolidated on unique A-P and lateral radiographs to generate a frequency map of the most accurate targets for fluoroscopic guided GN-RFA in our sample.
Identification process of the exact points to accurately target the trunk of the SMGN and SLGN in each specimen for fluoroscopic-guided RF ablation. (A) Medial aspect of the thigh and the knee, adductor canal opened: SMGN (white arrows) originating from the NVM, running distally along the adductor magnus tendon (green arrows) towards the AT. (B) A white radio-opaque thumbtack is fixed through the nerve at its bony contact before its distribution to the knee joint. (C) and (D) Posterior-lateral aspect of the distal thigh and the knee, Bi cut and retracted: SLGN (black arrows) establishing a bony contact in the area connecting the posterior cortical of the lateral shaft of the femur (green line) to the posterior-lateral edge of the lateral condyle (green curve), and then dividing in its two terminal branches. A white radio-opaque thumbtack is fixed through the nerve before its bifurcation to capture both terminal branches. (E) Antero-posterior and (F) lateral radiographs of the knee after dissection, showing the location of the thumbtacks. AT, adductor tubercle; Bi, biceps femoris muscle; CFN, common fibular nerve; LE, lateral epicondyle; ME, medial epicondyle; NVM, nerve to the vastus medialis; P, patella; SLGN, superior lateral genicular nerve; SMGN, superior medial genicular nerve; SN, saphenous nerve; TN, tibial nerve; VM, vastus medialis muscle.
Post-procedural anatomical dissections. (A) Medial aspect of the distal thigh and the knee, vastus medialis muscle is partially lifted. Lesioning of the SMGN (white arrows) using revised target: The nerve is in contact with black mark (yellow thick arrow). (B) Medial aspect of the distal thigh and the knee, vastus medialis muscle is partially lifted. Lesioning of the SMGN (white arrows) using current target: The SMGN is too far from the black mark. Even its transversal terminal branch (white arrowheads) cannot be lesioned. (C) Posterior-lateral view of the knee, biceps femoris (BI) cut and retracted. Lesioning of the SLGN using revised target: In this specimen, the bony contact of the SLGN is lower, on the posterior edge of lateral condyle (LC). Only the ascending terminal branch is captured. (D) Posterior-lateral view of the knee, biceps femoris (BI) cut and retracted. With current targets, only the distal part of the ascending terminal branch could be captured. (E) Anterior-medial view of the knee: IPBSN with two terminal branches (yellow arrowheads) running within both extremities of the treatment line. (F) Lateral view of the knee and proximal leg, the proximal part of the muscles has been removed. RFN (orange arrowheads) captured by the black mark at its distal extremity, far away from the CFN. Bi,biceps femoris muscle; CFN, common fibular nerve; F, posterior cortex of thelateral femoral shaft; GT, Gerdy’s tubercle; LC, lateral condyle; P, patella; VM, vastus medialis muscle.
Statistical analysis was performed using SPSS 25.0 (Chicago, Illinois, USA). The various distances are described as median (range). The accuracy of the cannula placement for each nerve with both techniques was described as proportions and compared using the Fisher’s exact test. We used the Mann-Whitney U test to compare the distance from the nerve to the black mark in both techniques. A-P value <0.05 was taken to indicate a significant difference.
Results
The median distance from the top of the active tip to the targeted nerve was significantly lower using revised targets than classical targets for the SMGN (0.0 mm vs 19.0 mm, p=0.01) and the descending branch of the SLGN (2.0 mm vs 24.0 mm, p=0.02) (table 1).
Median distance and range from genicular nerve to the active tip of RF cannula: comparison of classical targets versus revised targets with Mann-Whitney U test
Moreover, for the SMGN, the median shortest distance from the 10 mm active tip to the targeted nerve was significantly lower using revised targets than classical targets (0.0 mm vs 18.0 mm, p=0.02). It was also the case of the descending branch of the SLGN (2.0 mm vs 24.0 mm, p=0.02) (table 1).
In our specimens, the accuracy rate of the simulated RF monopolar lesions was higher with revised targets than classical targets for the SMGN (100% vs 0%) and the SLGN (64% vs 35%) (table 2).
Comparison of the accuracy rate of a monopolar RF lesion with classical versus revised targets
The accuracy rate for the IMGN was 100% in both techniques. There was no difference between the mean distances from the IMGN to the top of the active tip in both techniques.
The IPBSN was found in all the specimens and the revised targets was 100% accurate.
With revised targets, the RFN was captured in 100% of specimens, without any lesion to the common fibular nerve.
In the 14 specimens, the SMGN originated from the nerve to the vastus medialis (figures 2 and 3), and did not give an articular branch above the AT (figure 4). In eight cases, a small branch was detached from the nerve above the AT, and ran to the distal part of the vastus medialis muscle. We found that the current targets capture the transversal vessels detached from the descending genicular artery (at the periosteal areas connecting the shaft to the medial condyle), which were not accompanied with a terminal branch of the SMGN (figure 4).
(A) Medial aspect of the thigh and the knee of an embalmed specimen, adductor canal opened: SMGN (white arrows) dissected all along its course. (B) Same view in the same specimen, with the distal part of the VM partially retracted anteriorly: evidence that the transversal branches (red arrowheads) at the junction of femoral shaft and medial condyle are vessels from the descending genicular artery (red arrows) with no contribution of the SMGN (white arrows). FA, femoral artery; P, patella; ME, medial epicondyle; NVM, nerve to the vastus medialis; SN, saphenous nerve; VM, vastus medialis muscle.
The SLGN originated from the sciatic nerve in 6/14 cases and common fibular nerve in 8/14 cases, and divided into two terminal branches around the area connecting the posterior edge of the femoral shaft and the lateral femoral condyle (figures 2 and 3). Its ascending (transversal) branch joined the superior lateral genicular vessels (SLGV) in 10/14 cases, but its descending branch was not accompanied by vessels. In four cases, the transversal branch of the nerve was found about 1 cm below the SLGV.
The frequency map of the most accurate targets for the SLGN and the SMGN during fluoroscopic-guided procedures is presented in figure 5. The location of the SMGN was constant in all the specimens; revised targets would have captured the nerve in 14/14 knees. There was a little variability in the most accurate target for the SLGN; the revised landmarks would have captured the trunk of the nerve in 10/14 knees (whereas the classical targets wouldn’t have captured it). In fact, there are two specimens in which the target should have been more proximal and two others more distal. To get close to 100% accuracy with revised targets for the SLGN, a proximal and a distal lesion should be added to take this variability in account.
Scatter plot representing, on single lateral and anterior-posterior X-ray views, the aggregation of all the exact target points of the 14 specimens for accurate capture of the trunk of the SMGN (A and B) and SLGN (C and D). Each dot represents the most accurate target point to capture the specific genicular nerve in one specimen, on the corresponding X-ray view. SLGN, superior lateral genicular nerve; SMGN, superior medial genicular nerve.
Discussion
This study demonstrates that the revised targets for RFA of the SMGN and SLGN are more accurate than the classical targets. Moreover, the revised targets are 100% accurate on the RFN and the IPBSN, which play an important role in the sensory innervation of inferior lateral and inferior medial quadrants of the knee joint, but are not targeted by the current procedure. It shows once more that there is a real need to move forward toward a better technique, with accurate and validated anatomical targets.20 The frequency map (figure 5) of the accurate targets points to capture the trunk of the SMGN and SLGN before their distribution to the knee joint capsule is a unique tool that gives precise and reliable locations of the nerves. We believe that it should guide physicians during RF procedures under fluoroscopic guidance.
The anatomical bases for the classical targets for RF ablation of GNs are somewhat weak. Choi et al 1 who first described the procedure dissected only two specimens. They chose to target only three nerves and acknowledged that denervation of other articular branches would probably give different results. Concerning the SMGN, the current target was based on an early description of its origin and course,21 describing the SMGN as a branch of the tibial nerve, running transversally from posterior to anterior in the same trajectory as the superior medial genicular vessels. Despite the fact that the subsequent anatomical dissection studies22 23 did not find such results, and the few studies24 25 that repeated it did not show a single illustration, that description has been widely admitted. Choi et al chose the current cannula placement ‘to be parallel to the target nerve’, but this was based on incorrect anatomical information. Almost all the recent anatomical studies5–7 9 10 25 26 found that the SMGN descends vertically along the adductor magnus tendon toward the AT, which is the best target for the SMGN.10 25 The SMGN divides distally to the AT into its terminal branches. Thus, a lesion made above or just in front of the AT would capture the trunk of the nerve (figure 6A), but a lesion performed distally to the AT may not capture all terminal branches. In the current study, most of the thumbtacks fixed on the SMGN were above the AT. Since the course of the SMGN has been clarified by recent studies and it has been demonstrated that it descends completely posteriorly, the classical target is no more relevant. Given the limited size of the RF lesion and the average A-P depth of the distal part of the femoral shaft, an active tip which is at midpoint or at 60% from the anterior cortex of the femoral shaft would not capture the SMGN (figure 6A). To cross the trajectory of the SMGN and ensure its capture, the active tip should be advanced totally posteriorly, above or just in front of the AT (figure 6A). Moreover, the present study shows that the current target does not even capture the transversal terminal branch of the SMGN (figure 4B), which runs at the level of medial epicondyle, and is accompanied by vessels.
Summary of the different targets with the main course of the nerves, on lateral radiographs. (A) A lesion performed above or just in front of the AT would capture the SMGN whereas classical target would not. (B) Due to the variability of the bony contact of the SLGN, a palisade of lesions (three blue circles) would improve the capture rate of the SLGN and all its terminal branches. AT, adductor tubercle; SELC-PFS; superior edge of the lateral condyle - posterior femoral cortical surface; SLGN, superior lateral genicular nerve; SMGN, superior medial genicular nerve.
The accuracy of cannula placement for the SLGN using revised target was 64% (vs 35% for current target). This can be due to two reasons. First, in two cases, the nerve established its bony contact at the posterior edge of the lateral condyle (figures 3C and 5C) and then, only one of its terminal branches was captured (figure 3C). Second, in two cases, the nerve bifurcated proximally, before its bony contact, and then, only the transversal branch was captured. To improve the capture of both terminal branches of the SLGN, we propose not only one lesion, but a palisade of lesions (two to three contiguous lesions), starting from the area connecting the posterior cortex of femoral shaft to the lateral femoral condyle and moving backwards and distally to the posterior edge of the lateral condyle (figure 6B). In addition to better accuracy, the revised target avoids lesioning the SLGV, contrary to current targets. In addition to the SLGN, Tran et al 5 reported in all their specimens, a ‘long articular branch that originated just inferior to the bifurcation of the sciatic nerve to the superior border of the lateral condyle’. Thus, they reported two constant articular nerves, both found in all the specimens, originating from common fibular nerve/sciatic nerve, and innervating the superior-lateral aspect of the knee joint. We did not find those two simultaneous nerves, although we performed systematical dissections. To our knowledge, no other author who performed cadaveric dissection described such findings.
The frequency map provided in this study only concerns the SLGN and SMGN, which are controversial among authors. The anatomical target for the IMGN is not debated as all the authors have found that the current landmark is accurate. Concerning the IPBSN and RFN, we previously defined accurate clinical targets, easily identifiable and reproducible. Building a fluoroscopic-based frequency map for the IPBSN and RFN as we have done for the SMGN and SLGN would not be relevant. Ultrasound could be helpful, but the use of fluoroscopy is not clinically justified for the IPBSN. It should be noted that, since the IPBSN is subcutaneous, the cannula should be inserted deep enough to avoid skin lesion when using conventional RF and that pulsed RF could be preferable. The physician should be careful when lesioning the inferior-medial quadrant of the knee (IMGN and IPBSN), which requires multiple lesioning compared with other quadrants.
Even though clinical studies have shown positive results,1 3 4 these would probably be enhanced with better anatomical targets. When looking carefully at a few recent anatomical studies that support the current landmarks, there is one striking remark. The illustration of their dissections does not allow to identify either the origin and the course of the GNs27 or their ending at the level of the knee.28 This is a huge bias since it is not easy to visually discriminate between very small nerves and vessels in a cadaver, a fortiori in an embalmed cadaver. To illustrate that, we did a direct dissection through a ‘window’ at the site of the current target for the SLGN and found a structure that closely resembled a nerve (figure 7A). Based on that, we would have concluded that the current target was accurate. But when we realized a total dissection of that branch until its origin, we realized that it was a vessel (figure 7B). The ‘figure 3’ presented in recent anatomical study of Vanneste et al 27 illustrates that possible bias. Based on the dissections presented on that figure, it is not possible to be certain that the highlighted structures are GNs. This could be misleading as it maintains the doubt among physicians about the real need to improve the current landmarks. To avoid confusion, we suggest that further studies conducted to demonstrate a GN should include either dissection of the nerve since its original trunk (which is big and easily identifiable) to its termination at the level of the knee, or vascular injection to highlight the small vessels, or histological analysis. Only such a rigorous methodology shared by all the researchers, would irrefutably demonstrate the best targets.
(A) Lateral view of the knee, the vastus lateral muscle and the fascia lata are left in place: Dissection performed through a ‘window’ directly at the level of the current target for SLGN. A transversal structure that resembles a nerve is found. (B) Same view on the same knee: when performing a complete dissection of that ‘nerve’ (supposed to be the SLGN) to check for its origin, it is rather the superior lateral genicular artery originating from popliteal artery. ITT, iliotibial tractus covering vastus lateralis muscle; LC, lateral condyle; P, patella; PA, popliteal artery SLGN, superior lateral genicular nerve.
It is worth reporting that a volume of 0.1 mL (10 times less than what is currently used for prognostic nerve blocks) dyed lidocaine spreads beyond the limits of the larger RF lesion. This may explain why a prognostic GN block using a local anesthetic volume of 1 mL for subsequent RFA eligibility does not improve the treatment success.15 Based on the findings of our pilot study, we should use a volume of 0.1 mL or less.
This study has some limitations. First, the accuracy of cannula placement would have been more definitely assessed if the RF lesions were directly appreciated during dissection. This was not possible despite a rigorous methodology to simulate clinical conditions (fresh warmed cadavers, injection of fluids…). However, the use of non-diffusible ink to create just a black mark, rather that injecting fluids like in other studies,9 25 27 29 is a reliable alternative. Second, a larger sample size would have allowed to create a more exhaustive figure map representing the entire spectrum of variations of the target points, especially for the SLGN, which was the only nerve to have less than 100% accuracy with revised targets. Nonetheless, our methodology for the exact localization of target points for each nerve on A-P and lateral fluoroscopic image seems to be the most robust and would be of great interest for physicians in their daily practice.
Conclusion
This study found that the revised targets are more accurate than current targets for RFA of the SLGN and SMGN. In addition, the revised targets are 100% accurate for the RFN and the IPBSN, which are not targeted by the current procedure. The current technique, targeting only three nerves, based on confusing anatomical foundations, definitely needs to be improved. Our map presents the localization of exact target points to accurately capture the SLGN and SMGN, and provides an optimal design to guide needle placement during GN-RFA. Further studies must be conducted to compare the effectiveness and safety of both techniques for RF ablation of the GNs in patients suffering from chronic knee pain.
Acknowledgments
The authors thank Martial Vergauwen, Nicolas Charles-Pirlot and Bernard Caelen for their important technical assistance. They also thank the donors of the cadavers and their families.
References
Footnotes
Contributors LF: conception and design, cadaveric dissections, data collection and analysis, manuscript writing. CB: conception and design, cadaveric dissections, manuscript revision. AS: conception and design, manuscript revision. J-EKK: assistance in cadaveric dissections, manuscript revision. CD: data analysis, manuscript revision. BLPDW: conception and design, manuscript revision. OC: project manager, conception and design, manuscript revision.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient consent for publication Not required.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement All data relevant to the study are included in the article or uploaded as supplementary information.