Eur Spine J DOI 10.1007/s00586-014-3598-y
ORIGINAL ARTICLE
Diffusion-weighted magnetic resonance (DW-MR) neurography of the lumbar plexus in the preoperative planning of lateral access lumbar surgery Cristiano Magalha˜es Menezes • Luciene Mota de Andrade • Carlos Fernando Pereira da Silva Herrero Helton Luiz Defino • Marcos Antonio Ferreira Ju´nior • William Blake Rodgers • Marcello Henrique Nogueira-Barbosa
•
Received: 14 January 2014 / Revised: 23 September 2014 / Accepted: 23 September 2014 Ó Springer-Verlag Berlin Heidelberg 2014
Abstract Purpose Magnetic resonance (MR) neurography has been used to evaluate entire nerves and nerve bundles by providing better contrast between the nerves and the surrounding tissues. The purpose of the study was to validate diffusion-weighted MR (DW-MR) neurography in visualizing the lumbar plexus during preoperative planning of lateral transpsoas surgery. Methods Ninety-four (188 lumbar plexuses) spine patients underwent a DW-MR examination of the lumbar plexus in relation to the L3–4 and L4–5 disc spaces and superior third of the L5 vertebral body. Images were reconstructed in the axial plane using high-resolution
C. M. Menezes M. A. Ferreira Ju´nior Servic¸o de Cirurgia de Coluna, Hospital Ortope´dico/Lifecenter and Hospital Sa˜o Francisco de Assis, Belo Horizonte, Brazil C. M. Menezes (&) Minimally Invasive Spine Surgery Department, Hospital Ortope´dico/Lifecenter, Rua Prof. Ota´vio Coelho de Magalha˜es, 111 Mangabeiras, Belo Horizonte, MG CEP 30210-300, Brazil e-mail: [email protected] L. M. de Andrade Axial Centro de Imagem, Belo Horizonte, Brazil C. F. P. da Silva Herrero H. L. Defino Departamento de Biomecaˆnica, Medicina e Reabilitac¸a˜o do Aparelho Locomotor da Faculdade de Medicina de Ribeira˜o Preto da USP, Ribeira˜o Preto, Brazil W. B. Rodgers Spine Midwest, St Mary’s Health Center, Jefferson City, USA M. H. Nogueira-Barbosa Divisa˜o de Radiologia e Diagno´stico por Imagem da Faculdade de Medicina de Ribeira˜o Preto da USP, Ribeira˜o Preto, Brazil
Maximum Intensity projection (MIP) overlay templates at the disc space and L3–4 and L4–5 interspaces. 10 and 22 mm MIP templates were chosen to mimic the working zone of standard lateral access retractors. The positions of the L4 nerve root and femoral nerve were analyzed relative to the L4–5 disc in axial and sagittal planes. Third-party radiologists and a senior spine surgeon performed the evaluations, with inter- and intraobserver testing performed. Results In all subjects, the plexus was successfully mapped. At L3–4, in all but one case, the components of the plexus (except the genitofemoral nerve) were located in the most posterior quadrant (zone IV). The L3 and L4 roots coalesced into the femoral nerve below the L4–5 disc space in all subjects. Side-to-side variation was noted, with the plexus occurring in zone IV in 86.2 % right and only 78.7 % of left sides. At the superior third of L5, the plexus was found in zone III in 27.7 % of right and 36.2 % of left sides; and in zone II in 4.3 % right and 2.1 % left sides. Significant inter- and intraobserver agreement was found. Conclusions By providing the surgeon with a preoperative roadmap of the lumbar plexus, DW-MR may improve the safety profile of lateral access procedures. Keywords XLIF Complication Neuromonitoring Electromyography Plexopathy
Introduction The less disruptive, transpsoas approach to the lumbar spine allows access to the intervertebral disc for decompression and reconstruction, either by fusion or arthroplasty [1, 2]. By preserving the anterior longitudinal ligament (ALL) and the posterior muscular and ligamentous
123
Eur Spine J
structures, this approach affords excellent access to the spine while minimizing collateral tissue damage. Despite being less disruptive, this retroperitoneal approach nonetheless involves a possible risk of injury to local anatomic structures. Due to the limited visualization necessitated by the technique, the lumbar plexus is at risk during the operative traverse of the psoas muscle [1, 3–17]. In recent years, this approach has rapidly grown in popularity, yet concerns about safety persist. Recent papers have documented that injury to the components of the lumbar plexus is a rare, but concerning complication associated with the transpsoas retroperitoneal approach; a risk that is reduced, but not eliminated, by the use of real-time, directional neuromonitoring technologies [18–24]. Heretofore, accurate visualization of the branches of the lumbar plexus preoperatively was not possible as traditional magnetic resonance T2-weighted and fat suppressed images do not allow for accurate differentiation between the similar signal intensities of neural and vascular structures [25, 26]. Magnetic resonance neurography (MRN) has been shown to be helpful to evaluate abnormal conditions of entire nerves and nerve bundles by providing better contrast between the nerves and the surrounding tissues. It has been used in clinical practice to evaluate peripheral neuropathies and has obvious advantages over other electrodiagnosis studies, such as electromyography and nerve conduction studies [25–27]. More recently, diffusion-weighted magnetic resonance neurography (DW-MR) has been introduced into use in the field, and uses the diffusion technique to enable the suppression of the signal from surrounding tissues (muscles and vessels), yielding improved contrast between muscular, adipose, and neural structures, thereby highlighting neural anatomy [25–31]. Three-dimensional diffusion-weighted reversed imaging with steady-state precession (3D DW-PSIF) provides adequate vascular signal suppression with an acceptable signal to noise ratio [28]. The 3D DW-PSIF sequence is isotropic, allowing for multiplanar reconstructions in any plane without significant loss of spatial resolution [28, 29]. Maximum intensity projections (MIP) may be obtained from 3D DW-PSIF and could nicely depict nerve anatomy [29]. Thus, DW-MR holds great promise as an adjunct to the preoperative planning for lateral retroperitoneal surgery. By providing a detailed roadmap of the lumbar plexus, DW-MR may be used by surgeons to improve the safety profile of lateral access procedures. The object of this study was to evaluate the utility and replicability of DW-MR in the visualization of the position lumbar nerve roots and plexus relevant to the lateral transpsoas approach to L3–4 and L4–5.
123
Methods Study design/patient sample A prospective observational study of consecutive patients ([18 years old) presenting with complaints of low back pain to the spine surgery center at the Ortope´dico/Lifecenter Hospital in Belo Horizonte, Brazil between January 2012 and March 2013 was included in the study. The study protocol was approved by the local Ethics Committee on Research, Clinics Hospital, Ribeirao Preto Medicine School at Sa˜o Paulo University and all patients signed detailed informed consent paperwork. Exclusion criteria included the presence of spondylolisthesis grade greater than II [30]; scoliosis greater than 10°; previous surgery with lumbar spine instrumentation and/or fusion; and/or other diseases and/or clinical conditions that rendered it impossible to obtain magnetic resonance imaging (MRI) studies. All patients were initially assessed using the institution’s standard protocol for the evaluation of low back pain (including MRI). In addition to the standard MR sequences, the neurographic study by diffusion technique using the three-dimensional (3D) DW-PSIF sequence, for assessing the position of components of the lumbar plexus. These additional studies where then randomly assigned into groups of ten and presented to the radiologist for a new evaluation and data acquisition for the study. In addition, the studies were evaluated by a senior spine surgeon to assess intraobserver and interobserver variability. Imaging evaluation/magnetic resonance (MR) Reversed fast image with steady-state precession (3D DW-PSIF) was used for MR neurography. Patients were all scanned with a 1.5 T MRI unit (Magneton Essenza, Siemens Healthcare) in the supine position using an 8-channel spine coil. The following parameters were used: 80–90 s/mm2 b value, TR = 17.99 ms, TE = 5.95 ms, flip angle 30°, 1.0 mm slice thickness, 256 mm field of view, 256 9 80 % matrix size, receiver bandwidth 140 Hz/pixel, fat suppression, acquisition time 8 min 37 s. One of the challenges of any radiographic modality involves the necessity of depicting three-dimensional structures, such as the lumbar plexus coursing through the psoas muscle, in two dimensions. To obtain the most pertinent clinically applicable data from the DW-MR images, this 3D to 2D approximation reconstructed the axial plane images using high-resolution maximum intensity projection (MIP) overlay templates at the level of the disc spaces of L3–4 and L4–5 of 10 and 22 mm. These reconstructive overlays were selected based on the working zone of
Eur Spine J
Fig. 1 Lateral radiographic showing approximate 10 and 22 mm reconstructive templates (windows) used during neurography to assess the potential presence of neural structures during a lateral approach using different retractors
standard lateral access retractors (10 mm for MaXcessÒ retractor (NuVasive, Inc., San Diego, CA), 22 mm for METRxÒ (Medtronic Sofamor Danek, Memphis, TN), and used to assess the potential presence of neural structures within this window with respect to these baseline retractor sizes (Fig. 1). MIP images at the disc space and at the working zone of a possible surgical retractor were reconstructed by a radiologist with extensive experience in muscle and spine diseases studies as well as in diffusion images and MIP reconstructions. L4 spinal nerve and femoral nerve position were determined in relation to the L4–5 disc and this position was analyzed in the sagittal planes using a 30 mm MIP reconstructive template (RT). The area between the anterior and posterior edges of the disc space was divided into four quartiles (zones) in the axial and sagittal planes, following the same pattern described in previous anatomic studies [8, 14, 17]. Zones were defined as: zone I (anterior quartile), zone II (middle anterior quartile), zone III (middle posterior quartile), zone IV (posterior quartile). Nerve root position was defined according to their location in each quartile as relating to the anteroposterior diameter of the vertebral body. Statistical analysis Given that these results required dependent data analysis (correlated samples), and that the same individual was measured six times (one in each group, and in both sides, right and left, in every group), marginal logistic regression using alternating logistic regression (ALR) was used to analyze the influence of groups, sides, sex, age, the presence or absence of transitional vertebrae, the last rib
(LR) arising from T12, and L4–5 listhesis for the presence of nerve in zones II or III. A hierarchical structure based in three variables in its association structure was used to model the association between measures obtained in a same individual: measures in same individual in different groups and sides, measures in same side in different groups and measures in the same group in different sides. When selecting the possible significant variables for the nerve presence at zones II or III, univariate marginal logistic regression was initially performed. A significance level of 25 % was employed. All variables selected in univariate analysis were used in multivariate marginal logistic regression, reaching the final model through backward algorithm at a 5 % significance level. The software applied for analysis was R version 2.15.3 (R Foundation for Statistical Computing, Viena, Austria). To assess intra- and interobserver variability, a Kappa coefficient was applied, and to compare intra- and interobserver Kappa indexes, as well as to compare 10 and 22 mm groups, bootstrap percentile 95 % confidence intervals were used. The software applied for analysis was R version 2.15.2 (R Foundation for Statistical Computing, Viena, Austria).
Results Ninety-four patients (188 lumbar plexuses) were analyzed. There were 46 (48.9 %) female and 48 (51.1 %) male subjects. Lumbosacral transitional vertebrae were found in 12 individuals (12.8 %). In 80 (85.1 %) patients, the last rib was positioned at the T12 level. Bilateral plexus mapping using DW-MR neurography was successfully achieved in all patients. In all subjects, the L3 and L4 roots coalesced into the femoral nerve below the L4–5 disc space. The constitution and course of the femoral nerve were documented in all patients (Fig. 2). One patient presented with a Grade I L3–4 spondylolisthesis, and 13 (15.5 %) showed grade I or II L4–5 spondylolisthesis (Table 1). Table 2 shows the results of L3–4 related to groups and sides. In one patient with an axial MIP reconstructive template (axial RT) of 22 mm and a sagittal reconstructive template (sagittal RT) of 30 mm the L4 spinal nerve entered zone III (on the right side only). When assessing the data from L4–5, it was noted that the 22 mm axial reconstructive template was significantly more likely to overlay nerves in zones II and III than the 10 mm axial template (Fig. 3), implying that a larger retractor (22 mm diameter) would be more likely to contact the neural structures of the plexus. There was also a trend towards a greater prevalence of nerves in zone III at left side compared to the right side (Table 2 and Fig. 4).
123
Eur Spine J Table 1 Variables frequence: sex, transitional vertebra and type, last rib T12, listhesis: L3– L4 and listhesis: L4–L5
Variables
N
%
Sex Female
46
48.9
Male
48
51.1
Total
94
100.0
No
82
87.2
Yes
12
12.8
Total
94
100.0
6
54.5
Transition vertebra
Type L5 sacralized S1 lumbarized
5
45.5
11
11.7
No
14
14.9
Yes Total
80 94
85.1 100.0
Total Last rib at T12
Spondylolisthesis at L3–L4 No
82
Yes
1
98.8 1.2
Total
83
100.0
Spondylolisthesis at L4–L5 Fig. 2 Lumbar plexus mapped on sagittal reconstruction over L3–L4 and L4–L5 intervertebral spaces. Oblique femoral nerve pattern is evidence at the L4–L5 level
Nerves in zones II and II were combined for the purpose of analysis (since there were so few nerves in zone II) and comparison with nerves in zone IV. Table 2 and Fig. 4 show comparisons between the presence of the nerve in zones II or III and at zone IV, between groups and sides. Nerves were more frequently found at zones II or III in the axial 22 mm RT and sagittal RT groups, compared to the axial disc 10 mm group. They were also more frequently located in zones II or III on the left side when compared to the right side. When the analysis of the L4–5 level was stratified by group (axial disc, axial 22 mm template, and sagittal 30 mm template) and side (left/right) related to the different variables (sex; transition vertebra; last rib T12; listhesis: L3–4; listhesis: L4–5), regardless of images and side studied, there was a trend towards femoral nerve location in zones II or III in male patients as well as in patients in which the transitional vertebra (TV) was present (Table 3). Interestingly, in patients with a last rib arising from T12, (LR T12), the nerve was found more frequently in zones II or III if the axial RT was 22 mm and the sagittal RT was 30 mm; this trend was reversed if the last rib did not arise from T12, and a 10 mm axial RT was employed. Only one patient presented with L3–4 listhesis, and in the axial disc 10 mm group the nerve was located in zone
123
No
71
Yes
13
84.5 15.5
Total
84
100.0
IV. Except for right side in axial disc 10 mm group, a trend towards nerve location at zones II or III was observed in those cases with L4–5 listhesis (Table 3). Figure 5 shows that the variable groups, sides, sex and transitional vertebrae were selected for the multivariate model by p values \ 0.25. Table 4 presents the multivariate model adjusted to all selected variables from the univariate analysis. Sex did not represent a significant predictor of the presence of nerves at zones II or III. The final model showed in Table 4 was thus obtained excluding the variable sex. Significant intra- and interobserver agreement in both groups was found [31]. Based on confidence intervals, there was no significant intra- and interobserver variance. Table 5 shows the intra- and interobserver agreement, using Kappa coefficient for each group and both sides.
Discussion The minimally disruptive lateral retroperitoneal transpsoas approach for anterior lumbar interbody fusion in the lumbar spine avoids many of the same potential complications of other anterior approaches and its safety and reproducibility have been repeatedly documented [23, 32–34].
Eur Spine J Table 2 L3–L4 and L4–L5 contingency table stratifying by groups between side and zones Groups
Sides
L3–L4 II
Axial disc 10 mm Axial 22 mm RT Sagittal 30 mm RT
Groups
III
Right
0
0.0 %
0
0.0 %
94
100.0 %
Left
0
0.0 %
0
0.0 %
94
100.0 %
Right
0
0.0 %
1
1.1 %
93
98.9 %
Left
0
0.0 %
0
0.0 %
94
100.0 %
Right
0
0.0 %
1
1.1 %
93
98.9 %
Left
0
0.0 %
0
0.0 %
94
100.0 %
Sides
L4–L5 II
Axial disc 10 mm Axial 22 mm RT Sagittal 30 mm RT
Groups
IV
III
IV
Right
1
1.1 %
12
12.8 %
81
86.2 %
Left
0
0.0 %
20
21.3 %
74
78.7 %
Right Left
4 2
4.3 % 2.1 %
26 34
27.7 % 36.2 %
64 58
68.1 % 61.7 %
Right
4
4.3 %
26
27.7 %
64
68.1 %
Left
2
2.1 %
37
39.4 %
55
58.5 %
Sides
L4–L5 II or III
Axial disc 10 mm Axial 22 mm RT Sagittal 30 mm RT
IV
Right
13
13.8 %
81
86.2 %
Left
20
21.3 %
74
78.7 %
Right
30
31.9 %
64
68.1 %
Left
36
38.3 %
58
61.7 %
Right
30
31.9 %
64
68.1 %
Left
39
41.5 %
55
58.5 %
Complications related to transpsoas retroperitoneal approach, despite intraoperative nervous monitoring, have been well documented [18–20, 24, 35]. Iatrogenic injuries of the lumbar plexus can result from direct mechanical compression, laceration, traction or stretching and/or indirect ischemia. Knowledge of the precise location of components of the lumbar plexus by the surgeon may afford greater safety for the procedure and could, potentially, enhance the utility of directional electroneurography [23, 24]. Several cadaveric studies have been performed to better define the precise location of the plexus and its components and, thereby, have attempted to define so-called ‘‘safe working zones’’ to gain information on and avoid neural structures during the lateral approach [1, 3, 4, 6, 8–12, 14, 15, 22]. Benglis et al. [4], in a study of three cadavers, found that 0, 11, 18, and 28 % of the posterior aspect of lateral disc space at L1–2, L2–3, L3–4, and L4–5, was covered by the nerves of the lumbar plexus, respectively. This leaves approximately the anterior 100, 89, 82, and 72 % of the
lateral disc spaces free of motor nerves in a cadaveric model. While these results suggest adequate area for docking on the lateral disc space for the lateral approach, the authors also noted the possibility of genitofemoral nerve (GFN) injury with a more anterior approach, but did not address the location of the GFN in their study. Uribe et al. [17], when studying five fresh male cadavers (with dissection of psoas muscle, lumbar plexus and nerve roots in each disc space, in 20 lumbar segments), divided out the area between anterior and posterior edges of vertebral body/intervertebral disc in four equal zones (zones I–IV, from anterior to posterior). All lumbar plexus components, including nerve roots, were contained within the substance of the psoas muscle, dorsal to posterior quarters (zones III and IV) of vertebral body. Zone II was not posterior to any part of the plexus, except for the genitofemoral nerve, a sensory branch originating from L1 and L2 roots that obliquely traverses psoas muscle from its origin to distal location, traveling to the lateral surface of the psoas muscle at approximately the L3–4 level. The study concluded that anatomical safe zones at L1–2 to L3–4 disc spaces are at the middle posterior quartile (zone III), and at or near the midline of the vertebral body and anterior at the L4–5 space (zone II–III). Moreover, there was a risk of direct genitofemoral nerve injury in zone II of L2–3 disc space and in zone I at inferior lumbar levels L3–4 and L4–5. Additionally avoidance of the ilioinguinal, iliohypogastric and lateral femoral cutaneous nerves within retroperitoneal space, where they traverse obliquely, inferiorly and anteriorly to reach iliac crest and abdominal wall is recommended [5]. The results by Uribe et al. [17] largely corroborate those results of Moro et al. [14], who analyzed 30 cadavers to establish safe working zones for transpsoas retroperitoneal endoscopic surgery. Excluding the position of the genitofemoral nerve, at least the anterior half of the disc space was free of motor nerves from L4–5 superiorly. It was also noted that the psoas muscle is often bifid at the L3 or L4 vertebral bodies and thus there is a risk of genitofemoral nerve injury. These studies, however, are limited by several factors. First, in a cadaveric model there are often limitations on number of samples collected (increasing variability), as well as by potential for displacement of the lumbar plexus during dissection. Also, cadaveric studies only account for the baseline anatomic position of the plexus. Flexion of the hip and bending of the table during the procedure that may alter the position of the plexus and the ability to retract nerves are not considered in findings of cadaveric studies. From a methodological standpoint, as well, the use of quartiles in determining zones likely underestimates the true (more posterior) position of the lumbar plexus, as corroborated by the current study and other studies which
123
Eur Spine J
Fig. 3 T2-weighted axial magnetic resonance imaging (MRI) showing L4–L5 intervertebral spaces with standard psoas muscle anatomy (a, b). MR neurography in the same patient demonstrating anteriorization of the lumbar plexus to zone II on the right (c, d)
measure the absolute position of the lumbar plexus. The previously mentioned Benglis et al. [4] study is one such study, finding the lumbar plexus at 28 % of the vertebral body from the posterior aspect. In a categorical study (quartiles), this would be defined as zones I and II being free of motor nerves despite the fact that, on average, 88 % (22/25) % of the anterior aspect of zone III would be free of motor nerves. A similar study by Regev et al. [15] found nearly identical results, with the lumbar plexus occupying 26 % of the posterior disc space at L4–5 (95 % confidence interval). Guerin et al. [9] found this measure at L4–5 to be 35.8 %. These findings suggest that the quadrant system is overly broad. Imaging by magnetic resonance is a useful tool in preoperative assessment of spine anatomy, especially in lateral approach spine surgery, helping to identify the anatomic location of nerve roots and retroperitoneal vessels as they relate to the intervertebral disc, and may aid in avoiding vascular and neurologic injuries [16]. Regev et al. [15], when reviewing 100 spine magnetic resonances from patients treated for different spine pathologies, observed that the safe surgical corridor for
123
discectomy and intervertebral spacer insertion narrows from L1–2 to L4–5, where the risk of nerve root or retroperitoneal vessel injury is increased, given the more ventral location of nerve root which forces the discectomy window ventral and thereby raising the risk of injury to the contralateral vessels. However, these MRIs were taking in a supine position and, in particular, the vessels and peritoneal structures migrate significantly anteriorly when in the lateral decubitus position, thus limiting the risk of injury [7]. In this study, preoperative MRI was recommended for assessing the position of adjacent neurovascular structures near the inferior endplate of the vertebra at each level [15]. This study uses, however, conventional magnetic resonance, without properly distinguishing between lumbar plexus and epidural and perineural vessels. This reduces the surgeon’s ability to interpret preoperative images in a useful way. DW-MR is able to suppress signal from adjacent tissues (muscles and vessels), highlighting neural anatomy and offering better contrast with other soft tissues, and is thus superior to standard MRI [25–27]. It must be noted, however, that the anatomic findings in the study were not confirmed by surgical exploration.
Eur Spine J Table 3 Contingency table of L4–L5 segments related to sex, transitional vertebra, last rib T12, listhesis: L3–L4 and listhesis; stratified by group and side Stratifications
Variables
L4–L5 II or III
IV
Axial disc 10 mm Right Left
Female
6
13.0 %
40
87.0 %
Male
7
14.6 %
41
85.4 %
Female
7
15.2 %
39
84.8 %
Male
13
27.1 %
35
72.9 %
Axial 22 mm RT Right Left
Female
13
28.3 %
33
71.7 %
Male
17
35.4 %
31
64.6 %
Female
15
32.6 %
31
67.4 %
Male
21
43.8 %
27
56.3 %
Sagittal 30 mm RT Right Left Fig. 4 Bar charts depicting the L4–L5 level stratified by registration template between groups, zones, and sides
The number of lumbosacral and thoracolumbar transition segments in this sample did not differ from that usually observed among the general population. The presence of a transitional vertebra may correlate to a so-called ‘‘tear-drop psoas’’, an indication that the plexus may be ventrally located [16]. One important implication of the current study is in relation to the use of 22 mm retractors for lateral lumbar access. Not only do these larger diameter retractors place the plexus at greater risk of injury when the neural elements are in Zone III at L4–5 (40.7 % of images analyzed) but even puts nerves in zone II as some small risk (5 %). The possible usefulness of DW-MR at preoperative planning is thus reinforced. Furthermore, by assessing side-to-side (right/left) variations in the same patient, a surgeon could choose the best surgical side, considering not only the iliac crest height or even the concave/convex side of an adult degenerative deformity, but also estimate the lumbar plexus location and the risk of injury to its components at each level being treated. One limitation of the study was that no comparisons were made between the diagnostic performance of DWMR and surgical or cadaveric data. Thus, it was not possible to assess the absolute accuracy of this classification by MRN. This limitation is relative and we believe that it does not weaken the value of the results achieved, as the goal of the study was to evaluate the ability of MRN to visualize the lumbar plexus preoperatively.
Female
13
28.3 %
33
71.7 %
Male
17
35.4 %
31
64.6 %
Female Male
15 24
32.6 % 50.0 %
31 24
67.4 % 50.0 %
Axial disc 10 mm Right Left
TV = no
10
12.2 %
72
87.8 %
TV = yes
3
25.0 %
9
75.0 %
TV = no
14
17.1 %
68
82.9 %
TV = yes
6
50.0 %
6
50.0 %
Axial 22 mm RT Right Left
TV = no
25
30.5 %
57
69.5 %
TV = yes
5
41.7 %
7
58.3 %
TV = no
29
35.4 %
53
64.6 %
TV = yes
7
58.3 %
5
41.7 %
Sagittal 30 mm RT Right Left
TV = no
25
30.5 %
57
69.5 %
TV = yes
5
41.7 %
7
58.3 %
TV = no TV = yes
31 8
37.8 % 66.7 %
51 4
62.2 % 33.3 %
Axial disc 10 mm Right Left
LR T12 = no
2
14.3 %
12
85.7 %
LR T12 = yes
11
13.8 %
69
86.3 %
LR T12 = no
4
28.6 %
10
71.4 %
LR T12 = yes
16
20.0 %
64
80.0 %
Axial 22 mm RT Right Left
LR T12 = no
3
21.4 %
11
78.6 %
LR T12 = yes
27
33.8 %
53
66.3 %
LR T12 = no
4
28.6 %
10
71.4 %
LR T12 = yes
32
40.0 %
48
60.0 %
Sagittal 30 mm RT Right Left
LR T12 = no
3
21.4 %
11
78.6 %
LR T12 = yes
27
33.8 %
53
66.3 %
LR T12 = no
4
28.6 %
10
71.4 %
LR T12 = yes
35
43.8 %
45
56.3 %
123
Eur Spine J Table 3 continued Stratifications
Variables
L4–L5 II or III
IV
Axial disc 10 mm Right
Listhesis: L3L4 = no
11
13.4 %
71
86.6 %
0
0.0 %
1
100.0 %
Left
Listhesis: L3L4 = yes Listhesis: L3L4 = no Listhesis: L3L4 = yes
17
2.7 %
65
79.3 %
0
0.0 %
1
100.0 %
Axial 22 mm RT Right
Left
Listhesis: L3L4 = no
26
31.7 %
56
68.3 %
Listhesis: L3L4 = yes
1
100.0 %
0
0.0 %
Listhesis: L3L4 = no
31
37.8 %
51
62.2 %
Listhesis: L3L4 = yes
1
100.0 %
0
0.0 %
Listhesis: L3L4 = no
26
31.7 %
56
68.3 %
Listhesis: L3L4 = yes
1
100.0 %
0
0.0 %
Listhesis: L3L4 = no
34
41.5 %
48
58.5 %
Listhesis: L3L4 = yes
1
100.0 %
0
0.0 %
Sagittal 30 mm RT Right
Left
Fig. 5 Graphic showing estimated odds ratio and 95 % confidence interval estimates by univariate logistic marginal regression
Table 4 Multivariate logistic marginal regression for the occurence in Zones II and III at L4–L5 Medium structure
b
E. P (b)
Intercept Group = axial 22 mm RT Group = sagittal 30 mm RT Side = left Transitional vertebra— yes
-1.918 0.953
0.271 \0.001 0.191 \0.001
– 2.59
[1.78–3.77]
1.034
0.202 \0.001
2.81
[1.89–4.18]
0.428 0.949
0.204 0.486
1.53 2.58
[1.03–2.29] [1.00–6.70]
Axial disc 10 mm Right
Left
Listhesis: L4L5 = no
10
14.1 %
61
85.9 %
Listhesis: L4L5 = yes
1
7.7 %
12
92.3 %
Listhesis: L4L5 = no
14
19.7 %
57
80.3 %
3
23.1 %
10
76.9 %
Listhesis: L4L5 = yes Axial 22 mm RT Right
Left
Listhesis: L4L5 = no
22
31.0 %
49
69.0 %
Listhesis: L4L5 = yes
5
38.5 %
8
61.5 %
Listhesis: L4L5 = no
27
38.0 %
44
62.0 %
Listhesis: L4L5 = yes
5
38.5 %
8
P value
0.036 0.050
O. R
I.C.—95 %
Association structure
a
E. P(a)
P valor
P.O.R
I.C.—95 %
61.5 %
Same subject Same side vs. exams Same exam vs. sides
1.974 3.822 0.456
0.445 1.393 0.254
\0.001 0.006 0.069
7.20 45.7 1.58
[3.01–17.2] [2.98–700.5] [0.98–2.60]
Table 5 Coefficient and Kappa test for intra- and interobserver variability index at 10 and 22 mm groups in both sides
Sagittal 30 mm RT Right
Listhesis: L4L5 = no Listhesis: L4L5 = yes
22 5
31.0 % 38.5 %
49 8
69.0 % 61.5 %
Left
Listhesis: L4L5 = no
29
40.8 %
42
59.2 %
Listhesis: L4L5 = yes
6
46.2 %
7
53.8 %
Source 10 mm 22 mm
Obtaining a reference pattern in this case is complicated, since the study of cadavers to identify the nerves in the psoas muscle is hard to achieve without distortion of the anatomical trajectory of the nerves. Surgical confirmation would also imply a larger selection bias in the studied
123
Kappa
P value
I.C.—95 % [0.46–0.84]
Interobserver
0.693
\0.001
Intraobserver
0.668
\0.001
[0.45–0.83]
Interobserver
0.680
\0.001
[0.50–0.81]
Intraobserver
0.782
\0.001
[0.62–0.88]
population, since it would not be possible to proceed with the investigation of patients without indication for surgery. Furthermore, there was substantial interobserver and
Eur Spine J
intraobserver reproducibility of the classification assessing the location of the elements of the lumbar plexus using with DW-MR neurography.
Conclusion To our knowledge, this represents the first report on mapping the lumbar plexus using DW-MR. The correlation between these findings and intraoperative anatomy remains to be established, as a noninvasive technique that uses neither dye nor ionizing radiation, DW-MR holds great promises as an adjunct to the preoperative planning of lateral transpsoas lumbar surgery. Acknowledgments The authors would like to thank Kyle Malone, MS for his editorial assistance. Conflict of interest
None.
References 1. Park DK, Lee MJ, Lin EL, Singh K, An HS, Phillips FM (2010) The relationship of intrapsoas nerves during a transpsoas approach to the lumbar spine: anatomic study. J Spinal Disord Tech 23:223–228 2. Pimenta L, Oliveira L, Schaffa T, Coutinho E, Marchi L (2011) Lumbar total disc replacement from an extreme lateral approach: clinical experience with a minimum of 2 years’ follow-up. J Neurosurg Spine 14:38–45 3. Banagan K, Gelb D, Poelstra K, Ludwig S (2011) Anatomic mapping of lumbar nerve roots during a direct lateral transpsoas approach to the spine: a cadaveric study. Spine (Phila Pa 1976) 36:E687–E691 4. Benglis DM, Vanni S, Levi AD (2009) An anatomical study of the lumbosacral plexus as related to the minimally invasive transpsoas approach to the lumbar spine. J Neurosurg Spine 10:139–144 5. Dakwar E, Vale FL, Uribe JS (2011) Trajectory of the main sensory and motor branches of the lumbar plexus outside the psoas muscle related to the lateral retroperitoneal transpsoas approach. J Neurosurg Spine 14:290–295 6. Davis TT, Bae HW, Mok JM, Rasouli A, Delamarter RB (2011) Lumbar plexus anatomy within the psoas muscle: implications for the transpsoas lateral approach to the L4–L5 disc. J Bone Joint Surg Am 93:1482–1487 7. Deukmedjian AR, Le TV, Dakwar E, Martinez CR, Uribe JS (2012) Movement of abdominal structures on magnetic resonance imaging during positioning changes related to lateral lumbar spine surgery: a morphometric study: clinical article. J Neurosurg Spine 16:615–623 8. Guerin P, Obeid I, Gille O et al (2011) Safe working zones using the minimally invasive lateral retroperitoneal transpsoas approach: a morphometric study. Surg Radiol Anat 33:665–671 9. Guerin P, Obeid I, Bourghli A et al (2012) The lumbosacral plexus: anatomic considerations for minimally invasive retroperitoneal transpsoas approach. Surg Radiol Anat 34:151–157 10. Hu WK, He SS, Zhang SC et al (2011) An MRI study of psoas major and abdominal large vessels with respect to the X/DLIF approach. Eur Spine J 20:557–562
11. Kepler CK, Bogner EA, Herzog RJ, Huang RC (2011) Anatomy of the psoas muscle and lumbar plexus with respect to the surgical approach for lateral transpsoas interbody fusion. Eur Spine J 20:550–556 12. Lu S, Chang S, Zhang YZ, Ding ZH, Xu XM, Xu YQ (2011) Clinical anatomy and 3D virtual reconstruction of the lumbar plexus with respect to lumbar surgery. BMC Musculoskelet Disord 12:76 13. Moller DJ, Slimack NP, Acosta FL Jr, Koski TR, Fessler RG, Liu JC (2011) Minimally invasive lateral lumbar interbody fusion and transpsoas approach-related morbidity. Neurosurg Focus 31:E4 14. Moro T, Kikuchi S, Konno S, Yaginuma H (2003) An anatomic study of the lumbar plexus with respect to retroperitoneal endoscopic surgery. Spine (Phila Pa 1976) 28:423–428 15. Regev GJ, Chen L, Dhawan M, Lee YP, Garfin SR, Kim CW (2009) Morphometric analysis of the ventral nerve roots and retroperitoneal vessels with respect to the minimally invasive lateral approach in normal and deformed spines. Spine (Phila Pa 1976) 34:1330–1335 16. Smith WD, Youssef JA, Christian G, Serrano S, Hyde JA (2012) Lumbarized sacrum as a relative contraindication for lateral transpsoas interbody fusion at L5-6. J Spinal Disord Tech 25:285–291 17. Uribe JS, Arredondo N, Dakwar E, Vale FL (2010) Defining the safe working zones using the minimally invasive lateral retroperitoneal transpsoas approach: an anatomical study. J Neurosurg Spine 13:260–266 18. Ahmadian A, Abel N, Uribe JS (2013) Functional recovery of severe obturator and femoral nerve injuries after lateral retroperitoneal transpsoas surgery. J Neurosurg Spine 18:409–414 19. Ahmadian A, Deukmedjian AR, Abel N, Dakwar E, Uribe JS (2013) Analysis of lumbar plexopathies and nerve injury after lateral retroperitoneal transpsoas approach: diagnostic standardization. J Neurosurg Spine 18:289–297 20. Houten JK, Alexandre LC, Nasser R, Wollowick AL (2011) Nerve injury during the transpsoas approach for lumbar fusion. J Neurosurg Spine 15:280–284 21. Ozgur BM, Aryan HE, Pimenta L, Taylor WR (2006) Extreme lateral interbody fusion (XLIF): a novel surgical technique for anterior lumbar interbody fusion. Spine J 6:435–443 22. Sofianos DA, Briseno MR, Abrams J, Patel AA (2012) Complications of the lateral transpsoas approach for lumbar interbody arthrodesis: a case series and literature review. Clin Orthop Relat Res 470:1621–1632 23. Tohmeh AG, Rodgers WB, Peterson MD (2011) Dynamically evoked, discrete-threshold electromyography in the extreme lateral interbody fusion approach. J Neurosurg Spine 14:31–37 24. Uribe JS, Vale FL, Dakwar E (2010) Electromyographic monitoring and its anatomical implications in minimally invasive spine surgery. Spine (Phila Pa 1976) 35:S368–S374 25. Takahara T, Hendrikse J, Yamashita T et al (2008) Diffusionweighted MR neurography of the brachial plexus: feasibility study. Radiology 249:653–660 26. Taouli B, Koh DM (2010) Diffusion-weighted MR imaging of the liver. Radiology 254:47–66 27. Chhabra A, Andreisek G, Soldatos T et al (2011) MR neurography: past, present, and future. AJR Am J Roentgenol 197:583–591 28. Chhabra A, Soldatos T, Subhawong TK et al (2011) The application of three-dimensional diffusion-weighted PSIF technique in peripheral nerve imaging of the distal extremities. J Magn Reson Imaging 34:962–967 29. Chhabra A, Subhawong TK, Bizzell C, Flammang A, Soldatos T (2011) 3T MR neurography using three-dimensional diffusionweighted PSIF: technical issues and advantages. Skeletal Radiol 40:1355–1360
123
Eur Spine J 30. Meyerding HW (1956) Spondylolisthesis; surgical fusion of lumbosacral portion of spinal column and interarticular facets; use of autogenous bone grafts for relief of disabling backache. J Int Coll Surg 26:566–591 31. Landis JR, Koch GG (1977) The measurement of observer agreement for categorical data. Biometrics 33:159–174 32. Rodgers WB, Gerber EJ, Patterson J (2011) Intraoperative and early postoperative complications in extreme lateral interbody fusion: an analysis of 600 cases. Spine (Phila Pa 1976) 36:26–32 33. Smith WD, Christian G, Serrano S, Malone KT (2012) A comparison of perioperative charges and outcome between open and
123
mini-open approaches for anterior lumbar discectomy and fusion. J Clin Neurosci 19:673–680 34. Youssef JA, McAfee PC, Patty CA et al (2010) Minimally invasive surgery: lateral approach interbody fusion: results and review. Spine (Phila Pa 1976) 35:S302–S311 35. Arnold PM, Anderson KK, McGuire RA Jr (2012) The lateral transpsoas approach to the lumbar and thoracic spine: a review. Surg Neurol Int 3:S198–S215
ORIGINAL ARTICLE
Diffusion-weighted magnetic resonance (DW-MR) neurography of the lumbar plexus in the preoperative planning of lateral access lumbar surgery Cristiano Magalha˜es Menezes • Luciene Mota de Andrade • Carlos Fernando Pereira da Silva Herrero Helton Luiz Defino • Marcos Antonio Ferreira Ju´nior • William Blake Rodgers • Marcello Henrique Nogueira-Barbosa
•
Received: 14 January 2014 / Revised: 23 September 2014 / Accepted: 23 September 2014 Ó Springer-Verlag Berlin Heidelberg 2014
Abstract Purpose Magnetic resonance (MR) neurography has been used to evaluate entire nerves and nerve bundles by providing better contrast between the nerves and the surrounding tissues. The purpose of the study was to validate diffusion-weighted MR (DW-MR) neurography in visualizing the lumbar plexus during preoperative planning of lateral transpsoas surgery. Methods Ninety-four (188 lumbar plexuses) spine patients underwent a DW-MR examination of the lumbar plexus in relation to the L3–4 and L4–5 disc spaces and superior third of the L5 vertebral body. Images were reconstructed in the axial plane using high-resolution
C. M. Menezes M. A. Ferreira Ju´nior Servic¸o de Cirurgia de Coluna, Hospital Ortope´dico/Lifecenter and Hospital Sa˜o Francisco de Assis, Belo Horizonte, Brazil C. M. Menezes (&) Minimally Invasive Spine Surgery Department, Hospital Ortope´dico/Lifecenter, Rua Prof. Ota´vio Coelho de Magalha˜es, 111 Mangabeiras, Belo Horizonte, MG CEP 30210-300, Brazil e-mail: [email protected] L. M. de Andrade Axial Centro de Imagem, Belo Horizonte, Brazil C. F. P. da Silva Herrero H. L. Defino Departamento de Biomecaˆnica, Medicina e Reabilitac¸a˜o do Aparelho Locomotor da Faculdade de Medicina de Ribeira˜o Preto da USP, Ribeira˜o Preto, Brazil W. B. Rodgers Spine Midwest, St Mary’s Health Center, Jefferson City, USA M. H. Nogueira-Barbosa Divisa˜o de Radiologia e Diagno´stico por Imagem da Faculdade de Medicina de Ribeira˜o Preto da USP, Ribeira˜o Preto, Brazil
Maximum Intensity projection (MIP) overlay templates at the disc space and L3–4 and L4–5 interspaces. 10 and 22 mm MIP templates were chosen to mimic the working zone of standard lateral access retractors. The positions of the L4 nerve root and femoral nerve were analyzed relative to the L4–5 disc in axial and sagittal planes. Third-party radiologists and a senior spine surgeon performed the evaluations, with inter- and intraobserver testing performed. Results In all subjects, the plexus was successfully mapped. At L3–4, in all but one case, the components of the plexus (except the genitofemoral nerve) were located in the most posterior quadrant (zone IV). The L3 and L4 roots coalesced into the femoral nerve below the L4–5 disc space in all subjects. Side-to-side variation was noted, with the plexus occurring in zone IV in 86.2 % right and only 78.7 % of left sides. At the superior third of L5, the plexus was found in zone III in 27.7 % of right and 36.2 % of left sides; and in zone II in 4.3 % right and 2.1 % left sides. Significant inter- and intraobserver agreement was found. Conclusions By providing the surgeon with a preoperative roadmap of the lumbar plexus, DW-MR may improve the safety profile of lateral access procedures. Keywords XLIF Complication Neuromonitoring Electromyography Plexopathy
Introduction The less disruptive, transpsoas approach to the lumbar spine allows access to the intervertebral disc for decompression and reconstruction, either by fusion or arthroplasty [1, 2]. By preserving the anterior longitudinal ligament (ALL) and the posterior muscular and ligamentous
123
Eur Spine J
structures, this approach affords excellent access to the spine while minimizing collateral tissue damage. Despite being less disruptive, this retroperitoneal approach nonetheless involves a possible risk of injury to local anatomic structures. Due to the limited visualization necessitated by the technique, the lumbar plexus is at risk during the operative traverse of the psoas muscle [1, 3–17]. In recent years, this approach has rapidly grown in popularity, yet concerns about safety persist. Recent papers have documented that injury to the components of the lumbar plexus is a rare, but concerning complication associated with the transpsoas retroperitoneal approach; a risk that is reduced, but not eliminated, by the use of real-time, directional neuromonitoring technologies [18–24]. Heretofore, accurate visualization of the branches of the lumbar plexus preoperatively was not possible as traditional magnetic resonance T2-weighted and fat suppressed images do not allow for accurate differentiation between the similar signal intensities of neural and vascular structures [25, 26]. Magnetic resonance neurography (MRN) has been shown to be helpful to evaluate abnormal conditions of entire nerves and nerve bundles by providing better contrast between the nerves and the surrounding tissues. It has been used in clinical practice to evaluate peripheral neuropathies and has obvious advantages over other electrodiagnosis studies, such as electromyography and nerve conduction studies [25–27]. More recently, diffusion-weighted magnetic resonance neurography (DW-MR) has been introduced into use in the field, and uses the diffusion technique to enable the suppression of the signal from surrounding tissues (muscles and vessels), yielding improved contrast between muscular, adipose, and neural structures, thereby highlighting neural anatomy [25–31]. Three-dimensional diffusion-weighted reversed imaging with steady-state precession (3D DW-PSIF) provides adequate vascular signal suppression with an acceptable signal to noise ratio [28]. The 3D DW-PSIF sequence is isotropic, allowing for multiplanar reconstructions in any plane without significant loss of spatial resolution [28, 29]. Maximum intensity projections (MIP) may be obtained from 3D DW-PSIF and could nicely depict nerve anatomy [29]. Thus, DW-MR holds great promise as an adjunct to the preoperative planning for lateral retroperitoneal surgery. By providing a detailed roadmap of the lumbar plexus, DW-MR may be used by surgeons to improve the safety profile of lateral access procedures. The object of this study was to evaluate the utility and replicability of DW-MR in the visualization of the position lumbar nerve roots and plexus relevant to the lateral transpsoas approach to L3–4 and L4–5.
123
Methods Study design/patient sample A prospective observational study of consecutive patients ([18 years old) presenting with complaints of low back pain to the spine surgery center at the Ortope´dico/Lifecenter Hospital in Belo Horizonte, Brazil between January 2012 and March 2013 was included in the study. The study protocol was approved by the local Ethics Committee on Research, Clinics Hospital, Ribeirao Preto Medicine School at Sa˜o Paulo University and all patients signed detailed informed consent paperwork. Exclusion criteria included the presence of spondylolisthesis grade greater than II [30]; scoliosis greater than 10°; previous surgery with lumbar spine instrumentation and/or fusion; and/or other diseases and/or clinical conditions that rendered it impossible to obtain magnetic resonance imaging (MRI) studies. All patients were initially assessed using the institution’s standard protocol for the evaluation of low back pain (including MRI). In addition to the standard MR sequences, the neurographic study by diffusion technique using the three-dimensional (3D) DW-PSIF sequence, for assessing the position of components of the lumbar plexus. These additional studies where then randomly assigned into groups of ten and presented to the radiologist for a new evaluation and data acquisition for the study. In addition, the studies were evaluated by a senior spine surgeon to assess intraobserver and interobserver variability. Imaging evaluation/magnetic resonance (MR) Reversed fast image with steady-state precession (3D DW-PSIF) was used for MR neurography. Patients were all scanned with a 1.5 T MRI unit (Magneton Essenza, Siemens Healthcare) in the supine position using an 8-channel spine coil. The following parameters were used: 80–90 s/mm2 b value, TR = 17.99 ms, TE = 5.95 ms, flip angle 30°, 1.0 mm slice thickness, 256 mm field of view, 256 9 80 % matrix size, receiver bandwidth 140 Hz/pixel, fat suppression, acquisition time 8 min 37 s. One of the challenges of any radiographic modality involves the necessity of depicting three-dimensional structures, such as the lumbar plexus coursing through the psoas muscle, in two dimensions. To obtain the most pertinent clinically applicable data from the DW-MR images, this 3D to 2D approximation reconstructed the axial plane images using high-resolution maximum intensity projection (MIP) overlay templates at the level of the disc spaces of L3–4 and L4–5 of 10 and 22 mm. These reconstructive overlays were selected based on the working zone of
Eur Spine J
Fig. 1 Lateral radiographic showing approximate 10 and 22 mm reconstructive templates (windows) used during neurography to assess the potential presence of neural structures during a lateral approach using different retractors
standard lateral access retractors (10 mm for MaXcessÒ retractor (NuVasive, Inc., San Diego, CA), 22 mm for METRxÒ (Medtronic Sofamor Danek, Memphis, TN), and used to assess the potential presence of neural structures within this window with respect to these baseline retractor sizes (Fig. 1). MIP images at the disc space and at the working zone of a possible surgical retractor were reconstructed by a radiologist with extensive experience in muscle and spine diseases studies as well as in diffusion images and MIP reconstructions. L4 spinal nerve and femoral nerve position were determined in relation to the L4–5 disc and this position was analyzed in the sagittal planes using a 30 mm MIP reconstructive template (RT). The area between the anterior and posterior edges of the disc space was divided into four quartiles (zones) in the axial and sagittal planes, following the same pattern described in previous anatomic studies [8, 14, 17]. Zones were defined as: zone I (anterior quartile), zone II (middle anterior quartile), zone III (middle posterior quartile), zone IV (posterior quartile). Nerve root position was defined according to their location in each quartile as relating to the anteroposterior diameter of the vertebral body. Statistical analysis Given that these results required dependent data analysis (correlated samples), and that the same individual was measured six times (one in each group, and in both sides, right and left, in every group), marginal logistic regression using alternating logistic regression (ALR) was used to analyze the influence of groups, sides, sex, age, the presence or absence of transitional vertebrae, the last rib
(LR) arising from T12, and L4–5 listhesis for the presence of nerve in zones II or III. A hierarchical structure based in three variables in its association structure was used to model the association between measures obtained in a same individual: measures in same individual in different groups and sides, measures in same side in different groups and measures in the same group in different sides. When selecting the possible significant variables for the nerve presence at zones II or III, univariate marginal logistic regression was initially performed. A significance level of 25 % was employed. All variables selected in univariate analysis were used in multivariate marginal logistic regression, reaching the final model through backward algorithm at a 5 % significance level. The software applied for analysis was R version 2.15.3 (R Foundation for Statistical Computing, Viena, Austria). To assess intra- and interobserver variability, a Kappa coefficient was applied, and to compare intra- and interobserver Kappa indexes, as well as to compare 10 and 22 mm groups, bootstrap percentile 95 % confidence intervals were used. The software applied for analysis was R version 2.15.2 (R Foundation for Statistical Computing, Viena, Austria).
Results Ninety-four patients (188 lumbar plexuses) were analyzed. There were 46 (48.9 %) female and 48 (51.1 %) male subjects. Lumbosacral transitional vertebrae were found in 12 individuals (12.8 %). In 80 (85.1 %) patients, the last rib was positioned at the T12 level. Bilateral plexus mapping using DW-MR neurography was successfully achieved in all patients. In all subjects, the L3 and L4 roots coalesced into the femoral nerve below the L4–5 disc space. The constitution and course of the femoral nerve were documented in all patients (Fig. 2). One patient presented with a Grade I L3–4 spondylolisthesis, and 13 (15.5 %) showed grade I or II L4–5 spondylolisthesis (Table 1). Table 2 shows the results of L3–4 related to groups and sides. In one patient with an axial MIP reconstructive template (axial RT) of 22 mm and a sagittal reconstructive template (sagittal RT) of 30 mm the L4 spinal nerve entered zone III (on the right side only). When assessing the data from L4–5, it was noted that the 22 mm axial reconstructive template was significantly more likely to overlay nerves in zones II and III than the 10 mm axial template (Fig. 3), implying that a larger retractor (22 mm diameter) would be more likely to contact the neural structures of the plexus. There was also a trend towards a greater prevalence of nerves in zone III at left side compared to the right side (Table 2 and Fig. 4).
123
Eur Spine J Table 1 Variables frequence: sex, transitional vertebra and type, last rib T12, listhesis: L3– L4 and listhesis: L4–L5
Variables
N
%
Sex Female
46
48.9
Male
48
51.1
Total
94
100.0
No
82
87.2
Yes
12
12.8
Total
94
100.0
6
54.5
Transition vertebra
Type L5 sacralized S1 lumbarized
5
45.5
11
11.7
No
14
14.9
Yes Total
80 94
85.1 100.0
Total Last rib at T12
Spondylolisthesis at L3–L4 No
82
Yes
1
98.8 1.2
Total
83
100.0
Spondylolisthesis at L4–L5 Fig. 2 Lumbar plexus mapped on sagittal reconstruction over L3–L4 and L4–L5 intervertebral spaces. Oblique femoral nerve pattern is evidence at the L4–L5 level
Nerves in zones II and II were combined for the purpose of analysis (since there were so few nerves in zone II) and comparison with nerves in zone IV. Table 2 and Fig. 4 show comparisons between the presence of the nerve in zones II or III and at zone IV, between groups and sides. Nerves were more frequently found at zones II or III in the axial 22 mm RT and sagittal RT groups, compared to the axial disc 10 mm group. They were also more frequently located in zones II or III on the left side when compared to the right side. When the analysis of the L4–5 level was stratified by group (axial disc, axial 22 mm template, and sagittal 30 mm template) and side (left/right) related to the different variables (sex; transition vertebra; last rib T12; listhesis: L3–4; listhesis: L4–5), regardless of images and side studied, there was a trend towards femoral nerve location in zones II or III in male patients as well as in patients in which the transitional vertebra (TV) was present (Table 3). Interestingly, in patients with a last rib arising from T12, (LR T12), the nerve was found more frequently in zones II or III if the axial RT was 22 mm and the sagittal RT was 30 mm; this trend was reversed if the last rib did not arise from T12, and a 10 mm axial RT was employed. Only one patient presented with L3–4 listhesis, and in the axial disc 10 mm group the nerve was located in zone
123
No
71
Yes
13
84.5 15.5
Total
84
100.0
IV. Except for right side in axial disc 10 mm group, a trend towards nerve location at zones II or III was observed in those cases with L4–5 listhesis (Table 3). Figure 5 shows that the variable groups, sides, sex and transitional vertebrae were selected for the multivariate model by p values \ 0.25. Table 4 presents the multivariate model adjusted to all selected variables from the univariate analysis. Sex did not represent a significant predictor of the presence of nerves at zones II or III. The final model showed in Table 4 was thus obtained excluding the variable sex. Significant intra- and interobserver agreement in both groups was found [31]. Based on confidence intervals, there was no significant intra- and interobserver variance. Table 5 shows the intra- and interobserver agreement, using Kappa coefficient for each group and both sides.
Discussion The minimally disruptive lateral retroperitoneal transpsoas approach for anterior lumbar interbody fusion in the lumbar spine avoids many of the same potential complications of other anterior approaches and its safety and reproducibility have been repeatedly documented [23, 32–34].
Eur Spine J Table 2 L3–L4 and L4–L5 contingency table stratifying by groups between side and zones Groups
Sides
L3–L4 II
Axial disc 10 mm Axial 22 mm RT Sagittal 30 mm RT
Groups
III
Right
0
0.0 %
0
0.0 %
94
100.0 %
Left
0
0.0 %
0
0.0 %
94
100.0 %
Right
0
0.0 %
1
1.1 %
93
98.9 %
Left
0
0.0 %
0
0.0 %
94
100.0 %
Right
0
0.0 %
1
1.1 %
93
98.9 %
Left
0
0.0 %
0
0.0 %
94
100.0 %
Sides
L4–L5 II
Axial disc 10 mm Axial 22 mm RT Sagittal 30 mm RT
Groups
IV
III
IV
Right
1
1.1 %
12
12.8 %
81
86.2 %
Left
0
0.0 %
20
21.3 %
74
78.7 %
Right Left
4 2
4.3 % 2.1 %
26 34
27.7 % 36.2 %
64 58
68.1 % 61.7 %
Right
4
4.3 %
26
27.7 %
64
68.1 %
Left
2
2.1 %
37
39.4 %
55
58.5 %
Sides
L4–L5 II or III
Axial disc 10 mm Axial 22 mm RT Sagittal 30 mm RT
IV
Right
13
13.8 %
81
86.2 %
Left
20
21.3 %
74
78.7 %
Right
30
31.9 %
64
68.1 %
Left
36
38.3 %
58
61.7 %
Right
30
31.9 %
64
68.1 %
Left
39
41.5 %
55
58.5 %
Complications related to transpsoas retroperitoneal approach, despite intraoperative nervous monitoring, have been well documented [18–20, 24, 35]. Iatrogenic injuries of the lumbar plexus can result from direct mechanical compression, laceration, traction or stretching and/or indirect ischemia. Knowledge of the precise location of components of the lumbar plexus by the surgeon may afford greater safety for the procedure and could, potentially, enhance the utility of directional electroneurography [23, 24]. Several cadaveric studies have been performed to better define the precise location of the plexus and its components and, thereby, have attempted to define so-called ‘‘safe working zones’’ to gain information on and avoid neural structures during the lateral approach [1, 3, 4, 6, 8–12, 14, 15, 22]. Benglis et al. [4], in a study of three cadavers, found that 0, 11, 18, and 28 % of the posterior aspect of lateral disc space at L1–2, L2–3, L3–4, and L4–5, was covered by the nerves of the lumbar plexus, respectively. This leaves approximately the anterior 100, 89, 82, and 72 % of the
lateral disc spaces free of motor nerves in a cadaveric model. While these results suggest adequate area for docking on the lateral disc space for the lateral approach, the authors also noted the possibility of genitofemoral nerve (GFN) injury with a more anterior approach, but did not address the location of the GFN in their study. Uribe et al. [17], when studying five fresh male cadavers (with dissection of psoas muscle, lumbar plexus and nerve roots in each disc space, in 20 lumbar segments), divided out the area between anterior and posterior edges of vertebral body/intervertebral disc in four equal zones (zones I–IV, from anterior to posterior). All lumbar plexus components, including nerve roots, were contained within the substance of the psoas muscle, dorsal to posterior quarters (zones III and IV) of vertebral body. Zone II was not posterior to any part of the plexus, except for the genitofemoral nerve, a sensory branch originating from L1 and L2 roots that obliquely traverses psoas muscle from its origin to distal location, traveling to the lateral surface of the psoas muscle at approximately the L3–4 level. The study concluded that anatomical safe zones at L1–2 to L3–4 disc spaces are at the middle posterior quartile (zone III), and at or near the midline of the vertebral body and anterior at the L4–5 space (zone II–III). Moreover, there was a risk of direct genitofemoral nerve injury in zone II of L2–3 disc space and in zone I at inferior lumbar levels L3–4 and L4–5. Additionally avoidance of the ilioinguinal, iliohypogastric and lateral femoral cutaneous nerves within retroperitoneal space, where they traverse obliquely, inferiorly and anteriorly to reach iliac crest and abdominal wall is recommended [5]. The results by Uribe et al. [17] largely corroborate those results of Moro et al. [14], who analyzed 30 cadavers to establish safe working zones for transpsoas retroperitoneal endoscopic surgery. Excluding the position of the genitofemoral nerve, at least the anterior half of the disc space was free of motor nerves from L4–5 superiorly. It was also noted that the psoas muscle is often bifid at the L3 or L4 vertebral bodies and thus there is a risk of genitofemoral nerve injury. These studies, however, are limited by several factors. First, in a cadaveric model there are often limitations on number of samples collected (increasing variability), as well as by potential for displacement of the lumbar plexus during dissection. Also, cadaveric studies only account for the baseline anatomic position of the plexus. Flexion of the hip and bending of the table during the procedure that may alter the position of the plexus and the ability to retract nerves are not considered in findings of cadaveric studies. From a methodological standpoint, as well, the use of quartiles in determining zones likely underestimates the true (more posterior) position of the lumbar plexus, as corroborated by the current study and other studies which
123
Eur Spine J
Fig. 3 T2-weighted axial magnetic resonance imaging (MRI) showing L4–L5 intervertebral spaces with standard psoas muscle anatomy (a, b). MR neurography in the same patient demonstrating anteriorization of the lumbar plexus to zone II on the right (c, d)
measure the absolute position of the lumbar plexus. The previously mentioned Benglis et al. [4] study is one such study, finding the lumbar plexus at 28 % of the vertebral body from the posterior aspect. In a categorical study (quartiles), this would be defined as zones I and II being free of motor nerves despite the fact that, on average, 88 % (22/25) % of the anterior aspect of zone III would be free of motor nerves. A similar study by Regev et al. [15] found nearly identical results, with the lumbar plexus occupying 26 % of the posterior disc space at L4–5 (95 % confidence interval). Guerin et al. [9] found this measure at L4–5 to be 35.8 %. These findings suggest that the quadrant system is overly broad. Imaging by magnetic resonance is a useful tool in preoperative assessment of spine anatomy, especially in lateral approach spine surgery, helping to identify the anatomic location of nerve roots and retroperitoneal vessels as they relate to the intervertebral disc, and may aid in avoiding vascular and neurologic injuries [16]. Regev et al. [15], when reviewing 100 spine magnetic resonances from patients treated for different spine pathologies, observed that the safe surgical corridor for
123
discectomy and intervertebral spacer insertion narrows from L1–2 to L4–5, where the risk of nerve root or retroperitoneal vessel injury is increased, given the more ventral location of nerve root which forces the discectomy window ventral and thereby raising the risk of injury to the contralateral vessels. However, these MRIs were taking in a supine position and, in particular, the vessels and peritoneal structures migrate significantly anteriorly when in the lateral decubitus position, thus limiting the risk of injury [7]. In this study, preoperative MRI was recommended for assessing the position of adjacent neurovascular structures near the inferior endplate of the vertebra at each level [15]. This study uses, however, conventional magnetic resonance, without properly distinguishing between lumbar plexus and epidural and perineural vessels. This reduces the surgeon’s ability to interpret preoperative images in a useful way. DW-MR is able to suppress signal from adjacent tissues (muscles and vessels), highlighting neural anatomy and offering better contrast with other soft tissues, and is thus superior to standard MRI [25–27]. It must be noted, however, that the anatomic findings in the study were not confirmed by surgical exploration.
Eur Spine J Table 3 Contingency table of L4–L5 segments related to sex, transitional vertebra, last rib T12, listhesis: L3–L4 and listhesis; stratified by group and side Stratifications
Variables
L4–L5 II or III
IV
Axial disc 10 mm Right Left
Female
6
13.0 %
40
87.0 %
Male
7
14.6 %
41
85.4 %
Female
7
15.2 %
39
84.8 %
Male
13
27.1 %
35
72.9 %
Axial 22 mm RT Right Left
Female
13
28.3 %
33
71.7 %
Male
17
35.4 %
31
64.6 %
Female
15
32.6 %
31
67.4 %
Male
21
43.8 %
27
56.3 %
Sagittal 30 mm RT Right Left Fig. 4 Bar charts depicting the L4–L5 level stratified by registration template between groups, zones, and sides
The number of lumbosacral and thoracolumbar transition segments in this sample did not differ from that usually observed among the general population. The presence of a transitional vertebra may correlate to a so-called ‘‘tear-drop psoas’’, an indication that the plexus may be ventrally located [16]. One important implication of the current study is in relation to the use of 22 mm retractors for lateral lumbar access. Not only do these larger diameter retractors place the plexus at greater risk of injury when the neural elements are in Zone III at L4–5 (40.7 % of images analyzed) but even puts nerves in zone II as some small risk (5 %). The possible usefulness of DW-MR at preoperative planning is thus reinforced. Furthermore, by assessing side-to-side (right/left) variations in the same patient, a surgeon could choose the best surgical side, considering not only the iliac crest height or even the concave/convex side of an adult degenerative deformity, but also estimate the lumbar plexus location and the risk of injury to its components at each level being treated. One limitation of the study was that no comparisons were made between the diagnostic performance of DWMR and surgical or cadaveric data. Thus, it was not possible to assess the absolute accuracy of this classification by MRN. This limitation is relative and we believe that it does not weaken the value of the results achieved, as the goal of the study was to evaluate the ability of MRN to visualize the lumbar plexus preoperatively.
Female
13
28.3 %
33
71.7 %
Male
17
35.4 %
31
64.6 %
Female Male
15 24
32.6 % 50.0 %
31 24
67.4 % 50.0 %
Axial disc 10 mm Right Left
TV = no
10
12.2 %
72
87.8 %
TV = yes
3
25.0 %
9
75.0 %
TV = no
14
17.1 %
68
82.9 %
TV = yes
6
50.0 %
6
50.0 %
Axial 22 mm RT Right Left
TV = no
25
30.5 %
57
69.5 %
TV = yes
5
41.7 %
7
58.3 %
TV = no
29
35.4 %
53
64.6 %
TV = yes
7
58.3 %
5
41.7 %
Sagittal 30 mm RT Right Left
TV = no
25
30.5 %
57
69.5 %
TV = yes
5
41.7 %
7
58.3 %
TV = no TV = yes
31 8
37.8 % 66.7 %
51 4
62.2 % 33.3 %
Axial disc 10 mm Right Left
LR T12 = no
2
14.3 %
12
85.7 %
LR T12 = yes
11
13.8 %
69
86.3 %
LR T12 = no
4
28.6 %
10
71.4 %
LR T12 = yes
16
20.0 %
64
80.0 %
Axial 22 mm RT Right Left
LR T12 = no
3
21.4 %
11
78.6 %
LR T12 = yes
27
33.8 %
53
66.3 %
LR T12 = no
4
28.6 %
10
71.4 %
LR T12 = yes
32
40.0 %
48
60.0 %
Sagittal 30 mm RT Right Left
LR T12 = no
3
21.4 %
11
78.6 %
LR T12 = yes
27
33.8 %
53
66.3 %
LR T12 = no
4
28.6 %
10
71.4 %
LR T12 = yes
35
43.8 %
45
56.3 %
123
Eur Spine J Table 3 continued Stratifications
Variables
L4–L5 II or III
IV
Axial disc 10 mm Right
Listhesis: L3L4 = no
11
13.4 %
71
86.6 %
0
0.0 %
1
100.0 %
Left
Listhesis: L3L4 = yes Listhesis: L3L4 = no Listhesis: L3L4 = yes
17
2.7 %
65
79.3 %
0
0.0 %
1
100.0 %
Axial 22 mm RT Right
Left
Listhesis: L3L4 = no
26
31.7 %
56
68.3 %
Listhesis: L3L4 = yes
1
100.0 %
0
0.0 %
Listhesis: L3L4 = no
31
37.8 %
51
62.2 %
Listhesis: L3L4 = yes
1
100.0 %
0
0.0 %
Listhesis: L3L4 = no
26
31.7 %
56
68.3 %
Listhesis: L3L4 = yes
1
100.0 %
0
0.0 %
Listhesis: L3L4 = no
34
41.5 %
48
58.5 %
Listhesis: L3L4 = yes
1
100.0 %
0
0.0 %
Sagittal 30 mm RT Right
Left
Fig. 5 Graphic showing estimated odds ratio and 95 % confidence interval estimates by univariate logistic marginal regression
Table 4 Multivariate logistic marginal regression for the occurence in Zones II and III at L4–L5 Medium structure
b
E. P (b)
Intercept Group = axial 22 mm RT Group = sagittal 30 mm RT Side = left Transitional vertebra— yes
-1.918 0.953
0.271 \0.001 0.191 \0.001
– 2.59
[1.78–3.77]
1.034
0.202 \0.001
2.81
[1.89–4.18]
0.428 0.949
0.204 0.486
1.53 2.58
[1.03–2.29] [1.00–6.70]
Axial disc 10 mm Right
Left
Listhesis: L4L5 = no
10
14.1 %
61
85.9 %
Listhesis: L4L5 = yes
1
7.7 %
12
92.3 %
Listhesis: L4L5 = no
14
19.7 %
57
80.3 %
3
23.1 %
10
76.9 %
Listhesis: L4L5 = yes Axial 22 mm RT Right
Left
Listhesis: L4L5 = no
22
31.0 %
49
69.0 %
Listhesis: L4L5 = yes
5
38.5 %
8
61.5 %
Listhesis: L4L5 = no
27
38.0 %
44
62.0 %
Listhesis: L4L5 = yes
5
38.5 %
8
P value
0.036 0.050
O. R
I.C.—95 %
Association structure
a
E. P(a)
P valor
P.O.R
I.C.—95 %
61.5 %
Same subject Same side vs. exams Same exam vs. sides
1.974 3.822 0.456
0.445 1.393 0.254
\0.001 0.006 0.069
7.20 45.7 1.58
[3.01–17.2] [2.98–700.5] [0.98–2.60]
Table 5 Coefficient and Kappa test for intra- and interobserver variability index at 10 and 22 mm groups in both sides
Sagittal 30 mm RT Right
Listhesis: L4L5 = no Listhesis: L4L5 = yes
22 5
31.0 % 38.5 %
49 8
69.0 % 61.5 %
Left
Listhesis: L4L5 = no
29
40.8 %
42
59.2 %
Listhesis: L4L5 = yes
6
46.2 %
7
53.8 %
Source 10 mm 22 mm
Obtaining a reference pattern in this case is complicated, since the study of cadavers to identify the nerves in the psoas muscle is hard to achieve without distortion of the anatomical trajectory of the nerves. Surgical confirmation would also imply a larger selection bias in the studied
123
Kappa
P value
I.C.—95 % [0.46–0.84]
Interobserver
0.693
\0.001
Intraobserver
0.668
\0.001
[0.45–0.83]
Interobserver
0.680
\0.001
[0.50–0.81]
Intraobserver
0.782
\0.001
[0.62–0.88]
population, since it would not be possible to proceed with the investigation of patients without indication for surgery. Furthermore, there was substantial interobserver and
Eur Spine J
intraobserver reproducibility of the classification assessing the location of the elements of the lumbar plexus using with DW-MR neurography.
Conclusion To our knowledge, this represents the first report on mapping the lumbar plexus using DW-MR. The correlation between these findings and intraoperative anatomy remains to be established, as a noninvasive technique that uses neither dye nor ionizing radiation, DW-MR holds great promises as an adjunct to the preoperative planning of lateral transpsoas lumbar surgery. Acknowledgments The authors would like to thank Kyle Malone, MS for his editorial assistance. Conflict of interest
None.
References 1. Park DK, Lee MJ, Lin EL, Singh K, An HS, Phillips FM (2010) The relationship of intrapsoas nerves during a transpsoas approach to the lumbar spine: anatomic study. J Spinal Disord Tech 23:223–228 2. Pimenta L, Oliveira L, Schaffa T, Coutinho E, Marchi L (2011) Lumbar total disc replacement from an extreme lateral approach: clinical experience with a minimum of 2 years’ follow-up. J Neurosurg Spine 14:38–45 3. Banagan K, Gelb D, Poelstra K, Ludwig S (2011) Anatomic mapping of lumbar nerve roots during a direct lateral transpsoas approach to the spine: a cadaveric study. Spine (Phila Pa 1976) 36:E687–E691 4. Benglis DM, Vanni S, Levi AD (2009) An anatomical study of the lumbosacral plexus as related to the minimally invasive transpsoas approach to the lumbar spine. J Neurosurg Spine 10:139–144 5. Dakwar E, Vale FL, Uribe JS (2011) Trajectory of the main sensory and motor branches of the lumbar plexus outside the psoas muscle related to the lateral retroperitoneal transpsoas approach. J Neurosurg Spine 14:290–295 6. Davis TT, Bae HW, Mok JM, Rasouli A, Delamarter RB (2011) Lumbar plexus anatomy within the psoas muscle: implications for the transpsoas lateral approach to the L4–L5 disc. J Bone Joint Surg Am 93:1482–1487 7. Deukmedjian AR, Le TV, Dakwar E, Martinez CR, Uribe JS (2012) Movement of abdominal structures on magnetic resonance imaging during positioning changes related to lateral lumbar spine surgery: a morphometric study: clinical article. J Neurosurg Spine 16:615–623 8. Guerin P, Obeid I, Gille O et al (2011) Safe working zones using the minimally invasive lateral retroperitoneal transpsoas approach: a morphometric study. Surg Radiol Anat 33:665–671 9. Guerin P, Obeid I, Bourghli A et al (2012) The lumbosacral plexus: anatomic considerations for minimally invasive retroperitoneal transpsoas approach. Surg Radiol Anat 34:151–157 10. Hu WK, He SS, Zhang SC et al (2011) An MRI study of psoas major and abdominal large vessels with respect to the X/DLIF approach. Eur Spine J 20:557–562
11. Kepler CK, Bogner EA, Herzog RJ, Huang RC (2011) Anatomy of the psoas muscle and lumbar plexus with respect to the surgical approach for lateral transpsoas interbody fusion. Eur Spine J 20:550–556 12. Lu S, Chang S, Zhang YZ, Ding ZH, Xu XM, Xu YQ (2011) Clinical anatomy and 3D virtual reconstruction of the lumbar plexus with respect to lumbar surgery. BMC Musculoskelet Disord 12:76 13. Moller DJ, Slimack NP, Acosta FL Jr, Koski TR, Fessler RG, Liu JC (2011) Minimally invasive lateral lumbar interbody fusion and transpsoas approach-related morbidity. Neurosurg Focus 31:E4 14. Moro T, Kikuchi S, Konno S, Yaginuma H (2003) An anatomic study of the lumbar plexus with respect to retroperitoneal endoscopic surgery. Spine (Phila Pa 1976) 28:423–428 15. Regev GJ, Chen L, Dhawan M, Lee YP, Garfin SR, Kim CW (2009) Morphometric analysis of the ventral nerve roots and retroperitoneal vessels with respect to the minimally invasive lateral approach in normal and deformed spines. Spine (Phila Pa 1976) 34:1330–1335 16. Smith WD, Youssef JA, Christian G, Serrano S, Hyde JA (2012) Lumbarized sacrum as a relative contraindication for lateral transpsoas interbody fusion at L5-6. J Spinal Disord Tech 25:285–291 17. Uribe JS, Arredondo N, Dakwar E, Vale FL (2010) Defining the safe working zones using the minimally invasive lateral retroperitoneal transpsoas approach: an anatomical study. J Neurosurg Spine 13:260–266 18. Ahmadian A, Abel N, Uribe JS (2013) Functional recovery of severe obturator and femoral nerve injuries after lateral retroperitoneal transpsoas surgery. J Neurosurg Spine 18:409–414 19. Ahmadian A, Deukmedjian AR, Abel N, Dakwar E, Uribe JS (2013) Analysis of lumbar plexopathies and nerve injury after lateral retroperitoneal transpsoas approach: diagnostic standardization. J Neurosurg Spine 18:289–297 20. Houten JK, Alexandre LC, Nasser R, Wollowick AL (2011) Nerve injury during the transpsoas approach for lumbar fusion. J Neurosurg Spine 15:280–284 21. Ozgur BM, Aryan HE, Pimenta L, Taylor WR (2006) Extreme lateral interbody fusion (XLIF): a novel surgical technique for anterior lumbar interbody fusion. Spine J 6:435–443 22. Sofianos DA, Briseno MR, Abrams J, Patel AA (2012) Complications of the lateral transpsoas approach for lumbar interbody arthrodesis: a case series and literature review. Clin Orthop Relat Res 470:1621–1632 23. Tohmeh AG, Rodgers WB, Peterson MD (2011) Dynamically evoked, discrete-threshold electromyography in the extreme lateral interbody fusion approach. J Neurosurg Spine 14:31–37 24. Uribe JS, Vale FL, Dakwar E (2010) Electromyographic monitoring and its anatomical implications in minimally invasive spine surgery. Spine (Phila Pa 1976) 35:S368–S374 25. Takahara T, Hendrikse J, Yamashita T et al (2008) Diffusionweighted MR neurography of the brachial plexus: feasibility study. Radiology 249:653–660 26. Taouli B, Koh DM (2010) Diffusion-weighted MR imaging of the liver. Radiology 254:47–66 27. Chhabra A, Andreisek G, Soldatos T et al (2011) MR neurography: past, present, and future. AJR Am J Roentgenol 197:583–591 28. Chhabra A, Soldatos T, Subhawong TK et al (2011) The application of three-dimensional diffusion-weighted PSIF technique in peripheral nerve imaging of the distal extremities. J Magn Reson Imaging 34:962–967 29. Chhabra A, Subhawong TK, Bizzell C, Flammang A, Soldatos T (2011) 3T MR neurography using three-dimensional diffusionweighted PSIF: technical issues and advantages. Skeletal Radiol 40:1355–1360
123
Eur Spine J 30. Meyerding HW (1956) Spondylolisthesis; surgical fusion of lumbosacral portion of spinal column and interarticular facets; use of autogenous bone grafts for relief of disabling backache. J Int Coll Surg 26:566–591 31. Landis JR, Koch GG (1977) The measurement of observer agreement for categorical data. Biometrics 33:159–174 32. Rodgers WB, Gerber EJ, Patterson J (2011) Intraoperative and early postoperative complications in extreme lateral interbody fusion: an analysis of 600 cases. Spine (Phila Pa 1976) 36:26–32 33. Smith WD, Christian G, Serrano S, Malone KT (2012) A comparison of perioperative charges and outcome between open and
123
mini-open approaches for anterior lumbar discectomy and fusion. J Clin Neurosci 19:673–680 34. Youssef JA, McAfee PC, Patty CA et al (2010) Minimally invasive surgery: lateral approach interbody fusion: results and review. Spine (Phila Pa 1976) 35:S302–S311 35. Arnold PM, Anderson KK, McGuire RA Jr (2012) The lateral transpsoas approach to the lumbar and thoracic spine: a review. Surg Neurol Int 3:S198–S215
