Facet Joint

Detailed descriptions of the complex anatomy of lumbar paravertebral muscles and definitions regarding the spatial distribution of MFI on axial MRI are limited [37 –40 (link)]. Published images demonstrating investigators’ definition of ROI for these muscles predominantly depict the lower lumbar levels, with limited identification of separate muscles. Further, descriptions lack details towards acknowledging the complex three-dimensional structure that produces a changing spatial relationship observed across lumbar segmental levels. The lumbar paravertebral muscles typically examined in such studies include: multifidus (MF) as the largest lumbar spinotransverse muscles; erector spinae (ES) including lumbar longissimus and iliocostalis; and less frequently, psoas (including major and minor), and quadratus lumborum (see Fig. 1). This paper intentionally focuses on MF and ES as these are presumed to have the greatest clinical significance. However, other paravertebral muscles exist in the lumbar spine (e.g. the lumbar interspinales and intertransversarii, and thoracic semispinalis), yet they are generally not mentioned in descriptive investigations. This may relate to a lack of image resolution with available sequences, making it challenging to accurately delineate individual muscles from adjacent structures, and it therefore remains unclear how they should be treated when defining ROIs.Fig. 1

Axial E12 plastinated sections (a, c) and schematic illustrations (b, d) at approximately L1 (a, b) and L4 (c, d) highlighting anatomical structures at these vertebral levels. b, d Dotted lines and shading, Green - psoas major muscle; Blue – quadratus lumborum muscle; Purple – erector spinae muscles; Red – spinotransverse muscles. b round white dotted regions (bilateral) denote 12th rib. d square dotted box surrounds enlarged inset; round dotted circle indicates morphological feature of interest (ILB fatty ‘tent’). Legend: A – aorta; ES – erector spinae muscles; ESA – erector spinae aponeurosis; ILB – iliocostalis – longissimus boundary and indentation; ISL – interspinous ligament; IT – intertransversarii muscle; IVC – inferior vena cava; K – kidney; L – liver; P – psoas major muscle; QL – quadratus lumborum muscle; SAF – superior articular facet; SP – spinous process; SPC – spinal canal; SPT – spinotransverse muscle group; ZJ – zygapophysial joint

Our proposed method outlined in the results section, provides a foundational solution for the problem of how to measure muscles traversing the lumbar spine, and includes suggestions on operational characteristics for acquiring MR images. While we offer this starting point for a common methodology to facilitate accurate definition of lumbar muscle ROI, we are cognisant that the method is not a definitive end-point on ‘how to’. We hope that with time and new research findings these methods will be modified, expanded, and refined.

The procedures for developing the FE model of lumbar spine and the materials used in this model were introduced in the author’s previous study on the response of spine under dynamic loading [30 ]. Briefly, a nonlinear FE model of complete lumbar spinal segments (L1-S1) was generated based on geometrical reconstruction from computer tomography (CT) scan. The image segmentation and reconstruction of geometrical model were finished in a medical image processing software (Mimics 10.0; Materialise Technologies, Leuven, Belgium) and a reverse engineering and scanning software (Geomagic studio 10.0; Geomagic Inc., North Carolina, USA), respectively. The mesh was generated in a CAE pre-processoring software (Hypermesh 11.0; Altair Engineering Corp, Michigan, USA). A vertebra consists of a cortical wall, a cancellous bone, endplates and post elements. The intervertebral disc was made up of the nucleus and annulus, and the annulus was assumed to be composed of a ground substance reinforced by a collagen fiber network. The fibers embedded in the ground substance were oriented at an average angle of ±30° to the endplates. The seven ligaments: the anterior (ALL) and posterior (PLL) longitudinal ligaments, the intertransverse (ITL), flavum (FL), supraspinous (SSL), interspinous (ISL) and capsular (CL) ligaments were included. Facet cartilage was represented by three layers of hex elements with inhomogeneous thickness and the gap between the facet articular surfaces was about 0.1 mm. The distribution of cartilage thickness in each facet was in accordance with the findings of Woldtvedt et al. [24 (link), 26 (link)]. The complete FE model of lumbar L1-S1 segments is shown in Fig. 1.Fig. 1

FE model of lumbar spine with the details of the disc and the facet cartilage

All the components of vertebrae were considered as isotropic homogeneous elastic materials. The annulus ground substance, nucleus pulposus and facet cartilage were simulated to be nearly incompressible and hyper-elastic [28 (link), 30 ]. Ligaments and fibers were simulated as tension-only spring with nonlinear properties according to the research of Shirazi-Adl [31 (link)]. All the material properties and elements information (types and number) for each parts of this lumbar model were listed in Table 1.Table 1

Material properties and elements information of the model used in this study

Components name	Young’s modulus (MPa)	Poisson’s ratio	Element type	Element no.
Cortical bone	14000	0.30	Hex	2585
Cancellous bone	100	0.2	Tetra	129931
Posterior elements	3500	0.25	Tetra	250978
Endplate	10000	0.25	Hex	4921
Sacrum	5000	0.2	Tetra	200295
Facet cartilage	Neo-Hookean, C10 = 2		Hex	7293
Annulus	Mooney–Rivlin, C1 = 0.18, C2 = 0.045		Hex	6000
Nucleus pulpous	Mooney–Rivlin, C1 = 0.12, C2 = 0.03		Hex	7200
Fiber	Calibrated stress–strain curves		Spring	14400
Ligament	Calibrated deflection–force curves		Spring	234

For the PTED group, the surgical procedure (based on the L4–L5 segment of DLS) was performed following methods reported in the literature [18 (link)]. The following steps were performed: (1) part of the superior articular process (SAP) of L5 was removed. A soft pillow was placed under the patients' waist, while the patient was in the lateral decubitus position with their knee and hip flexed. The incision was located 8–12 cm from the midline horizontally and 1–3 cm above the iliac on the side with leg pain. The mixed local anesthetic, which consisted of 30 mL 1:200,000 epinephrine and 20 mL 2% lidocaine, was only used in PTED group. After 5 mL of the mixed anesthetic was inserted into the skin at the entry point, 20 mL was inserted into the trajectory, 15 mL was inserted into the articular process, and 10 mL was inserted into the foramen. Then, 0.8–1.0 cm of skin and the subcutaneous fascia were incised. Drills were used to resect the ventral osteophytes on the SAP. The PTED system (Hoogland Spine Products, Germany) was inserted (Fig. 1). (2) Parts of the ipsilateral ligamentum flavum, perineural scar, and extruded lumbar disc material were completely resected with endoscopic forceps (Fig. 2). (3) The superior endplate of the L5 vertebral body was removed by endoscopic micro punches and a bone knife. Therefore, 270-degree decompression of the traversing nerve root was achieved (Fig. 3). The drainage tube was placed after hemostasis was reached.Fig. 1

Fluoroscopic views. A, B The drill was inserted to resect the LF and the ventral osteophytes on the SAP. C, D The working cannula was placed

Fig. 2

Endoscopic views. A Endoscopic view of the hypertrophic posterior longitudinal ligament, extruded disc material, and perineural scar. B–G After the endoscopic instruments were used to carefully remove the vertebral body, ventral decompression of the traversing nerve root (L5) was completed. H The dura mater was torn

Fig. 3

Illustrations of the 270-degree PTED. A, B Specific pathologic features of LRS-DLS. C, D Final view of the nerve 270-degree decompression status and the restoration of the lateral recess

For the MIS-TLIF group, the surgical procedure was performed in accordance with methods reported in the literature [19 (link)]. After successful general anesthesia with tracheal intubation, the patient was placed in a prone position with chest and hip pads, and the L4–L5 intervertebral space was marked with X-ray fluoroscopy. The skin and subcutaneous fascia were cut; a trans-muscular surgical corridor was created with two micro-laminectomy retractors docking on the facet joint complex. After exposing the bony structure, part of the lamina and inferior articular process of L4 and the upper L5 articular process were removed with the rongeur on the ipsilateral side, and the hypertrophic ligamentum flavum was peeled backward. If MRI showed contralateral lateral recess stenosis, then predecompression was performed on the contralateral side. After decompression on the dorsal side, the nucleus pulposus and endplate cartilage were removed with forceps. An appropriate cage (Medtronic) filled with autograft from laminectomy was placed in the center of the intervertebral space via the Kambin’s triangle area. After adequate hemostasis was achieved, two drainage tubes were placed and removed when the drainage volume was < 50 mL/d.
Postoperatively, patients was treated with oral nonsteroidal anti-inflammatory drugs and antibiotics for 3 days. All patients were encouraged to perform straight leg raising 1 day postoperatively, and moderate off-bed activity with a brace 2–3 days postoperatively. On the third postoperative day, patients were allowed to go home if their lower extremity pain symptoms were effectively relieved with no evidence of infection. The patient demographics and perioperative outcomes were compared. The VAS score, ODI, and modified Macnab criteria were used to evaluate the clinical outcomes [20 (link)].

All procedures and protocols were conducted in accordance with the principles of Helsinki Declaration, written informed consent was obtained from all participants in accordance with standard operative procedures. The study was approved by the Clinical Hospital Center Zemun- Belgrade by the number of approval 8946.
The present study was an observational cohort study of 60 patients who were prospectively recruited and divided into two groups. Patients were hospitalized at the Department of Neurosurgery at Clinical Hospital Center Zemun between 2020 and 2021 and were operated with surgical indications of lumbar disc herniation (LDH) and lumbar spinal stenosis (LSS). All patients fulfilled the following criteria: 1. Age > 18 years, 2. No previous surgery on spine, 3. Diagnosis was verified by magnetic resonance imaging. Patients with history of osteoporosis, immunosuppression, chronic corticosteroid use, intravenous drug use, fever of unknown origin, history of malignancy, unexplained weight loss, or progressive/disabling symptoms were excluded from the study. All patients were operated by one neurosurgeon (V. A.). The LF samples were obtained from the 60 patients randomized in 2 groups. The first group underwent micro-discectomy for LDH and included LF samples from 30 patients (LDH group). The second group underwent decompressive surgery without instrumented fusion for LSS and included LF samples from 30 patients (LSS group). In the patients with multisegmental stenosis, samples were taken from the radiologically determined site of greatest stenosis. While every effort was made to remove the LF en-bloc, in the majority of cases, the LF was removed piecemeal.
Demographic and clinical data were obtained using a pre-prepared questionnaire as well as data from medical history. Morphological/radiological data were obtained by measuring specific parameters on magnetic resonance imaging—T2 sequences, performed by two experienced radiologists, after several repeated measurements. The examined morphological/radiological parameters measured on the sagittal image projection of the lumbosacral spine region were presence of Schmorl's nodes, vertebral body hemangioma, spondylolisthesis, and value of lumbar lordosis angle. Other measurements were performed at the axial image section where the degree of discal herniation or spinal stenosis were most prominent and included: interfacet distance, thickness of LF on both sides, dural laterolateral (LL) diameter and anteroposterior (AP) diameter of dural sac, average facet joint angle, and dural sac surface. The scoliosis angle was also determined using the standard Cobbs method on the coronary sections of spine magnetic resonance imaging of the patients^{30 (link)}. Spondylolisthesis was determined as a percentage of vertebral body slippage. Lumbar lordosis angle was also determined using the standard Cobbs method^{30 (link)}. The determination of the other mentioned parameters is shown in Fig. 2.Figure 2

Measurement of (A) ligamentous interfacet distance, (B) anteroposterior diameter of dural sac, (C) laterolateral diameter of dural sac, (D) thickness of LF on both sides, (E) average facet joint angle measured according to formula: (a + b)/2, (F) dural sac surface.