The study was performed in accordance with the Declaration of Helsinki after approval by the Ethics Committee of our institution (approval number: 262-1074). All participants received written and verbal explanations of the study and provided informed consent before participation.
Participants
We enrolled consecutive patients with LSS aged between 51 and 79 years who visited our university hospital. The inclusion criteria were as follows: symptoms in one or both legs, with or without LBP, radiographic evidence of spinal stenosis or cauda equina compression, and clinical signs of nerve root affection. The exclusion criteria were as follows: infection, neoplasm, history of spinal surgery, acute trauma, history of spinal fracture, and spondylolisthesis with obvious instability, which was defined as a sagittal translation of ≥ 3 mm, segmental mobility of ≥ 20°, or posterior opening of ≥ 5° on flexion/extension radiographs. LBP was defined as pain, discomfort, and stiffness in the lower back, extending from the 12th rib to the lumbar or lumbosacral area, lasting more than three months. All participants rated their LBP on the visual analog scale (VAS; 0−100 mm). Based on the findings of previous reports, a VAS score of > 30 mm was defined as moderate or severe pain, whereas a VAS score of ≤ 30 mm was defined as no or mild pain [20,21].
Classification of LES
As described in a previous report [22], LES were classified into cauda equina, radicular pain, and mixed.
Measurement of BMI and BMD
We measured the patients’ height and body weight and calculated their BMI (kg/m2) [23]. BMD was measured at L2, L3, and L4 using a dual-energy X-ray absorptiometry scanner.
Measurement of ROM
We acquired dynamic flexion-extension radiographs of the participants in the standing position. The angle between the superior endplates of L1 and S1 was termed the ROM.
Measurement of the sagittal spinopelvic radiologic parameters
As described in a previous report, we obtained full-length spinal and pelvic radiographs of the participants in the standing position and used them to calculate several parameters [11]. The following sagittal spinal parameters were measured on sagittal-view spinal radiographs: lumbar lordosis (LL; the superior endplate of L1 to the superior endplate of S1; Fig. 1a), thoracic kyphosis (TK; the superior endplate of T4 to the inferior endplate of T12; Fig. 1a), and sagittal vertical axis (SVA; the horizontal offset from the posterosuperior edge of S1 to the center of the body of C7; Fig. 1b). The following sagittal pelvic parameters were measured on sagittal-view pelvic radiographs: sacral slope (SS; the angle between the horizontal and the superior sacral endplate; Fig. 1c), pelvic tilt (PT; the angle between the vertical axis and the line running from the midpoint of the sacral plate to the center of the femoral head axis; Fig. 1d), pelvic incidence (PI; the angle between a line perpendicular to the superior sacral endplate at its midpoint and the line connecting this point to the center of the femoral head axis; Fig. 1e), and PI-LL. Two investigators blinded to the study assessed the intraobserver and interobserver reliability of the measurements of the spinopelvic parameters (observer 1, I.O., and observer 2, H.T.). The κ values for intraobserver and interobserver reliability were as follows: LL, 0.85 and 0.91; TK, 0.89 and 0.92; SVA, 0.84 and 0.91; SS, 0.83 and 0.90; PT, 0.85 and 0.88; PI, 0.81 and 0.87; and PI-LL, 0.80 and 0.84, respectively.
Measurement of the CSA and FI of the Mm
We used the Signa HDx 1.5T magnetic resonance imaging (MRI) system (GE Healthcare, Milwaukee, WI, USA) with a spine coil to obtain T2-weighted MRI images. The CSA and FI of the Mm at the L3–L4, L4–L5, and L5-S levels were measured using axial T2-weighted MRI. CSA was assessed by manually tracing the fascial border of the Mm, as previously described [14]. We analyzed the histograms of signal intensity in the regions of interest for the areas using digitized image-processing software (Image J; National Institutes of Health, Bethesda, MD, USA). We measured the percentage area with FI using the software’s pseudo-coloring tool, using which pixels representing fat tissue appeared red. We then calculated the percentage of the muscle compartment that was red. The CSA and FI data were averaged between the right and left Mms. The κ values for intraobserver and interobserver reliability were 0.88 and 0.92 for CSA and 0.82 and 0.89 for FI, respectively.
Assessment of IVDD using T2 mapping
We performed MRI T2 mapping using a protocol described in previous studies [24-28]. Sagittal images were acquired with the patients in the supine position, and T2 maps were created on a pixel-by-pixel basis. We used the T2 values of the midsagittal section, which was centered on the lumbar midline, with an optimized 8-echo multi-spin echo sequence obtained using the Advantage Workstation (version 4.4, Functool; GE Healthcare, Milwaukee, WI, USA) with the following parameters: repetition time, 1000 ms; first echo time [TE], 14.8 ms; last TE, 118.6 ms; receiver bandwidth, ± 15.63 kHz; field of view, 22 cm; matrix, 320 × 256; slice thickness/gap, 4 mm/4 mm; number of slices, 5; number of excitations, 2; and total scan time, 8 min and 34 s. We did not use the first echo from the multi-spin to minimize the effect of the stimulated echo. The T2 map was calculated for each pixel from the signal intensity in the respective TE using the following formula: SI (TE) = e –TE/T2.
The intervertebral discs at L3–L4, L4–L5, and L5-S were divided into five equal areas each. We measured the mean T2 values at the first, middle, and last fifth areas, which were the anterior annulus fibrosus (AF), the center of the nucleus pulposus (NP), and the posterior AF, respectively [24-28] (Fig. 2) A total of 300 levels were evaluated. The T2 values were measured using MedCalc (version 10.2.0.0; MedCalc Software, Mariakerke, Belgium) by a PhD researcher (H.T.) with 15 years of experience in spine MRI analysis.
Assessment of MC
MC were evaluated from L1–L2 to L5–S1 and classified as none, types 1, type 2, or type 3 according to their signal patterns on T1- and T2-weighted sagittal MRI [29]. Type 1 MC were hypointense on T1-weighted images and a hyperintense on T2-weighted images. Type 2 MC were hyperintense on both T1- and T2-weighted images. Type 3 MC were hypointense on both T1- and T2-weighted images. The intraobserver and interobserver reliability were excellent, with κ values of 0.81 and 0.84, respectively.
Assessment of FJD
To evaluate FJD, we acquired axial images at three lumbar levels (L3–L4, L4–L5, and L5−S1) using computed tomography (Aquilion PRIME, Toshiba, Japan). As described in a previous report, FJD was classified into four grades: grade 0, normal; grade 1, mild degenerative disease; grade 2, moderate degenerative disease; and grade 3, severe degenerative disease [30]. If there was a difference in FJD severity between the right and left sides at the same lumbar level, the worse grade was recorded. All patients were categorized as either grade 0–1 or grade 2–3. The intraobserver and interobserver reliability were excellent, with κ values of 0.80 and 0.81, respectively.
Statistical analyses
We compared the LES, BMI, BMD, spondylolisthesis, ROM, spinopelvic alignment, CSA and FI of the Mm, IVDD, MC, and FJD between the high and low groups using the Mann−Whitney U test and chi-square test. We performed multiple logistic regression analysis with the high and low groups as dependent variables. Variables found to be associated with LBP (p < 0.10 in the univariate logistic regression analysis) were entered into the multivariate logistic regression models using forward selection (likelihood ratio). To determine the boundary values of the VAS score for LBP, we performed receiver operating characteristic (ROC) analysis of the significant variables. Statistical significance was set at p < 0.05. We used SPSS (version 27.0; IBM Corp., Armonk, NY, USA) for all statistical analyses. Numerical data were expressed as mean ± standard error of the mean.