Lumbosacral spinal compression device with the use of a cushion back support in supine MRI

Abstract

Background

We hypothesized that axial-loaded magnetic resonance imaging (MRI), modified with the use of a cushion placed behind the lower back (i.e. BS-MRI method), would simulate the standing position more accurately than an axial-loaded MRI without a cushion back support (BS).

Purpose

To determine whether the BS-MRI method demonstrated similar morphologies on intervertebral disc (IVD), dural sac, and spinal curvature as those detected on 90° standing MRIs in individuals with suspected spinal stenosis.

Material and Methods

Twenty-five subjects underwent a BS-MRI, as well as axial-loaded and standing MRI studies. Outcome measures were four radiographic parameters of the lumbar spine: IVD height (DH); dural sac cross-sectional area (DCSA); and spinal curvature (i.e. lumbar lordosis [LL] and L1-L3-L5 angle [LA]).

Results

Major differences (>5%) between standing MRI and BS-MRI methods were observed in DCSA, DH, and LL. Major differences between standing and axial loaded MRIs were observed only in DCSA and LA. Although BS-MRIs demonstrate an image of the lumbar spine curvature (i.e. LA) which is closer to that when standing than axial-loaded MRIs, it is likely to overestimate both narrowing of dural sac and extent of LL.

Conclusion

Using a compression device with a BS to simulate weight-bearing on the lumbar spine is not recommended due to: (i) overestimation of the narrowing of the dural sac and extent of LL; and (ii) underestimation of loss of disc height. Supine axial-loading produced DCSA and DH which were strongly correlated with those detected with standing MRIs. Exceptions were that LL and LA were underestimated.

Keywords

Axial loading low back pain upright magnetic resonance imaging spinal stenosis cushion back support lumbar lordosis

Introduction

Supine magnetic resonance imaging (MRI) may be combined with the use of a lumbosacral spinal compression device (i.e. axial-loaded MRI) as a tool for the diagnosis of lumbar spinal stenosis when the findings of conventional supine MRI do not match the clinical signs of stenosis evident in some patients. Several studies confirm that such axial loading devices can add valuable information (especially in patients with neurogenic claudication), impact the treatment plan for degenerative disorders (1,2), increase the specificity of a diagnosis of spinal stenosis (3 –5), and reveal both changes of dural sac size (3) and decreases in disc height (6).

However, the effect of gravity is not fully simulated in supine axial-loaded MRI scans (7,8). Lumbar lordosis (LL) is underestimated when compared with standing MRI (7,8). Many studies suggest using a cushion for back support (BS) during axial loading in the supine position to produce extension of the lumbar spine (3,9 –12) but there are no reports assessing this modification. Thus, we investigated the hypothesis that an axial-loaded MRI, modified with the use of a BS (i.e. BS-MRI method), would simulate the standing position more accurately than the axial-loaded MRI without a BS. The aim of the present study was to determine whether the BS-MRI method would produce morphological changes in the dural sac, degree of lumbar spinal curvature, and height of intervertebral disc (IVD) closer to those detected when 90° standing MRIs are used in individuals with suspected spinal stenosis. Morphological changes (i.e. IVDs, spinal canal, and spine curvature) detected with the BS-MRI were compared against both those found when using axial-loaded and standing MRIs.

Material and Methods

Participants

The sample size was calculated based on the standard deviation (SD) of LL reported in Wessberg et al. (8). In order to achieve a significance level of 5% and power of test of 0.8, a sample size of at least 14 would be required.

Twenty-five adults, who had clinical symptoms of spinal stenosis (i.e. low back pain and/or neurogenic claudication) but had never been diagnosed and had no bone fractures, were prospectively recruited for the study. All participants stated that their symptoms were worse when standing and/or sitting and decreased when lying down. All individuals had body circumferences < 35 cm (due to the limited opening of the standing MRI scanner), with heights and weights not exceeding 1.80 cm and 100 kg, respectively (due to the constraints of the lumbosacral spinal compression device).

The study was performed in accordance with the Declaration of Helsinki and was approved by the Committee on Human Rights related to Research involving Human Volunteers at the Faculty of Medicine, Ramathibodi Hospital (COA. MURA2020/1158).

Compression device

A non-magnetic, SpineMAC lumbosacral spinal compression device, developed by the authors, was used to apply axial compression.

Experimental protocol

The lumbar spine of each participant was examined three times, by different methods of MRI. First, they were examined in the supine position with axial loading (axial-loaded MRI). Second, they were examined in the supine position with axial loading after placement of the BS under the lower back (BS-MRI). Third, they were examined in a 90° standing position.

Axial-loaded MRI

For the first scan, the study was performed using a 3.0-T system (Phillips Ingenia, Phillips Healthcare, Best, The Netherlands) and a total spine coil. The SpineMAC was attached to the individual on the MRI table. All participants lay on a slip carpet that covered the MRI table to minimize friction forces between the body of the individual and the table during axial loading. The load applied was 50% of the participant’s body weight (16,18), divided equally between the two foot plates (25% of body weight on each foot), and anchored to a body vest (Fig. 1a). Sagittal and axial T1-weighted (T1W) and T2-weighted (T2W), turbo spin echo images of the lumbar spine were acquired. Slice thicknesses were 3.0 mm (field of view [FOV] = 180 mm) for axial images and 4.0 mm (FOV = 320 mm) for sagittal images. The box for transverse slices was placed parallel to the IVDs. Total scan time was approximately 20 min. Then, patients rested their lumbar spine by staying on the MRI bed in the supine position without compression for 10 min before starting the next scan.

BS-MRI

Fig. 1.

Supine axial-loaded MRI (a), supine axial-loaded MRI with back support (b), and standing MRI (c) examinations. MRI, magnetic resonance imaging.

For the second scan (BS-MRI), a BS was placed under the lower back of the participant to produce extension of the lumbar spine while in the supine position (Fig. 1b). The BS was made from a rubber semi-circular pillow (Fig. 1b, inset). Its height and length were 7.5 cm and 50 cm, respectively. Then, a spinal compression load (50% of the individual’s body weight) was applied (Fig. 1b). MRI images were acquired, performed as with the first scan. Total scan time was approximately 20 min.

Upright standing MRI

For the last scan, a 90° standing MRI scan was performed in the month after the supine MRI scan. A 0.25-T system (Esaote G-scan Brio, Genova, Italy) with a four-channel lumbar spine coil was used to perform the weight-bearing MRI. The participant entered the scanner facing upward with their back on the table. From a horizontal position of 180°, the MRI table was adjusted to a vertical position (90° table tilt) so that the individual was then standing. The table tilt of 90° was a real upright position and participants were asked to stand normally without moving or leaning on the sides of the magnet nor on the back of the scanner (Fig. 1c). Sagittal and axial T1W spin echo and T2W fast spin echo images of the lumbar spine were acquired. Slice thicknesses were 3.0 mm (FOV = 160 mm) for axial images and 4.0 mm (FOV = 280 mm) for sagittal images. Total scan time was approximately 40 min in this standing position. The box for transverse slices was placed parallel to the IVDs and as close as possible to the same position as was used in the supine axial-loaded MRI studies.

Image analysis and evaluation

All images were exported as DICOM files for analysis on Philips Intellispace Portal 5.3 software. MR images were visually analyzed by a radiologist. Quantitative measurement was performed manually on supine axial-loaded MRIs (with and without a BS) and standing MRIs using a digitized tool. Cross-sectional area (CSA) of the dural sac (DCSA) was measured at the L4-L5 and L5-S1 levels in axial T2W images (Fig. 2a), while height of the IVD (DH) was measured at the same levels in sagittal T1W images (Fig. 2b). Other radiographic parameters, including LL and lumbar angle (LA), were measured in mid-sagittal slices of T2W images (Fig. 2c). In order to avoid bias, all radiographic parameters were measured by two experienced radiological technologists who were each blinded to the other’s assessments. The mean of each radiological parameter (from observers 1 and 2) was used for statistical analyses.

Fig. 2.

Measurements of DCSA in an axial T2W image (a), DH at L4-L5 (DH45) and L5-S1 (DH51) in a sagittal T1W image (b), LL and LA in a sagittal T2W image (c). DCSA, cross-sectional area of dural sac; DH, disc height defined as the distance between superior and inferior border of intervertebral discs exclusive of endplates; LA, angle of the center of L1-L3-L5; LL, lumbar lordosis measured from the angle between superior endplate of L1 and superior endplate of S1; T1W/T2W, T1-weighted/T2-weighted.

Statistical analysis

Interrater reliability of all measurements

To investigate the inter-observer variability of the measurements, intraclass correlation coefficients (ICCs) were assessed with SPSS version 18.0 (IBM Corp., Armonk, NY, USA). Distribution was checked for normality using a Kolmogorov–Smirnov test.

Comparisons between the MRI methods

A dependent-sample t test was used for comparisons of radiographic parameters between axial-loaded and standing MRIs, and comparisons between BS and standing MRIs. A P value < 0.05 was defined as denoting significance. Statistical analyses and plots were performed with SPSS version 18.0.

Agreement between the MRI methods

Brand–Altman analysis (26) was used to visually assess systemic differences, and estimate mean differences and limits of agreement (LOA) between axial-loaded and standing MRIs and between BS-MRI and standing MRIs. Bland–Altman plots, which presented the differences between the two methods versus the mean values from standing MRI with the representation of the LOA (from –1.96 SD to +1.96 SD), were constructed and evaluated. For determining the agreement between the two methods, ICC estimates (27) and their 95% confidence intervals (CIs) were calculated based on absolute-agreement and two-way random effects models. Statistical analyses and plots were performed with SPSS version 18.0.

Results

Participant characteristics

Twenty-five adults (9 men, 16 female; average age = 38 years; age range = 21–59 years) with suspected spinal stenosis enrolled in the study and were prospectively examined with axial-loaded MRI, BS-MRI, and standing MRI studies. Their average body mass index was 25.3 ± 1.9 kg/m². The median time interval between supine axial-loaded MRI and standing MRI studies was 0.3 months (range = 0.07–0.95 months).

Narrowing of dural sac

Comparison of narrowing of the dural sac obtained from BS-MRIs against the other two methods (axial-loaded and upright MRIs) is presented as DCSA measured in axial images at the L4-L5 and L5-S1 levels. For BS-MRIs, very good agreement at L4-L5 (ICC =0.905) and good agreement at L5-S1 (ICC = 0.874) were found with standing MRIs (Fig. 3a and 3b); mean differences were 15.7 ± 20.9 mm² and 10.8 ± 28.2 mm², respectively (Fig. 4a and 4b). For axial-loaded MRIs, very good agreement at L4-L5 (ICC =0.952) and good agreement at L5-S1 (ICC = 0.866) were found with standing MRIs (Fig. 3c and 3d); mean differences were 9.1 ± 15.9 mm² and 5.7 ± 30.3 mm², respectively (Fig. 4c and 4d). In comparison with standing MRIs, both BS-MRIs (P = 0.001) and axial-loaded MRIs (P = 0.01) overestimated narrowing of the dural sac at L4-L5 (Fig. 5a). In addition, BS-MRI overestimated narrowing of the dural sac (DCSA) at L5-S1 (P = 0.03), while axial-loaded MRIs showed them to be not different from standing MRIs (P = 0.44) (Fig. 5a).

Fig. 3.

Scatter diagrams show the values of DCSA on standing MRI vs. those values observed on BS-MRI at L4-L5 (a) and L5-S1 (b); the values of DCSA on standing MRI vs. those values observed on axial-loaded MRI at L4-L5 (c) and L5-S1 (d) for each patient and ICCs with 95% CI. BS, back support; DCSA, cross-sectional area of dural sac; MRI, magnetic resonance imaging.

Fig. 4.

Bland–Altman plots with data points for each patient. The y-axis shows the difference (diff) in the values of DCSA between standing and BS-MRI at L4-L5 (a) and L5-S1 (b) levels; and DCSA between standing and axial-loaded MRI at the L4-L5 (c) and L5-S1 (d) levels. The x-axis shows the mean values of the DCSA on standing MRI. The red solid line indicates the mean difference (value and percentage of the value expressed on the red solid line). The dashed lines indicate the limits of agreement. BS, back support; DCSA, cross-sectional area of dural sac; MRI, magnetic resonance imaging.

Fig. 5.

Comparison of mean DCSA (a), DH (b), LL and LA (c) obtained from axial-loaded MRIs with a cushion BS (light gray bars) against those from standing (black bars) and axial-loaded (dark gray bars) MRIs. BS, back support; DCSA, cross-sectional area of dural sac; DH, height of intervertebral disc; LA, lumbar angle; LL, lumbar lordosis; MRI, magnetic resonance imaging.

Figs. 6 and 7 are examples of axial T2W images from BS-MRI (Figs. 6a and 7a), standing (Figs. 6b and 7b), and axial-loaded (Figs. 6c and 7c) MRIs at the L4-L5 (Fig. 6) and L5-S1 (Fig. 7) levels of the same individual. DCSA was smallest in BS-MRIs (Figs. 6a and 7a) when compared to standing (Figs. 6b and 7b) and axial-loaded (Figs. 6c and 7c) MRIs at both levels.

Fig. 6.

MRI examinations in a 29-year-old woman with a five-year history of low back pain without neurogenic claudication. Axial T2W images compare DCSA at L4-L5 in BS-MRI (a, DCSA = 64 mm²) against standing (b, DCSA = 95 mm²) and axial-loaded (c, DCSA = 73 mm²) MRI examinations. BS, back support; DCSA, cross-sectional area of dural sac; MRI, magnetic resonance imaging; T2W, T2-weighted.

Fig. 7.

MRI examinations in a 47-year-old man with a one-year history of low back pain with neurogenic claudication in the left leg. Axial T2W images compare DCSA at L5-S1 in BS-MRI (a, DCSA = 174 mm²) against standing (b, DCSA = 191 mm²) and axial-loaded (c, DCSA = 184 mm²) MRI examinations. Focal annular fissure is observed at the posterior-central portion of the L5-S1 disc. BS, back support; DCSA, cross-sectional area of dural sac; MRI, magnetic resonance imaging; T2W, T2-weighted.

The height of IVDs from BS-MRI compared with the two other methods is presented as DH measured in sagittal images at the L4-L5 and L5-S1 levels. During BS-MRI, very good agreement at L4-L5 (ICC = 0.98) and good agreement at L5-S1 (ICC = 0.895) were found between BS and standing MRIs (Fig. 8a and 8b), with mean differences of –0.2 ± 0.42 mm and –0.5 ± 0.71 mm, respectively (Fig. 9a and 9b). During axial-loaded MRIs, very good agreements were found between BS and standing MRIs at both the L4-L5 (ICC = 0.96) and L5-S1 (ICC = 0.918) levels (Fig. 8c and 8d), with mean differences of 0.15 ± 0.63 mm and –0.3 ± 0.67 mm, respectively (Fig. 9c and 9d). In comparison with standing MRIs, we found that BS-MRIs underestimated loss of disc height at L4-L5 (P = 0.03), while axial-loaded MRIs showed no difference (P = 0.24). In addition, both BS-MRIs (P = 0.002) and axial-loaded MRIs (P = 0.03) underestimated loss of disc height at L5-S1 (Fig. 5b).

Fig. 8.

Scatter diagrams show the values of DH in standing MRIs vs. those observed in BS-MRIs at L4-L5 (a) and L5-S1 (b), and the values of DH in standing MRIs vs. those observed in axial-loaded MRIs at L4- L5 (c) and L5-S1 (d) for each patient, and ICCs with 95% CI. BS, back support; CI, confidence interval; DH, height of intervertebral disc; ICC, intraclass correlation coefficient; MRI, magnetic resonance imaging.

Fig. 9.

Bland–Altman plots with data points for each patient. The y-axis shows the difference (diff) in the values of DH between standing and BS-MRIs at the L4-L5 (a) and L5-S1 (b) levels; and DH between standing and axial-loaded MRIs at the L4-L5 (c) and L5-S1 (d) levels. The x-axis shows the mean values of the DH in standing MRIs. The red solid line indicates the mean difference (value and percentage of the value expressed on the red solid line). The dashed lines indicate the limits of agreement. BS, back support; DH, height of intervertebral disc; MRI, magnetic resonance imaging.

Fig. 10 is an example of sagittal T1W images from a BS-MRI (Fig. 10a), standing (Fig. 10b), and axial-loaded (Fig. 10c) MRIs at the L4-L5 and L5-S1 levels of the same participant. DH was greater in BS-MRI (Fig. 10a) than that in both standing (Fig. 10b) and axial-loaded (Fig. 10c) MRIs at the two levels.

Fig. 10.

MRI examinations in a 47-year-old man with a one-year history of low back pain with neurogenic claudication in the left leg. Sagittal T1W images compare DH at L4-L5 and L5-S1 in a BS-MRI (a) against standing (b) and axial-loaded (c) MRI examinations. BS, back support; DH, height of intervertebral disc; MRI, magnetic resonance imaging T1W, T1-weighted.

Spine curvature

Lumbar spine curvature from BS-MRI compared to the other two methods is presented as LL and LA measured in sagittal images. The LL found in BS-MRIs was in moderate agreement (ICC = 0.712) with that found in standing MRIs (Fig. 11a), with a mean difference of –6.9° ± 6.94° (Fig. 12a). The LA observed in axial-loaded MRIs was in good agreement (ICC = 0.875) with that observed in standing MRIs (Fig. 11b), with a mean difference of 0.12° ± 3.07° (Fig. 12b). The LL found in axial-loaded MRIs was in good agreement (ICC = 0.777) with that found in standing MRIs (Fig. 11c), with a mean difference of 4.7° ± 7.47° (Fig. 12c). The LA observed in axial-loaded MRIs was in moderate agreement (ICC =0.65) with that observed in standing MRIs (Fig. 11d), with a mean difference of –4.4° ± 3.92° (Fig. 12d). In comparison with standing MRIs, BS-MRI overestimated LL (P < 0.001), while axial-loaded MRI underestimated LL (P = 0.005) (Fig. 5c). In addition, there was no difference of LA between BS and standing MRIs (P = 0.847), but axial-loaded MRI underestimated decreases in LA during loading (P < 0.001).

Fig. 11.

Scatter diagrams show the values of LL (a) and LA (b) in standing MRIs vs. those observed on BS-MRIs, and the values of LL (c) and LA (d) in standing MRIs vs. those observed on axial-loaded MRIs for each patient, and ICCs with 95% CI. BS, back support; CI, confidence interval; ICC, intraclass correlation coefficient; LA, lumbar angle; LL, lumbar lordosis; MRI, magnetic resonance imaging.

Fig. 12.

Bland–Altman plots with data points for each patient. The y-axis shows the difference (diff) in the values of LL (a) and LA (b) between standing MRI and BS-MRI, and LL (c) and LA (d) between standing and axial-loaded MRIs. The x-axis shows the mean values of the LL (a, c) and LA (b, d) in standing MRIs. The red solid line indicates the mean difference (value and percentage of the value expressed on the red solid line). The dashed lines indicate the limits of agreement. BS, back support; LA, lumbar angle; LL, lumbar lordosis; MRI, magnetic resonance imaging.

Fig. 13 is an example of sagittal T2W images from a BS-MRI (Fig. 13a) and standing (Fig. 13b) and axial-loaded (Fig. 13c) MRIs of the same individual. LL was larger in the BS-MRI (Fig. 13a) than in the standing (Fig. 13b) and axial-loaded (Fig. 13c) MRIs, but LA was smallest in the standing MRI. BS-MRIs generated C-shaped, lumbar spinal curvatures (Fig. 13a) that were virtually identical to those observed in the standing MRIs (Fig. 13b), while axial-loaded MRIs tended to decrease lumbar spine curvatures (Fig. 13c).

Fig. 13.

MRI examinations of a 47-year-old man with a one-year history of low back pain with neurogenic claudication in the left leg. Sagittal T2W images compare LL (angle between red solid lines) and LA (angle between yellow dashed lines) in a BS-MRI (a) vs. standing (b) and axial-loaded (c) MRI examinations. BS, back support; LA, lumbar angle; LL, lumbar lordosis; MRI, magnetic resonance imaging.

Discussion

It was found that the axial-loaded MRI with a cushion providing BS can be used to visualize morphological abnormalities of the dural sac, the lumbar spinal curvature, and the height of the IVD similar to those detected when using a standing MRI with some limitations. First, BS-MRI may overestimate narrowing of the dural sac at both the L4-L5 and L5-S1 levels, as well as underestimate loss of disc height at both of these levels. These findings disagree with those of Madsen et al. (13) who reported that the DCSA at L4-L5 observed with BS-MRIs (138 mm²) was increased (1.5%) when compared with that of standing MRIs (136 mm²). However, our results agree with the report of Madsen et al. (13), that DCSA observed with BS-MRIs (132 mm²) was decreased (9.6%) when compared with standing MRI (146 mm²) at the L5-S1 level. Second, although BS-MRIs provided good agreement of LA with those observed in standing MRIs, LL was overestimated (12.9%). This concurs with Madsen et al. (13) who reported that the extent of LL observed with BS-MRIs (49.6°) was increased (13%) compared with that in standing MRIs (43.9°). It is worth noting that there are a few limitations regarding the comparison of our results with those of Madsen et al. First, we do not know the exact height of the cushion used in their study. We are the first group to report an experimental study using a cushion BS during axial-loading. Second, Madsen et al. (13) performed vertical MRI examinations while patients were instructed to lean slightly backwards against the examination bench and to rest their arms on a cross bar (to secure immobility). Thus, participants may have assumed a more relaxed and flexed posture of the lumbar spine than normally occurs with standing. In contrast, the standing MRIs in our study were performed with 90° upright standing.

We do not recommend using a cushion BS when applying axial loading in the supine position. Although BS-MRIs demonstrate an image of the lumbar spine curvature (i.e. LA) which is closer to that when standing than axial-loaded MRIs, it is likely to overestimate both narrowing of the dural sac and extent of LL. In addition, it may underestimate loss of disc height when compared with that observed while standing. Supine axial-loaded MRIs without using a cushion BS are clinically useful for the diagnosis of spinal stenosis and loss of disc height while the BS-MRI method is not. We found that there was no significant difference between axial-loaded and standing MRIs in DCSA at L5-S1 and DH at L4-L5. However, axial-loading may overestimate narrowing of the dural sac at L4-L5 (6.3%) and slightly underestimate loss of disc height at L5-S1 (4.3%).

The limitation of this study should be mentioned. We used only a single size of cushion BS for all individuals. Customized cushion BS for each patient would theoretically be better for simulating spine curvature as seen in the upright position. However, the cushion used in the present study provided good correlation with L1-L3-L5 spine curvature.

In conclusion, supine axial-loaded MRIs using a cushion BS increased extension of the lumbar angle as in the standing posture, but overestimated narrowing of the dural sac and extent of LL, and underestimated loss of disc height. We recommend using a lumbosacral spinal compression device without using a cushion BS to best simulate the morphology of the lumbar spine in patients when performing supine MRI examinations.

Footnotes

Acknowledgments

The authors thank Dr. Chulaluk Komontri from the Faculty of Medicine, Siriraj Hospital, for statistical consultation and Dr. Arthur Brown, Research Consultant at the Faculty of Medical Technology, Mahidol University, for language editing.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received the following financial support for the research, authorship, and/or publication of this article: This study was supported by Mahidol University.

ORCID iD

Khaisang Chousangsuntorn

References

Ali

Saleh

Lumbar spine MRI axial loading in patients with degenerative spine pathologies: Its impact on the radiological findings and treatment decision. Egypt J Radiol Nucl Med 2015; 46:1065–1069.

Willén

Danielson

The diagnostic effect from axial loading of the lumbar spine during computed tomography and magnetic resonance imaging in patients with degenerative disorders. Spine 2001; 26:2607–2614.

Kanno

Endo

Ozawa

, et al. Axial loading during magnetic resonance imaging in patients with lumbar spinal canal stenosis: does it reproduce the positional change of the dural sac detected by upright myelography? Spine 2012; 37:E985–E992.

Lorenc

Palczewski

Wójcik

, et al. Diagnostic benefits of axial-loaded magnetic resonance imaging over recumbent magnetic resonance imaging in obese lower back pain patients. Spine 2018; 43:1146–1153.

Willén

Danielson

Gaulitz

, et al. Dynamic effects on the lumbar spinal canal: axially loaded CT-myelography and MRI in patients with sciatica and/or neurogenic claudication. Spine 1997; 22:2968–2976.

Kimura

Steinbach

Watenpaugh

, et al. Lumbar spine disc height and curvature responses to an axial load generated by a compression device compatible with magnetic resonance imaging. Spine 2001; 26:2596–2600.

Hioki

Miyamoto

Sakai

, et al. Lumbar axial loading device alters lumbar sagittal alignment differently from upright standing position: a computed tomography study. Spine 2010; 35:995–1001.

Wessberg

Danielson

Willén

Comparison of Cobb angles in idiopathic scoliosis on standing radiographs and supine axially loaded MRI. Spine 2006; 31:3039–3044.

Danielson

Willén

Axially loaded magnetic resonance image of the lumbar spine in asymptomatic individuals. Spine 2001; 26:2601–2606.

10.

Hansson

Suzuki

Hebelka

, et al. The narrowing of the lumbar spinal canal during loaded MRI: the effects of the disc and ligamentum flavum. Eur Spine J 2009; 18:679–686.

11.

Kanno

Aizawa

Ozawa

, et al. An increase in the degree of olisthesis during axial loading reduces the dural sac size and worsens clinical symptoms in patients with degenerative spondylolisthesis. Spine J 2018; 18:726–33.

12.

Kanno

Ozawa

Koizumi

, et al. Changes in lumbar spondylolisthesis on axial-loaded MRI: do they reproduce the positional changes in the degree of olisthesis observed on X-ray images in the standing position? Spine J 2015; 15:1255–1262.

13.

Madsen

Jensen

Pope

, et al. The effect of body position and axial load on spinal canal morphology: an MRI study of central spinal stenosis. Spine 2008; 33:61–67.