Abstract
Purpose
-
To report accuracy, repeatability, and agreement of Cobb angle measurements on radiographs and/or stereo-radiographs (EOS) compared against one another or against other imaging modalities.
Methods
-
This review follows Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) guidelines. A literature search was conducted on 21 July 2021 using Medline, Embase, and Cochrane. Two researchers independently performed title/abstract/full-text screening and data extraction. Studies were eligible if they reported Cobb angles, and/or their repeatability and agreement, measured on radiographs and/or EOS compared against one another or against other imaging modalities.
Results
-
Of the 2993 records identified, 845 were duplicates and 2212 were excluded during title/abstract/full-text screening. Two more relevant studies were identified from references of eligible studies, leaving 14 studies for inclusion. Two studies compared Cobb angles from EOS vs CT, while 12 compared radiographs vs other imaging modalities: EOS, CT, MRI, digital fluoroscopy, or dual-energy x-ray absorptiometry. Angles from standing radiographs tended to be higher than those from supine MRI and CT, and angles from standing EOS tended to be higher than those from supine or prone CT. Correlations across modalities were strong (R = 0.78–0.97). Inter-observer agreement was excellent for all studies (ICC = 0.77–1.00), except one (ICC = 0.13 radiographs and ICC = 0.68 for MRI).
Conclusion
-
Differences of up to 11º were found when comparing Cobb angles across combinations of imaging modalities and patient positions. It is not possible, however, to determine whether the differences observed are due to the change of modality, position, or both. Therefore, clinicians should be careful when utilizing the thresholds for standing radiographs across other modalities and positions for diagnosis and assessment of scoliosis.
Introduction
Scoliosis is a three-dimensional (3D) spinal deformity in the coronal, sagittal, and/or transverse planes (1). Nonetheless, in daily clinical practice, scoliosis is commonly assessed by measuring spinal curvature in the coronal plane only, by means of standing antero-posterior (AP) radiographs (2, 3, 4). The Cobb angle measured on two-dimensional (2D) AP radiographs in the standing position is used to quantify scoliosis in the coronal plane, usually with angles above 10° indicating mild scoliosis, and angles above 30° or 40° indicating severe scoliosis (4, 5, 6).
With technological advancements, stereo-radiography (EOS) is increasingly used for the diagnosis of scoliosis because it enables the assessment of spinal deformities in 3D in the standing position with minimal exposure to radiation (7, 8). The authors’ clinics, as most spine deformity centers worldwide, are equipped with low-dose EOS machines that expose patients to less radiation than conventional radiographs (9, 10), therefore having important safety implications for patients who require frequent monitoring (11). However, it is not clear whether Cobb angles measured using stereo-radiographs are equivalent to those measured on plain radiographs or other imaging modalities. Moreover, even though imaging modalities that assess spine deformity during standing are the reference standard, knowing the differences in Cobb angles between standing and supine positions is relevant when assessing curve flexibility, which can influence decision-making regarding treatment.
There is no systematic review to date that describes the imaging modalities available to diagnose scoliosis and that synthesizes their repeatability and agreement. Therefore, the purpose of this systematic review is to report on the accuracy, repeatability, and agreement of Cobb angles measured on radiographs and/or EOS compared against one another or against other imaging modalities. These findings could help ascertain thresholds for assessment of scoliosis to each imaging modality and patient position and hence enable clinicians to adjust their diagnostic criteria depending on the technology and procedures available at their centers.
Materials and methods
The protocol for this systematic review was submitted to PROSPERO prior to commencement (registration number CRD42021264063) and follows the guidelines established by the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) (12). The original PROSPERO registration stated that the purpose would be to report on the repeatability and accuracy of Cobb angles measured across different imaging modalities and/or different measurement techniques. Due to the large number of studies reporting on different measurement techniques, the purpose of this systematic review was altered compared to that submitted to PROSPERO, to only include studies reporting on plain radiographs and/or EOS compared against one another or against other imaging modalities.
Search strategy
The authors conducted a structured electronic literature search on 21 July 2021 using the Medline, Embase, and Cochrane Central Register of Controlled Trials databases, applying the keywords and medical subject heading (MeSH) terms presented in the Supplementary file (see section on supplementary materials given at the end of this article). After the removal of duplicate records, two researchers (SB, SRP) each screened the titles and abstracts to determine the suitability for the review against predefined eligibility criteria:
Inclusion criteria
-
Clinical studies reporting Cobb angles, as well as the repeatability and agreement of Cobb angles, measured on radiographs and/or EOS compared against one another or against other imaging modalities, where the population is patients assessed for scoliosis.
Exclusion criteria
-
Studies that did not report on live human participants (e.g. cadavers, phantoms, or animals).
-
Studies that reported on computer simulations.
-
Studies that compared radiographs or EOS against surface imaging modalities, such as Moire topography, photogrammetry, raster-stereography, and pictures of the spine using smartphones.
-
Studies that did not report on the Cobb angle but instead measured other coronal angles to assess deformity, such as the spinous process angle or the transverse process angle.
-
Studies that did not report on coronal Cobb angle but instead measured sagittal Cobb angle (lumbar lordosis or thoracic kyphosis).
-
Case reports, reviews, editorials, expert opinions, or conference proceedings.
-
Studies written in languages other than English, French, or Spanish to avoid translation errors.
Study selection
Full-text review of studies meeting the criteria in the initial screening was carried out by two researchers (SB, SRP), and any disagreement about the final eligibility of studies was resolved by review and consensus. Subject experts (HdA, MS, JCLH) were consulted to further establish relevant studies not captured by the database searches.
Data extraction and quality assessment
Data extraction was performed by two researchers (SB, SRP) independently, and their results were compared to ensure accuracy. Where there was disagreement in the documented value, the true value was ascertained by a simultaneous review of the data in question by both researchers. The following data were extracted from the included studies: author(s), journal, year of publication, ethical approval, and conflicts of interest. Cohort characteristics were retrieved, including the type of cohort (e.g. scoliotic and volunteers), number of subjects, sex, age, and body mass index (BMI). Furthermore, the following were also retrieved: imaging modalities used, time between imaging modalities, subject position, number of observers, number of measurements per observer, type of curves measured (e.g. main and secondary), Cobb angle for each combination of modality and position, difference in Cobb angle between these combinations, Cobb angle thresholds, inter- and intra- observer repeatability, correlation coefficients, and linear regression analyses. The methodological quality of the eligible studies was assessed by two researchers (SB, SRP) according to the QUADAS evidence-based tool for diagnostic accuracy studies (13). Where there was disagreement between the researchers, consensus was achieved by discussion and review, or by involving a third researcher (HdA).
Data analysis
When available in the original articles, outcomes were tabulated. Continuous outcomes were reported as means, standard deviations, and ranges. Categorical outcomes were reported as proportions. Cobb angles were presented in graphical format when possible. Observer agreement was interpreted according to Cicchetti et al. (14) - poor: 0–0.40; fair: 0.40–0.59; good: 0.60–0.74 excellent: 0.75–1.00. When authors presented measurements for different types of curves (all, main, secondary, thoracic, and lumbar), only the main curve was considered, while other curve types are presented in the Supplementary file. Similarly, when measurements were reported for an overall cohort and subgroups, only the overall cohort was considered, while the subgroup data were presented in the Supplementary file. Furthermore, when intra-observer repeatability was presented for more than one observer, only the data for the first observer were considered, and any additional data are presented in the Supplementary file.
Results
The systematic search returned 2993 records, of which 845 were duplicates, leaving 2148 records for screening. A total of 2112 records were excluded by examining their titles and/or abstracts, leaving 36 records for full-text review. Twenty studies (15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34) were excluded after full-text review because they did not report coronal Cobb angles, two studies (35, 36) were excluded because they were not comparative, one study (37) was excluded because it reported on cadavers, and one study (38) was excluded because it reported on a surface imaging modality (ultrasound) combined with radiographs. This left 12 eligible studies (7, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49), and by checking their references, 2 more relevant studies (50, 51) were identified. No further studies were identified from discussion with subject experts (HdA, MS, JCLH). Therefore, 14 studies were included in this systematic review (Table 1, Fig. 1). Of the 14 studies, nine had low overall risk of bias, three had some concerns regarding the overall risk of bias, and two had high overall risk of bias (Fig. 2).
Flowchart describing inclusion and exclusion of articles for the systematic review.
Citation: EFORT Open Reviews 8, 6; 10.1530/EOR-23-0032
Methodological quality of the eligible studies was assessed independently by two researchers according to the QUADAS evidence-based tool for diagnostic accuracy studies.
Citation: EFORT Open Reviews 8, 6; 10.1530/EOR-23-0032
Information of the comparative studies included in the systematic review.
| Reference | Study design | Ethical approval | COI | Funding | Country | Patient condition | Imaging modality 1 | Imaging modality 2 | Time between imaging modalities | ||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Modality | Position | Modality | Position | ||||||||
| Al-Aubaidi et al. (47) | R | Yes | None | NA | NA | Scoliosis | EOS | Standing | CT | Supine | <6 months (39 ± 36.5 days) |
| Brink et al. (49) | R | NR | None | Yesa | Netherlands | AIS | X-ray | Standing | CT | Prone | 81 ± 51 days |
| X-ray | Standing | CT | Prone | 3 ± 17 days | |||||||
| Diefenbach et al. (46) | P | Yes | None | None | USA | AIS | X-ray (PF) | Standing | MRI (dynamic) | Standing | <1 week |
| Geijer et al. (51) | NA | Yes | NA | NA | Sweden | NS | X-ray (D) | Standing | Digital fluoroscopy | Standing | NA |
| Hirsch et al. (7) | P | Yes | None | NA | France | AIS | X-ray (C) | SLD | EOS | SLD | NA |
| Lechner et al. (41) | R | NR | None | Yesb | Austria | Scoliosis (IP or NP) | X-ray (PF) | Supine | CT (3D patient plane) | Supine | NA |
| Lee et al. (45) | R | Yes | None | None | USA | AIS | X-ray | Standing | MRI | Supine | <6 months (1 ± 1) |
| Schell et al. (39) | R | Yes | Yes | Yesc | USA | With and without scoliosis | X-ray | Standing | DEXA | Supine | <6 months |
| Shi et al. (44) | R | Yes | None | None | China | AIS | X-ray | Standing | MRI | Supine | <6 months (3.2 ± 2.0) |
| Tauchi et al. (43) | R | Yes | None | None | NA | Congenital scoliosis | X-ray (PF) | Supine | CT (3D) | Supine | NA |
| Vavruch & Tropp (42) | RO | Yes | None | None | Sweden | AIS (requiring surgery) | X-ray (PF) | Standing | CT (2D. scout view) | Supine | <1 h |
| Wessberg et al. (48) | P | NA | Yes | Yesd | Sweden | AIS (brace treatment) | X-ray | Standing | MRI | Supine | <2 weeks |
| Yazici et al. (50) | A | NA | NA | NA | Turkey | AIS (requiring surgery) | X-ray | Standing | CT (2D. scout view) | Supine | NA |
| Yeung et al. (40) | P | Yes | None | Yese | Hong kong | AIS (requiring surgery) | EOS (3D patient plane) | Standing | CT (3D patient plane) | Prone | Same day |
A, ambispective; AIS, adolescent idiopathic scoliosis; C, cassette; COI, conflicts of interest; D, digital; DEXA, dual-energy x-ray absorptiometry; EOS, stereo-radiography; IP, idiopathic; NA, not available; NP, neuropathic; NR, not required; P, prospective; PF, plain film; R, retrospective; RO, retrospective observational; SLD, supine lateral bending.
aBrink received funding from the Alexandre Suerman MD/PhD program, and Castelein from a Medtronic research grant and a K2M research grant; bFunded with internal funds from the research institution. No payments were received for this study; cSupported by the Clinical and Translational Science Award (CTSA) program, through the NIH National Center for Advancing Translational Sciences (NCATS), grant UL1TR002373; dSupported by the Bertha Neuberg Research Fund and the Wallenberg Foundation; eGrants from the Research Grants Council of Hong Kong Special Administrative Region, China (Project no. CUHK 411811 and 14206716) and Fondation Yves Cotrel de l’Institut de France (W.C.W.C.).
Study characteristics
The 14 included studies reported on a combined total of 646 individuals (Table 2), of which 11 studies (7, 39, 40, 41, 42, 44, 46, 47, 48, 49, 50) reported mean patient age, which ranged from 13.6 to 18.4 years in 10 studies but was 69.2 years in 1 study (39). Twelve studies (7, 39, 40, 42, 43, 44, 45, 46, 47, 48, 49, 50) reported on sex distribution, with the percentage of males ranging between 0 and 55%. Only two of the studies (40, 45) reported on BMI, with mean values of 19.5 and 23.0 kg/m2. Nine studies (7, 40, 42, 44, 45, 46, 48, 49, 50) reported on individuals with adolescent idiopathic scoliosis (AIS), one (41) reported on individuals with idiopathic or neuropathic scoliosis, one (43) reported on individuals with congenital scoliosis, one (47) reported on individuals with scoliosis, one (39) reported on individuals with and without scoliosis, and lastly one (51) did not specify patient conditions.
Characteristics of the individuals participating in the studies included in the systematic review.
| Reference | Cohort size | Age | Males, n (%) | |
|---|---|---|---|---|
| Mean ± s.d. | Range | |||
| Al-Aubaidi et al. (47) | 7 | 15.4 (±4.1)* | 10.0–18.0 | 3 (43) |
| Brink et al. (49) | 62 | 15.6 ± 2.5 | 10.0–23.0 | 56 (90) |
| 62 | 15.6 ± 2.5 | 10.0–23.0 | 56 (90) | |
| Diefenbach et al. (46) | 25 | 14.6 (±1.6)* | 12.0–18.0 | 9 (36) |
| Geijer et al. (51) | 10 | |||
| Hirsch et al. (7) | 50 | 15.6 ± 1.9 | 10 (20) | |
| Lechner et al. (41) | 55 | |||
| Lee et al. (45) | 70 | 14.0 ± 2.0 | 10.0–18.0 | 30 (43) |
| Schell et al. (39) | 51 | 69.2 ± 10.3 | 4 (8) | |
| Shi et al. (44) | 80 | 14.4 ± 1.8 | 10.5–17.1 | 0 (0) |
| Tauchi et al. (43) | 20 | 11 (55) | ||
| Vavruch & Tropp (42) | 128 | 15.5 | 11.0–26.0 | 31 (24) |
| Wessberg et al. (48) | 30 | 13.6 | 12.0–16.0 | 1 (3) |
| Yazici et al. (50) | 25 | 14.2 | 10.0–16.0 | 8 (32) |
| Yeung et al. (40) | 33 | 18.4 ± 4.2 | 13.0–31.0 | 7 (21) |
*Values in parentheses have been calculated by the authors of the present study.
Of the 14 studies, only three compared Cobb angles in the same position using different imaging modalities: standing radiographs vs standing dynamic MRI (46), standing radiographs vs standing digital fluoroscopy (51), and supine radiographs vs supine CT (41). The remaining 10 studies compared Cobb angles across imaging modalities in different positions: one (7) study compared supine lateral bending radiographs vs standing lateral bending EOS, three (42, 43, 50) studies compared standing radiographs vs supine CT, three (44, 45, 48) studies compared standing radiographs vs supine MRI, two (40, 47) studies compared standing EOS vs supine or prone CT, (7) one (39) study compared standing radiographs vs supine dual-energy x-ray absorptiometry (DEXA), and, finally, one (49) study compared standing radiographs vs supine MRI vs prone CT.
Cobb angles
Of the 14 studies, 11 (7, 39, 40, 41, 42, 44, 45, 47, 48, 50, 51) compared Cobb angles across two imaging modalities (Table 3, Figs. 3, 4 and 5), one (49) compared Cobb angles across three imaging modalities, and six (39, 42, 45, 48, 50, 51) reported the mean difference in Cobb angle between two imaging modalities. Of the studies that compared Cobb angles in the same position using different imaging modalities: one study (51) found similar angles using standing radiographs and standing digital fluoroscopy (14.8 ± 0.9 vs 14.4 ± 0.9, P = 0.64), while one study (41) found angles from supine radiographs to be 9° lower than those from supine CT (63.6 ± 32.0 vs 72.8 ± 28.5, P < 0.001). Of the studies that compared Cobb angles in different positions using different imaging modalities: one study (7) found similar angles using supine lateral bending radiographs and standing lateral bending EOS (23.5 ± 15.5 vs 22.7 ± 17.0, P > 0.05) (41, 42, 49, 50); four studies (44, 45, 48, 49) found angles from standing radiographs to be >7° higher than those from supine MRI (68.2 ± 15.4 vs 56.7 ± 13.5, P < 0.001; 58.0 ± 14.0 vs 48.0 ± 14.0, no P-value; 29.4 ± 13.7 vs 22.4 ± 14.3, P < 0.001; 30.8 ± 5.5 vs 23.1±5.3, no P-value), though differences across modalities disappeared when the supine MRI was loaded (30.8 ± 5.5 vs 30.5 ± 6.3, no P-value) (48); three studies (42, 49, 50) found angles from standing radiographs to be >10° higher than those from supine or prone CT (68.2 ± 15.4 vs 53.9 ± 14.8, P < 0.001; 59.2 ± 11.5 vs 48.1 ± 11.7, no P-value; 55.7 ± 16.8 vs 39.4 ± 10.8, no P-value); one study (39) found angles from standing radiographs to be nearly 3° higher than those from supine DEXA (10.9 ± 9.5 vs 8.1 ± 7.3, no P -value); two studies (40, 47) found angles from standing EOS to be >5° higher than those from supine or prone CT (32.0 ± 15.8 vs 27.0 ± 12.3, P = 0.56; 62.9 ± 9.3 vs 47.3 ± 1 0.0, P < 0.001). Some of the authors also presented Cobb angles for scoliotic and non-scoliotic subgroups, as well as for different curve types (Appendix).
Cobb angle measurements for studies comparing radiographs vs CT, presented as mean ± s.d.
Citation: EFORT Open Reviews 8, 6; 10.1530/EOR-23-0032
Cobb angle measurements for studies comparing radiographs vs MRI, presented as mean ± s.d.
Citation: EFORT Open Reviews 8, 6; 10.1530/EOR-23-0032
Cobb angle measurements for studies comparing other imaging modalities, presented as mean ± s.d.
Citation: EFORT Open Reviews 8, 6; 10.1530/EOR-23-0032
Cobb angle measurements from the studies included in the systematic review.
| Reference | Curve measured | Cobb angle: modality 1 | Cobb angle: modality 2 | Difference | Linear regression analysis | ||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Mean ± s.d. | Range | Mean ± s.d. | Range | Mean ± s.d. | P-value | R | 95% CI | P-value | |||
| Al-Aubaidi et al. (47) | Main | 32.0 (±15.8)† | 15–78 | 27.0 (±12.3)† | 16–65 | 0.56 | 0.95–0.98 | <0.001 | |||
| Brink et al. (49) | Main | 68.2 ± 15.4 | 56.7 ± 13.5 | <0.001 | 0.97 | 0.95–0.98 | <0.001 | ||||
| Main | 68.2 ± 15.4 | 53.9 ± 14.8 | <0.001 | 0.97 | 0.01 | X-ray = 1.01 × MRI + 10.8 | |||||
| Diefenbach et al. (46) | Main | 0.90 | X-ray = 0.96 × CT + 16.6 | ||||||||
| Geijer et al. (51) | Pre-selected levels | 14.8 ± 0.9 | 14.4 ± 0.9 | 0.64 | |||||||
| Hirsch et al. (7) | Main | 23.5 ± 15.5 | 22.7 ± 17.0 | >0.05 | |||||||
| Lechner et al. (41) | Main | 63.6 (±32.0)† | 72.8 (±28.5)† | <0.001 | |||||||
| Lee et al. (45) | All curves | 24–152 | 30–144 | 0.5 | |||||||
| Main | 58.0 ± 14.0 | 48.0 ± 14.0 | |||||||||
| Schell et al. (39) | L1–L4 | 10.9 ± 9.5 | 8.1 ± 7.3 | 9.2 | 0.94 | 0.80–0.93 | X-ray = 1.05 × MRI + 5.51 | ||||
| Shi et al. (44) | Structural and non-structural | 29.4 ± 13.7 | 22.4 ± 14.3 | 3–74 | 10.0 | 0.88 | 0.96 | ||||
| Vavruch & Tropp (42) | Main | 59.2 ± 11.5 | 10–79 | 48.1 ± 11.7 | 23–94 | 2.8 | 0.91 | <0.01 | |||
| Wessberg et al. (48) | All curves | 41–116 | 0.90 | X-ray = 0.87 × MRI + 9.93 | |||||||
| Main | 30.8 (±5.5)† | 23–45 | 30.5 (±6.3)† | 16–41 | 11.1 ± 5.2 | 0.78 | CT = 0.91 × x-ray − 5.96 | ||||
| Main | 30.8 (±5.5)† | 23–45 | 23.1 (±5.3)† | 12–33 | −0.6 ± 4.9 | ||||||
| Yazici et al. (50) | Structural | 55.7 (±16.8)† | 35–102 | 39.4 (±10.8)† | 18–61 | 0.0 ± 5.2 | <0.001 | ||||
| Yeung et al. (40) | Main | 62.9 ± 9.3 | 47.3 ± 10.0 | 30%a | 0.86 | CT = 0.93 × EOS−11.47 | |||||
Notes on Cobb angle measurements: Lechner et al. selected end vertebrae by the senior author on x-ray and CT; Lee et al. selected the same end vertebrae for x-ray and MRI; Shi et al. selected end vertebrae on x-ray for MRI; Wessberg et al. selected end vertebrae on x-ray for MRI.
†Values in parentheses have been calculated by the authors of the present study; aScoliotic curves were on average 29.8% (0–55°) greater on x-ray compared to CT.
Correlations between modalities and repeatability of Cobb angles
Of the 14 studies, eight (39, 40, 41, 43, 44, 45, 46, 49) calculated correlation coefficients of Cobb angles between modalities, as well as intra- and/or inter-observer agreement. All eight studies reported strong correlation coefficients (R > 0.75) between modalities (Table 4). Five studies reported excellent inter-observer agreement for various imaging modalities: three (41, 43, 49) for standing radiographs (ICC > 0.85) and supine or prone CT (ICC > 0.89), two (46, 49) for standing radiographs (ICC = 0.96) and standing or supine MRI (ICC = 0.90), one (39) for standing radiographs (ICC = 0.975) and supine DEXA (ICC = 0.962), and one (40) for standing EOS (ICC = 0.99) and prone CT (ICC = 1.00). One study (45) reported excellent inter-observer agreement for radiographs (ICC = 0.77) and good agreement for MRI (ICC = 0.70). Lastly, one study (44) reported poor-to-good inter-observer agreement for radiographs (ICC = 0.13–0.68) and fair-to-good agreement for MRI (ICC = 0.47–0.70). For studies that reported more than one coefficient of intra-observer agreement, the coefficients tended to be consistent across observers (Appendix).
Repeatability in Cobb angle measurements for the studies included in the systematic review.
| Reference | Observers, n | Measurements, n | Time between tests | Model | Definition | Data presented | Cobb angle for modality 1 | Cobb angle for modality 2 | Intra-observer | Inter-observer | |||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| A | B | ICC | 95% CI | ICC | 95% CI | ICC | 95% CI | ICC | 95% CI | ||||||
| Brink et al. (49) | 2 | 10 | 10 | NA | 2-way mixed | NA | NA | 0.99 | 0.97–1.00 | 0.97 | 0.89–0.99 | 1.00 | 0.98–1.00 | 0.97 | 0.90–0.99 |
| 2 | 10 | 10 | NA | 2-way mixed | NA | NA | 0.99 | 0.97–1.00 | 0.97 | 0.89–0.99 | 1.00 | 0.99–1.00 | 1.00 | 0.98–1.00 | |
| Diefenbach et al. (46) | 2 | 25 | 25 | NA | 2-way | Consistency | Observational | 0.97 | 0.96 | 0.97 | 0.90 | ||||
| Lechner et al. (41) | 5 | 55 | 55 | 3 weeks | 2-way mixed | Agreement | Observational | 0.98 | 0.97–0.99 | 0.96 | 0.93–0.97 | 0.98 | 0.96–0.99 | 0.94 | 0.92–0.96 |
| Lee et al. (45) | 2 | 20 | 20 | 2 weeks | NA | NA | NA | 0.90 | 0.77 | 0.88 | 0.70 | ||||
| Schell et al. (39) | NA | NA | NA | NA | NA | NA | NA | 0.97 | 0.95–0.98 | 0.98 | 0.96–0.99 | 0.97 | 0.94–0.98 | 0.96 | 0.94–0.98 |
| Shi et al. (44) | 2 | 27 | 27 | 2 weeks | NA | NA | Mean of observers | 0.04 | 0.13 | 0.68 | 0.47 | ||||
| 0.66 | 0.64 | 0.68 | 0.66 | ||||||||||||
| 0.84 | 0.68 | 0.88 | 0.70 | ||||||||||||
| Tauchi et al. (43) | 13 | 20 | 20 | 1 week | NA | NA | NA | 0.94 | 0.85 | 0.74–0.93 | 0.96 | 0.89 | 0.81–0.95 | ||
| Yeung et al. (40) | 2 | 33 | 33 | NA | NA | NA | NA | 0.99 | 0.93–1.00 | 0.99 | 0.87–1.00 | 1.00 | 0.96–1.00 | 0.98 | 0.82–1.00 |
Appendix: Measurements presented for different curve types (secondary and thoracic), different subgroups (scolioitic and non-scoliotic), and different observers.
Thresholds used to define scoliosis
Of the 14 studies, only three (39, 44, 48) reported on Cobb angle thresholds used for defining scoliosis. Schell et al. (39) defined scoliosis as a Cobb angle ≥10° when comparing radiographs vs DEXA, while Wessberg et al. (48) defined scoliosis as >20° when comparing radiographs vs MRI. Furthermore, Shi et al. (44) defined moderate scoliosis as >20° and severe scoliosis as >40° when comparing radiographs vs MRI.
Discussion
This systematic review identified 14 relevant studies that compared Cobb angles measured on radiographs and/or EOS, against one another or against other imaging modalities. Only two studies directly compared Cobb angle measurements in the same position using different imaging modalities, of which one found similar angles using standing radiographs and standing digital fluoroscopy (51) and the other found angles from supine radiographs to be 9° lower than those from supine CT (41). For studies comparing Cobb angles across different imaging modalities and patient positions, four studies found them to be 7.0–11.5° higher on standing radiographs compared to supine MRI (44, 45, 48, 49), three studies found them to be 11.1–16.3° higher on standing radiographs compared to supine/prone CT (42, 49, 50), and two studies found them to be 5.0–15.6° higher on standing EOS compared to supine/prone CT (40, 47). It is not possible, however, to determine whether the differences observed are due to the change of modality, or position, or both. Therefore, clinicians should be careful when utilizing the thresholds for standing radiographs across other modalities and positions for diagnosis and assessment of scoliosis.
Surprisingly, only one study compared Cobb angles measured on radiographs vs EOS, albeit in different positions, and found similar angles across modalities (respectively, 23.5 ± 15.5 vs 22.7 ± 17.0, P > 0.05) (7); radiographs were acquired in supine lateral bending, while EOS images were acquired in standing lateral bending. It is worth noting that a previous study (52) that compared Cobb angles from radiographs in standing vs supine positions found a statistically significant difference between measurements (respectively, 26±12° vs 21±12°, P < 0.01), as gravity exacerbated the deformity. In contrast, another study (53) that compared segmental Cobb angles from MRI in standing vs supine positions found that the L5-S1 Cobb angle significantly decreased in standing (P < 0.05), while there were no significant differences at L1–L2, L2–L3, and L3–L4 (P > 0.05).
With technological advancements, EOS and other EOS machines are most commonly used in spine deformity centers worldwide for the diagnosis of scoliosis, because they enable the assessment of spinal deformities in 3D in the straight standing position with minimal exposure to radiation while also measuring spino-pelvic parameters. However, since none of the studies in this systematic review compared Cobb angles in a straight standing position, from radiographs vs EOS, it is not possible to determine whether angles from these two imaging modalities are equivalent or not.
Differences in Cobb angles across modalities varied per study, though as expected, angles from standing radiographs were higher than those from supine MRI and CT, and similarly, angles from standing EOS tended to be higher than those from supine or prone CT. However, there was one exception to this, Wessberg et al. (48) found no differences in Cobb angle when comparing standing radiographs and supine loaded MRI. Interestingly, when comparing Cobb angles across different imaging modalities in the same position, Lechner et al. found that Cobb angles on supine radiographs were on average 9.2° smaller than on supine CT and concluded that this was because CT was performed in 3D while radiographs in 2D. Geijer et al. (51) found no differences in Cobb angle between standing radiographs vs standing digital fluoroscopy and concluded that digital fluoroscopy had the added advantage of lower radiation.
Studies that measured the correlation between modalities always found them to be strong (R = 0.78–0.97), though it is worth noting that a strong correlation does not necessarily imply that measurements are equal. Furthermore, inter-observer agreement was excellent for all studies (ICC = 0.77–1.00), with the exception of Shi et al. (44), who reported coefficients of 0.13 and 0.47 for patients with mild scoliosis (Cobb angle: < 20°) on radiographs and MRI, respectively. This implies that inter-observer agreement is low in patients with little-to-no scoliosis, where selecting end vertebrae is more challenging.
Interestingly, only three of the 14 included studies reported on Cobb angle thresholds used for defining scoliosis, with one study defining scoliosis as a Cobb angle ≥ 10°, another as a Cobb angle > 20°, and the last one defining moderate scoliosis as a Cobb angle > 20° and severe scoliosis as a Cobb angle > 40°. It is therefore possible that the thresholds for assessment of scoliosis specific to each imaging modality may need to be adjusted to avoid misdiagnosis due to inappropriate assumptions and definitions.
This systematic review has a number of limitations. First, there is a low number of published studies comparing Cobb angle measurements across imaging modalities; only seven studies reported mean Cobb angle measurement differences across modalities, and only two of these reported mean and s.d. values; therefore a meta-analysis could not be performed. Second, there was some heterogeneity across the included studies in terms of the patient condition, with some studies not specifying the type of scoliosis, other studies including patients with and without scoliosis, and other studies not at all specifying the patient condition. Third, one of the 14 studies included patients without scoliosis, and the repeatability of Cobb angle measurements in non-scoliotic patients is difficult, since there is no deformity, and thus no obvious upper and lower vertebrae. Fourth, for certain studies, the reader was not blinded when measuring the Cobb angle, and the end vertebrae were selected beforehand, or by the senior author, or on radiographs for measuring the Cobb angle on MRI or CT. Finally, in terms of the quality of the included studies, two of the 14 studies had a high overall risk of bias, while three had some concerns regarding the overall risk of bias.
Conclusions
This systematic review identified 14 relevant studies that compared Cobb angles measured on radiographs and/or EOS against one another or against other imaging modalities and found differences of up to 11° when comparing Cobb angles across combinations of imaging modalities and patient positions. It is not possible, however, to determine whether the differences observed are due to the change of modality, or position, or both. Therefore, clinicians should be careful when utilizing the thresholds for standing radiographs across other modalities and positions for diagnosis and assessment of scoliosis.
Supplementary materials
This is linked to the online version of the paper at https://doi.org/10.1530/EOR-23-0032.
ICMJE conflict of interest statement
HdA: consultancy fees and royalties from Clariance. SB: no conflicts of interest. SRP: no conflicts of interest. MS: consultancy fees and royalties from Clariance, and consultancy fees from Zimmer. JCLH: consultancy fees from Medtronic and fees for presentations and/or educational events from BD Bard.
Funding
This work was supported by ‘GCS Ramsay Santé pour l’Enseignement et la Recherche’, which provided funding for data collection and manuscript preparation.
References
- 1.↑
Stokes IA. Three-dimensional terminology of spinal deformity. A report presented to the Scoliosis Research Society by the Scoliosis Research Society Working Group on 3-D terminology of spinal deformity. Spine (Phila Pa 1976) 1994 19 236–248.
- 2.↑
Bettany-Saltikov J, Weiss HR, Chockalingam N, Taranu R, Srinivas S, Hogg J, Whittaker V, Kalyan RV, & Arnell T. Surgical versus non-surgical interventions in people with adolescent idiopathic scoliosis. Cochrane Database of Systematic Reviews 2015 2015 CD010663. (https://doi.org/10.1002/14651858.CD010663.pub2)
- 3.↑
Dunn J, Henrikson NB, Morrison CC, Blasi PR, Nguyen M, & Lin JS. Screening for adolescent idiopathic scoliosis: evidence report and systematic review for the US Preventive Services Task Force. JAMA 2018 319 173–187. (https://doi.org/10.1001/jama.2017.11669)
- 4.↑
Asher MA, & Burton DC. Adolescent idiopathic scoliosis: natural history and long term treatment effects. Scoliosis 2006 1 2. (https://doi.org/10.1186/1748-7161-1-2)
- 5.↑
Reamy BV, & Slakey JB. Adolescent idiopathic scoliosis: review and current concepts. American Family Physician 2001 64 111–116.
- 6.↑
Kane WJ. Scoliosis prevalence: a call for a statement of terms. Clinical Orthopaedics and Related Research 1977 126 43–46. (https://doi.org/10.1097/00003086-197707000-00006)
- 7.↑
Hirsch C, Ilharreborde B, & Mazda K. Flexibility analysis in adolescent idiopathic scoliosis on side-bending images using the EOS imaging system. Orthopaedics and Traumatology, Surgery and Research 2016 102 495–500. (https://doi.org/10.1016/j.otsr.2016.01.021)
- 8.↑
Ilharreborde B, Steffen JS, Nectoux E, Vital JM, Mazda K, Skalli W, & Obeid I. Angle measurement reproducibility using EOS three-dimensional reconstructions in adolescent idiopathic scoliosis treated by posterior instrumentation. Spine (Phila Pa 1976) 2011 36 E1306–E1313. (https://doi.org/10.1097/BRS.0b013e3182293548)
- 9.↑
Hamzian N, Afereydoon S, Ghorbani M, Abrisham SMJ, Roozmand Z, Abdollahi-Dehkordi S, Sepehr Javan M, & Reza Deevband M. Equivalent dose and risk of exposure induced cancer death of different organs due to various image techniques of EOS imaging system. Journal of Biomedical Physics and Engineering 2021 11 289–296. (https://doi.org/10.31661/jbpe.v0i0.2012-1242)
- 10.↑
Hui SC, Pialasse JP, Wong JY, Lam TP, Ng BK, Cheng JC, & Chu WC. Radiation dose of digital radiography (DR) versus micro-dose x-ray (EOS) on patients with adolescent idiopathic scoliosis: 2016 SOSORT- IRSSD "John Sevastic Award" Winner in Imaging Research. Scoliosis and Spinal Disorders 2016 11 46. (https://doi.org/10.1186/s13013-016-0106-7)
- 11.↑
Doody MM, Lonstein JE, Stovall M, Hacker DG, Luckyanov N, & Land CE. Breast cancer mortality after diagnostic radiography: findings from the U.S. Scoliosis Cohort Study. Spine (Phila Pa 1976) 2000 25 2052–2063. (https://doi.org/10.1097/00007632-200008150-00009)
- 12.↑
McInnes MDF, Moher D, Thombs BD, McGrath TA, Bossuyt PM, and the PRISMA-DTA Group Clifford T, Cohen JF, Deeks JJ, Gatsonis C, et al.Preferred reporting items for a systematic review and meta-analysis of diagnostic test accuracy studies: the PRISMA-DTA statement. JAMA 2018 319 388–396. (https://doi.org/10.1001/jama.2017.19163)
- 13.↑
Sterne JA, Hernán MA, Reeves BC, Savović J, Berkman ND, Viswanathan M, Henry D, Altman DG, Ansari MT, Boutron I, et al.Robins-I: a tool for assessing risk of bias in non-randomised studies of interventions. BMJ 2016 355 i4919. (https://doi.org/10.1136/bmj.i4919)
- 14.↑
Cicchetti DV, Showalter D, & Rosenheck R. A new method for assessing interexaminer agreement when multiple ratings are made on a single subject: applications to the assessment of neuropsychiatric symtomatology. Psychiatry Research 1997 72 51–63. (https://doi.org/10.1016/s0165-1781(9700095-4)
- 15.↑
Pang H, Wong YS, Yip BH, Hung AL, Chu WC, Lai KK, Zheng Y, Chung TW, Sharma G, Cheng JC, et al.Using ultrasound to screen for scoliosis to reduce unnecessary radiographic radiation: A prospective diagnostic accuracy study on 442 schoolchildren. Ultrasound in Medicine and Biology 2021 47 2598–2607. (https://doi.org/10.1016/j.ultrasmedbio.2021.05.020)
- 16.↑
Lee TT-Y, Lai KK-L, Cheng JC-Y, Castelein RM, Lam TP, & Zheng YP. 3D ultrasound imaging provides reliable angle measurement with validity comparable to X-ray in patients with adolescent idiopathic scoliosis. Journal of Orthopaedic Translation 2021 29 51–59. (https://doi.org/10.1016/j.jot.2021.04.007)
- 17.↑
Lee TT, Lai KK, Cheng JC, Castelein RM, Lam TP, & Zheng YP. Investigation of the phenomenon of coronal-sagittal curvature coupling on curve progression: an exploratory study using 3-D ultrasound. Ultrasound in Medicine and Biology 2021 47 2202–2212. (https://doi.org/10.1016/j.ultrasmedbio.2021.03.037)
- 18.↑
de Reuver S, Brink RC, Lee TTY, Zheng YP, Beek FJA, & Castelein RM. Cross-validation of ultrasound imaging in adolescent idiopathic scoliosis. European Spine Journal 2021 30 628–633. (https://doi.org/10.1007/s00586-020-06652-9)
- 19.↑
Lv P, Chen J, Dong L, Wang L, Deng Y, Li K, Huang X, & Zhang C. Evaluation of scoliosis with a commercially available ultrasound system. Journal of Ultrasound in Medicine 2020 39 29–36. (https://doi.org/10.1002/jum.15068)
- 20.↑
Główka P, Politarczyk W, Janusz P, Woźniak Ł, & Kotwicki T. The method for measurement of the three-dimensional scoliosis angle from standard radiographs. BMC Musculoskeletal Disorders 2020 21 475. (https://doi.org/10.1186/s12891-020-03494-w)
- 21.↑
Wong YS, Lai KK, Zheng YP, Wong LL, Ng BK, Hung AL, Yip BH, Chu WC, Ng AW, Qiu Y, et al.Is radiation-free ultrasound accurate for quantitative assessment of spinal deformity in idiopathic scoliosis (IS): A detailed analysis with EOS radiography on 952 patients. Ultrasound in Medicine and Biology 2019 45 2866–2877. (https://doi.org/10.1016/j.ultrasmedbio.2019.07.006)
- 22.↑
Trac S, Zheng R, Hill DL, & Lou E. Intra- and interrater reliability of Cobb angle measurements on the plane of maximum curvature using ultrasound imaging method. Spine Deformity 2019 7 18–26. (https://doi.org/10.1016/j.jspd.2018.06.015)
- 23.↑
Ferràs-Tarragó J, Valencia JMM, Belmar PR, Vergara SP, Gómez PJ, Hermida JLB, Hermida PB, & Hermida TB. Cobb angle measurement with a conventional convex echography probe and a smartphone. European Spine Journal 2019 28 1955–1961. (https://doi.org/10.1007/s00586-019-06030-0)
- 24.↑
Takács M, Orlovits Z, Jáger B, & Kiss RM. Comparison of spinal curvature parameters as determined by the ZEBRIS spine examination method and the Cobb method in children with scoliosis. PLoS One 2018 13 e0200245. (https://doi.org/10.1371/journal.pone.0200245)
- 25.↑
Brink RC, Wijdicks SPJ, Tromp IN, Schlösser TPC, Kruyt MC, Beek FJA, & Castelein RM. A reliability and validity study for different coronal angles using ultrasound imaging in adolescent idiopathic scoliosis. Spine Journal 2018 18 979–985. (https://doi.org/10.1016/j.spinee.2017.10.012)
- 26.↑
Zheng YP, Lee TT, Lai KK, Yip BH, Zhou GQ, Jiang WW, Cheung JC, Wong MS, Ng BK, Cheng JC, et al.A reliability and validity study for Scolioscan: a radiation-free scoliosis assessment system using 3D ultrasound imaging. Scoliosis and Spinal Disorders 2016 11 13. (https://doi.org/10.1186/s13013-016-0074-y)
- 27.↑
Zheng R, Young M, Hill D, Le LH, Hedden D, Moreau M, Mahood J, Southon S, & Lou E. Improvement on the accuracy and reliability of ultrasound coronal curvature measurement on adolescent idiopathic scoliosis with the aid of previous radiographs. Spine 2016 41 404–411. (https://doi.org/10.1097/BRS.0000000000001244)
- 28.↑
Zheng R, Chan ACY, Chen W, Hill DL, Le LH, Hedden D, Moreau M, Mahood J, Southon S, & Lou E. Intra- and inter-rater reliability of coronal curvature measurement for adolescent idiopathic scoliosis using ultrasonic imaging method - A pilot study. Spine Deformity 2015 3 151–158. (https://doi.org/10.1016/j.jspd.2014.08.008)
- 29.↑
Young M, Hill DL, Zheng R, & Lou E. Reliability and accuracy of ultrasound measurements with and without the aid of previous radiographs in adolescent idiopathic scoliosis (AIS). European Spine Journal 2015 24 1427–1433. (https://doi.org/10.1007/s00586-015-3855-8)
- 30.↑
Cheung C-WJ, Zhou GQ, Law SY, Lai KL, Jiang WW, & Zheng YP. Freehand three-dimensional ultrasound system for assessment of scoliosis. Journal of Orthopaedic Translation 2015 3 123–133. (https://doi.org/10.1016/j.jot.2015.06.001)
- 31.↑
Ovadia D, Bar-On E, Fragnière B, Rigo M, Dickman D, Leitner J, Wientroub S, & Dubousset J. Radiation-free quantitative assessment of scoliosis: a multi center prospective study. European Spine Journal 2007 16 97–105. (https://doi.org/10.1007/s00586-006-0118-8)
- 32.↑
Ramírez N, Olivella G, Cuneo A, Carrazana L, Ramírez N, & Iriarte I. Prevalence and clinical relevance of underlying pathological conditions in painful adolescent idiopathic scoliosis: a MRI-based study. Spine Deformity 2020 8 663–668. (https://doi.org/10.1007/s43390-020-00065-w)
- 33.↑
Bonanni PG. Contour and angle-function based scoliosis monitoring: relaxing the requirement on image quality in the measurement of spinal curvature. International Journal of Spine Surgery 2017 11 22. (https://doi.org/10.14444/4022)
- 34.↑
Pang H, Wong YS, Yip B, Hung A, Chu W, Lai K, Zheng YP, Chung T, Sharma G, Cheng J, et al.Using ultrasound for screening scoliosis to reduce unnecessary radiographic radiation - a prospective diagnostic accuracy study on 442 schoolchildren. Studies in Health Technology and Informatics 2021 280 106–108. (https://doi.org/10.3233/SHTI210445)
- 35.↑
Wu HD, Chu WC, He CQ, & Wong MS. Assessment of the plane of maximum curvature for patients with adolescent idiopathic scoliosis via computed tomography. Prosthetics and Orthotics International 2020 44 298–304. (https://doi.org/10.1177/0309364620941297)
- 36.↑
Adam CJ, Izatt MT, Harvey JR, & Askin GN. Variability in Cobb angle measurements using reformatted computerized tomography scans. Spine (Phila Pa 1976) 2005 30 1664–1669. (https://doi.org/10.1097/01.brs.0000169449.68870.f8)
- 37.↑
Nguyen DV, Vo QN, Le LH, & Lou EHM. Validation of 3D surface reconstruction of vertebrae and spinal column using 3D ultrasound data - A pilot study. Medical Engineering and Physics 2015 37 239–244. (https://doi.org/10.1016/j.medengphy.2014.11.007)
- 38.↑
Zheng R, Hill D, Hedden D, Mahood J, Moreau M, Southon S, & Lou E. Factors influencing spinal curvature measurements on ultrasound images for children with adolescent idiopathic scoliosis (AIS). PLoS One 2018 13 e0198792. (https://doi.org/10.1371/journal.pone.0198792)
- 39.↑
Schell TL, Krueger D, Binkley N, Hetzel S, Bernatz JT, & Anderson PA. Opportunistic use of dual-energy X-ray absorptiometry to evaluate lumbar scoliosis. Archives of Osteoporosis 2021 16 38. (https://doi.org/10.1007/s11657-021-00898-6)
- 40.↑
Yeung KH, Man GCW, Lam TP, Ng BKW, Cheng JCY, & Chu WCW. Accuracy on the preoperative assessment of patients with adolescent idiopathic scoliosis using biplanar low-dose stereoradiography: a comparison with computed tomography. BMC Musculoskeletal Disorders 2020 21 558. (https://doi.org/10.1186/s12891-020-03561-2)
- 41.↑
Lechner R, Putzer D, Dammerer D, Liebensteiner M, Bach C, & Thaler M. Comparison of two- and three-dimensional measurement of the Cobb angle in scoliosis. International Orthopaedics 2017 41 957–962. (https://doi.org/10.1007/s00264-016-3359-0)
- 42.↑
Vavruch L, & Tropp H. A comparison of Cobb angle: standing versus supine images of late-onset idiopathic scoliosis. Polish Journal of Radiology 2016 81 270–276. (https://doi.org/10.12659/PJR.895949)
- 43.↑
Tauchi R, Tsuji T, Cahill PJ, Flynn JM, Flynn JM, Glotzbecker M, El-Hawary R, Heflin JA, Imagama S, Joshi AP, et al.Reliability analysis of Cobb angle measurements of congenital scoliosis using X-ray and 3D-CT images. European Journal of Orthopaedic Surgery and Traumatology: Orthopedie Traumatologie 2016 26 53–57. (https://doi.org/10.1007/s00590-015-1701-7)
- 44.↑
Shi B, Mao S, Wang Z, Lam TP, Yu FWP, Ng BKW, Chu WC, Zhu Z, Qiu Y, & Cheng JC. How does the supine MRI correlate with standing radiographs of different curve severity in adolescent idiopathic scoliosis? Spine 2015 40 1206–1212. (https://doi.org/10.1097/BRS.0000000000000927)
- 45.↑
Lee MC, Solomito M, & Patel A. Supine magnetic resonance imaging Cobb measurements for idiopathic scoliosis are linearly related to measurements from standing plain radiographs. Spine (Phila Pa 1976) 2013 38 E656–E661. (https://doi.org/10.1097/BRS.0b013e31828d255d)
- 46.↑
Diefenbach C, Lonner BS, Auerbach JD, Bharucha N, & Dean LE. Is radiation-free diagnostic monitoring of adolescent idiopathic scoliosis feasible using upright positional magnetic resonance imaging? Spine (Phila Pa 1976) 2013 38 576–580. (https://doi.org/10.1097/BRS.0b013e318286b18a)
- 47.↑
Al-Aubaidi Z, Lebel D, Oudjhane K, & Zeller R. Three-dimensional imaging of the spine using the EOS system: is it reliable? A comparative study using computed tomography imaging. Journal of Pediatric Orthopedics. Part B 2013 22 409–412. (https://doi.org/10.1097/BPB.0b013e328361ae5b)
- 48.↑
Wessberg P, Danielson BI, & Willén J. Comparison of Cobb angles in idiopathic scoliosis on standing radiographs and supine axially loaded MRI. Spine (Phila Pa 1976) 2006 31 3039–3044. (https://doi.org/10.1097/01.brs.0000249513.91050.80)
- 49.↑
Brink RC, Colo D, Schlösser TPC, Vincken KL, van Stralen M, Hui SCN, Shi L, Chu WCW, Cheng JCY, & Castelein RM. Upright, prone, and supine spinal morphology and alignment in adolescent idiopathic scoliosis. Scoliosis and Spinal Disorders 2017 12 6. (https://doi.org/10.1186/s13013-017-0111-5)
- 50.↑
Yazici M, Acaroglu ER, Alanay A, Deviren V, Cila A, & Surat A. Measurement of vertebral rotation in standing versus supine position in adolescent idiopathic scoliosis. Journal of Pediatric Orthopedics 2001 21 252–256. (https://doi.org/10.1097/01241398-200103000-00025)
- 51.↑
Geijer H, Beckman K, Jonsson B, Andersson T, & Persliden J. Digital radiography of scoliosis with a scanning method: initial evaluation. Radiology 2001 218 402–410. (https://doi.org/10.1148/radiology.218.2.r01ja32402)
- 52.↑
Yang C, Li Y, Zhao Y, Zhu X, Li M, & Liu G. Adult degenerative scoliosis: can Cobb angle on a supine posteroanterior radiograph be used to predict the Cobb angle in a standing position? Medicine (Baltimore) 2016 95 e2732. (https://doi.org/10.1097/MD.0000000000002732)
- 53.↑
Weber CI, Hwang CT, van Dillen LR, & Tang SY. Effects of standing on lumbar spine alignment and intervertebral disc geometry in young, healthy individuals determined by positional magnetic resonance imaging. Clinical Biomechanics (Bristol, Avon) 2019 65 128–134. (https://doi.org/10.1016/j.clinbiomech.2019.04.010)
