1 Introduction

Spine specialists have recognized that the famous “gold standard” given by the measurement of the Cobb angle is not sufficient for evaluating scoliosis deformities.¹ Assessment of deformities in patients with adolescent idiopathic scoliosis (AIS) with only the Cobb angle is not enough, as the Cobb angle only provides information about the coronal deformation of the spine. In the meantime postsurgical evaluation of the sagittal and transversal plane is recommendable but often neglected.¹ However, on lateral standing full spine radiographs, the circumstantial analysis of the sagittal profile can be performed. Certainly, for detailed information on the transversal plane, the use of 3D analysis with stereoradiography is a potential option. Recently, the clinical relevance of three-dimensional (3D) analysis in patients with AIS was emphasised by the Scoliosis Research Society (SRS).^2–4 As reported by Imrie et al., the processing and application of modern pedicle screw instrumentation systems can lead to outstanding coronal corrections while deteriorating the sagittal profile at the same time.⁵ The problems of the satisfactory restoration of thoracic kyphosis (TK) that were observed in AIS patients treated with posterior fusion were pointed out when vertebral derotation was used for correction in pedicle screw constructs.^5,6 Modern posterior fusion procedures for AIS correction demonstrated rather satisfying results in the coronal and sagittal planes, but data on their efficacy in the transversal plane are sparse.⁷

Based on images obtained by using George Charpak's low-dose X-ray technology, which was honoured with the Nobel Prize in Physics (1992), Dubousset et al. presented an innovative biplanar radiographic system (EOS® system) that allowed 3D reconstructions of spinal deformities.⁸ Recent progress with this device has allowed its use in daily routine environments,⁹ and it is now considered a standard diagnostic tool for AIS patients in our institution. The sterEOS (EOS® imaging, Paris, France) software allows the 3D reconstruction of the whole spine based on anatomic references defined by the user, and it provides specific parameters.^8,9 The reliability and precision of 3D reconstructions obtained from biplanar X-rays for mild scoliosis^9–11 and for severe scoliosis with indications for surgery¹² were tested.

To study the effects of posterior fusion on AIS patients, we conducted a retrospective study to analyse 3D reconstructions of spine deformities. To the best of our knowledge, studies of 3D correction analyses in AIS patients with such a large study cohort are sparse. Our primary research objective was to investigate the magnitude of correction of the spine profile in the coronal, sagittal, and transverse planes due to the primary curve topography (for thoracic, thoracolumbar, and lumbar curves). Furthermore, intervertebral axial rotation and detorsion was investigated in all three subgroups.

2 Methods

2.1

2.1 Study cohort

This is a retrospective single-centre study of patients with AIS and indications for surgery (Cobb angle greater than 45°). The study cohort was surgically treated with posterior fusion in our institution between 2010 and 2020. A further inclusion criterion was the availability of full-spine EOS® radiographs in an anterior–posterior and lateral view. Patients with congenital, neuromuscular, syndromic scoliosis, and patients with former surgeries, were not considered. In addition, patients without their full spine visible on radiographs were excluded. Fig. 1 illustrates the inclusion process. This study was approved by the ethics committee of Heidelberg University (permission No. S-872/2019).

Close

Fig. 1 Flow diagram describing the exclusion and inclusion process for the study population.

Export to PPT

2.3

2.3 Data collection and analysis

Radiographs were acquired with a stereo radiography device,⁸ which is a low-dose biplanar imaging system that provides simultaneous anterior–posterior and lateral views in an upright posture. All images were obtained with standardized methods. All patients were barefoot and had their hands on their clavicles to avoid the overlay of the upper extremities on the upper part of the thoracic spine.¹⁴ All patients were requested to look straight ahead and to stand in a relaxed position. Radiographs that were performed immediately before and 6 months after surgery were used for measurements. Semiautomatic 3D reconstructions of the spine and pelvis were performed by the second author (C.L.), who is a research fellow with a medical background who was not involved in the treatment of those patients. Measurements were executed via validated and dedicated software (EOS-imaging®, Paris, France).⁹ Full-spine method measurements (Fig. 2) were chosen for all radiographs. The radiographic parameters included the following:

Close

Fig. 2 (A) Pre- and (B) post-surgery anterior–posterior and lateral radiographs illustrating the analysed spinal parameters.

Export to PPT

Coronal parameters: Cobb angle, pelvic obliquity, and coronal vertebral tilt (from T1 to L5).

Sagittal spinopelvic parameters: regional parameters—thoracic kyphosis (TK (T1-T12) and TK (T4-T12)), lumbar lordosis (LL (L1-L5) and LL (L1-S1)), sagittal vertebral tilt (from T1 to L5), pelvic incidence (PI), pelvic tilt (PT), and sacral slope (SS); global alignment parameters—sagittal vertical axis (SVA), spinosacral angle (SSA), T1—spinopelvic inclination (T1SPi), and T9— spinopelvic inclination (T9SPi).

Transverse plane parameters: apical vertebral rotation (AVR) and axial vertebral rotation (from T1 to L5). The intervertebral axial rotation (IAR) was calculated in relation to the axial rotation of the upper vertebra in the plane of the lower vertebra.^14–16 The torsion index (Ti) was estimated as described by Steib et al. (the mean of the two sums of the IAR from the lower junction to the apex and then from the apex to the upper junction)¹⁷ (Fig. 3). Finally, a “detorsion index” ((Ti pre − Ti post)/Ti pre × 100) was computed to quantify the surgical correction of the torsion of the structural curve in terms of percentage.

Close

Fig. 3 Parameters of the transverse plane including intervertebral axial rotation (IAR) and the torsion index (Ti) as described by Steib et al.

Export to PPT

3 Results

3.2

3.2 Analysis of the segmental vertebral tilt by curve type and plane

3.2.1

3.2.1 Coronal tilt

Concerning the segmental coronal vertebral tilt, the surgical posterior correction revealed statistically significant improvements in all mentioned subgroups (p ≤ 0.05). Vertebral tilt was significantly improved from T2 to L5 in the thoracic group, from T2 to L5 the thoracolumbar group, and from T4 to L5 in the lumbar group. Moreover, in all subgroups, numerous parts of the significantly improved levels showed a change in the vertebral tilt in a range of at least or greater than 50 %. The thoracic curves also revealed a decrease in the coronal vertebral tilt of at least 50 % or more in T2–L2, as well as in the thoracolumbar curves in T2-L2 and in the lumbar curves in in T6-L4 (Fig. 4.)

Close

Fig. 4 These diagrams illustrate pre-to postoperative changes in the vertebral coronal tilt from T1 to L5 for the study subgroups regarding the structural curves (thoracic, thoracolumbar, and lumbar). Statistically significant changes (p ≤ 0.05) in the vertebral tilt are delineated in a red colour.

Export to PPT

3.2.2

3.2.2 Sagittal tilt

The analysis of the vertebral sagittal tilt in the lumbar subgroup revealed a significant increase in the vertebral tilt in T9-T11 (vertebral reclination) and L3-L5 (vertebral inclination). Moreover, in L3 and L4, posterior fusion led to tilt cross-over from the posterior to the anterior tilt (inclination). Significant loss of tilt was seen in L1-L2. For the thoracic and thoracolumbar groups, there were a significant but clinically irrelevant changes in the tilt in T5-T6, L2, and T10-T11, respectively (Fig. 5).

Close

Fig. 5 These diagrams illustrate pre-to postoperative changes in the vertebral sagittal tilt from T1 to L5 for the study subgroups regarding the structural curves (thoracic, thoracolumbar, and lumbar). Statistically significant changes (p ≤ 0.05) in the vertebral tilt are delineated in a red colour.

Export to PPT

3.2.3

3.2.3 Axial vertebral rotation and detorsion index

Concerning vertebral rotation before and after surgery, AVR 1 was reduced in all subgroups, reaching a high level of significance (thoracic curves: AVR 1 pre vs. post: 17.6 ± 9.3 vs. 5.7 ± 4.3, p = 0.001; thoracolumbar curves: AVR 1 pre vs. post: 22.5 ± 5.9 vs. 9.4 ± 6.4, p = 0.001; lumbar curves: AVR 1 pre vs. post: 22.8 ± 4.5 vs. 9.3 ± 3.3, p = 0.001) (Table 3). Furthermore, the lower IAR, upper IAR, and torsion index were significantly reduced in all groups after the surgical procedure. Vertebral detorsion between 55 % and 69 % could be achieved with posterior correction and fusion (Table 3). Regarding the vertebral axial rotation, the thoracic subgroup had the greatest derotation in the middle of the thoracic spine, the thoracolumbar subgroup had the greatest derotation in the lower thoracic spine, and the lumbar subgroup had the greatest derotation in the thoracolumbar junction (Fig. 6).

Table 3 Analysis of vertebral rotation in the axial plane for thoracic, thoracolumbar and lumbar group.

		Thoracic Group	Thoracolumbar Group	Lumbar Group
Lower IAR°	Pre	24.9 ± 10.6	18.7 ± 5.8	6.9 ± 4.3
	Post	7.7 ± 7.0	7.1 ± 5.1	3.2 ± 2.3
	Diff	17.2 ± 7.8	11.6 ± 7.5	3.8 ± 6.3
	P	0.001	0.001	0.002
Upper IAR°	Pre	15.1 ± 11.4	18.4 ± 8.4	16.2 ± 6.3
	Post	3.8 ± 3.0	6.9 ± 5.2	5.9 ± 3.4
	Diff	11.3 ± 11.0	11.5 ± 8.7	10.4 ± 8.7
	p	0.001	0.001	0.003
AVR°	Pre	17.6 ± 9.3	22.5 ± 5.9	22.8 ± 4.5
	Post	5.7 ± 4.3	9.4 ± 6.4	9.3 ± 3.3
	Diff	11.9 ± 7.5	13.0 ± 6.3	13.5 ± 6.3
	p	0.001	0.001	0.001
Torsion index	Pre	20.0 ± 7.0	18.6 ± 5.3	11.6 ± 3.2
	Post	5.8 ± 3.7	7.0 ± 4.6	4.5 ± 2.4
	Diff	14.3 ± 6.7	11.5 ± 6.6	7.1 ± 4.8
	p	0.001	0.001	0.001
Detorsion index		69 %	59 %	55 %

Close

Fig. 6 These diagrams illustrate pre-to postoperative changes in vertebral axial rotation from T1 to L5 for the study subgroups regarding the structural curve (thoracic, thoracolumbar, and lumbar). Statistically significant changes (p ≤ 0.05) in vertebral rotation are delineated in a red colour.

Export to PPT

4 Discussion

The posterior surgical procedure for the correction of scoliosis is a multifactorial act that has a direct impact on patients’ outcomes. Over the last decade, the improvement of surgical procedures and the development of implant systems has led to satisfactory correction results in the coronal plane, as well as in the sagittal profile. In the past, the study of the axial plane was often restricted to the analysis of the axial vertebral rotation of the apical vertebra in the primary and secondary curves. The use of 3D full-spine EOS reconstruction technology enables surgeons to obtain even more detailed information for deformity analysis than in the past. One fundamental advantage of conducting an analysis with the EOS system is the possibility of elucidating pre-to postoperative rotational changes for each measured vertebra (T1-L5) in all planes. To the best of our knowledge, this is the largest study in the literature to use 3D full-spine EOS reconstruction for measurements of rotational changes in each vertebra in surgically treated AIS patients with posterior fusion.

The global analysis of the study cohort revealed that with the posterior fusion that was performed, the Cobb angle correction that was achieved for the primary and secondary curves exceeded 50 % (Table 1). As expected, the Cobb angle of the structural curve was significantly decreased in all study subgroups. Merely one patient had no relevant correction of the compensatory curve (less than 5°) (Table 2). The Cobb angle correction in this study was shown to have similar results to the correction outcomes mentioned in other studies.^18,19 However, data on the coronal vertebral tilt of each vertebra after AIS correction are still sparsely available. Full-spine EOS measurements provide a better understanding of the pre-to postoperative reorientation of the segmental tilt. The coronal tilt correction had a specific concentration due in each study subgroup; the thoracic group was revealed to have the highest coronal tilt correction at levels T5 and T11, the thoracolumbar curves had the highest correction at levels T6 and L1, and the lumbar curves had the highest correction at levels T10 and L3 (Fig. 4). The highest levels of coronal tilt correction identified in the subgroups could be explained with the morphologies of the scoliosis types and with the manoeuvres performed for surgical correction. If vertebrae were more horizontally aligned, the surgical procedure affected coronal vertebral tilt at the apex level less than in the cephalad or caudal vertebrae of the aforementioned apex. Perhaps 3D analysis of the spine may be of additive value to the gold standard biplanar measurement of Cobb angle for verification of surgical correction. The measurement of the segmental coronal tilt provides the possibility of evaluating the specific correction for each measured vertebra. For that, it could allow more detailed identification on coronal correction to be obtained.

A further purpose of surgical AIS correction is to preserve a normal sagittal profile as much as possible. Alterations in the regional and global sagittal parameters as a result of the performance of posterior fusion are feasible and have already been elucidated in other studies.^20–22 The global analysis of the sagittal parameters in this study cohort revealed that there were no significant changes in any of the measured regional spinal and pelvic parameters. In only the lumbar subgroup, T1SPi and T9SPi were significantly changed, indicating that there were slight profile changes in the thoracic shape. The lack of relevant changes in the sagittal parameters in this study group might be explained by two factors. Firstly, the study cohort had predominantly a normal sagittal profile, so no relevant corrections of the sagittal shape were needed. Secondly, the posterior correction that was undertaken was performed while respecting the normal sagittal profile. Therefore, the implantation of bent rods did not cause any relevant alterations in the sagittal profile. The clinical importance of preserving a normal sagittal profile in AIS patients with posterior fusion was already revealed in several studies.^22–25 Furthermore, the study of the sagittal tilt of each measured vertebra can give a wider view of changes in the sagittal alignment. In the lumbar subgroup of this study population, on some levels, we noticed significant reorientation due to the sagittal vertebral tilt. In this group, postoperatively, the lower lumbar spine (L4-L5) had significantly higher vertebral inclination, the upper lumbar spine (L1-L3) had significantly lower reclination, and the lower thoracic part (T9-T11) had significantly higher reclination. The interplay of the changed lumbar vertebral in- and reclination caused alterations in the lumbar shape without significant changes in the LL. The changes in the lumbar shape in this subgroup might be an explanation for the slight shift in the global parameters, such as T1SPi and T9SPi, without significant changes in TK, SVA, or SSA.

Today, the study of the parameters of the transverse plane belongs to the standard analysis of the outcomes of correction in AIS patients.²⁶ As a diagnostic device, stereoradiography has provided a new opportunity to evaluate the transverse rotation of all measured vertebral levels in subjects with spinal deformities and to evaluate the correction magnitude postoperatively. The analysis of axial rotation in the subgroups in this study revealed a specific derotation pattern in each group. On average, preoperatively, the maximum of the axial rotation in the thoracic group was at T8; in thoracolumbar group, it was at T9, and in the lumbar group, it was at L1. In all groups, significant axial correction could be achieved over at least eight levels according to the analysis of the structural curve (Fig. 6). These findings show that axial vertebral changes and detorsion occurred in all of this study's subgroups. With 69 %, the highest detorsion could be reached in the thoracic primary curves, and the “lowest” was achieved with 55 % in the lumbar primary curves. In this study group, with posterior fusion, the detorsion that was reached was on a good level surgical quality in comparison with the results of other studies.^1,3 Furthermore, our data demonstrate that significant detorsion also occurred from the upper and lower junctional intervertebral derotation (Table 3). Remarkably, the higher level of detorsion that was reached in the thoracic curves than in the lumbar curves might be explained by the chosen construct length. For sufficient derotation during surgery, a solid anchoring above and below the primary curve is crucial. For that, the accurate definition of the upper instrumented vertebra (UIV) and lower instrumented vertebra (LIV) is essential. On the other hand, an important aspect of surgical strategy is to preserve as many vertebral segments as possible. Lower lumbar segments should especially be preserved from fusion due to their high functional motion. It is possible that, in some cases, the LIV in the lumbar subgroup was chosen with a functional purpose and not with the purpose of achieving the best derotation results. There are several derotation techniques and implant systems^27–30 that were developed over decades and are used for treatment of patients with AIS. Sufficient spinal derotation can have positive effect on rib hump or lumbar bulge reduction, that is of crucial importance for a predominant number of patients. The circumstantial study of derotation magnitude as with this mentioned method, gives new possibilities to investigate and to compare the established derotation techniques.

There are several weaknesses in this study. The generalisation of the study results is limited by the retrospective study design and the relatively small study cohort. Therefore, the performance of studies with much more representative study cohorts should be the focus of further investigations. Secondly, because the surgeries were performed only the two senior authors, these results cannot be generalised. Furthermore, the absence of a control group should be seen as a limitation. However, performing radiography with no medical indication is prohibited due to ethical considerations. Moreover, the stratification according to the classification established by Lenke³¹ or Scoliosis Research Society (SRS)³² were not practical because it would result in subgroups that were too small for statistical analysis. Furthermore, immediate postoperative correction does not directly translate into surgical effectiveness. The measurements obtained with the EOS software were not compared with measurements from other validated software, since the aim of this research was not the evaluation of the reliability of the EOS system. This has already been reported.^11,12 Finally, this purely radiological study with no functional scoring was performed to reveal changes in radiological parameters and not to evaluate the clinical outcomes of the study subjects. Nevertheless, a correlation of radiological spinal correction magnitude with clinical outcome as pain score and quality of life score should be in focus in further studies.

5 Conclusion

This study confirms that posterior fusion effectively corrects AIS in all three planes, achieving significant coronal correction and substantial axial derotation (55–69 %) while preserving sagittal balance. The maintained sagittal alignment minimizes the risk of long-term complications, while axial derotation may improve cosmetic outcomes, particularly reducing rib hump prominence. These findings highlight the clinical importance of 3D assessment in surgical planning, as Cobb angle alone does not fully capture correction outcomes. By integrating stereoradiography, surgeons can optimize correction strategies for better functional and aesthetic results. Future research should explore how these radiological improvements translate into patient-reported outcomes and long-term quality of life.

Informed consent statement

Patient consent was waived due to retrospective study design and permission by the Ethics Committee. Analysed radiographs were obtained while daily praxis. No radiographs were obtained explicitly for this study.

Author contributions

Conceptualization, W.P. and S.H.; methodology, W.P an S.H.; software, C.L.; validation, C.L., W.P. and S.H.; formal analysis, T.B., C.L. and W.P.; investigation, C.L. and W.P.; resources, W.P.; data curation, C.L.; writing—original draft preparation, S.H.; writing—review and editing W.P., M.A., and S.H.; visualization, C.L. and W.P.; supervision, W.P. and S.H.; project administration, W.P.; All authors have read and agreed to the published version of the manuscript.

Institutional review board statement

The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Ethics Committee of University Heidelberg (permission No. S-872/2019).

Data availability

The datasets generated during and analysed during the current study are available from the corresponding author on reasonable request.

Declaration of generative AI and AI-assisted technologies in the writing process

No artificial intelligence assisted technology was used to carry out this study.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.