EOS® imaging: Concept and current applications in spinal disorders

Abstract

EOS® imaging is a proprietary imaging technology that was launched in 2007. Based on a gaseous particle detector with a multi-wire proportional chamber, it offers several advantages over other imaging modalities: low dose of radiation, ability to create 3D reconstructions, ability to conduct whole body imaging, high reproducibility in measuring various parameters of alignment and faster imaging time. EOS® imaging is slowly gaining widespread acceptance as its applications in various disorders continue to evolve. It has been found to be particularly useful and has opened up new avenues of research in the field of spinal deformities. This narrative review seeks to provide an overview of the proprietary technology behind EOS® imaging, compare it to existing imaging modalities, summarize its current applications in various spinal disorders and outline its limitations.

1. Introduction

Imaging of the spine has evolved greatly in the last century – beginning from its humble origins in conventional radiography, passing through invasive investigations like discography and myelography, and culminating in the present day ‘standard of care’ modalities like computed tomography (CT) and magnetic resonance imaging (MRI). However, a quest to improve has always fueled innovations and improvisations in healthcare – the EOS® imaging is the brainchild of such a quest. Being a recent addition to the armamentarium of the spine surgeons, EOS® imaging is slowly gaining acceptance across the world as more and more of its applications are being discovered.¹ The need for an imaging modality like EOS® stemmed from the realization that the spine is not an independent anatomical and functional entity – but is rather intricately linked to the pelvis and lower limbs.² Several studies have reported on the influence of sagittal balance or lack thereof on patient reported outcomes in disorders like adult spinal deformity, paediatric spinal deformity, ankylosing spondylitis, cervical deformity, neuromuscular disorders and Parkinsonism.3, 4, 5, 6, 7 The evaluation of sagittal alignment and compensatory mechanisms occurring in the pelvis and lower limbs to restore its imbalance is best done in the functionally loaded standing position.⁸

EOS® (EOS® imaging, France) is essentially a low dose biplanar radiography, based on the pioneering work of Professor Georges Charpak, who invented a gaseous particle detector with a multi-wire proportional chamber.⁹ The magnitude of his contribution can be gauged by the fact that he was awarded the Nobel Prize in Physics in 1992 for this invention. The EOS® imaging machines were first tested in 2002 in France and have been commercially available since 2007. EOS® has the advantages of low radiation dose compared to digital radiography (DR), ability to reconstruct three dimensional (3D) images and image the whole body including the spine and lower limbs in the functional standing position. In an EOS® imaging system, a multi-wire gas chamber is placed between the object being radiographed and the detector. The emergent rays from the object cause the release of a second wave of photons in the gas chamber, which is detected by the detector. This second wave of photons is the prime reason for the need of low dose of primary x-ray beam.¹⁰

An EOS® imaging system has two pairs of X-ray tubes and detectors which are placed at 90° to each other (Fig. 1). Simultaneous orthogonal images are captured and spatially calibrated to each other. Each detector has a field of 180 × 45 cm which means that it can scan the whole body in an acquisition time of 30–45 s. Scanning of just the spine is even faster and takes between 5 and 15 s¹⁰ 3D reconstruction is done with the help of a special in-built software, known as ‘ster-EOS®‘.¹⁰ The difference in radiation exposure is important in conditions like paediatric spinal deformity where repeated radiographic examinations are required throughout the growing age of the child to monitor the progress of the deformity and make critical treatment decisions. These unique strengths of EOS® imaging place it at a very favourable position for becoming a preferred imaging modality in several spinal disorders. The purpose of this narrative review is to provide the reader with a brief overview of the EOS® system and summarize its current applications in spinal disorders.

Fig. 1.

An EOS® imaging system with two pairs of tubes and detectors placed orthogonal (at 90°) to each other.

1.1. EOS® and digital radiography (DR)

Since the advent of EOS® imaging, many studies have compared it with DR. When compared to DR, the average dose of radiation to the skin has been seen to be 6 to 9 times lesser in the thoracoabdominal region with EOS®. Moreover, the image quality is significantly better with EOS® for all structures in both posteroanterior (PA) and lateral views except for the visualization of lumbar spinous processes.¹¹ The mean time for examination is also significantly lower for the EOS® system. The dose of radiation has been reported to be even lower for the lumbar region.¹² Luo et al.¹³ compared EOS® and DR over the whole period of treatment till skeletal maturity in patients with adolescent idiopathic scoliosis (AIS). An average of 20.9 radiographic examinations was needed per patient throughout the course of treatment from presentation to skeletal maturity. The mean cumulative dose was 50% less for EOS® (2.66 milli- Sievert (mSv)) compared to DR (5.38 mSV).This amounted to a mean reduction of 2.72 mSv or 0.91 years of background radiation. Chung et al. noted comparable intra-observer and inter-observer reliability with both DR and EOS® imaging for measurement of Cobb angle in 32 spinal phantoms.¹⁴

1.2. EOS® and CT scans

Considering the ability of EOS® to generate a 3D reconstructed image, it has been compared to CT which has been the time-tested modality for getting 3D reconstructed images of an anatomical part. 3D reconstruction of images with EOS® is based on two simultaneously taken orthogonal images – however, it requires a skilled operator who can use the SterEOS® software (EOS Imaging, France) to create a 3D reconstructed image. This also means that in contrast to CT, variation in patient positioning may influence the accuracy of the 3D reconstruction.¹⁵ Creation of 3D reconstructed images using EOS® is a semi-automated process which does not require any additional specific change in positioning of the system or the patient. The simultaneous acquisition of biplanar images in two orthogonal planes allows them to be spatially calibrated to each other – the proprietary SterEOS® software uses algorithms based on bone shape recognition and parametric modeling of longitudinal and transversal inferences to create 3D reconstructed images from biplanar images.¹⁶ Glaser et al. compared the accuracy of 3D reconstruction by standing EOS® imaging and supine CT in a study performed on scoliotic phantoms. The mean vertebral body shape accuracy for 3D reconstructed images with EOS® imaging was within 1.1 ± 0.2 mm when compared to CT scan. Different anatomical regions modeled well, the position and orientation accuracy were high and there was no significant difference in any of the measured parameters.¹⁷ Hasegawa et al. compared whole spine alignment between standing EOS® imaging and supine CT imaging. They noted that the mean interclass correlation coefficients (ICC) between the 2 techniques for intra-rater and inter-rater reliabilities were excellent.¹⁸

1.3. Micro dose EOS® protocol

In an attempt to further decrease the radiation dose with EOS®, a new microdose protocol was introduced.¹⁹ Microdose EOS® has been shown to reduce the cumulative dose by 26 times compared to DR and up to 45 times compared to computed radiography (CR).^19,20 This translates into a 5–7 times dose reduction when compared to conventional EOS®.¹⁹ The image quality in microdose EOS® has been shown to be similar to DR and slightly lower than conventional EOS® but adequate enough for the clinical measurements to be made reliably.¹⁹ A further step in regard to radiation dose reduction is the ‘reduced microdose protocol’ and the ‘nano-dose protocol’. The ‘reduced microdose protocol’ decreases the radiation dose by 58% and acceptable 3D reconstructions can be obtained compared to microdose protocol.²¹ Another study compared the EOS® standard dose, microdose and nano-dose protocols and found that the dose area product in the nano-dose protocol was 6 times less compared to microdose protocol and 40 times less compared to the standard dose protocol. The measured Cobb angle varied less than 5° amongst these techniques.²²

2. Current application of EOS® in spinal disorders

2.1. Adolescent idiopathic scoliosis (AIS)

A major section of existing literature on EOS® is devoted to its varied uses in AIS. Being a three-dimensional deformity with considerable implications on coronal and sagittal alignment of the trunk, it is quite understandable that EOS® has tremendous potential is revolutionizing the decision-making and treatment paradigms in scoliosis.

2.1.1. General applications

The feasibility of side bending films in EOS® imaging has been established and no significant difference between reduction of Cobb angle in the standing side bending films and supine CR films was found for all Lenke curve types.²³ It was also noted that the radiation dose was 5 times lesser with EOS® imaging. However, 3D reconstruction is not possible with this method as the patient has to stand in corner of the machine due to space constraints. 3D imaging is possible with another methods such as the ‘EOS® suspension test’ in which the patient is suspended using a frame around the head and mandible allowing spatially calibrated images to be taken with patient standing in the center.²⁴ Lau et al.²⁵ tested the feasibility and validity of taking simultaneous hand and spine radiographs in AIS patients. They found excellent intra-rater and inter-observer reliability for EOS® hand radiographs as far as computing the thumb ossification composite index (TOCI) was concerned. The ICC was ‘excellent’ for both TOCI and Sanders classification between DR and EOS® imaging. This simultaneous evaluation of skeletal maturity along with spine imaging improves efficiency and saves time.

2.1.2. 3D EOS® imaging and AIS (Fig. 2, Fig. 3)

Fig. 2.

2D biplanar Whole body EOS® imaging.

Fig. 3.

3D reconstruction from a 2D biplanar image using the SterEOS® software.

(a) Identification and digitalization of anatomical landmarks and shaping every vertebral body.

(b) Creation of a 3D whole spine model.

(c) Stagnara derotation view of the 3D reconstructed model.

(d) A typical SterEOS® report providing multiple measurements in coronal, sagittal and axial planes.

Multiple authors have established the intra-rater and inter-rater reliability of 3D reconstruction in AIS patients including rotation for each vertebra.^26,27 However, it was worth noting that reliability of axial rotation in the upper and middle thoracic spine was lower. This finding has been attributed to the obtuse angulation of pedicle and spinous processes in the frontal view which make their identification and delineation difficult.²⁶ Ilharreborde et al. noted excellent inter-rater and intra-rater reliability between 3 different operators for measurements of Cobb angle, kyphosis, lumbar lordosis and pelvic parameters in both preoperative and postoperative periods. 3D reconstructions are not affected by the presence of orthopaedic implants and can be easily used in the postoperative scans as well.²⁸

The intra-observer and inter-observer reliability of sagittal measurements including global sagittal parameters using 3D EOS® imaging have been well established.²⁹ Recent studies have suggested that 2D sagittal measurements overestimate the thoracic kyphosis which can be attributed to the coexistence of axial rotation. 3D measurements have been deemed to be necessary to estimate true kyphosis.³⁰ Newton et al. confirmed with the usage of EOS® that thoracic kyphosis was over-estimated on 2D images (mean 18°±13°) when compared to 3D images (mean 6°±14°).³¹ Efforts have also been directed towards developing tools or mathematical formulae to predict 3D kyphosis using measurements made on 2D imaging. One such formula developed by Parvaresh et al. had an average model error of ±7°.³² EOS® imaging also allows the measurement of rotational profile of each vertebra. Kato et al. in their study on 153 AIS patients measured apical vertebrae rotation (AVR) in preoperative and postoperative EOS® images and noted that 3D EOS® reconstruction provides an excellent method for measurement of AVR. They found high intra-rater and inter-rater reproducibility of measurements and attributed it to bilateral acetabular center line being used as reference which remains unaffected by correction in surgery.³³

A novel vertebral vector system has been recently described wherein each vertebra was mapped using a vector in all 3 dimensions. The measurement between the two acetabular centers was assigned a value of 200 with their midpoint acting as origin. A higher intra-rater and inter-rater reliability for vector based methods was seen when compared to 2D methods for coronal and thoracic Cobb angle measurement.³⁴ The efficacy of this method has been established in post-operative Cobb angle measurements too. Using this method, a representation of any spinal deformity true to its size, position, shape and rotation can be made and represented as 2D orthogonal images and vice versa. This is suggested to be a valuable tool to develop a 3D classification for scoliosis.³⁵ Another potential application of the 3D EOS® reconstruction is planning of scoliosis surgeries using the SpineEOS® software (EOS Imaging, France). In a preliminary clinical study, this novel 3D stereoradiography based planning software was found to perform well and accurately simulate the postoperative alteration in global alignment.³⁶

2.1.3. Pulmonary function

The pulmonary function in AIS patients is known to be compromised.This is routinely assessed and documented by conducting preoperative pulmonary function tests. With the ability of EOS® imaging to form 3D reconstructed images of the spine and the rib cage, its role in the prediction of pulmonary function in scoliotic patients has been suggested. Machino et al. compared EOS® with CT scan for evaluation of preoperative ribcage parameters in AIS patients. No significant difference was found in the measurements of any of the rib cage parameters like endothoracic hump ratio (EHR), surface spinal penetration index (sSPI), rib-vertebral angle difference (RVAD) and vertebra-sternum angle (VSA).³⁷ Poor pulmonary function was associated with increasing coronal thoracic Cobb angle, thoracic vertebral rotation and decreasing 3D thoracic kyphosis - Cobb angle >80°, thoracic lordosis >20° and AVR >25° were associated with severe pulmonary function impairment.³⁸

2.1.4. Whole body EOS® for scoliosis

It has been increasingly recognized that scoliosis should not be considered to be simply a spinal deformity – it also has considerable implications on the pelvis and lower limbs as the body responds and compensates to the alteration in three dimensional alignment and asymmetric body weight distribution.^2,3,8,39 A study of these factors was previously limited due to the fact that simultaneous imaging of the spine and the lower limbs was difficult with DR and carried additional risk of radiation exposure. But with advent of EOS® imaging, studying these parameters has become easier. Burkus et al. compared 320 scoliosis patients with 350 controls and used 3D reconstruction of spine with femur to compare proximal femoral parameters. They found a small but significant decrease in the neck shaft angle and higher femoral mechanical axis–anatomical axis angle in the scoliosis group.⁴⁰ Whole body EOS® has also been used to correlate limb lengths, mechanical and anatomical axis, femoral anteversion and tibial torsion with scoliosis.⁴¹ This feature of EOS® is likely to open up numerous such avenues in the future.

3. EOS® in adult spinal deformity (ASD) and degenerative spinal disorders

EOS® has found multiple applications in ASD and degenerative spine disorders. The intra-observer and inter-observer reliability of EOS® for measuring spinopelvic and sagittal parameters has been established in adult degenerative spinal disorders. Besides being better for sagittal measurements, 3D EOS® scans have also proved beneficial in assessment of rotatory deformities. Ferrero et al. investigated the role of 3D EOS® imaging in diagnosing rotatory subluxation and correlated them to other pelvic parameters in ASD. As many as 27% of patients who had a rotational deformity >10° were missed when only 2D images were reviewed – thus validating the role of 3D imaging in early diagnosis of rotatory subluxations in ASD.⁴²

The importance of sagittal and coronal balance can be understood well through the concept of ‘cone of economy’, which highlights the importance of maintaining these parameters within the cone to maintain balance and minimize energy expenditure. If the sagittal balance falls outside this cone - as is often associated with ASD and degenerative scoliosis - it can cause increased muscle exertion and eventual fatigue leading to pain and disability.⁴³ Whole body EOS® provides an opportunity to study the sagittal balance in toto. It has the advantages of lower radiation, no conical projection and simultaneous imaging of the whole body in the functional standing position. Poor sagittal balance measurements have been correlated with poor quality of life indices scores using EOS® imaging.^3,44 The unique ability of EOS® to scan the whole body from head to feet has led to exploration of cranial sagittal vertical axis (CrSVA) as a measure of global sagittal balance and its comparison with the C7SVA. The distance between the plumb line from the cranial center to posterior corner of S1 (CrSVA-S), and to the centers of the hip (CrSVA-H), knee (CrSVA-K) and ankle (CrSVA-A) have been found to correlate well with ODI and SRS-22 scores and outperform C7SVA in this aspect.⁴⁵ Such a parameter linking the cranium with the feet could have only been possible with EOS® and proves the worthiness of EOS® in sagittal balance evaluation.

4. Osteoporosis

In light of its ability to provide whole body imaging without stitching, EOS® permits assessment in the weight bearing position of the global spinal deformity and sagittal imbalance which is commonly seen in long-standing osteoporosis. In addition, the role of EOS imaging in the diagnosis of vertebral osteoporosis and prediction of fracture risk is an exciting prospect under investigation. At present bone mineral density (BMD) measurement by Dual Energy X-ray Absorptiometry (DEXA) is the gold standard for diagnosis - a low BMD however only explains 20% to 70% loss in vertebral strength leading to exploration of other methods to quantify vertebral strength and predict fracture risk.⁴⁶ A low BMD is not entirely predictive of future fracture risk – mathematical models in the past have combined 3D geometrical parameters, subject-specific mechanical properties such as the Young’s modulus and BMD to predict the bone strength.⁴⁷ The subject specific finite element analysis (FEM) values using EOS® imaging correlate with quantitative CT based models - highlighting the potential of EOS® for testing vertebral strength.⁴⁸ The trabecular bone score (TBS) obtained with EOS® has been shown to be significantly lower in osteoporotic fracture patients. The inability to perform this technique in up to one third of the patients due to technical issues like scoliosis, presence of artifacts, bad image quality, high tissue thickness and gas projections is a major drawback.⁴⁹ Osteoporotic vertebral fractures are indicators of osteoporosis and predict future fracture risk. EOS® has been shown to have a good sensitivity (79.7%), excellent specificity (91.6%) and negative predictive value (99%) and better agreement amongst users compared to vertebral fracture assessment tool (VFA) of DEXA scan.⁵⁰

5. Ankylosing spondylitis

Patients with ankylosing spondylitis present to spine surgeons with a global kyphotic deformity and operative intervention is often warranted to restore horizontal gaze and improve global spinal alignment. However, since there are multiple compensatory mechanisms in the pelvis and lower limbs such as pelvic retroversion and flexion at hips and knees which are at play in such patients, a true estimate of their deformity can only be made when all these parameters are taken into account simultaneously. Le Huec et al.⁶ described a full balanced integrated (FBI) technique to estimate the surgical correction needed in such cases – the technique involves three separate angular measurements namely the C7 translation angle (C7TA), the pelvic tilt correction angle (PTCA) and the femoral obliquity angle (FOA). A modality like EOS® makes the use of this technique considerably easy by enabling the measurement of all the angles simultaneously. Were it not for EOS®, these patients normally require multiple and repeated radiographs – EOS® acts as a one-stop shop for imaging in these patients and provides composite information about the degree of spinal kyphosis as well as pelvic retroversion, femoral obliquity and knee flexion. Though EOS® has an excellent agreement with DR for assessing the spinal alignment and extent of deformity; it fares poorly as far as assessment of sacroilitis is concerned.⁵¹

6. Cervical spine disorders

Cervical sagittal parameters measured from whole spine EOS® imaging have been shown to be as reliable as lateral cervical spine x-rays.⁵² Le Huec et al. utilized SterEOS® reconstructions to measure sagittal cervical and global parameters and noted that cervical lordosis is associated with a higher C7 slope and one third of the asymptomatic population had cervical kyphosis.⁵³ Yoshida et al. utilized whole-body EOS® to study the global spine alignment in 88 patients of cervical myelopathy and recommended assessment of global spinal alignment for preoperative planning.⁵

7. Miscellaneous applications

Hasegawa et al. evaluated the sagittal balance of the axial skeleton with reference to gravity line in 136 healthy volunteers using whole body EOS®. They stated that with age the stooping of the trunk increases, which is compensated by increase in cervical lordosis, pelvic tilt and knee flexion.⁵⁴ In another study, Hey et al. evaluated the radiological differences in whole spine sagittal alignment in directed and natural relaxed postures of 60 healthy young volunteers using whole body EOS® scans. They showed the presence of more kyphosis and decreased lumbar lordosis in relaxed positions.⁵⁵ Another potential emerging application is measurement of virtual chest capacity in small children as conventional pulmonary function tests are not feasible in this population - this is still in developmental phase currently.

7.1. Limitations

While the proponents of EOS® vex eloquently about its touted benefits, there is still a long way to go before it becomes a routine part of imaging in patients with spinal disorders. EOS® imaging requires the patient to stand still in the EOS® machine which may be difficult for children and the elderly. Motion artifacts may affect the image quality, but their effect on radiographic measurements is not yet known. Simons et al. found presence of motion artifacts in 39/198 (19.7%) EOS whole spine radiographs. Despite this, they did not find any significant difference in measurements between EOS® imaging and routine radiographs and the intra-observer and inter-observer reliability continued to be high.⁵⁶ The EOS® imaging does not have an option for supine imaging limiting its applicability in patients with neuromuscular disorders and non-ambulatory patients. A few specially designed chairs (Fig. 4) have been used for imaging of non-ambulatory patients with neuromuscular scoliosis with some success but they are not readily available commercially.⁵⁷ The 3D reconstruction with EOS® is not rendered automatically – it requires a trained operator who can use the designated software. Also, this software is made for adult bones – and hence, 3D reconstruction is not possible in children less than 6 years of age.¹⁵ Since the 3D reconstructions in EOS® are based on 2 radiographic images, they cannot provide the information which axial cuts of CT scan can.¹⁵ 3D reconstruction is also not feasible currently for congenital spinal deformities. The high initial costs and the overall cost-effectiveness of EOS® has also been a matter of contention. Dietrich et al. noted that EOS® system required significantly higher number of yearly radiographic examinations (4077/year) compared to DR (2602/year) despite the lowered labour cost per examination due to lower examination time.⁵⁸ Faria et al. analyzed the cost effectiveness of EOS® and compared it with standard radiography. They noted that the health benefits from the reduction of radiation were very small. They worked out that for EOS® to be cost effective, either its use should lead to better patient outcomes or a minimum output of 15,100 radiographic examinations/year should be obtained.⁵⁹ The upfront cost of installing an EOS® system includes the cost of the machine and its corresponding software for image acquisition, 2D processing and 3D reconstruction – and currently amounts to a staggering 500,000 euros.¹⁵ Presently, there are over 350 EOS® systems installed worldwide – with the majority being in Europe and North America. To extend its reach worldwide and become the new ‘gold standard’, possibility of cost optimization by indigenous manufacturing in developing countries must be explored. The founders of the technology and researchers worldwide are engaged in bringing about improvements to the EOS® system – but demonstration of clear-cut advantages over other imaging techniques in terms of applicability, benefits in diagnosis and treatment planning, radiation dose, patient comfort and financial break-even point for hospitals is necessary before its widespread usage will be seen in research and clinical setting.

Fig. 4.

A radiolucent chair for EOS® imaging in non-ambulatory patients or patients unable to stand.

8. Conclusion

EOS® imaging is a revolutionary technology – low radiation dose, ability to generated 3D reconstructed images and scan the entire skeleton in a short span of time are its most obvious advantages. Presently, EOS® imaging has been found extensive application in the area of spinal deformity. With its current limitations notwithstanding, we believe that this technology will continue to evolve in the future and researchers and clinicians alike will discover more ways to harness its tremendous potential in spinal disorders.

Declaration of conflict of interest

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

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Informed consent

Informed written consent was taken from each patient included in this study.

Submission declaration

The work described has not been published, is not under consideration for publication elsewhere, its publication is approved by all authors and tacitly or explicitly by the responsible authorities where the work was carried out, if accepted, it will not be published elsewhere in the same form, in English or in any other language, including electronically without the written consent of the copyright-holder.