Accuracy on the preoperative assessment of patients with adolescent idiopathic scoliosis using biplanar low-dose stereoradiography: a comparison with computed tomography

Although computed tomography (CT) is commonly used to diagnose the scoliotic spine in patients with adolescent idiopathic scoliosis (AIS) preoperatively, it is limited by the high radiation and prone scanning position. Recently, a new biplanar stereoradiography (EOS) was used to image the scoliotic spine in an upright posture with significantly less radiation in non-severe AIS subjects. However, its reliability to assess preoperative AIS patients remains unreported. Hence, the purpose of this study is to compare the scoliotic curvature between prone (CT) and upright positions (EOS) in preoperative AIS patients.

Thirty-three pre-operative AIS patients (mean age:18.4 ± 4.2) were recruited. EOS was used to scan the whole thoracic spine at upright position. Whereas on the same day, a conventional CT scan was used to evaluate the spine in prone position. The three-dimensional reconstruction of EOS and CT of the spine were then generated. Using previous validated techniques, multiple scoliotic parameters in both modalities were determined. The agreement between the two modalities was compared using the Bland-Altman test, whereas the correlation was assessed by the intraclass correlation coefficient (ICC).

The mean ICC (prone and upright) of intra-rater/inter-rater reliabilities for the measured parameters were 0.985,0.961/0.969,0.903, respectively. Thoracic Cobb angles, intervertebral wedging and lumbar lordosis correlated significantly between upright EOS imaging radiographs (62.9 ± 9.3°,6.4 ± 2.9° and 48.8 ± 12.4°) and prone CT (47.3 ± 10.0°,5.8 ± 2.7° and 27.9 ± 11.4°; P < 0.001). The apical vertebral wedging and apical intervertebral disc wedging showed a good correlation among the two modalities (upright, 6.5 ± 3.5° and 6.4 ± 2.9°; prone, 6.5 ± 3.6° and 5.8 ± 2.7°; R² ≥ 0.94; P < 0.01). Similarly, there was significant correlation in apical intervertebral rotation (R² = 0.834; P < 0.01) between the prone CT (3.4 ± 3.0°) and upright EOS (3.8 ± 3.2°). In addition, the Cobb angle was significantly larger in upright EOS (62.9 ± 9.3°) than in prone CT (47.3 ± 10.0°, P < 0.01) position. There was significant underestimation on scoliotic severity in the prone position when compared with upright position.

Importantly, the image acquisition and reconstruction from EOS can better provide accurate three-dimensional spinal representations of the scoliotic curvature in preoperative AIS patients. Moreover, our findings suggested that scoliotic curvatures in preoperative AIS patients can be largely represented by both imaging modalities despite the difference in body positioning.