Difference in whole spinal alignment between supine and standing positions in patients with adult spinal deformity using a new comparison method with slot-scanning three-dimensional X-ray imager and computed tomography through digital reconstructed radiography

Spinal deformity is a three-dimensional abnormality. Alignment is usually assessed in standing radiographs in both the sagittal and coronal planes using the Cobb angle [4]. When a patient presents with neurologic symptoms, details of the abnormality are usually assessed by referring to computed tomography (CT) images or magnetic resonance images (MRI) obtained with the patient in a supine position. Previous studies of adolescent idiopathic scoliosis (AIS) report major Cobb angles 7–10° smaller in the supine position than in the standing position due to changes in the gravitational loading direction [1, 21, 28, 30]. Therefore, to understand the pathology of the whole spine, it is important to determine the difference in the curve measurement between the supine and standing positions, including alignment. A drawback of previous reports is the potential error related to different measurement methods used for X-ray and other imaging modalities (CT or MRI) to compare the Cobb angle, which may lead to misinterpretation or controversy [1, 20, 21, 28, 30].

To our knowledge, a precise comparison of supine and standing measurements using the same modality and measurement method to overcome this drawback has not been reported. Whole spine alignment in both the coronal and sagittal planes, including the pelvic parameters, with regard to a positional difference has not been investigated. Furthermore, previous studies investigated positional differences in the Cobb angle only in young patients with AIS. The difference in alignment has never been investigated in a population of patients with adult spinal deformity (ASD). In contrast to AIS, the chief complaint of patients with ASD is usually mechanical back pain with or without radicular pain to the lower extremities. In most cases, the patients are aware of the symptoms when standing and tend to stoop with back pain, whereas the symptoms disappear when lying on a bed. Therefore, it is important to elucidate the positional differences in the deformity in older adults.

The purposes of this study were to: 1) Establish a method of comparing whole spine alignment between the supine and standing positions with minimal measurement error; and 2) Clarify the difference in the whole spine alignment between the supine position and the weight-bearing standing position in adult-to-elderly patients with spinal deformity.

Adults with spinal deformities generally have a lower health-related quality of life. The present findings are consistent with findings of previous studies in that the patients in our population had poor low-back pain-related quality of life. Fairbank et al. [12] reported that the mean ODI score in a normal Caucasian population is 10.2. In addition, Tonosu et al. [27] reported that the mean ODI in a normal Japanese population is 8.73 (8.80 in men, 8.66 in women) when corrected for age, and a tendency for the ODI to gradually increase with advancing age. The mean ODI score in the present study (31.4%) was much greater than these values and mean SRS-22 score in our study (2.9) was lower than the normal value [5]. To clarify the pathology leading to the lower quality of life in patients with ASD, a precise comparison of supine and standing measurements using the same modality and measurement method is needed.

The difference in curve magnitude between standing and supine positions was studied previously in AIS patients. Keenan et al. [20] investigated the difference in the Cobb angle between the supine position (reconstructed CT) and standing position (conventional X-ray) in 52 patients with AIS having a mean age of 14.6 years. They reported a mean Cobb angle on standing radiographs of 51.9°, which was a significantly greater value than the mean Cobb angle on supine CT images of 40.5° [20]. The Cobb angle measurement, however, includes a fundamental error related to the selection of the vertebral endplates. In another study [24], spinal parameters were measured on 50 anteroposterior radiographs of scoliotic spines, on 6 separate occasions each by 4 orthopedic surgeons using the Cobb method; they reported that the 95% confidence limit for intraobserver variability was 4.9° and that for interobserver variability was 7.2°. Thus, due to the inherent inter- and intra-observer error in conventional X-ray measurements, it is difficult to conclude whether there are quantitative differences in the curve magnitude due to the positioning of the subjects.

In the present study, we transformed the CT dataset (supine position images) into an EOS-like dataset with the DRR technique, which generates two-dimensional X-ray-like images using the same calibration parameters and geometry as the EOS cabin. The projected anteroposterior and lateral DRRs were used as inputs for routine-like EOS stereoradiographic spine modeling with sterEOS software. The EOS data (standing position) and the CT-generated DRRs (supine position) were then compared. In the present study, we confirmed the reliability of this comparison method using repeatability tests with ICC in the 24 subjects, and found that the ICCs were excellent in all the values, both in CT-generated DRRs and EOS measurements. Thus, the inherent error due to measurement and different imaging modalities was overcome by the comparison with the EOS 3D measurement system.

In adult-to-elderly patients with spinal deformity, sagittal malalignment is more important than coronal alignment [14]. In contrast to AIS, adult-to-elderly patients with spinal deformity complain that they tend to stoop with back pain or radicular pain of the lower extremities, and the symptoms disappear when they lie down. Therefore, the positional change of the 3D whole spine alignment should be clarified, especially in aging adults. In the present study with patients having a mean age of 60.1 years, the Cobb angle and Rotation angle of the major curve, mostly the thoracolumbar area, were significantly greater in the standing position than in the supine position. Lumbar lordosis in the standing position was significantly more kyphotic and the pelvis was significantly more retroverted, with a smaller SS and greater PT, compared with that in the supine position. The axial skeleton of human beings is aligned as a chain of balance in the standing position with the “cone of economy” in healthy subjects [8]. The standing full-body sagittal alignment and balance in reference to the gravity line were recently described in a healthy adult population using EOS radiographs [17]. The results of the present study suggest that the “cone of economy” principle deteriorates in patients with spinal deformity, making it difficult to maintain the standing position due to back pain with or without radicular pain.

PI is a fundamental parameter for spinopelvic standing alignment [10, 11, 22], and is believed to be a constant value in each individual. The value was, however, significantly greater in the standing position than in the supine position. Mangion et al. [23] measured PI on radiographs of 30 fetuses, 30 children, and 30 adults, and found that PI considerably increases during the first few months of life, continues to increase during the early years, and stabilizes at around the age of 10 years. In our previous study [16], PI was positively correlated with age, even in healthy adult subjects. PI increased around 10° on average from 20 to 70 years of age, probably due to sacroiliac osteoarthritis [16]. This finding is compatible with a previous review article on sagittal pelvic alignment parameters, which described that PI tends to increase with age in both normal and scoliotic subjects [29]. In the present study, we demonstrated a significant difference in PI between the supine and standing positions. The reconstructed coronal CT image of a case with degenerative kyphoscoliosis (Fig. 4a) shows osteoarthritis of the sacroiliac joint with a vacuum phenomenon, subchondral osteosclerosis, and cyst formation (Fig. 5a). The preoperative PI was 57.1° in standing EOS (49.8° in supine CT) and the PI decreased to 50.2° in standing EOS (52.2° in supine CT) following spinopelvic correction surgery with bilateral S2 alar-iliac fixations (Fig. 4b) [25]. Furthermore, the vacuum in the sacroiliac joint disappeared and the cyst became smaller after surgery as a result of fixation by the screws stabilizing the sacroiliac joints (Fig. 5b). Sacroiliac joints are the only sites that can move between the base of the sacrum and the acetabulum, thereby affecting the PI value. Therefore, our data suggest that the sacroiliac joint moves abnormally due to osteoarthritic changes in ASD cases, leading to changes in the PI between the supine and standing positions. The high ICC for both the supine and standing position values suggests that the difference is a true difference, and not due to measurement variability. PI is considered a set characteristic of the individual spinopelvic shape, but it can change due to degenerative processes of the spine. Thus, it is important to remember that the aging and degenerative processes affect the whole spinopelvic alignment and balance in reference to patient positioning.

The findings in the present study offer a basic pathology of ASD with positional change in thoraco-lumbar to pelvic alignment, suggesting that a main focus of the correction and fusion surgery is the lumbar to pelvic area and further extension of the fusion to upper thoracic area is not necessary. Furthermore, we believe that the results contribute not only for spine surgeons but also for the professionals involved in the conservative management of patients with back pain and spine deformities by keeping in mind the sagittal spinopelvic profile in the supine position which could represent a starting point. Potential limitations of this study are, however, the wide age range of the patients but relatively small sample size. Aging and spondylotic change affect the spino-pelvic flexibility. Thus, although the results in the present study is statistically accurate, we need to continue the investigation with further sample size.

We established a method for comparing spinopelvic alignment between the supine and standing positions by converting CT DICOM data into an EOS-like dataset, a DRR technique. Comparison revealed that the Cobb angle and axial apical rotation of the major curve, mostly in the thoracolumbar area, were significantly greater in the standing position than in the supine position. Lumbar lordosis in the standing – weight-bearing – position was significantly less lordotic and the pelvis was significantly more retroverted, with a smaller SS, greater PT, and even greater PI, compared with that in the supine position.