Keywords

Introduction

Radiographic evaluation is essential in the management of scoliosis. X-rays provide objective insight into a patient’s structural deformity, often validating a proper yet subjective history and physical. Radiographic measurements from posteroanterior and lateral standing films provide the language we use to communicate about patients and compare results. Since the advent of the Risser sign and the Cobb angle, through the evaluation of spinopelvic alignment and the sagittal plane, radiographic measurements have provided reliable, objective measurements for the diagnosis and treatment of scoliosis.

The radiographic analysis of scoliosis in the twentieth century concentrated primarily on coronal deformities; coronal alignment continues to occupy a position of primacy in the evaluation and treatment of childhood scoliosis. In the management of adult deformity, however, emphasis has shifted toward the correction of sagittal malalignment. Analyzing the sagittal plane is more complex than analyzing the coronal or axial planes, owing to the natural kyphosis and lordosis of the spine. This complexity has driven the development of parameters to simplify and guide the management of adult deformity. The work of Roussouly and others have characterized the normal curvatures of the spine and, importantly, its relationship to the pelvis [1, 2, 10, 12, 15, 16, 18, 25, 38]. Building upon this, parameters defining pathological alignment in the sagittal plane were evaluated using patient-reported outcome studies, leading to the development of the SRS-Schwab classification system for adult scoliosis [7, 28–30, 32].

History of Radiographic Parameters in Scoliosis

X-ray measurements have been a keystone in the evaluation of scoliosis since the advent of the Risser and Cobb measurements in the 1950s. The Risser sign, a measurement of iliac ossification, has been used to evaluate skeletal maturity and has persisted in the study of adolescent idiopathic scoliosis. Likewise, John Cobb’s end plate-to-end plate angular measurement still serves as the primary radiographic finding in coronal deformity and is used to diagnose, discuss, classify, and treat these curves. The Cobb measurement, in particular, has been used in multiple classification systems designed to predict the natural history and surgical outcome from the angle and location of coronal curves. Ponseti and Friedman; James, Collis, and Ponseti; and Harrington combined Cobb angles with other factors, e.g., curve location, rotation, progression, and length, as well as patient maturity, to form distinct classification systems intended to guide management [9, 21].

In 1983, King published a classification system based entirely on posteroanterior upright and bending x-rays of the thoracolumbar spine, combining Cobb angle measurements with curve patterns, locations, relative flexibilities, and vertebral axial rotations [13]. It also required more than just the Cobb angle, codifying many of the terms used in deformity evaluation today, e.g., the center sacral line, stable and neutral vertebrae, and a “flexibility index” derived from comparing lateral bending in thoracic and lumbar curves. This system was designed to guide selection of fusion levels in adolescent idiopathic scoliosis and was the first classification system to be widely adopted.

The widespread adoption of the King classification offered an excellent opportunity to study a large population of deformity patients. Systematic examination ultimately exposed the weaknesses in the classification; more significant than its reliability pitfall was its lack of consideration for the sagittal plane [36]. Several subsequent AIS classification schemes improved on the King system, adopting its attention to the coronal curve but adding parameters to characterize pathologic sagittal alignment. The Lenke classification accounted for the chief shortcomings of the King system, improving reproducibility and adding a modifier for lordosis as measured on lateral films [17]. The Lenke Classification for AIS served as a starting point for the radiographic examination and classification of adult deformity, although the disease processes and important measures for each would prove very different.

Adult Deformity and the Cone of Economy

The study of adult deformity, separate from its juvenile counterpart, has grown rapidly over the past few decades. The application of key radiographic parameters and classification systems used in AIS and other juvenile scoliotic diseases has proved largely ineffective [7]. Emphasis has shifted away from coronal realignment—frequently the primary goal of juvenile scoliosis surgery—toward alignment correction in the sagittal plane.

Spinal alignment is more complicated in the sagittal plane than it is in the coronal or axial planes. Whereas the goal of coronal and axial correction is to straighten and de-rotate, correction in the sagittal plane must account for the natural spinal lordosis and kyphosis. Appropriate alignment in the sagittal plane has been shown to improve outcomes in the adult scoliotic population [14, 32]. As such, the parameters that constitute pathologic sagittal malalignment, including compensatory measures outside the thoracolumbar spine, have been the subject of increasing study [19, 20, 22, 24, 26, 33, 34].

The “cone of economy” as published by Dubousset in 1994 describes the range of standing postures in which the body can remain balanced without support and with minimum energy expenditure [3]. Those unable to maintain a standing posture in the center of the cone demand the muscles and joints of the spine and legs to compensate, which can result in fatigue, pain, and disability. Many of these patients require external aids such as walkers or canes to stand. Studies on flatback syndrome have noted the clinical sequela of iatrogenic sagittal malalignment since the 1970s. That, with the quantification of normal and pathologic spinal curvatures, has driven the development of many radiographic parameters [5].

Multiple studies have attempted to characterize radiographic alignment in the sagittal plane. Stagnara, in 1982, proposed normal reference values for thoracolumbar lordosis and kyphosis, as well as for sacral slope [35]. His findings—that there were wide and irregular variations between healthy subjects for both values, belying the idea of a “normal” lumbar lordosis or thoracic kyphosis—have been born out in subsequent studies. The study did note the intra-patient relationships between lordosis, kyphosis, and sacral slope, which would also be a theme of sagittal analysis going forward.

Quantitative Radiographic Evaluation for Sagittal Plane

A landmark study by Jackson et al. in 1994 compared healthy adult volunteers with patients reporting low back pain, noting a wide but largely similar range of values for lordosis ad kyphosis in healthy patients, as well as similar C7 plumbline values, between the two groups [11]. However, they noted a critical proximal shift in segmental lordosis and a decrease in sacral inclination in back pain patients, representing possible compensatory mechanisms for any loss of lordosis at the lower lumbar levels in these patients.

An emphasis on sagittal alignment led to widespread use of the sagittal vertical axis (SVA), determined by measuring the AP translation relative to S1 of a cephalad vertebrae. Gelb et al. examined the horizontal distance between a plumbline dropped from the middle of the C7 vertebral body to the anterior superior corner of the sacrum on a standing lateral x-ray, noting the tendency for SVA to move anteriorly in older subjects, while sagittal alignment remained neutral in asymptomatic patients [6]. Van Royen et al. examined the horizontal distance between a plumbline dropped from the tip of the C7 spinous process to the anterior superior S1 vertebral body in a single patient with an ankylosed spine to isolate the relationship between posture and SVA (Fig. 3.1) [37]. They pointed out that small angular adjustments in the lower extremities resulted in significant changes to SVA measurements, implying that SVA ought to be considered in the context of compensatory postural mechanisms. Further studies pointed out inadequacies in SVA measurements: a dependence on arm position, a lack of correlation to “functional” standing position, and a poor correlation between a cervical plumbline and the true center of gravity. Still, poor clinical outcomes have been shown to correlate linearly with increasing sagittal malalignment as measured with a C7 plumbline, indicating SVA as an important parameter for health-related quality of life.

Fig. 3.1
figure 1

Schematic diagram for sagittal vertical axis (SVA)

Full size image

The incorporation of pelvic parameters led to a fuller understanding of sagittal alignment and its contribution to quality of life outcomes. In 1998, Legaye and Duval-Beaupere et al. proposed pelvic incidence (PI), a measure quantifying the interface between the spine and the pelvis [4, 16]. Defined as the angle between the line from the femoral head axis to the midpoint of the superior S1 end plate and the line perpendicular to the S1 end plate, PI is morphologically unique to each individual and is independent of postural changes. PI, a fixed value, correlated well with LL; patients with a high PI were also likely to have a high LL. They postulated that a chain of interdependence existed between the pelvic and spinal parameters. Other parameters proposed by Legaye include sacral slope (SS), defined as the angle between the S1 end plate and the horizontal on a lateral standing x-ray, and pelvic tilt (PT), defined as the angle between the line from the mid-axis of the femoral heads to the midpoint of the superior S1 end plate and the vertical on a lateral standing x-ray (Fig. 3.2).

Fig. 3.2
figure 2

Schematic diagrams for pelvic parameters

Full size image

Attention to the pelvic parameters revealed the importance of pelvic compensation for sagittal malalignment. Earlier papers had characterized the effect of small, angular changes in posture around the hip axis on the SVA, but in the late 1990s and early 2000s, efforts were made to quantify this compensation [1, 12].

Pelvic Parameters and the Sagittal Plane

The high degree of patient-to-patient variability in spinal sagittal alignment complicates the study of pathologic malalignment. Roussouly et al., in 2005, published a classification system describing categories of lumbar lordosis in relation to curve apices and spinopelvic relationships in 160 normal subjects [25]. In addition to describing an association between PI and LL, they found a reciprocal relationship between the sacral slope and pelvic tilt and established the equation: SS + PT = PI. Relating spinal sagittal curves to pelvic parameters lends meaning to these measurements that otherwise vary so wildly as to make radiographic identification of pathology, in many cases, difficult if not impossible.

Spinopelvic alignment criteria have been shown to correlate with patient-reported outcomes. Previous studies sought to delineate, without success, a relationship between coronal deformity and clinical outcomes. However in the sagittal plane, Glassman et al. demonstrated that positive sagittal malalignment is predictive of poor clinical health status; their two studies revealed that symptom severity increased linearly with worsening positive sagittal malalignment and that restoring normal sagittal alignment improved clinical symptoms [7, 8].

The identification of sagittal alignment as a primary driver in adult scoliosis patient satisfaction, both pre- and post-op, set the stage for the establishment of the SRS-Schwab classification system, which has undergone several iterations since the early 2000s. Based originally on a prospective analysis of 95 patients, the initial study in 2002 identified L3 and L4 end plate obliquity in the frontal plane, lateral olisthesis, lumbar lordosis, and thoracolumbar kyphosis as radiographic parameters that correlated with increased pain [29]. This led to the first SRS-Schwab classification system, which grouped patients into three categories based on lumbar lordosis and L3 coronal obliquity. The system was then expanded; the curves were further characterized by their coronal deformity apex, degree of lordosis, and intervertebral subluxation. Coronal curve categories were prescriptive—different curve types demanded tailored surgical approaches—while the lordosis and subluxation modifiers stratified patients into clinical groups, with higher grades indicating worsening HRQOL.

The work of Glassman et al. led to the inclusion of a global sagittal balance modifier in later iterations [8]. Ultimately, outcome-driven criteria led to refining the SRS-Schwab classification system to include a coronal curve modifier and three sagittal alignment modifiers: PI-LL, SVA, and pelvic tilt (Fig. 3.3). The coronal modifier describes the coronal curve type: T for thoracic only, L for thoracolumbar or lumbar only curves, D for double curves (T and TL/L curves both >30°), and N for no coronal curves>30°. The three sagittal modifiers, stratifying patients by clinical symptomatology, were established based on HRQOL studies:

Fig. 3.3
figure 3

SRS-Schwab classification for adult spinal deformity

Full size image

PI-LL, calculated by subtracting the lumbar lordosis from pelvic incidence: 0 (non-pathologic) for PI-LL < 10°, + (moderate deformity) for PI-LL between 10° and 20°, and ++ (marked deformity) for PI-LL>20°

Global alignment, assessed by measuring the translational distance from the posterior superior S1 body to a plumbline dropped from the middle of the C7 vertebral body: 0 (non-pathologic) for SVA< 4 cm, + (moderate deformity) for SVA between 4 and 9.5 cm, and ++ (marked deformity) for SVA<9.5 cm

Pelvic tilt, measured as the angle between the line from the mid-axis of the femoral heads to the midpoint of the S1 plate and a vertical line: 0 (non-pathologic) < 20°, + (mild deformity) between 20° and 30°, and ++ (marked deformity) > 30°

The SRS-Schwab classification provides a framework for interpreting radiographic parameters by incorporating the current base of knowledge regarding sagittal alignment, spinopelvic parameters, and compensatory measures [27]. The classification has been validated using patient-reported outcomes for both operative and nonoperative patients [30, 31]. When combined with clinical judgment, the SRS-Schwab classification can guide treatment in adult scoliosis patients. Prospective studies have validated the classification in follow-up studies, relating improvement in SRS-Schwab classification with higher HRQOL scores [32].

Future Directions

Sagittal alignment and spinopelvic parameters have allowed surgeons to pursue evidence-based radiographic goals anchored in patient-reported outcomes. Still, complications persist, and outcomes are not perfect. Several parameters show promise with regard to predicting complications and patient dissatisfaction beyond those described by the SRS-Schwab classification. Patients with severe sagittal malalignment, unsurprisingly, have poorer outcomes than those with mild or moderate deformities. High preoperative PT and SVA have been specifically shown to increase the risk of poor surgical outcomes. Poor postoperative alignment is a common cause of patient dissatisfaction and low HRQOLs; careful and adequate planning is critical in providing the proper degree of sagittal correction tailored to each individual patient. Postsurgical reciprocal changes, e.g., alterations in TK after lumber realignment surgery, have been observed. Surgical planning will need to account for these changes, although they are currently still difficult to predict.

Staying true to the global nature of malalignment, concomitant cervical deformity is also not uncommon in adult thoracolumbar disease. 53% of thoracolumbar deformity patients have cervical deformity, either as a compensatory mechanism or as a primary disease process [33]. New cervical deformity has also been found in 48% of post-op patients, as has improvements in preoperative cervical deformity following thoracolumbar realignment [19, 20, 22, 34]. This is a logical extension of the chain of interdependence connecting the pelvis and thoracolumbar spine. Radiographic parameters to quantify and predict cervical deformity are currently being studied, including T1 angle, T1 spinopelvic inclination, C2-T1 SVA, and cervical lordosis. T1 spinopelvic inclination also correlates with HRQOL outcome scores in adult scoliosis patients [23, 26]. Caudal to the spinopelvic axis, studies are being directed at knee flexion, another compensatory mechanism with similar biomechanics to pelvic tilt.

Predicting outcomes from adult scoliosis surgery has proven difficult. Patients on either end of the disease spectrum tend to improve after surgery; it is those who fall between the extremes—the majority of patients—that have mixed results. Poor outcomes occur even after a successful sagittal realignment. This emphasizes the need for further studies to determine if there are radiographic parameters that can be further optimized to increase chances of obtaining good clinical results.

Conclusion

Radiographic parameters, clinically backed with patient-reported outcomes, are both useful in the baseline evaluation of and the treatment selection for adult spinal deformity patients. With the spinopelvic parameters and the SRS-Schwab classification in mind, a framework has been established to deliver a more personalized surgical approach, resulting in better clinical outcomes and greater patient satisfaction.