2 Basic Principles
2.1 Definitions and Terminology
Over the last several decades a spinal deformity language has evolved with which we all need to be familiar so that we can communicate with each other either by written or spoken word. This is based on the excellent work of the Scoliosis Research Society. 1 Some of the terms we have come to use have no obvious meaning, such as “nonstructural” scoliosis, and, indeed, we should probably look at many of these different terms and consider a better use of words.
2.1.1 Planes and Deformities
The spine has three planes: the frontal or coronal plane, the sagittal or lateral plane, and the transverse or axial plane (▶ Fig. 2.1). In the coronal plane the spine should be straight and if it is not then a scoliosis (lateral spinal curvature) is present (▶ Fig. 2.2). Meanwhile, in the sagittal plane there are four natural curvatures (▶ Fig. 2.2): cervical and lumbar lordoses (curvatures convex anteriorly) and thoracic and sacral kyphoses (curvatures convex posteriorly) (▶ Fig. 2.3). These curvatures in the sagittal plane develop in the early years of life. The cervical lordosis develops as infants raise their heads when they begin to look around their environment whereas the lumbar lordosis develops later when they start sitting and standing up. In the transverse plane, the vertebrae should be symmetrical but are not, even in the so-called “normal” spine.
Fig. 2.1 The planes of the body. (Reproduced with permission from Newton P, O’Brien M, Shufflebarger H, et al. Idiopathic Scoliosis: The Harms Study Group Treatment Guide. Stuttgart/New York: Thieme; 2010: 17.)
Fig. 2.2 Diagrammatic representation of deformities in the coronal and sagittal planes.
(Reproduced with permission from Greenspan A. Scoliosis and anomalies with general effect on the skeleton. In: Orthopaedic Radiology. A Practical Approach. 2nd ed. London: Gower Medical Publishing; 1992: Figure 28.3.)
Fig. 2.3 The four natural curvatures of the spine in the sagittal plane: cervical and lumbar lordoses and thoracic and pelvic kyphoses. (Reproduced with permission from Newton P, O’Brien M, Shufflebarger H, et al. Idiopathic Scoliosis: The Harms Study Group Treatment Guide. Stuttgart/New York: Thieme; 2010: 17.)
2.1.2 Nonstructural and Structural Deformities
These terms do not describe the deformities that bear their name in any meaningful way. Nonstructural deformities would include those secondary to a leg length inequality, or secondary to muscle spasm provided by a painful focus in the spine (e.g., osteoid osteoma or adolescent disk hernia) (▶ Fig. 2.4). Structural spinal deformities are due primarily to a spinal problem and the etiological classification described later lists all the conditions associated with structural deformities. Perhaps structural deformities could be termed “primary” spinal deformities and nonstructural deformities “secondary” to some other pathological problem. There is another rather poor differentiation between nonstructural and structural deformities. Structural deformities are regarded as having a rotational component in addition to the lateral curvature and nonstructural deformities do not—however, this is often not so. Whereas structural deformities due to idiopathic scoliosis do have rotation (▶ Fig. 2.5), those with a congenital hemivertebra as the keystone to the curvature are often not associated with rotation (▶ Fig. 2.6) while leg length inequality produces a secondary nonstructural scoliosis in the lumbar spine where there is already a natural lordosis and so there is obligatory spinal rotation 2 (▶ Fig. 2.4a) (biplanar asymmetry). 3
Fig. 2.4 Nonstructural scoliosis. (a) PA radiograph of a nonstructural lumbar curve secondary to a leg-length inequality. (b) A teenage boy with a nonstructural lumbar scoliosis caused by muscle spasm from an adolescent disk hernia. Note the erythema ab igne from the use of heat pads because of severe left lumbar pain. (Reproduced with permission from Newton P, O’Brien M, Shufflebarger H, et al. Idiopathic Scoliosis: The Harms Study Group Treatment Guide. Stuttgart/New York: Thieme; 2010: 18.)
Fig. 2.5 PA radiograph of an idiopathic lumbar curve. The spinous processes have been labelled with black triangles, and it can be clearly seen that they rotate toward the concavity of the curve, as is always the case with structural scoliosis. The back of the spine is therefore shorter than the front and so all structural scolioses are lordoscolioses. (Reproduced with permission from Newton P, O’Brien M, Shufflebarger H, et al. Idiopathic Scoliosis: The Harms Study Group Treatment Guide. Stuttgart/New York: Thieme; 2010: 18.)
Fig. 2.6 Scoliosis due to a hemivertebra showing no rotation.
2.2 Curve Characteristics
2.2.1 Named Vertebrae
▶ Fig. 2.7 shows the typical appearances of a moderate idiopathic thoracic scoliosis. It is customary to look at posteroanterior (PA) X-rays as we look at our patients mainly from the back (which, incidentally, reduces the radiation dosage to the developing breast and thyroid in females by first travelling through the thickness of the neck and torso 4). The vertebrae maximally tilted at each end of the curvature are referred to as the end vertebrae, in this case T5 above and T11 below, whereas the apical vertebra is that at the center of the scoliotic curve, in this case T8. If there happens to be an even number of vertebrae in the curve then the two central vertebrae are both apical; if there is an odd number of vertebrae then there is one apical vertebra at the center of the curvature, as in this case. However, the true structural curvature extends from the first neutral (nonrotated) vertebra above to the first neutral vertebra below and these are above and below the upper and lower end vertebrae, respectively, in this case T3 above and T12 below.
Fig. 2.7 PA radiograph of an idiopathic thoracic curve. Again, being a structural curve, there is rotation with the spinous processes turning toward the curve concavity. The end vertebrae for Cobb angle measurement (the most tilted vertebrae at the top and bottom of the curve) are T5 above and T11 below. The first neutral vertebra above is T3, two above the upper end vertebra; T12 is the lower neutral vertebra. The apical vertebra is T8. (Reproduced with permission from Newton P, O’Brien M, Shufflebarger H, et al. Idiopathic Scoliosis: The Harms Study Group Treatment Guide. Stuttgart/New York: Thieme; 2010: 19.)
2.2.2 Curve Size
Although there is more than one method of measuring curve size—such as the Ferguson 5 method of drawing longitudinal lines at the top and bottom of the curve and measuring the angle of intersection—the universal method of measuring curve size is that of Cobb. 6 Lines are drawn along the upper end plate of the upper end vertebra and the lower end plate of the lower end vertebra and these lines are then produced until they intersect and that is the Cobb angle (▶ Fig. 2.8). Of course, with mild to moderate curves these lines do not intersect on the X-ray film and thus it has been traditional to drop perpendiculars and measure the angle of intersection in that way. But it is quite unnecessary to do that and possibly damages the X-ray emulsion and so a Cobbometer is strongly recommended (based upon the Oxford Orthopaedic Engineering Centre Cobbometer, you can construct one yourself) 7 (▶ Fig. 2.9). You should have one in your pocket to measure curve size in clinic. It is not necessary to draw any lines on the X-ray film itself. Furthermore, it is far more accurate than when measurements are made drawing lines and then applying a protractor. However, nowadays, using the Picture Archiving and Communications System (PACS) or alternative X-ray system technique, there are tools available to make all the measurements which simplifies matters greatly with considerable precision.
Fig. 2.8 PA radiograph of a right thoracic curve with lines drawn along the lower borders of the vertebrae below the apex and the upper borders of the vertebrae above the apex showing that it is easy to pick out maximally tilted end vertebrae, namely T6 and L1.
Fig. 2.9 Measuring the Cobb angle with the Oxford Cobbometer (a protractor with a vertical free hanging needle). The upper border of the instrument is first aligned with the upper surface of the upper end vertebra, and the protractor dial is set to zero. When the upper border of the instrument is then aligned with the lower surface of the lower end vertebra, the needle gives the Cobb angle. (Reproduced with permission from Newton P, O’Brien M, Shufflebarger H, et al. Idiopathic Scoliosis: The Harms Study Group Treatment Guide. Stuttgart/New York: Thieme; 2010: 19.)
Of course these measurements are made on a PA radiograph of the patient with each vertebra progressively more rotated toward the apex. This clearly means that the curve has rotated out of the frontal plane, mostly at the apex. The less the curve size, the closer it is to the frontal plane and the less the amount of apical rotation. A curve of 60 degrees is therefore much more than twice as big as a curve of 30 degrees because it is seen less en face. Stagnara realized this and in his scoliosis series of X-rays the really important ones were the true plan d’élection views 8 (▶ Fig. 2.10). If the apical region is rotated, say, 30 degrees from the frontal plane, then the patient is turned or the X-ray beam is turned 30 degrees so that a true AP plan d’élection view is taken. Similarly, a lateral view of the patient is not a true lateral projection and the patient or X-ray beam must now be rotated 90 degrees to the AP plan d’élection view to obtain the true lateral. Lateral views of the patient therefore appear to be more and more kyphotic with increasing curve size when of course there is no such thing as kyphoscoliosis as we shall see again later. Rotating a coat hanger and looking at it or taking photographs is a useful way of understanding the problems of spinal rotation (▶ Fig. 2.11). In simple terms the Cobb angle changes its apparent magnitude simply by changing the plane of projection. Notwithstanding, inevitably, all publications about the size of a scoliosis before and after treatment use arithmetic data, means, and percentage changes that, because of the effect of rotation, are effectively meaningless. A curve that changes from 60 to 30 degrees as a result of treatment represents much more than a 50% improvement.
Fig. 2.10 Radiograph of a right thoracic curve. (a) PA view of the patient. (b) A true PA plan d’élection view showing a much bigger curve. (c) Lateral view of the patient showing the spurious appearance of kyphosis. (d) True lateral view, showing the essential lordosis. These are all views of the same deformity! (Reproduced with permission from Newton P, O’Brien M, Shufflebarger H, et al. Idiopathic Scoliosis: The Harms Study Group Treatment Guide. Stuttgart/New York: Thieme; 2010: 20.)
Fig. 2.11 (a) PA view of a coat hanger, showing an angle of 60 degrees. (b) The coat hanger has been rotated 45 degrees and now the Cobb angle registers only 30 degrees because the coat hanger is not being seen en face as it is in (a). (c) When the coat hanger is turned a further 45 degrees there is no angle at all. (Reproduced with permission from Newton P, O’Brien M, Shufflebarger H, et al. Idiopathic Scoliosis: The Harms Study Group Treatment Guide. Stuttgart/New York: Thieme; 2010: 21.)
2.2.3 Vertebral Rotation
This is an important measurement because it provides useful information as to the severity of the deformity particularly in terms of its flexibility and thus correctability. A thoracic curve of 40 degrees with a lot of apical rotation is a much more severe clinical deformity in terms of rib hump and stiffness than a 40-degree curve with less rotation.
The method of Nash and Moe was an early technique described for recording apical vertebral rotation 9 (▶ Fig. 2.12). It is still useful and probably the best method of measuring rotation other than on three-dimensional (3D) magnetic resonance imaging (MRI) or low dose computed tomography (CT) scans using the measurement tools available on the machine. If a vertebra is not rotated, then the pedicles will be equidistant from the sides of the vertebral body. Rotation with a lordoscoliosis always occurs with the posterior elements turning toward the concavity of the curve (concordant rotation) (▶ Fig. 2.5) and therefore the convex pedicle moves farther away from the side of the vertebral body while the concave pedicle is seen progressively less well—rather like an eclipse of the sun. Nash and Moe’s method is to grade the amount of rotation in pluses but it is more precise to measure the distance between the lateral side of the convex pedicle and the convex side of the vertebral body and express it as a percentage of the total width of the vertebral body. 1
Fig. 2.12 Nash-Moe method of measuring vertebral rotation. The distance that the convex pedicle has moved from the convex side of the vertebral body is graded from 0 to 4 (0, normal; 1, +; 2, ++; 3, +++; 4, ++++).
(Reproduced with permission from Greenspan A. Scoliosis and anomalies with general effect on the skeleton. In: Orthopaedic Radiology. A Practical Approach. 2nd ed. London: Gower Medical Publishing; 1992: Figure 28.14.)
Another method of measuring rotation is using the protractor devised by Perdriolle 10 (▶ Fig. 2.13). There are a number of longitudinal lines on the protractor and the line that bisects the convex pedicle indicates the degree of rotation of that vertebra when read off from the bottom of the protractor. It has not been validated for the whole length of the spine indicating that Nash and Moe’s method is preferred. Nowadays rotation can be readily measured by MRI scanning using the inbuilt measuring tools (▶ Fig. 2.14).
Fig. 2.13 Measuring rotation using Perdriolle’s protractor. The protractor comprises diverging lines and is laid over the PA radiograph, with the side lines of the protractor aligned with the sides of the vertebral body. The line that bisects the convex pedicle indicates the degree of rotation of that vertebra. (Reproduced with permission from Newton P, O’Brien M, Shufflebarger H, et al. Idiopathic Scoliosis: The Harms Study Group Treatment Guide. Stuttgart/New York: Thieme; 2010: 21.)
Fig. 2.14 Measuring vertebral rotation by MRI scanning. (a) Measuring the rotation of the upper end vertebra. This is the reference plane to measure apical rotation. (b) Measuring the rotation of the apical vertebra; the lines are drawn by the MRI software which will also calculate the angle of apical rotation—in this case, 10 degrees. Remember that this is measured in the supine position whereas Nash and Moe can be measured on any PA radiographs, erect or supine, lateral bending, or maximum stretch films.
(With thanks to Dr. J Rankine, Consultant Spine Radiologist, The Leeds Hospitals.)
Another method of measuring rotation is that described by Min Mehta many years ago and is really only applicable to early onset (infantile) idiopathic scoliosis 11 (▶ Fig. 2.15). At the curve apex, the angle of the neck of the ribs subtended against a line along the vertical axis of the vertebra is measured on each side. These are referred to as the rib vertebra angles (RVAs). On the convex side the ribs droop more and thus the convex RVA is smaller than on the concave one. If the difference in the angles (the rib vertebra angle difference [RVAD]) is bigger than 20 degrees, this suggests the likelihood of a progressive curve, along with a bigger Cobb angle. This is in no way surprising, the bigger the curve and the more rotation then the more likely the deformity will worsen. However, double structural infantile curves are always progressive as in ▶ Fig. 2.15c,d.
Fig. 2.15 (a) Infantile idiopathic thoracic scoliosis. Although the Cobb angle is 30 degrees, the RVAD is only 5 degrees. (b) Two years later, the deformity has almost resolved. (c) Infantile idiopathic double-structural scoliosis to the right in the thoracic region and to the left in the thoracolumbar region. Although the Cobb angles measure only 28 and 25 degrees, respectively, and the RVAD of the thoracic curve measures only 6 degrees, infantile double-structural curves are always progressive unless treated. (d) One year later both curves have increased significantly, particularly the lower one. Unfortunately, no therapeutic action was taken in the interim. (Reproduced with permission from Newton P, O’Brien M, Shufflebarger H, et al. Idiopathic Scoliosis: The Harms Study Group Treatment Guide. Stuttgart/New York: Thieme; 2010: 22.)
Going back to late-onset (adolescent) idiopathic scoliosis, another method of perceiving the amount of rotation is to measure the degree of pseudokyphosis on the lateral radiograph of the patient. As soon as the lordotic area at the apex of the curve moves out of the frontal plane it points progressively farther backward and gives the spurious appearance of kyphosis when the lateral X-ray of the patient is viewed (▶ Fig. 2.10c). It should be remembered that kyphosis means that the back of the spine points backward and in structural scoliosis it is the front of the spine moving progressively more backward as the lordosis twists out of the frontal plane. It can be readily appreciated that curves with a bigger Cobb angle have a bigger pseudokyphosis on the lateral X-ray of the patient. Thus, the lateral view of the patient is an index of the amount of scoliosis and rotation. If the PA Cobb angle is, say, 40 degrees and the amount of apical rotation 10 degrees versus a Cobb angle of 40 degrees and apical rotation of 20 degrees then the one with the apical rotation of 20 degrees will have a much bigger pseudokyphosis on the lateral X-ray of the patient and a bigger rib hump if the curve is thoracic. Thus when the Lenke classification brings in the amount of so-called kyphosis it loses its objectivity and much of its value. 12
2.2.4 Curve Patterns
There are single, double, and multiple curve patterns (▶ Fig. 2.16). With idiopathic scoliosis it is commonly thought that the right thoracic single curve pattern is the most common but on inspection of the radiographs this is not so with vertebrae above and below the so-called major curve being rotated in the opposite direction indicating the presence of a double or multiple curve pattern. The paper by Cruickshank et al. on this subject is well worth studying. 13 If the apex of the curve is in the thoracic region (T2 to T11) then it is described as a thoracic curve. It is nearly always at T7 to 9. Thoracolumbar curves have the apex at T12 or L1 whereas cervicothoracic curves have their apex at C7 or T1. Lumbar curves are apical from L2 to L4 (nearly always at L2) and the unusual lumbosacral curve is apical at L5 or S1. Double curves tend to be thoracic and thoracolumbar/lumbar whereas the common classic thoracic double major is a double curve with a thoracic curve and a cervicothoracic curve above 14 (▶ Fig. 2.17). Moe was the first to describe the importance of this curve because if there is a right thoracic curve as the major curve and a cervicothoracic above this in the opposite direction then the shoulders will be tilted with the left shoulder higher than the right (the signe d’épaule 15). It would be inappropriate just to deal with the thoracic curve surgically and, thus, by improving that only, there would necessarily be further tilting of the cervicothoracic curve above, and an ugly higher left shoulder.
Fig. 2.16 A selection of common curve patterns in idiopathic scoliosis. (a) Right thoracic curve; these are always apical between T7 and T9. (b) Right thoracolumbar curve: these are apical at T12 or L1. (c) Right thoracic and left lumbar double-structural curve; lumbar curves are always apical at L2. (Reproduced with permission from Newton P, O’Brien M, Shufflebarger H, et al. Idiopathic Scoliosis: The Harms Study Group Treatment Guide. Stuttgart/New York: Thieme; 2010: 23.)
Fig. 2.17 PA radiograph showing a double thoracic curve. There is a right thoracic curve and then above it a left cervicothoracic curve so that the left shoulder is higher than the right (the signe d’épaule). The Lenke classification of this pattern of curvature is Type 2 AN.
The direction of a curve is described according to its convexity. Thus right thoracic curves or left thoracolumbar or lumbar curves are common.
Spinal balance (compensation) can be easily appreciated if there is a single curve—say, a single right thoracic curve. Above and below this are compensatory curves bringing the head up straight and the lumbar spine and pelvis straight down below so that the head is centered over of the pelvis. This can be assessed using a plumb line dropped down from the vertebra prominens and see whether it cuts through the middle of the natal cleft. Perhaps a more precise method can be done radiologically 16 (▶ Fig. 2.18). For spinal balance, the upper and lower compensatory curves should have the same magnitude; if the lower one is bigger than the upper one, then spinal balance has been lost and this is referred to as decompensation (▶ Fig. 2.19).
Fig. 2.18 Measuring spinal balance (compensation). The sum of the upper and lower compensatory curves is always equal to the size of the structural curve. (Left) When the spine is in perfect balance, the upper and lower compensatory curves are of equal magnitude. (Right) When the spine lists to the side of the convexity of the curve (decompression), it does so because the lower compensatory curve is bigger than the upper compensatory curve. In this case, the spine is decompensated by 10 degrees. (Reproduced with permission from Newton P, O’Brien M, Shufflebarger H, et al. Idiopathic Scoliosis: The Harms Study Group Treatment Guide. Stuttgart/New York: Thieme; 2010: 23.)
Fig. 2.19 Decompensation. There is a marked list of the torso to the side of the convexity of the curve.
It is worth exploring these compensatory curves in a little more detail. 13 For example, a radiograph of a right thoracolumbar curve shows that a vertebra or two above and below the major curve are still rotated in the direction of the concavity of the major curve. It is much easier to demonstrate this illustratively using a thoracolumbar or lumbar curve because the background air and abdominal gases provide much better contrast to the bony anatomy (▶ Fig. 2.20). The lower end vertebra is L3 and clearly below L3 is the compensatory curve. It can, however, be seen that in the major structural curve to the left the spinous processes are of course rotated toward the right, as with concordant rotation. But, as we go into the lower compensatory curve, the spinous processes are still rotated to the right and yet these compensatory curves are convex to the right and not the left. Therefore, there is discordant rotation at the beginnings of the upper and lower compensatory curves indicating that these, at least to begin with, are kyphoses (direction of rotation in the same direction as that of tilt). This is best shown with the lower compensatory curve where L4 at the beginning of the lower compensatory curve is clearly seen to be rotated to the right and indeed the L5 vertebra is to a much lesser extent. These compensatory curves are therefore kyphoscolioses in contradistinction to the structural curve which is a lordoscoliosis. Indeed, for 3D balance of the spine, the central lordoscoliosis must be balanced by kyphoscolioses above and below. We shall deal with this in more detail in the next chapter where there are important “must know” concepts.
Fig. 2.20 The lower part of a left thoracolumbar curve with the end vertebra at L3 and L4 and L5 still rotated in the same direction as the structural curve.