THE DEFORMITY OF IDIOPATHIC SCOLIOSIS MADE VISIBLE BY COMPUTER GRAPHICS

Size: px

Start display at page:

Download "THE DEFORMITY OF IDIOPATHIC SCOLIOSIS MADE VISIBLE BY COMPUTER GRAPHICS"

Frank Dean
5 years ago
Views:

1 THE DEFORMITY OF IDIOPATHIC SCOIOSIS MADE VISIBE BY COMPUTER GRAPHICS F. R. HOWE, R. A. DICKSON From St. James s University Hospital, eeds We present a method ofvisualising spinal deformities in three dimensions using conventional radiographs and computer graphics. The shape of the spinal column can be determined from the anteroposterior and lateral radiographs and displayed in any projection. In patients with adolescent idiopathic scoliosis, the fundamental lesion, an abnormal lordosis, can be demonstrated without the need for additional views. The method is applicable to other spinal deformities and may help to elucidate their three-dimensional shape. Radiography of spinal deformity displays its threedimensional structure in two dimensions, depending on the planes of projection. These are usually anteroposterior and lateral in reference to the patient, but the three-dimensional structure can be inferred from measurements of vertebral rotation (Nash and Moe 1969; Mehta 1973; Aaro and Dahlborn 1981). By combining the information from two radiographic views ofthe same deformity, a three-dimensional image can be reconstructed with the aid of a computer (Suh 197; Brown et al. 1976; De Smet et al. 1980; Hindmarsh, arsson and Mattsson 1980; Raso, Gillespie and McNiece 1980). This image can then be manipulated geometrically to be displayed in two dimensions, varying the apparent viewpoint of the observer. The conventional views of idiopathic scoliosis depict the deformity obliquely because of its axial rotation. The size of the deformity as measured by the Cobb angle (Cobb 198) is thus an underestimate of its true size which will be seen when the deformity is viewed en face (du Peloux et al ; Deacon, Flood and Dickson 198). Similarly, the lateral radiograph does not project the true lateral view of the apex of the curve. The computer system allows multiple views of the deformity to be reproduced from the information F. R. Howell, MA, FRCS Ed, Medical Research Council Research Fellow R. A. Dickson, MA, ChM, FRCS, Professor of Orthopaedic Surgery University Department oforthopaedic Surgery, St. James s University Hospital, eeds S9 7TF, England. Correspondence should be sent to Mr F. R. Howell British Editorial Society of Bone and Joint Surgery 030l-620X/89/308l $2.00 J Bone Joint Surg [Br] l989;7l-b: obtained from two standard radiographs, and avoids the need for multiple exposures. The results confirm that a consistent inappropriate lordosis is the underlying deformity in late-onset idiopathic scoliosis (Somerville 1952; Roaf 1966; Dickson et al. 198). MATERIAS AND METHODS The eeds MRC epidemiological survey of the natural history of idiopathic scoliosis has identified a group of 962 schoolchildren aged eight to 1 3 years with a degree of truncal asymmetry on forward bend testing, from a population of 15,793 children. This cohort is being followed longitudinally, both clinically and radiographically. Postero-anterior and lateral radiographs are taken each year, using the low-dose technique and standardised positioning ofardran et al. (1980), and these form the basis for this study. The computer method requires identification of the same anatomical points on both views of the spine (Suh 197). The points chosen are the bases of the pedicles, and the mid points of the upper and lower end plates (Figs 1 and 2); these have been shown to be the most consistent (Brown et al. 1976; De Smet et al. 1982). The points are marked on the radiograph, and, by means of a sonic digitising device, their co-ordinates are entered into a microcomputer for processing and storage. The data for each point from each pair of radiographs are merged to provide three Cartesian co-ordinates, for four points per vertebra, for all the vertebrae which can be visualised. A three-dimensional model of the spine is thus recorded. The projection of this model on to a plane produces a two-dimensional image, which may then be rotated in any desired direction by a mathematical transformation. VO. 71-B, No. 3, MAY

00 F. R. HOWE, R. A. DICKSON This can produce images relating to the postero-anterior and lateral radiographic projections, or to any other chosen viewpoint.

plane, xt, to the old, x, and to the third co-ordinate, z, as follows: xt = x. cosa - z. sin A The microcomputer is programmed to perform this using the basic language.

2 00 F. R. HOWE, R. A. DICKSON This can produce images relating to the postero-anterior and lateral radiographic projections, or to any other chosen viewpoint. The transformation itself is by a simple algebraic formula, which for rotation of a point x, y, z by an angle A around the vertical y-axis, will relate the horizontal projection in the new vertical plane, xt, to the old, x, and to the third co-ordinate, z, as follows: xt = x. cosa - z. sin A The microcomputer is programmed to perform this using the basic language. The limiting factor in the technique as we describe it is the identification of the chosen anatomical points. The errors involved have been quantified (Hindmarsh et al. 1980). Magnification and other problems have been overcome by other means. Validation. The anteroposterior and lateral views of mounted museum specimens from cases of idiopathic scoliosis were processed as above and a computer representation created. The same specimens were then radiographed at 10#{176} intervals of rotation around a vertical axis. The computer diagram at any given rotation and the radiograph taken at that same rotation can be compared, and the similarity is quite obvious (Figs 3 and ). This confirms that the two standard views contain all the necessary information to visualise the spine at any chosen degree of rotation. Fig. 1 Fig. 2 The identification of the four points to be digitised on the anteroposterior and the lateral radiographs. RESUTS In the cohort of 962 children followed up for two years, 69 have an idiopathic type of curve, with a Cobb angle greater then 10#{176} and rotation of the vertebral body towards the convexity ofthe curve. In all 69 cases there is an inappropriate lordosis at the apex of the curve. There may or may not be a frank lordosis visible on the lateral view, but when the computer model is rotated then an apical lordosis becomes obvious (see Fig. 5). This view corresponds to the Stagnara or eeds lateral projection (Dickson 1987). A normal, straight spine does not show any such lordosis, but a rotated view does produce artefactual curves. These could be confused with minor idiopathic type curves on a plain radiograph if viewed obliquely (see Fig. 6) If a progressive idiopathic curve is followed over a period of time, then the abnormal lordosis can be seen to increase on the oblique reconstruction in parallel with other measures of the deformity (see Fig. 7). DISCUSSION The deformity of idiopathic scoliosis is known to be a rotated lordoscoliosis, but single plane radiographs cannot adequately demonstrate its three-dimensional nature. Two orthogonal views do however contain all the information required to describe the deformity in any plane and the use of this can avoid the need for multiple radiographs and reduce the radiation dosage to a child during observation or treatment. In our cohort, we found a fundamental lordosis in every case of progressive idiopathic scoliosis. In some of these, the initial curve was small (5#{176} or less), and a diagnosis of early idiopathic scoliosis was not practical because of the size of the curve in relation to positional and other variables. However, these early curves declare themselves by progression, and reconstruction of their early deformities shows that an underlying lordosis is present. When this lordosis is followed, it increases and starts to rotate as the deformity progresses (Fig. 7). When there are greater degrees of vertebral rotation, the overall shape of the patient is spuriously kyphotic, as the vertebral bodies come to face backwards. We do not use this definition of kyphosis, which relates to the sagittal plane of the patient and not of the spine, because it has led to confusion in discussion of the pathogenesis of these curves in the past (De Smet et al. 198; Stokes, Bigalow and Moreland 1987). A normal straight spine, viewed obliquely, shows an apparent scoliosis which simulates a double structural thoracic and lumbar curve (Fig. 6). On a plain radio- THE JOURNA OF BONE AND JOINT SURGERY

$THE DEFORMITY OF IDIOPATHIC SCOIOSIS MADE VISIBE BY COMPUTER GRAPHICS 01 Fig. 3 \ > Fig.$

3 THE DEFORMITY OF IDIOPATHIC SCOIOSIS MADE VISIBE BY COMPUTER GRAPHICS 01 Fig. 3 \ > Fig. Radiographs of a museum specimen of a scoliotic spine taken at 0, 0, 90 and 1 30 of rotation, with computer reconstructions of the specimen at the same positions of rotation. VO. 71-B, No. 3, MAY 1989

4 02 F. R. HOWE, R. A. DICKSON 7 7 i :N 1 c: :: :: v t: c cz:i & A z1 n c; p E Fig. S Fig. 6 Reconstruction of a case with right thoracic idiopathic scoliosis: projections at 0, 35, and 90 of rotation, equivalent to anteroposterior, oblique and lateral views. An abnormal lordosis is visible in the oblique view. Reconstruction of a normal straight 50 oblique and lateral projections. in anteroposterior, r 1 : & c Fig. 7 Fig. 8 Reconstructions of the curve shown in Figure 5 over three successive years, all viewed at the same obliquity, showing progression of the lordosis. Top view of a left thoracic idiopathic scoliosis. graph, the lumbar component is indistinguishable from true idiopathic scoliosis, and is apparently a rotated lordosis with rotation of the vertebral body to the convexity of the curve. The thoracic component, however, can be differentiated because rotation is in the opposite direction; the displacement of the vertebral body is to the concavity of the curve, showing it to be a rotated kyphosis, not a rotated lordosis. This phenortienon may explain, in part at least, the observation of a group of curves within a screened population, in which vertebral rotation is non-standard (Armstrong et al. 1982). The concept of Eulerian buckling of a unstable flexible column as the cause of idiopathic scoliosis (Gordon 1978) would predict that these curves will not progress as the idiopathic type may well do. We have used rotation about the vertical axis to show the lordotic deformity in idiopathic scoliosis, but it is as easy to reproduce a top view of the spine (Fig. 8), and this has been claimed to monitor more accurately the progression of the deformity (De Smet et al. 1983). Again, the lordosis may be seen. Scoliotic deformity is complex and does not lend THE JOURNA OF BONE AND JOINT SURGERY

5 F. R. HOWE, R. A. DICKSON 03 itself readily to quantification. The Cobb angle at least has the merit that it is widely used. The computer method we have described goes no further in quantifying the deformity, nor is it intended that it should. It does, however, make the qualitative nature of the deformity easily visible. No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. REFERENCES Aaro S, Dahlborn M. Estimation of vertebral rotation and the spinal and rib cage deformity in scoliosis by computer tomography. Spine 1981 ;6:60-7. Ardran GM, Coates R, Dickson RA, Dixon-Brown A, Harding FM. Assessment of scoliosis in children : low dose radiographic technique. Br J Radiol 1980:53:16-7. Armstrong GWD, ivermore NB III, Suzuki N, Armstrong JG. Nonstandard vertebral rotation in scoliosis screening patients : its prevalence and relation to the clinical deformity. Spine 1982:7:50-. Brown RH, Burstein AH, Nash C, et al. Spinal analysis using a threedimensional radiographic technique. J Biomech 1976;9: Cobb JR. Outline for the study of scoliosis. Am Acad Orthop Surg!nstr Course ect 198;S: Deacon P, Flood BM, Dickson RA. Idiopathic scoliosis in three dimensions : a radiographic and morphometric analysis. J Bone Joint Surg [Br] 198:66-B: De Smet AA, Tarlton MA, Cook T, Fritz S, Dwyer SJ III. A radiographic method for three-dimensional analysis of spinal configuration. Radiology 1980;137 :33-8. De Smet AA, Burr DB, Cook T, Goin JE, Fritz S. Evaluation of radiographic landmarks for three-dimensional spinal analysis. In: Jacobs RR, ed. Pathogenesis of idiopathic scoliosis : Proceedings of an international conference. Conference held Sept Chicago: Scoliosis Research Society, 1982:5-59. De Smet AA, Tariton MA, Cook T, Bemdge AS, Asher MA. The top view for analysis of scoliosis progression. Radiologi 1983:17: De Smet AA, Asher MA, Cook T, et al. Three-dimensional analysis of right thoracic idiopathic scoliosis. Spine 198:9: Dickson RA, awton JO, Archer IA, Butt WP. The pathogenesis of idiopathic scoliosis : biplanar spinal asymmetry. J Bone Joint Surg [Br] 198;66-B:8-1S. Dickson RA. Scoliosis: how big are you? Orthopedics 1987:l0:88l-7. du Peloux J, Fauchet R, Faucon B, Stagnara P. e plan d #{233}lection pour l examen radiologique des cypho-scolioses. Re, Chir Orthop 1965:51 : Gordon JE. Structures: or win things don t/all down. Harmondsworth, Middlesex: Penguin Books td Hindmarsh J, arsson J, Mattsson 0. Analysis of changes in the scoliotic spine using a three-dimensional radiographic technique. J Biomech 1980; 13: Mehta MH. Radiographic estimation of vertebral rotation in scoliosis. J Bone Joint Surg [Br] 1973:SS-B:5l Nash C Jr, Moe JH. A study of vertebral rotation. J Bone Joint Surg [Am] 1969:51-A : Raso J, Gillespie R, McNeice G. Determination of the plane of maximum deformity in idiopathic scoliosis. Orthop Trans 1980::23. Roaf R. The basic anatomy of scoliosis. J Bone Joint Surg [Br] 1966:8-B : Somerville EW. Rotational lordosis: the development of the single curve. J Bone Joint Surg [Br] 1952:3-B:21-7. Stokes IA, Bigalow C, Moreland MS. Three-dimensional spinal curvature in idiopathic scoliosis. J Orthop Res 1987:5: Suh CH. The fundamentals of computer aided X-ray analysis of the spine. J Biomech 197:7: VO. 71-B, No. 3, MAY 1989

The effect of vertebral rotation of the lumbar spine on dual energy X-ray absorptiometry measurements: observational study

The effect of vertebral rotation of the lumbar spine on dual energy X-ray absorptiometry measurements: observational study JCY Cheng HL Sher X Guo VWY Hung AYK Cheung Key words: Absorptiometry, Bone density; Densitometry, X-ray; Lumbar vertebrae; Scoliosis "#$%& "# X HKMJ 2001;7:241-5 The Chinese University of Hong Kong, Prince