Comparison between different radiographic methods for evaluating the flexibility of scoliosis curves

OBJECTIVE: To compare different radiographic methods of spine evaluation to estimate the reducibility and flexibility of the scoliosis curves. METHODS: Twenty one patients with Lenke types I and III adolescent idiopathic scoliosis (AIS) were included. Radiographic evaluations were made preoperatively on the orthostatic, supine decubitus with lateral inclination to the right and left and supine positions with manual reduction, with support in the apex of each curve on the X-ray table. On the day of surgery, when the patient was anesthetized, radiography was taken with longitudinal traction through divergent forces, holding under the arms and ankles, and with translational force at the apex of the deformity for curve correction. After one week, a post-operative radiography was performed in orthostatic position. RESULTS: The correction and flexibility of the main thoracic and thoracic/lumbar curves were statistically different between the supine radiographs, manual reduction, modified traction under general anesthesia, lateral inclination and postoperatively. The modified maneuver for traction under general anesthesia is the one which showed greater flexibility, besides presenting higher radiographic similarity to postoperative aspects. CONCLUSION: Among the radiographic modalities evaluated the study under anesthesia with traction and reduction showed better correlation with postoperative radiographic appearance. Level of Evidence IV, Case Series.


INTRODUCTION
Surgical correction of scoliosis is aimed at obtaining coronal, sagittal and axial balance of the spine with the lowest fused levels. 1,2 The advantages of selective arthrodesis are: less blood loss, preservation of movable segments and decreased risk of pseudoartrose. 3 To this end, the preoperative analysis of the structure of curves and evaluation of its flexibility is a key part in surgical planning. 4 The preoperative flexibility of the curves can be evaluated using radiographs in lateral tilt when standing upright or supine radiographic study under traction. 5,6 Radiographs in supine position with maximum slopes for preoperative evaluation have been performed most commonly, since they are inexpensive and easy to interpret. 7,8 However, several other methods are also used, such as gradients in the standing position, longitudinal traction in supine position, tilt with fulcrum at the apex of the curve and sedated radiographic study. A standardized assessment of curve flexibility in patients with adolescent idiopathic scoliosis (AIS), preoperatively, allows better comparison of results from different studies, besides being a useful tool in surgical planning. 9 Therefore, the aim of this study is to evaluate the flexibility of scoliotic curves in patients with AIS by standard radiographs in supine and lateral tilt preoperatively and after anesthesia with manual correction, as well as its correlation with post-surgical corrective results.

MATERIALS AND METhODS
A convenience sample of 21 patients with AIS structured curves was studied according to the criteria of Lenke et al. 10 treated surgically with posterior instrumentation with pedicle screws 11 . Three male and 18 female patients were enrolled in this study, aged between 13 and 19 years old (mean age 15 years and 3 months old). The following inclusion criteria were used: patients with Lenke curve types I and III. Exclusion criteria were: Lenke curves types 2, 4, 5 and 6; patients who underwent reoperation, need for thoracoplasty, osteotomy and anterior approach. Patients with thoracic curve to the left were also excluded. All patients underwent surgical correction of scoliosis and instrumentation by posterior approach. In all cases, only monoaxial pedicle screws were used and no hooks or sublaminar strings were used in the assembly. Density of the implants (ratio between the number of implants used in the assembly and the total number of sites available for implants) was assessed, as recommended by Suk et al. 12 The method used for curve correction maneuver was "derotation" of the rod in the concavity, as described by Cotrel et al. 13 Before surgery, each patient underwent the first radiograph in the standing position and the second in supine decubitus with right and left leaning. The third radiograph was performed in supine position with manual reduction, with support in the summits of each curve in the X-ray table, as described by Kleinman et al. 14 All inclination radiographs and manual reduction were supervised by a spine surgeon during ambulatory follow up. On the day of surgery, with the patient under anesthesia, the fourth radiograph was performed, immediately before surgery, with the patient in supine position. For this radiograph longitudinal traction by divergent strength under armpits and ankles held by two spine surgeons. Translational force at the apex of the thoracic deformity was also held, for correction of the curve. One week after scoliosis correction surgery, the fifth radiographic study, in orthostatic position, was performed. In all analyzed radiographs, the Cobb angle was measured for primary thoracolumbar curve and thoracic curve. The lower limit for main thoracic curve was the apex above or equal to T11/T12 disc. Data was organized and tabulated, then subjected to statistical analysis to calculate significance. A level of significance of 5 % (p<0.050) was adopted. The Statistical Package for Social Sciences (SPSS) program was used in its version 19.0, to obtain the results. Tables 1 and 2 show flexibility and correction of main thoracic curve (T MAIN) and thoracic-lumbar/lumbar curve (TL/L). Table 1 shows that there was no statistical difference between the supine radiographs, manual reduction (RED MAN), modified traction under general anesthesia (TRA MOD), lateral tilt and postoperative (PO) in the main thoracic curves, as in Table 2 with the thoracic-lumbar/lumbar curve (TL /L). The modified traction maneuver under general anesthesia showed greater flexibility and reducibility of the main thoracic curves and also in thoracic-lumbar/lumbar curve (TL/L), besides resembling the postoperative radiographic study. Table 3 shows the results of post-hoc analysis with Wilcoxon signed-rank test, comparing the five variables pairwise tended to show differences. Regarding the main moderate thoracic curves (≤ 65 degrees), proportionally we found no statistically significant difference between the radiographs during the reduction maneuver. The same was observed in severe curves (65 degrees) as shown in Table 4. No statistically significant difference in thoracic-lumbar/ lumbar curves was found in both serious as in moderate difference, as shown in Table 5.

DISCUSSION
Flexibility radiographs have been recommended to help determine the surgical technique and the levels to be selected in the correction of scoliosis. Lateral bending radiographs were described for the accuracy of the surgical correction with first generation instruments (Harrington system) and are considered the gold standard for this purpose. However, with the use of pedicle screws, the predictive value of active side slope supine radiographs began to be questioned. Gotfryd et al. 7 demonstrated that the method is predictive of the correction achieved exclusively with pedicle screws. However, in some situations, it is difficult to obtain cooperation from patients to perform the exam, a fact that complicates the evaluation of flexibility of curves. 15,16 A false interpretation of the stiffness of a curve may induce the surgeon to make unnecessary fusions. Thus, other methods of evaluation of the curve flexibility have been proposed. Various types of radiographs have been advocated to assess the flexibility of the curves, such as manual reduction on the apex of curve. 14 Cheung and Luk 17 evaluated flexibility through radiographs with slope associated with corrective force, with the fulcrum at the apex of the curve, and compared these results with those obtained on radiographs with slope and postoperative correction achieved in thoracic curves after posterior fusion. However, the authors found no significant difference. 17, 18 Davis et al. 19 described the radiographic technique traction under general anesthesia in the operating environment, which, according to the author, best resemble the radiographic appearance after corrective surgery. In our study, we used three types of deformity correction techniques to evaluate the flexibility of the Table 3. Comparison between variables, pairwise through Wilcoxon signed-rank test, adjusted by Bonferroni correction*.

Variables pair p
Main   curve: X-ray tilt supine radiographs with manual reduction of the curve at the apex of the deformity and radiography with traction after anesthesia.
In this study, we performed radiography with manual reduction as described by Kleinman et al. 14 and observed no significant difference compared to other correction methods. However, when analyzed separately, the location of the curve, no significant difference for both main thoracic curves and for the thoracic-lumbar/lumbar was observed. This led us to suggest that the efficacy of the test was not affected by the location of the curve. Similarly, Kleinman et al. 14 observed that the effectiveness of manual reduction was also not altered by the location of the curve, pattern (single or double curve) or scoliosis etiology. Regarding supine radiographs, we believe that the main drawback is the difficulty to standardize the force to be exerted during correction, besides the degree of relaxation of the patient during the examination, factors that directly affect the degree of correction of the deformidade. 3 Manual reduction in pronation is still useful in predicting the behavior of the curves at the levels that will not be submitted to arthrodesis. Traditionally, in traction radiographs are performed in patients less capable of performing lateral inclination (not collaborative patient or in case of neuromuscular scoliosis). 1 Traction under general anesthesia is a relatively new technique, first reported by Davis et al. 19 The flexibility of curve in traction under general anesthesia is enhanced by muscle relaxation of the patient, there is no discomfort and does not required the patient's collaboration. 17 Through it, one can more easily standardize the degree of force being exerted by the examinator. 4 In present study, it was found that the tensile radiographs show greater similarity with surgical correction against the main thoracic curves. Regarding lumbar curves, the reducibility was larger than the correction achieved postoperatively, which may be explained by the very muscular relaxation obtained with the method. To our knowledge, no previous study had described this phenomenon. Regarding severe curves (greater than 65 Cobb degrees) no statistical difference was noted in traction under general anesthesia for both main thoracic curves as in thoracic-lumbar/lumbar curves. Radiography with traction under general anesthesia showed greater flexibility when compared to supine lateral tilt and tilt with fulcrum at the apex of the curve for high degree curves (Cobb angle>65 degrees) and rigid ones; however, this result was not statistically significant, possibly due the small sample of patients with curves of high degree. In a more recent study, traction under general anesthesia was compared to the same side slope in the supine and thoracic-lumbar/lumbar main thoracic curves and, when divided into severe curves (65 degrees) and moderate, they were again equivalent, with a tendency to greater flexibility for traction under general anesthesia in main severe thoracic curves. 20 A limiting factor for potential clinical application of this type of radiography is that the surgeon cannot give the patient a preoperative planning set before general anesthesia geral. 4 It is also difficult to obtain quality radiographs in the operating room for evaluation and measuring curves. In the present study, the sample was selected and a recruitment bias may be present.

CONCLUSION
It was shown that the radiographic postoperative appearance in Lenke's type I and III adolescent idiopathic scoliosis, treated with a posterior spinal fusion, can be better predicted from analysis of radiographic imaging under general anesthesia with traction and manual reduction.