Introduction: In situ contouring is meant to give the shape of the spine to the rod and then the shape of the rod to the spine. Thus, it is used in order to set up the instrumentation as well as to reduce the spinal deformity. This technique was born in 1993, when we presented our first scoliosis correction results (CT scan study of vertebral derotation) with the rod rotation technique during the French SRS (GES). Our great disappointment with the rod rotation technique forced us to try to find a different correction method.

Scoliosis is the consequence of vertebral rotation. Each vertebra turns about a different axis which results into a global torsion of the spine. This torsion will yield characteristic modifications. On the frontal x-ray view one can notice the maximum projection of the deformity, usually estimated by means of the Cobb angle, whereas on the sagittal x-ray view a flat back will be observed. Indeed, scoliosis flattens sagittal physiological curvatures. Hyperkyphosis may occur only between two scoliotic curves (two adjacent flat back segments) or in case of vertebral rotation higher than 90° when the sagittal projection corresponds to frontal structures. In this last case, the maximum deformity is projected on the sagittal view. The vertebral rotation will also pull on the ribs, thus creating the rib hump.

Classical Surgical Techniques: Nowadays there are several classical correction techniques for scoliosis treatment.

Over the last decades, Harrington developed the distraction-compression technique, then Eduardo Luque proposed the spinal translation technique, and latter on Cotrel and Dubousset developed the rod rotation method that revolutionised spine surgery.

By pulling on the concave side of the spine, the distraction-compression technique is intended to reduce the deformity shaft while dragging along the apex in a pure translation movement. The distraction is applied mainly onto the flexible segments, far from the apex. Therefore, the apex will hardly modify its relative position with regard to the other vertebrae. Besides, there is a high risk of spinal cord stretching on the concave side at the apex level.

Furthermore, this technique is often associated with a high rate of post-operative flat back and requires postoperative cast and brace wear as the fixation remains fragile. Last but not least, the traction technique does not solve the rotation problem. On the contrary, traction increases the torsion forces and leads to higher rotation constraints.

The spinal translation used to be performed by means of metallic wires passing under the lamina that were tightened around the rod. This technique of scoliosis correction was based on a totally different correction mechanism with regard to Harrington’s one. Indeed, medialisation of the apex results into a spontaneous increase of the intervertebral gap at the extreme levels of the curve. This distraction is automatic, and as a matter of fact it is impossible to apply it as the wires are sliding on the rod. The problem with spinal translation is that it cannot control rotation, neither with screws nor with hooks. Frontal X rays show that the anterior spine will always be located outside the rods pulling the posterior arch. This technique was improved by Asher and Chopin, who introduced screws and hooks. However it is still very difficult to decide on which side one should work, i.e. concave or convex side. The problem of rotation is still unsolved as anterior spine projection onto the x-rays is still next to and outside the rods.

Rod rotation is the most popular technique nowadays as it allows rather good global correction. However, the thoracic correction does not follow the pathology path and therefore has no impact of vertebral rotation. This technique allows only slight adjustments, often very difficult to perform especially in the frontal plane. In 1993, we reviewed 52 scolioses operated with the rod rotation technique. All patients had undergone pre- and post-operative CT scan, so we could estimate the rotation correction. The results were highly disappointing as the vertebral rotation at the LEV (lower end vertebra) decreased of only 2.3°, while at the UEV (upper end vertebrae) level it was 1.1° higher after surgery, and at the apex level it remained almost unchanged (0.4° smaller). In conclusion, correction was obtained by vertebral translation, horizontalisation, forward and backward pushing of the vertebra, without any derotation. Several examples clearly reflected this mechanism, proved by the mobilization of the vertebrae with regard to the aorta.

When looking at the path described by the vertebra, one can easily notice that the different techniques described above do not allow to follow the deformity path. Thus, the thoracic vertebra goes frontward and turns to the right. This circular movement has a posterior centre of rotation. Vertebral translation does not follow this path as it moved about the arch cord. The rod rotation performs a circular movement about an anterior centre of rotation. The correction and deformity paths describe an ellipse. We can conclude that these techniques will lead to high constraints within the spine. Hence the risk of neurological structures damage during correction manoeuvres.

At the lumbar level, the apex moves backwards and to the left. Thus, it will describe an arch about a posterior rotation centre. The vertebra traction will move along the cord of this arch while the rod rotation will strictly follow the reverse pathology path. As the convex rod is linked to a hook or a screw, it will lead to a combined force of internal traction and anterior push. This convex push increases rotation by turning the screw in the sense of pathologic deformity. Therefore, the projections of the screws on the frontal x-rays will be oriented outside the rod, while the normal axis of the pedicle is about 20° oblique oriented toward the inside.

In Situ Contouring: Given these conclusions on the failure of traditional methods, we tried to develop the in situ contouring for a more efficient scoliosis correction. We would like to remind surgeons that the technique was designed to give the shape of the spine to the rod in order to set the instrumentation up, then to give the shape of the rod to the spine in order to reduce the deformity. All surgeons performed in situ rod bending at least once in their practice without being aware of.

However, when performing in situ contouring, some security rules have to be strictly respected. First of all, the rod must be free to move, so implants must be closed around the rod but remain unlocked until the correction manoeuvres are finished. The rod mobility will allow the automatic spine stretching/shortening without dangerous constraints. Vertebrae must slide along the rod by means of the implants, i.e. screws and hooks, solidly attached to them. In other words, the spine, i.e. vertebrae, must be mobilised. To do so, the benders must be placed close to the implants. The other reason is to avoid high lever arms that would lead to high risky forces (loads).

The correction principles are based on the vertebrae movement in space in order to enable a frontal and sagittal correction while working into the axial plane. To do so, the rod must have specific mechanical features: initial short elastic and long plastic domains. The correction manoeuvres on the rod will modify this mechanical behaviour and at the end of the correction manoeuvres the plastic domain will decrease wile the elastic one will increase. An initially too elastic rod would require stronger manoeuvres with regard to the residual correction, which may present some supplementary risk for neurological structures.

The levels to be instrumented are selected as usual, as in situ contouring does not modify rules usually used in order to determine the strategic vertebrae. The strategic vertebrae are selected depending on the information provided by bending tests. Thus all discs that do not open in both directions will be included into the fused segment.

Thoracic scoliosis: Similar to the rod rotation, the working rod is concave one in the thoracic spine. In contrast with the rod rotation technique, every second vertebra will be instrumented, going as close as possible from the apex. As neither distraction nor compression is performed, laminar hooks are not compulsory anymore. In our practice, we use pedicular hooks above T10 and screws below T10.

The rod will be contoured towards the inside and backwards for all instrumented levels. These manoeuvres will allow the medialisation of the apex while restoring kyphosis. At the same time, these actions will lead to a derotation of the apex. The contouring manoeuvres are performed iteratively starting from the apex towards the limits of the curvature through successive manoeuvres in the frontal plane and in the sagittal plane. Contouring is over when required correction is obtained and when the rod modified its mechanical behaviour and became too elastic to allow further contouring.

The apex follows the deformity path. Thus the vertebra moves backwards and towards the inside, describing a circular movement similar to the deformity path in the opposite sense. Therefore, three-dimensional correction of both mild and severe (> 100°) thoracic scolioses. However, the purpose of the surgery should not be to have a straight vertical rod, but to obtain the best possible spinal balance with the best possible correction in the three planes.

Lumbar scoliosis: Lumbar scoliosis may also be treated by in situ contouring. In this case, the working rod is the convex one. This rod will be bent towards the inside and forwards, thus enabling the lordosis restoration and the medialisation of the apex. These combine manoeuvres should lead to derotation. However, similarly to the rod rotation technique, forward bending will make the apex move in the pathological direction, thus increasing rotation. Thereby, it is paramount that screws turn simultaneously with the contouring manoeuvres. Only this combination of movements will provide with a three-dimensional correction.

To do so, derotation blocks are placed on the screws heads so that the assistant can turn them while the surgeon is performing the forward contouring manoeuvres that will allow lordosis restoration. This mobilization perfectly follows the deformity path and replaces the spine between the rods. This technique may be used both for mild and severe scoliosis correction in the three planes.

To facilitate correction and to maintain it on a long term basis, posterior release and posterior fusion may not always enough. In this case, anterior release and grafting may be required. Anterior approach may be facilitated by video assistance. Thoracoscopy will be preferred between T3 and T11, while video assistance is recommended for the thoracolumbar and lumbar regions. Anterior release associated with in situ contouring does provide significant correction especially in severe scoliosis as well as in stiff curvatures in the adult.

Three Dimensional Validation in Practice: Theory is nice but validation in practice is compulsory in order to verify our hypotheses. To check the validity of our statements, Raphael Dumas, PhD, performed a biomechanical analysis of the surgical correction by in situ contouring technique. He studied 20 scoliotic patients by means of stereoradiographic three-dimensional reconstruction. The stereoradiographic reconstruction technique is based on the identification of anatomical landmarks on frontal and sagittal x-ray films, previously acquired in a calibrated radiological environment equipped with a 90° turning table. This method provided us with three-dimensional reconstructions of spines allowing for an accurate measurement of vertebral rotations. Indeed, vertebral rotations must be measured in standing position, especially in the pre-operative examination, and has to be expressed in a fixed referential. These requirements could not be met with the traditional methods, i.e. CT scan. Three-dimensional reconstructions also provide us with an axial view of the whole spine, while allowing a comparison between the post-operative and pre-operative vertebral rotations at each level. We also calculated the intervertebral rotation. This rotation is maximum at the end vertebrae levels and minimum at the apex level. It is totally independent from the reference axis as the trunk movement will not alter the relative position of adjacent vertebrae. We actually consider that intervertebral angles are paramount for the estimation of the deformation severity as well as of the obtained correction. In our series (20 scoliotic patients) we observed a maximal rotation at the thoracic apex level (17.3°) and at the lumbar apex one (19°). The correction gain obtained was 11.3° at both levels. The intervertebral rotation had a maximum value at the limit vertebrae levels, i.e. 8.9° for thoracic superior level, 11.3° at the thoraco-lumbar junction and 7.3° at the inferior lumbar level. The correction obtained in these three regions was respectively 7.2°, 8.9° and 6°. We developed a detorsion index that corresponds to the difference between post-operative and pre-operative sums of intervertebral rotations of vertebrae within the organic curvature pondered by the pre-operative sum. The detorsion index at the thoracic level is 52% while at the lumbar level it is 85%. One can note that the thoracic detorsion is quite disappointing.

We could consider that the pedicular hook prevents from important detorsion in the thoracic spine, as it will not allow important derotation of vertebrae. This is why we had to design a new pedicular implant that was meant to provide bilateral support during correction manoeuvres. The so called bipedicular implant is linked to the vertebra at the costo-vertebral joint level holds the pedicle on its lateral side. This new implant enables a double action, i.e. posterior traction combined with concave medialisation and convex push. Thus the vertebra moves as a wheel, describing a global movement of derotation.

We have used this implant for two years now and we had no particular drawbacks as far. No tolerance problems were noted either. Derotation blocs allow for the combination of rotation movements at the thoracic and lumbar levels while the rod is contoured to reach the best possible curve correction.

Conclusion: The in situ contouring has been used for 10 years now. It is not just a physical gesture; it is a whole new philosophy of reduction, a new way of thinking in spine surgery. In situ contouring replaces vertebra in its initial spatial context while replacing the surgeon in the best position to create after thorough reflection on the pathological mechanism. The in situ contouring may be successfully used not only for scoliosis correction but also for other deformities, especially sagittal ones such as hyperkyphosis, fractures, malunions and lumbar degenerative deformities.