Looking for information? You have questions?
Check out our questions / answers.
Q1: Does the brace simulator take the corpulence of the patient into account?
Yes, of course. The model of the customised trunk for each patient, the backbone of the brace simulator, is built from the patient's skeleton geometry taken from x-rays and the patient’s skin geometry taken from the external scan of the trunk. In the finite element model, these two internal and external geometries are mechanically connected by a layer representing the external soft tissues between the skin and the skeleton: epidermis, muscles, fat.
For an obese patient, the “empty space” between the two geometries (skeleton and skin) is intrinsically larger than that of a thin patient. The layer of external soft tissue will therefore be thicker in the finite element model of an obese patient. As a rule, the mechanical model of an obese patient’s trunk will therefore be more rigid.
Q2: Has BraceSim been validated? In other words, what is its level of precision?
A complete validation study of BraceSim has been conducted at the Saint-Justine hospital in Montreal through a prospective randomised study.
Patients randomly assigned to two groups, with one group receiving a traditionally designed brace and one group receiving a brace designed using simulation, to check that the brace simulator improves the average efficiency of braces.
The results showed that the standard braces reduced the main thoracic and lumbar by 29% and 40% respectively compared to 43% and 46% for the new braces which were also 50% thinner and had 20% less covering surface than than the standard braces. The new braces showed a clinically and statistically significantly greater main thoracic Cobb angle correction and similar lumbar correction
Preliminary studies conducted so far on a limited number of patients (twenty or so) have evaluated the brace simulator’s precision to be 5 degrees for Cobb angles.
Q3: Does the brace simulator take the patient’s flexibility into account?
BraceSim can take the flexibility of the patient into account, in a qualitative manner, since the software has a patient ‘flexibility’ parameter. The CPO can adjust this parameter on a scale from 1 to 5 (1: very stiff, 2: stiff, 3: normal, 4: flexible, 5: very flexible) depending on his own assessment, or the doctor's assessment of the patient's flexibility.
Depending on this parameter, the flexibility of the spine will be adjusted in the FE mechanical model of the trunk that forms the basis of BraceSim. More precisely, the rigidity (Young’s modulus) of intervertebral discs and ligaments will be multiplied by an appropriate factor. This factor has been calculated from empirical observations during preliminary validation tests.
This method is not perfect. It remains qualitative. Ideally, the flexibility of the patient should be measured objectively and experimentally and then, this flexibility must be adjusted / optimised in BraceSim for it to be truly customised.
However, there is no a good objective measurement of the patient’s flexibility. Objective customisation tests on the flexibility of patients have been conducted using ‘bending tests’ (lateral bending) which are commonly used to measure the reducibility of scoliosis curves before surgery. Nevertheless, the high variability inherent to these tests, the fact that they involve the active muscular action of patients and above all, their primary objective to measure reducibility and not flexibility, have limited the relevance of results obtained.
Q4: What is the best brace according to BraceSim? Is it the Cheneau or the Boston brace?
Without wanting to be ‘politically correct’, BraceSim has yet to provide a clear answer to this question, even if preliminary studies comparing different types of braces have been carried out.
It should be pointed out that a ‘Boston brace' or a ‘Cheneau brace’ for example, do not really exist but rather different variants of these braces based on a certain number of common general principles which are adapted or interpreted differently by each CPO. Therefore, a statement like ‘Cheneau braces are better than Boston braces’ is completely outdated.
Furthermore, the objectives of a ‘treatment by brace’ may be very different depending on CPOs or doctors: reduce frontal curvature as much as possible, maintain a good sagittal balance, reduce the transverse axial rotation, reduce protuberance, etc. Without a clearly defined and common object, it is very difficult to compare different braces.
Nevertheless, it is clear that BraceSim’s long term goal is to answer this question: what is the best brace? As different CPOs start to use this tool and their ‘feedback” is recorded, answers will eventually come together to define which is the 'optimum brace’.
Q5: For BraceSim, is it better to have an anterior or posterior opening for a brace?
As with the previous question, BraceSim has yet to provide a clear answer to this question, even if preliminary studies have been carried out. Again, this depends on the aim of the treatment as defined by the CPO or the doctor. Also, the practical aspect of the brace cannot be evaluated or simulated by the simulator. A brace with an anterior opening offers the big advantage that it can be fitted by the patient himself.
However, preliminary studies have shown that braces with an anterior opening seem to offer more potential to correct protuberance and axial rotations (twisting of the spine). The anterior position of attachment points of straps helps create more leverage on the posterior protuberance than with a posterior opening. However, braces with a posterior opening seem to be more capable of exerting high forces on the postero-lateral part of the lumbar spine to correct lumbar curves in the frontal plane. Attachment points of straps are closer to this region which allows a more efficient transfer of forces. Note that these results are preliminary and should be confirmed by a comprehensive study.
Q6: What brace design parameters can be modified in BraceSim?
Virtually all parameters of a brace can be modified in BraceSim: position and size of correction pads, the position of window openings, shape and size of the rigid shell, the sagittal profile of the brace, the number and position of straps, force/tension in straps, the position of the brace opening (posterior, anterior or other...), the materials used, etc. Therefore, all rigid braces, even experimental ‘fantasy’ braces, can be tested. It should be noted however, that here we are talking about a ‘rigid brace’. To date, flexible braces (SpineCor or Tria for example) cannot be simulated.
One of BraceSim’s strengths is its ability to be able to modify and test all imaginable design parameters. As such, the CPO can virtually test an infinite number of braces ‘for free’: by for free, here we mean that tests conducted by the CPO with BraceSim do not affect a real patient (no harmful X-rays, no waste of time for patients) and braces which are tested are not physically produced. Everything is virtual.
Q7: Can BraceSim be used to test night time braces?
Yes, of course. The software has an option to specify whether the effectiveness of the brace should be predicted in a standing or lying down position.
In a standing position, the gravitational forces exerted on the patient (his 'weight') are vertically downwards. The geometry of the spine is such as provided by X-rays taken in the standing position.
In the lying down position, a simulation is carried out to ‘guess” what the form of the spinal column will be in this position. To do this, we first apply the g-forces on the spine (we subtract the weight of the patient) then we re-apply the gravitational forces but in the anterior-posterior direction. Typically, the geometry of the spine is obtained showing a natural reduction of scoliosis curves (35% on average) due to the suppression of vertical gravitational forces. The effect of the night time brace can then be simulated in this position.
Charleston and Providence night time braces have been tested with BraceSim. But more generally, any brace (Cheneau, Boston or other) may be tested in day time mode or night time mode. In particular, this has allowed us to demonstrate that the effectiveness of braces which seemed ‘good’ during the day was more limited in the lying down position since the braces were not aggressive enough for this position. In a perfect world, patients should probably have 2 braces: A day time brace and a more aggressive night time brace, taking into account the natural reduction of curvature due to the lying down position.
Q8: Is BraceSim able to predict if a brace will be able to correct the patient’s shift in the frontal plane and restore “normal” sagittal curvature?
The question of posture and overall balance of a patient in-brace/out-of-brace is a fairly complex question. Indeed, it is related to neurological aspects and the patient’s brain and nervous system. The patient continually adjusts his posture by means of a retroaction loop according to one or several given objectives: minimise energy expenditure, reduce stresses on flexible anatomical components (e.g.: intervertebral discs), etc. These objectives may differ from one patient to another. Moreover, the adopted posture also depends on each patient’s own anatomy (shape of the pelvis, for example). Here, we have a ‘patient-dependent' neurological phenomena which is very difficult to simulate or predict via a generic model. More generally, any phenomenon which is not of the ‘passive mechanical’ type, where the brain actively intervenes (as for muscular activation, for example) is very difficult to predict through finite element models such as those used in BraceSim.
However a simplified approach has been introduced in BraceSim to help CPOs predict what effect his brace will have on the patient’s posture. By default, the T1 vertebra is constrained in the transverse plane. It does not move in the x and y directions but can translate along z (vertical axis). The simulator does not quantitatively predict what the coronal or sagittal shift will be for the patient in his brace, or by how many mm the patient may be unbalanced. But the simulation indicates the direction and the reaction force on T1. This force indicates on which side the brace will, on the whole, tend to ‘push’ the patient’s trunk. If this force is too great, this indicates that the brace may unbalance the patient’s trunk. The simulator can then provide a qualitative indication as to the effect of the brace on the patient’s posture.
Q9: Does the brace simulator take the “active” effect of the brace into account? Is the action of the patient’s muscles included in the model?
The active effect of the brace is linked to the patient’s muscular activation and therefore to neurological aspects, the patient’s brain and the nervous system. As a rule, muscular activation strategies (what muscle is activated to do what?) depend on each patient. As such, the active effect of the brace, if any, will be expressed differently depending on each patient: it will be to a greater or lesser extent and will result in different corrections. Here, we have a ‘patient-dependent' neurological phenomena which is very difficult to simulate or predict via a generic model.
More generally, any phenomenon which is not of the ‘passive mechanical’ type, where the brain actively intervenes, is very difficult to predict through finite element models such as those used in BraceSim.
This is obviously an important limitation of BraceSim. We can only model or predict the purely passive effect of the brace, as if the patient was a dummy without active muscles. In particular, this will limit any studies aiming to evaluate the absolute accuracy of the simulator. But, this limit probably only has a very minor effect on the usefulness of BraceSim.
Essentially, the value of BraceSim is more relative than absolute. It provides the CPO with an indication of whether one brace is better than another. In addition to knowing precisely what correction will be generated by a given brace, BraceSim is used to identify shortcomings in a given brace, differentiate between various braces, and optimise the design. Admittedly, these comparisons are made solely on the basis of the passive effect, but it appears likely that the active effect of the brace is proportional and correlated with its passive effect: as such, if a brace is more effective than another for the passive effect, it is likely to remain so when the active effect is added.