**3.2. Case n.2 scoliosis and leg length discrepancy: multi‐factorial static and gait analysis first evaluation versus control**

This describes a 13‐year‐old patient with scoliosis and leg length discrepancy (LLD). The LLD is corrected during measurements at the first evaluation by placing an underfoot wedge—the optimal value of which (10 mm under left foot) was determined as the one producing the best global posture outcome considering all the combined spine deformities and postural param‐ eters. In addition, the analysis of underfoot pressure distributions suggested the use of cus‐ tomized foot orthoses to reduce ankle pronation at both feet, complemented by the wedge correction under the left foot. The patient was measured after a period of 6 months during which she was wearing the recommended foot orthoses.

The mathematical details for the optimization of the procedure as well as the average gait

The final outcome is the mean gait cycle in which the average time course and associated standard deviation is defined per each variable of interest [8, 27, 28, 31]. Two main advantages can be enu‐ merated with the possibility to extract the mean characteristics of both static posture and cyclic motor task (gait): first, it allows to overcome the single measurements analysis limits by taking into account the ensemble behaviour improving the statistical reliability of the evaluation; second, it permits to obtain information about the repeatability and variability of the performed motor task, thus enlightening the subject's motor control capability. For the graphical representation as well as clinical parameter visualization and enlightening, a software package (named ASAP 3D Skeleton Model©) based on 3D graphic modelling has been developed. This latter is now available

as a commercial software package (Bioengineering & Biomedicine Company S.r.l. Italy).

Several studies are currently being carried out by our group about spine and posture disor‐ ders with the described methodology. A few examples are summarized below in order to show the capability of this multi‐factorial approach to process many different measurements,

**3.1. Case n.1 scoliosis 3D posture static and dynamic analysis versus X‐ray measurement**

The first example (**Figures 6** and **7**) represents the outcome of the procedure for the analysis of a patient with scoliosis. The left panel shows the description of the spinal deformity that

On the right panel, a full 3D skeletal posture reconstruction in both the frontal and sagittal planes is depicted. The end and apical vertebrae, Cobb angle values as well as the spinal and

In **Figure 7a**, the pelvis orientation is described together with the associated torsion as derived by the relative position of PSIS and ASIS landmarks (left panel); in addition, the relative rota‐ tions of shoulder‐pelvis‐feet on the horizontal plane are represented (a second illustration without skull and feet is also given on a side to highlight trunk torsion). In **Figure 7b**, the dissimilarities in the stiffness of different spinal segments along the vertebral column are depicted, due to position and magnitude of scoliotic curves, describing hyper/hypomobility

**3.2. Case n.2 scoliosis and leg length discrepancy: multi‐factorial static and gait analysis** 

This describes a 13‐year‐old patient with scoliosis and leg length discrepancy (LLD). The LLD is corrected during measurements at the first evaluation by placing an underfoot wedge—the optimal value of which (10 mm under left foot) was determined as the one producing the best

global offset values and their averages are automatically identified and computed.

cycle computation are beyond the limits of this chapter [8, 27, 28].

30 Innovations in Spinal Deformities and Postural Disorders

thereby allowing all the results to be combined in a unified view.

shows full agreement with the X‐ray measurement which is displayed.

during the performance of lateral bending tasks.

**first evaluation versus control**

**3. Results**

In **Figure 8**, the within‐sessions comparison (both at first evaluation and at control measurement session after 6 months) between a neutral‐standing posture and an underfoot wedge‐corrected neutral‐standing posture (upper row panels) is presented. At the control session, for the wedge‐ corrected standing position, instead of a simple heel rise the patient was measured with her cus‐ tomized foot orthoses. A cross‐session comparison between neutral orthostasis at first evaluation and wedge‐corrected neutral orthostasis at control is also displayed in the bottom‐row panels.

The model is able to indicate and explain the almost perfect realignment of the spine and trunk balance induced by foot orthoses/wedge correction worn during a 6‐month period, with a dramatic reduction of spine deformities in the frontal plane.

**Figure 8.** Composite figures representing comparison of posture, and related balance and spine morphology changes in the frontal and sagittal planes between indifferent (i.e. neutral) orthostasis (left side of a,b,c,d panels) and wedge‐corrected neutral orthostasis (right side of a,b,c,d panels), of a 13‐year‐old patient presenting with a leg length discrepancy (LLD), when 10‐mm heel rise was positioned under left foot. The same patient underwent a gait analysis test at both first and follow‐up (after about 6‐month) sessions.

For this patient, the study of gait biomechanics was also performed. **Figure 9** shows the mean gait cycle computed over 10 different strides. The mean trajectories of each considered land‐ mark together with the associated frame‐by‐frame standard deviations are represented. Each trajectory is depicted by a time series of spheres, each one being centred in the frame‐by‐frame computed mean 3D co‐ordinate and having radius given by the magnitude of the assessed standard deviation. In addition, the mean patterns of ground reaction forces are displayed.

It is interesting to note how the standard deviations of the trunk‐spine landmark trajectories are significantly smaller than those of the head and lower limb landmarks (as shown by the different dimensions of spheres). Such findings demonstrate, from a biomechanical stand‐ point, that the patient was preserving her mechanical energy and minimizing oscillations. The patient also had a very strict repetitive motor pattern for the trunk (where most of the body mass is amassed), while she was releasing more variability in the distal segments. Our group is currently conducting studies to further explore this phenomenon on a wider population.

In this case, the model allowed the investigators to also assess the ROMs of the spinal defor‐ mity angles as well as the variation of stresses acting on the spine during gait. Such outcomes have been compared to the values assessed during orthostasis.

Comparisons are presented in **Figures 10** and **11**, where the effects of LLD on both morphol‐ ogy modifications and stresses acting on the patient's spine during orthostasis and gait are put on evidence. **Figure 10** shows the cross‐session comparison between neutral orthostasis at first evaluation (left panel) and wedge‐corrected (customized orthoses) neutral orthostasis at

#### **Figure 9.** The mean gait cycle. Averaged 3D trajectories of each considered landmark and of ground reaction forces patterns, together with the associated frame‐by‐frame standard deviations (3D radius of each depicted sphere).

For this patient, the study of gait biomechanics was also performed. **Figure 9** shows the mean gait cycle computed over 10 different strides. The mean trajectories of each considered land‐ mark together with the associated frame‐by‐frame standard deviations are represented. Each trajectory is depicted by a time series of spheres, each one being centred in the frame‐by‐frame computed mean 3D co‐ordinate and having radius given by the magnitude of the assessed standard deviation. In addition, the mean patterns of ground reaction forces are displayed.

It is interesting to note how the standard deviations of the trunk‐spine landmark trajectories are significantly smaller than those of the head and lower limb landmarks (as shown by the different dimensions of spheres). Such findings demonstrate, from a biomechanical stand‐ point, that the patient was preserving her mechanical energy and minimizing oscillations. The patient also had a very strict repetitive motor pattern for the trunk (where most of the body mass is amassed), while she was releasing more variability in the distal segments. Our group is currently conducting studies to further explore this phenomenon on a wider population. In this case, the model allowed the investigators to also assess the ROMs of the spinal defor‐ mity angles as well as the variation of stresses acting on the spine during gait. Such outcomes

Comparisons are presented in **Figures 10** and **11**, where the effects of LLD on both morphol‐ ogy modifications and stresses acting on the patient's spine during orthostasis and gait are put on evidence. **Figure 10** shows the cross‐session comparison between neutral orthostasis at first evaluation (left panel) and wedge‐corrected (customized orthoses) neutral orthostasis at

**Figure 9.** The mean gait cycle. Averaged 3D trajectories of each considered landmark and of ground reaction forces patterns, together with the associated frame‐by‐frame standard deviations (3D radius of each depicted sphere).

have been compared to the values assessed during orthostasis.

32 Innovations in Spinal Deformities and Postural Disorders

**Figure 10.** Composite figures representing comparison of posture, spine morphology and related spine torques in frontal plane between neutral orthostasis (left panels) and orthostasis when 1‐cm wedge was positioned under left foot at first evaluation of a scoliotic patient (right panels).

**Figure 11.** Composite figures representing a mean gait cycle comparison of the above‐described scoliotic patient, during first evaluation (left panels) and at 6‐month control (follow‐up) (right panels). In each figure, the lower panels represent spine shape and computed spine torques at each inter‐vertebral level taken at frames in which the spinal curves reach their maximum (at thoracic levels). The upper panels show spinal curves full ranges throughout the mean gait cycle, arrows pointing at reached maximum angular value. From these results, it can be noted that the inter‐vertebral torque values, assessed at control session, fall below one‐third with respect to the first‐session values.

control (right panel). As can be seen, the spine shape remodelling and global posture rebalanc‐ ing induce a complete transforming of the spinal inter‐vertebral torques in the frontal plane, with a factor 4 reduction (i.e. the maximal values at follow‐up are below 25% of the maximal values at first evaluation) at the upper thoracic level where the major deformity was present.

Theoretically, during neutral orthostasis, no inter‐vertebral torques should be present in the frontal plane, if the spine was straight, aligned with gravity line and in the symmetry plane of the subject. Conversely, the occurrence of spinal deformities and postural unbalance deter‐ mine an amount of not‐negligible trunk torques loading each vertebra in an asymmetric way. Such behaviour could be dramatically worsened during a dynamic activity like walking. In fact, when static and dynamic values are compared (**Figures 10** vs. **11**), it is evident that during gait, spinal deformities increase and spine loads are higher. For instance, at thoracic level the Cobb angle value of the curve passes, from static to dynamic condition, from around 17° up to around 26°. Moreover, as it can be seen, also the values of assessed trunk torques, result dur‐ ing gait, of a magnitude of up to six times the values assessed during neutral erect standing. This phenomenon can be very important in the evolution of scoliosis inducing a spine defor‐ mities progression. In fact, taking into account the in vivo demonstration of the well‐known Heuter‐Volkmann principle, asymmetric loads on functional spinal unit (i.e. vertebrae plus inter‐vertebral discs) determine a wedging process on both inter‐vertebral disc and vertebrae during growth [39, 40]. In this way, any action inducing a reduction of such asymmetries results in beneficial effects. In the described case, in the comparison between the first session and follow‐up at 6 months, the positive effect of customized foot orthoses complemented by wedge correction is greatly confirmed both in the static measurements and during gait.

Finally, to analyse the changes induced on the biomechanics of walking of a patient after 6 months of the use of customized foot orthoses, complemented by the wedge correction under left foot, baropodometric analysis outcomes have been compared. **Figure 12** shows the comparison of the mean gait cycles (as assessed from 10 strides each) at first evaluation (left panels) and at control (right panels) in terms of pressure maps and derived vertical forces measured by baropodometric foot insoles device during locomotion on the floor. This kind of baropodometric device provides measurements of the in‐shoe direct interaction of the foot. So, to avoid obvious direct modification of pressure maps induced by customized foot ortho‐ ses at control session, the same neutral insoles have been used in patient's shoes for both sessions, but adding the 10‐mm left underfoot wedge correction during the control session.

In **Figure 12**, it is manifest that the use of underfoot wedge correction (and of the customized insoles during 6 months) demonstrated to contribute in underfoot load asymmetries reduction as well as a better underfoot pressure distribution. In fact, by the analysis of vertical forces pat‐ terns it results that, at first evaluation, a different load between feet was occurring, being the left one more loaded along all the stance phase (upper‐left panel). Moreover, a different heel to forefoot load transfer pattern is evident between the right and left foot showing that left foot stance phase was significantly longer than the contralateral, that is, the left foot pushes more and for longer time being more propulsive. In the pictures, to represent map pressures of stance phases, the so‐called 'peak frames' are used. They are defined as the maps obtained by assign‐ ing to each pressure cell the peak pressure value reached during stance. Peak pressure maps enlighten (left‐lower panel) the different loading patterns in each foot showing the pressure

control (right panel). As can be seen, the spine shape remodelling and global posture rebalanc‐ ing induce a complete transforming of the spinal inter‐vertebral torques in the frontal plane, with a factor 4 reduction (i.e. the maximal values at follow‐up are below 25% of the maximal values at first evaluation) at the upper thoracic level where the major deformity was present. Theoretically, during neutral orthostasis, no inter‐vertebral torques should be present in the frontal plane, if the spine was straight, aligned with gravity line and in the symmetry plane of the subject. Conversely, the occurrence of spinal deformities and postural unbalance deter‐ mine an amount of not‐negligible trunk torques loading each vertebra in an asymmetric way. Such behaviour could be dramatically worsened during a dynamic activity like walking. In fact, when static and dynamic values are compared (**Figures 10** vs. **11**), it is evident that during gait, spinal deformities increase and spine loads are higher. For instance, at thoracic level the Cobb angle value of the curve passes, from static to dynamic condition, from around 17° up to around 26°. Moreover, as it can be seen, also the values of assessed trunk torques, result dur‐ ing gait, of a magnitude of up to six times the values assessed during neutral erect standing. This phenomenon can be very important in the evolution of scoliosis inducing a spine defor‐ mities progression. In fact, taking into account the in vivo demonstration of the well‐known Heuter‐Volkmann principle, asymmetric loads on functional spinal unit (i.e. vertebrae plus inter‐vertebral discs) determine a wedging process on both inter‐vertebral disc and vertebrae during growth [39, 40]. In this way, any action inducing a reduction of such asymmetries results in beneficial effects. In the described case, in the comparison between the first session and follow‐up at 6 months, the positive effect of customized foot orthoses complemented by wedge correction is greatly confirmed both in the static measurements and during gait.

34 Innovations in Spinal Deformities and Postural Disorders

Finally, to analyse the changes induced on the biomechanics of walking of a patient after 6 months of the use of customized foot orthoses, complemented by the wedge correction under left foot, baropodometric analysis outcomes have been compared. **Figure 12** shows the comparison of the mean gait cycles (as assessed from 10 strides each) at first evaluation (left panels) and at control (right panels) in terms of pressure maps and derived vertical forces measured by baropodometric foot insoles device during locomotion on the floor. This kind of baropodometric device provides measurements of the in‐shoe direct interaction of the foot. So, to avoid obvious direct modification of pressure maps induced by customized foot ortho‐ ses at control session, the same neutral insoles have been used in patient's shoes for both sessions, but adding the 10‐mm left underfoot wedge correction during the control session.

In **Figure 12**, it is manifest that the use of underfoot wedge correction (and of the customized insoles during 6 months) demonstrated to contribute in underfoot load asymmetries reduction as well as a better underfoot pressure distribution. In fact, by the analysis of vertical forces pat‐ terns it results that, at first evaluation, a different load between feet was occurring, being the left one more loaded along all the stance phase (upper‐left panel). Moreover, a different heel to forefoot load transfer pattern is evident between the right and left foot showing that left foot stance phase was significantly longer than the contralateral, that is, the left foot pushes more and for longer time being more propulsive. In the pictures, to represent map pressures of stance phases, the so‐called 'peak frames' are used. They are defined as the maps obtained by assign‐ ing to each pressure cell the peak pressure value reached during stance. Peak pressure maps enlighten (left‐lower panel) the different loading patterns in each foot showing the pressure

**Figure 12.** Baropodographic analysis of the mean gait cycle. The mean gait cycle characteristics obtained when the patient was walking wearing an underfoot wedge (right panel) and without underfoot wedge (left panel) are compared. Averaged force patterns, averaged peak pressure maps and COP patterns are represented pointing out the reduction of differences and asymmetries when underfoot wedge was worn.

values in different foot regions, with left foot presenting a cavo‐valgus foot pattern due to ankle pronation during stance phase as it is evident by the absence of load on the lateral aspect of the foot between the heel and forefoot (isthmus). At control session, such asymmetries are really reduced. The stance phases of both feet present the same duration in time, vertical forces are almost superimposed and the lateral aspect of the left foot between the heel and forefoot (isthmus) started to be charged presenting a reduction of ankle pronation.

In this picture, a new introduced special graphical feature for gait analysis, the so‐called 'but‐ terfly' window, is displayed in the lower panels of the figure between the two peak pressure maps of the feet. This graphic is obtained by computing the frame‐by‐frame spatial weighted baricentre of both feet COPs actual positions (feet are considered as they were symmetrically positioned); such pattern allows an intuitive and immediate evaluation of the symmetry of subject's gait. In fact, at first evaluation, a strong asymmetry was evident between the left and right 'butterfly wings' of the diagram, pointing out different COP patterns for the left and right foot (both in their shape and in their time duration). Conversely, at control session, the two 'butterfly wings' resulted much more symmetrical, thus confirming the improvement obtained by underfoot wedge correction.
