Automatic Modeling of an Orthotic Bracing for Nonoperative

Correction of Pectus Carinatum

João L. Vilaça

1,5

, Pedro L. Rodrigues

1,2

, António H. J. Moreira

1,2

, João Gomes Fonseca

A. C. M. Pinho

, Jaime C. Fonseca

and Nuno Rodrigues

5,6

ICVS/3B’s - PT Government Associate Laboratory, Braga/Guimarães, Portugal

Algoritmi Center, School of Engineering, University of Minho, Guimarães, Portugal

Mechanical Department, University of Minho, Guimarães, Portugal

Industrial Electronics Department, University of Minho, Guimarães, Portugal

DIGARC – Polytechnic Institute of Cávado and Ave, Barcelos, Portugal

HASLab/INESC TEC, University of Minho, Braga, Portugal

Keywords: Pectus Carinatum, Nonoperative Treatment, 3D Image Processing, Mesh Processing.

Abstract: Pectus Carinatum is a deformity of the chest wall, characterized by an anterior protrusion of the sternum,

often corrected surgically due to cosmetic motivation. This work presents an alternative approach to the

current open surgery option, proposing a novel technique based on a personalized orthosis. Two different

processes for the orthosis’ personalization are presented. One based on a 3D laser scan of the patient chest,

followed by the reconstruction of the thoracic wall mesh using a radial basis function, and a second one,

based on a computer tomography scan followed by a neighbouring cells algorithm. The axial position where

the orthosis is to be located is automatically calculated using a Ray-Triangle intersection method, whose

outcome is input to a pseudo Kochenek interpolating spline method to define the orthosis curvature. Results

show that no significant differences exist between the patient chest physiognomy and the curvature angle

and size of the orthosis, allowing a better cosmetic outcome and less initial discomfort.

1 INTRODUCTION

Pectus Carinatum is a deformity of the chest wall

characterized by an anterior protrusion of the

sternum and ribs. It occurs in approximately one out

of every 1500 children (about 80% of patients are

male). Current treatment for Pectus Carinatum

includes both surgical and nonsurgical options.

Despite the success of operative approaches, the

related surgery complications (e.g. bleeding), scars

and postoperative pain, present considerable

drawbacks of these more invasive techniques

(Alexander et al., 2009).

Based on the fact that the anterior chest wall is

malleable during puberty, many authors advocate the

benefits of nonoperative approaches to induce chest

wall remodelling (Frey et al., 2006); (Goretsky et al.,

2004). Although this treatment option has been

shown to provide positive results over time, some

parents find it difficult to keep their children from

taking off the brace (most of the times due to

wearing discomfort), which reduces the overall

therapeutic process efficiency (Stephenson and Du

Bois, 2008).

The present work proposes two different

approaches for the automatic orthosis

personalization: one based on a 3D laser scan and

another on a computer tomography (CT) scan of the

chest.

Afterwards, the system developed in (Vilaça et

al., 2009) is applied to automatically model and

bend the orthosis bar, according to each patient

morphology.

2 METHODS AND RESULTS

This section describes all methods to automatic

determine the array points that are used in (Vilaça et

al., 2009).

L. Vilaça J., L. Rodrigues P., H. J. Moreira A., Gomes Fonseca J., C. M. Pinho A., C. Fonseca J. and Rodrigues N..

Automatic Modeling of an Orthotic Bracing for Nonoperative Correction of Pectus Carinatum.

DOI: 10.5220/0004221100710074

In Proceedings of the International Conference on Computer Vision Theory and Applications (VISAPP-2013), pages 71-74

ISBN: 978-989-8565-48-8

 2013 SCITEPRESS (Science and Technology Publications, Lda.)

2.1 Thoracic Wall Reconstruction

This stage reconstructs the anterior and posterior

thoracic wall according to two different sets of input

data, namely CT imagiology and 3D scanner dat.

Since CT scans remain the primary choice for

preoperative imaging of Pectus Carinatum, a CT

data set is available in most patients, providing

useful information about abnormal thoracic anatomy

and symmetry defects. Therefore, one has

implemented an algorithm that automatically

segments all skin points and generates a 3D mesh

from a CT data set. The process starts with a pre-

processing stage that performs a smooth operation

on the CT data (using a Gaussian filter) and

calculates its gradient volume magnitude. Then, all

image pixels are vertically and horizontally tracked

by recursively selecting all first voxels belonging to

the image gradient (skin points). These voxels are

input to a neighbouring cells algorithm, where each

voxel of a cubic grid cell is classified as lying above

or below an 800 iso-surface threshold. A tri-linear

interpolation is then used to reconstruct a skin

surface mesh from CT data (Figure 1 – B).

Figure 1: 3D Mesh representation from a 3D scan (A) and

using a 3D CT data set.

Although CT scans have been the main analysis

option to assess Pectus Carinatum severity,

sometimes these are not available or are deprecated.

For these patients, an inoffensive and radiation-free

3D laser scan using the FastSCAN system can be a

much better choice, avoiding the need of a CT scan,

and the related radiation, to reconstruct the thoracic

wall. Exploring such an option, one was able to

achieve a minimal error acquisition of

approximately 1 mm, by attaching the system

reference to the 10º left rib.

The unorganized point clouds acquired from the

3D scan, containing outlier’s points, noise and non-

uniformities in thickness and spacing were used to

reconstruct the thoracic wall mesh surface. The

algorithm used to reconstruct a smooth surface from

the raw points of the 3D scanner was based on

Radial Basis Functions (RBFs) (Savchenko et al.,

1995).

The data misalignments and surface holes (when

two or more samples overlap in the same Oxyz

plane) were eliminated by modifying the voxel size

and an adaptive kernel window during

reconstruction. In the end, a 3D mesh was created

representing the thoracic wall (Figure 1 - A).

After the mesh reconstruction, some surface

holes can still exist. This situation was more

frequent when the mesh was created using the 3D

scanner, due to patient movements during the

acquisition and light interference. To overcome this

problem, a correction filter was developed, that

searches for accentuated changes on the X and Y

components of the discrete signal derivative. For

each triangle T of the surface mesh whose partial

derivative value is higher than a given threshold,

when compared to its neighbourhood triangles

(within a defined radius R), the triangle vertices

coordinates within R were merged together

according to its neighbourhood maximum values.

The outcome of this process was input to a

Laplacian smooth method to reduce the shrinking

effect.

2.2 Highest Sternal Protrusion

The surface mesh created on the previous section

was input to a Ray-Triangle intersection method,

which was used to compute the minimal distance

between two triangle meshes: the thoracic wall mesh

(Ts) and a plane mesh reference (Tp) which size was

adjusted for the Ts size.

Figure 2: Illustration of the Ray-Triangle intersection

method results, used to automatically determine the region

of the greatest sternum protrusion.

In the end, the difference between each plane facet

triangle and the thoracic facet triangle was

determined. Based on these distances, the thoracic

triangle facets were coloured according to the

VISAPP2013-InternationalConferenceonComputerVisionTheoryandApplications

colour-map defined on the colour bar in Figure 2.

The most anterior prominent sternal protrusion (red

area, Figure 2), was chosen for orthosis placement

and for the sizing of the cushioned compression

plates of the anterior position.

2.3 Orthosis Automatic Modelling

The higher sternum protrusion point and a normal

along Oy were firstly used to define an axial plane to

place the orthosis. Then, a method was implemented

to cut the thoracic wall mesh that intersects this

plane.

The cutting result was a set of points that were

input to a method that determines a pseudo

Kochenek interpolating spline with a control

continuity and tension parameter (Kochanek and

Bartels, 1984).

Figure 3: Bar modelling representation.

The continuity/discontinuity of the first derivative of

at a key spline position was defined according to

Equation 1:

Out

Spline

=





∙

















∓

















(1)

where c is the continuity factor (experimentally

calculated – c0.2.

This kind of spline produced fewer oscillations

(when the thoracic mesh data is not smooth), had no

overshoots, and preserved the data monotonicity. By

using this spline it was possible to determine all

regions where the prosthesis is placed, avoiding skin

irregularities that are not relevant for the orthosis

modelling.

The following methodology was then used to

determine the final orthosis shape (Figure 3):

1) Determination of the rightist and leftist spline

points which defines a coronal plane along the

thoracic wall (red and green spheres, Figure 3).

Each of these locations defines two planar

regions (which size depends on the anterior-

posterior patient length) on the anterior and

posterior orthosis braces (LRA and LRP

respectively in Figure 4), allowing (1) the

placement of the linking area between the

anterior and posterior orthosis braces; (2) the

position of two mechanical systems for the

physician to adjust the correction chest pressure

to the desired level; and (3) the position of four

cushioned compression plates to increase patient

comfort;

2) Determination of the most anterior and posterior

spline points. Those define the area where two

cushioned plates are attached to the anterior and

posterior segment of the brace (Figure 3 and

Figure 4 (AC and PC, respectively)). The size of

each cushioned plates are also automatically

determined with respect to the thoracic

morphology and the deformity area (regions with

higher red intensity, Figure 3).

3) Bar modelling, taking into account (1) the

different positions determined at step 1 and 2; (2)

an offset curvature by scaling each spline point

along Oxyz normal, according to the cushioned

plate’s thicknesses (approximately 7 mm);

4) Finally, an array of points, containing the shape

information of each posterior and anterior bars, is

input to the (Vilaça et al., 2009) system to

automatically model the entire orthosis bar

(Figure 3 – spline points).

Figure 4: Final shape of the orthosis.

3 CONCLUSIONS

Current orthosis are modelled and bended using

AutomaticModelingofanOrthoticBracingforNonoperativeCorrectionofPectusCarinatum

control anatomical references that are identified

manually. Such procedure is time consuming, leaves

imperfections in the prosthesis surface and greatly

depends on the physician’s experience. Moreover,

most of these orthosis are modelled independently of

the symmetric defect and do not take into

consideration the thoracic wall shape. Consequently,

this process often produces non-uniform strength

distributions, misalignments and offsets between the

orthosis and the patient thoracic wall (which end up

producing initial discomfort), thoracic pain,

increased patient adaptation time and decreased

cosmetic outcome of the deformity. To overcome

current orthosis practice disadvantages, this work

proposes a systematic methodology for the

personalisation of a Pectus Carinatum orthosis. The

orthosis size and curvature are automatically

determined using information retrieved by CT

imagiology or a 3D scanner of the thoracic wall

shape.

An array of points containing this virtual model

information is input to the (Vilaça et al., 2009)

system to automatically bend two stainless steel

braces (AISI type 316LVM (low carbon vacuum

melt), corresponding to the anterior and posterior

orthosis braces. The differences between the virtual

and physical braces were determined by modelling

and bending 15 orthosis using 15 CT data sets of

patients with Pectus Carinatum, acquired at São

João Hospital of Porto (Portugal). To measure these

differences, an LVTD (linear variable differential

transformer) was used to compare, at different bend

times, the differences between the virtual point and

curvature, with the real values. Results show that no

significant differences exist concerning the curvature

angle and size of the orthosis, given that all errors

are below 10 µm.

Recently, this personalized orthosis has been

modelled for two patients whose thoracic wall

showed symmetric (Figure 1 - B) and asymmetric

defects (Figure 1 - A). For both patients, the orthosis

suitably fitted the thoracic wall shape, and the

pressure of the cushioned compression plates was

adequate to prevent slippage. The pressure of

treatment is controlled using an electronic pressure

sensor whose output value is shown on an LCD

(PMD in Figure 4). This permits to adjust the

correction pressure to the desired level and prevents

from making too much pressure, which could cause

pressure necroses.

Such results indicate a considerable step forward

that might decrease the need of open surgery for a

nonoperative approach in Pectus Carinatum

deformity correction. In addition, nonoperative

management offers a significant cost benefit since

with this new method, hospitalization time, per-

operative and post-surgical complications are

eliminated.

ACKNOWLEDGEMENTS

The authors acknowledge to Foundation for Science

and Technology (FCT) - Portugal for the fellowships

with the references: SFRH/BD/74276/2010;

SFRH/BD/68270/2010; UMINHO/BI/95/2012; and,

SFRH/BPD/46851/2008. This work was also

supported by FCT R&D project PTDC/SAU-

BEB/103368/2008.

REFERENCES

Alexander, A. F., Nury, M. S., William, A. A., Karen, E.

A., 2009. Anatomical, Histologic, and Genetic

Characteristics of Congenital Chest Wall Deformities.

Seminars in thoracic and cardiovascular surgery 21,

44-57.

Carr, J. C., Beatson, R. K., McCallum, B. C., Fright, W.

R., McLennan, T. J., Mitchell, T. J., 2003. Smooth

surface reconstruction from noisy range data,

Proceedings of the 1st international conference on

Computer graphics and interactive techniques in

Australasia and South East Asia. ACM, Melbourne,

Australia, pp. 119-ff.

Frey, A. S., Garcia, V. F., Brown, R. L., Inge, T. H.,

Ryckman, F. C., Cohen, A. P., Durrett, G., Azizkhan,

R. G., 2006. Nonoperative management of pectus

carinatum. J Pediatr Surg 41, 40-45; discussion 40-45.

Fritsch, F. N., Carlson, R. E., 1980. Monotone Piecewise

Cubic Interpolation. SIAM Journal on Numerical

Analysis 17, 238-246.

Goretsky, M. J., Kelly, R. E., Jr., Croitoru, D., Nuss, D.,

2004. Chest wall anomalies: pectus excavatum and

pectus carinatum. Adolesc Med Clin 15, 455-471.

Kochanek, D. H. U., Bartels, R. H., 1984. Interpolating

splines with local tension, continuity, and bias control.

SIGGRAPH Comput. Graph. 18, 33-41.

Savchenko, V. V., Pasko, A. A., Okunev, O. G., Kunii, T.

L., 1995. Function Representation of Solids

Reconstructed from Scattered Surface Points and

Contours. Comput Graph Forum 14, 181-188.

Stephenson, J. T., Du Bois, J., 2008. Compressive orthotic

bracing in the treatment of pectus carinatum: the use

of radiographic markers to predict success. Journal of

pediatric surgery 43, 1776-1780.

Vilaça, J., Pinho, A., Correia-Pinto, J., Fonseca, J.,

Peixinho, N., 2009. System for automatic and

personalized modelling/bending of surgical prosthesis

for correction of pectus excavatum based on pre-

surgical imagiology information, WO2009/035358.

VISAPP2013-InternationalConferenceonComputerVisionTheoryandApplications