Pesqui Odontol Bras
2003;17(1):24-8
Appli ca bi lity of 3D-CT fa ci al re cons truc ti on for fo ren sic in di vi du al iden ti fi ca ti on
Apli ca bi li da de da re cons tru ção fa ci al em 3D-TC para iden ti fi ca ção in di vi du al
fo ren se
Sara dos San tos Ro cha*
Dal ton Luiz de Pa u la Ra mos**
Mar ce lo de Gus mão Pa ra í so Ca val can ti***
ABSTRACT:
Com pu ted to mo graphy (CT) is used in se ve ral cli ni cal den tistry ap pli ca ti ons even by axi al sli ces and two and
three-di men si o nal re cons truc ted ima ges (2D-CT and 3D-CT). The pur po se of the cur rent study is to as sess the pre ci si -
on of li ne ar me a su re ments made in 3D-CT using cra ni o me tric pat terns for in di vi du al iden ti fi ca ti on in Fo ren sic Den -
tistry. Five ca da ver he ads were sub mit ted to a spi ral com pu ted to mo graphy using axi al sli ces, and 3D-CT re cons truc -
ti ons were ob ta i ned by vo lu me ren de ring tech ni que with com pu ter grap hics to ols. Ten (10) cra ni o me tric
me a su re ments were de ter mi ned in 3D-CT ima ges by two exa mi ners in de pen dently, twi ce each, and the stan dard er ror
of in tra- and in ter-exa mi ner me a su re ments was as ses sed. The re sults de mons tra ted a low stan dard er ror of tho se me -
a su re ments, from 0.85% to 3.09%. In con clu si on, the li ne ar me a su re ments ob ta i ned in os se ous and soft tis sue struc -
tu res were con si de red to be pre ci se in 3D-CT with high ima ging qua lity and re so lu ti on.
DESCRIPTORS:
To mo graphy, X-ray com pu ted; Skull; Fo ren sic ant hro po logy.
RESUMO:
A to mo gra fia com pu ta do ri za da (TC) tem sido uti li za da em di ver sas áre as clí ni cas da Odon to lo gia; uti li zam-se
tan to seus cor tes ori gi na is quan to as re cons tru ções em duas e três di men sões (2D-TC e 3D-TC). O pre sen te es tu do
pro põe ava li ar a pre ci são das me di das li ne a res re a li za das na 3D-TC, uti li zan do a cra ni o me tria, para fins de iden ti fi ca -
ção in di vi du al na Odon to lo gia Fo ren se. Cin co ca be ças de ca dá ve res fo ram sub me ti das a to mo gra fia com pu ta do ri za da
em es pi ral por meio de cor tes axi a is e re cons tru ções em 3D-TC fo ram ob ti das por meio da téc ni ca de vo lu me, uti li zan -
do re cur sos da com pu ta ção grá fi ca. Me di das cra ni o mé tri cas (n = 10) fo ram de ter mi na das nas ima gens em 3D-TC por
dois exa mi na do res in de pen den te men te, duas ve zes cada um, e uma aná li se de erro pa drão per cen tu al das me di das in -
tra- e in ter-exa mi na do res foi re a li za da. Os re sul ta dos de mons tra ram um erro pa drão per cen tu al ba i xo apre sen ta do
por es sas me di das, va ri an do en tre 0,85% e 3,09%. Em con clu são, as me di das li ne a res ob ti das nas es tru tu ras ós se as e
te gu men ta res fo ram con si de ra das pre ci sas em 3D-TC, com alta qua li da de e re so lu ção de ima gem.
DESCRITORES:
Tomografia computadorizada por raios X; Crânio; Antropologia forense.
IN TRO DUC TION
The study of anthropometric char ac ter is tics is
of fun da men tal im por tance to solve prob lems re -
lated to iden ti fi ca tion. Craniometric fea tures are
in cluded among these char ac ter is tics which are
closely con nected to fo ren sic den tistry, since they
can be used to aid in iden ti fy ing an in di vid ual from
a skull, found de tached from its skeleton
12
.
Craniometric anal y sis is per formed by lo cat ing
the anthropometric points and de ter min ing the
anthropometric mea sure ments, which have al -
ready been es tab lished in the lit er a ture and sub di -
vided into spe cific stud ies and can usu ally be ob -
tained by sim ple rul ers, cal i pers and other spe cific
tools.
6,7,11
.
The main ad van ta ge of com pu ted to mo graphy
(CT) is that over la ying struc tu res can be vi su a li -
zed, and less at te nu a ting struc tu res can be dis tin -
guis hed from more at te nu a ting struc tu res. Also
the axi al and co ro nal ima ges can be ac qui red, and
the ori gi nal data can be re for mat ted ba sed on dif fe -
rent vi ews
2,3,4,5
.
Asso ci a ted to the ad van ces in com pu ter grap -
hics, three-di men si o nal re cons truc ted ima ges
from CT (3D-CT) have been used in se ve ral dif fe -
rent cli ni cal ap pli ca ti ons, as tu mor me a su re -
ments, den tal im plants plan ning, eva lu a ti on of
cra ni o fa ci al de for mi ti es, and for ant hro po me tric
stu di es
1,2,3,4,5,8,10,13
.
Thus, the pur pose of this pa per is to pres ent an
in no va tive meth od ol ogy for fo ren sic in di vid ual iden -
ti fi ca tion, by de ter min ing anthropometric mea sure -
ments us ing 3D-CT com puter graph ics re sources.
24
Radiologia
*Gra du a te stu dent, Fo ren sic Den tistry; **Pro fes sor, De part ment of So ci al Den tistry; ***Pro fes sor, De part ment of Ra di o logy – Scho ol
of Den tistry, Uni ver sity of São Pa u lo.
MA TE RIAL AND METHODS
The study pop u la tion con sisted of five ca daver
heads that were scanned in a spi ral CT (Toshiba
X-Press) for re search pur poses, with 3 mm thick
ax ial slices with 2 mm of re con struc tion in ter val per
sec ond of time, 512 x 512 ma trix, 120 kVp and
100 mA. The orig i nal CT data were stored in a
CD-ROM and trans ferred to an in de pend ent work -
sta tion (Dell 420 Pre ci sion, with Win dows NT) us ing
a Vitrea 2.3 ver sion vol u met ric vi su al iza tion soft -
ware (Vi tal Im ages Inc., Plym outh, MN, USA) (Fig -
ure 1). Multiplanar (co ro nal and sagittal) and 3D re -
con struc tion of the bone and tegumentar pro to cols
were ob tained si mul ta neously, based on the orig i -
nal ax ial slices.
The anthropometric points (Ta ble 1) were lo -
cated and the cor re spon dent mea sure ments (Ta -
ble 2) were de ter mined elec tron i cally by two ob -
serv ers twice each, in de pend ently, us ing the
com puter soft ware tools (Fig ures 2, 3, 4 and 5). Vi -
su al iza tion, ma nip u la tion and anal y sis were pos -
si ble us ing pre es tab lished cri te ria for bone pro to -
col and soft tis sue pro to cols. Sev eral pa ram e ters
were used:
1. the side to lo ca te la te ral po ints (left);
2. the stan dar di za ti on of fron tal and la te ral po si ti -
ons using pro gram func ti ons;
3. the same head po si ti on to de ter mi ne the me a -
su re ments in the two types of pro to cols (bone
and soft tis sue); and
4. the ques ti on whet her or not a trans pa rency
should be used.
Af ter wards, a stan dard er ror per cent age anal y -
sis was con ducted for the bone and soft tis sue
mea sure ments – intra- and inter-ob serv ers – to
test the pre ci sion of these mea sure ments.
RE SULTS
Graph 1 shows the quan ti fi ca tion of stan dard
er ror per cent age for bone and soft tis sue
25
Ro cha S dos S, Ra mos DL de P, Ca val can ti M de GP. Appli ca bi lity of 3D-CT fa ci al re cons truc ti on for fo ren sic in di vi du al iden ti fi ca ti on.
Pes qui Odon tol Bras 2003;17(1):24-8.
TABLE 1 - Anthro po me tric po ints.
Nº
Abreviations
Name
1
Al
Ma xi mum width of na sal
ope ning
2
G
Gla bel la, most an te ri or po int
of the skull in sa git tal pla ne
3
M
Men ton, most in fe ri or and an te -
ri or po int of the men tal re gi on
4
N
Na si on, me di al po int bet we en
the na so fron tal su tu re and the
na sal ori gin in mid sa git tal pla ne
5
Ns
Na sos pi na le, hig hest po int on
na sal spi ne
6
Op
Opist ho cra ni on, most an te ri or
part of skull in mid sa git tal
pla ne
7
Po
Po ri on, hig hest po int of su pe -
ri or bor der of the in ter nal
au di tory ca nal
8
Zy
Zygi on, most la te ral po int of
zygo ma tic arch
TABLE 2 - Anthro po me tric me a su re ments.
Nº
Abreviations
Name
1
Al-Al
Na sal bre adth
2
G-Op
Skull length
3
N-M
Fa ci al he ight
4
N-Ns
Na sal he ight
5
Po-Al
Cam per’s pla ne
6
Po-G
Dis tan ce bet we en Po and G
7
Po-M
Dis tan ce bet we en Po and M
8
Po-N
Dis tan ce bet we en Po and N
9
Po-Ns
Dis tan ce bet we en Po and Ns
10
Zy-Zy
Fa ci al bre adth
FIGURE 1 - Inde pen dent work sta ti on (Dell) using Vi -
trea
®
soft wa re.
anthropometric mea sure ments, in or der to an a lyze
the pre ci sion of each ob server and to make an
interobserver com par i son. The stan dard er ror per -
cent age in all these eval u a tions was found to be
low, rang ing be tween 0.85% and 3.09%. The re -
sults show that the great est rate of er ror was found
be tween dif fer ent bone mea sure ments per formed
by ex am iner one, 3.09%, and the least rate of er ror
was that be tween dif fer ent soft tis sue mea sure -
ments per formed by ob server two, 0.85%. The
26
Ro cha S dos S, Ra mos DL de P, Ca val can ti M de GP. Appli ca bi lity of 3D-CT fa ci al re cons truc ti on for fo ren sic in di vi du al iden ti fi ca ti on.
Pes qui Odon tol Bras 2003;17(1):24-8.
FIGURE 2 - 3D-CT fron tal view of a bone me a su re ment:
N-Ns (53.4 mm).
FIGURE 4 - 3D-CT fron tal view of a soft tis sue me a su re -
ment: N-Ns (60.0 mm).
FIGURE 3 - 3D-CT la te ral view of a bone me a su re ment:
Po-N (100.3 mm).
FIGURE 5 - 3D-CT la te ral view of a soft tis sue me a su re -
ment: Po-N (125.5 mm).
mea sure ments of the two ob serv ers were com -
pared, and er ror fac tors of 1.95% and 1.61%, re -
spec tively, were found in the bone and soft tis sue
mea sure ments.
DIS CUS SION
Aiming at hu man iden ti fi ca tion through cranio -
facial anal y sis, and con sid er ing that craniometry
re quires that points can be lo cated with pre ci sion
and mea sure ments can be de ter mined with ac cu -
racy, we sug gest that 3D-CT is the more re pro duc -
ible method.
3D-CT im ag ing has been proven to be more ac -
cu rate in de ter min ing mea sure ments than im ag -
ing per formed di rectly on CT slices and 2D-CT im -
age reconstruction
3,8
. Hildelbolt, Vannier
10
, in
1990, per formed mea sure ments di rectly on con -
ven tional CT slices and on 3D re con struc tion,
com par ing with those per formed with a cal i per,
and stated that 3D-CT mea sure ments were more
ac cu rate than those taken from CT slices. Later, in
1998, Cavalcanti, Vannier
3
val i dated lin ear
craniofacial mea sure ments in 2D-CT and 3D-CT
im ages us ing a spi ral CT and com par ing them with
the cor re spon dent phys i cal mea sure ments. The
au thors con cluded that, in many cases, the lin ear
mea sure ments were con sid ered more ac cu rate in
3D-CT images
3
. In our cur rent re search,
craniofacial anal y sis, us ing a newer com puter
graph ics meth od ol ogy, of fered ad van ta geous fea -
tures that have shown to be fun da men tal:
a) Ima ge seg men ta ti on. This is very im por tant
when in ter nal po ints must be eva lu a ted
13
.
b) Qu ick ma ni pu la ti on of the ima ge may be re qui -
red, for exam ple, when de ter mi ning a me a su re -
ment by lo ca ting one po int that is fron tal and
anot her that is la te ral. The pro grams used to
take up to a few se conds to make a sim ple ro ta -
ti on
9
.
c) Ima ge qua lity. It of fers ex cel lent co lor sca le and
trans pa rency.
d) Eva lu a ti on of vo lu me, area, and both li ne ar and
an gu lar me a su re ments.
The 3D vol ume ren der ing tech nique was used
in our meth od ol ogy be cause of its great ad van tage
of be ing able to as sim i late all the in for ma tion from
the orig i nal CT slices. Fur ther more, it is able to
work with a dif fer ent range of col ors and trans par -
en cies, and is con sid ered a very re fined method for
vol ume ren der ing of 3D imaging
4
. The 3D vol ume
tech nique used the most mod ern method avail able
for vol ume re con struc tion and sur passed the 3D
sur face tech nique used in ear lier work
3
.
In a study to eval u ate vol ume mea sure ments
re gard ing oral tu mors us ing 3D-CT, Cavalcanti,
Vannier
4
stated that the same soft ware al lowed di -
rect vol ume re con struc tion, of fered better data
stor age, mak ing it eas ier to an a lyze three dimensio -
nally, and to ob tain a 3D im age of high qual ity. In
our study, we also ob served that this pro gram en -
abled us to lo cate anthropometric points, even
those which could be lo cated only by ob serv ing su -
tures, which was made pos si ble by us ing trans par -
ency scales. Ad di tionally, it was pos si ble to work
on the im age quickly with ro ta tion and trans la tion
move ments, to change the zoom view, or even to
seg ment points of in ter est.
Ac cord ing to the find ings, the stan dard er ror
per cent age was ad e quate for this type of anal y sis.
This is shown in Graph 1, where we ob served that
the great est rate of er ror was 3.09% for ex am iner
one, in bone mea sure ments, and the least rate of
er ror was 0.85% for ex am iner two, in soft tis sue
mea sure ments.
We be lieve that the in tro duc tion of this meth od -
ol ogy and tech nol ogy rep re sents a great in no va tion
in the field of Fo ren sic Den tistry for hu man iden ti -
fi ca tion, pri mar ily in ma jor ur ban cen ters where
there is a great con cen tra tion of peo ple and where
the speed and the ob jec tiv ity of the method are fac -
tors of fun da men tal im por tance.
None the less, a study with a greater num ber of
data and with com par i sons to phys i cal mea sure -
27
Ro cha S dos S, Ra mos DL de P, Ca val can ti M de GP. Appli ca bi lity of 3D-CT fa ci al re cons truc ti on for fo ren sic in di vi du al iden ti fi ca ti on.
Pes qui Odon tol Bras 2003;17(1):24-8.
GRAPH 1 - Assess ment of stan dard er ror of in tra- and
in te rob ser ver (n = 2) me a su re ments of bone and soft tis -
su es ana to mi cal po ints.
ments will be needed to de ter mine the va lid ity (ac -
cu racy and pre ci sion) of 3D-CT com puter graph ics
fo ren sic in di vid ual iden ti fi ca tion.
CON CLU SIONS
1. It was pos si ble to es ta blish a qua li ta ti ve stan -
dard for quan ti ta ti ve analy sis of 3D-CT ima ges,
in bone and soft tis sue pro to col, with high ima -
ging re so lu ti on and pro ces sing.
2. 3D-CT ima ging using the vo lu me ren de ring
tech ni que by com pu ter grap hics made it pos si -
ble to de ter mi ne the cra ni o fa ci al me a su re ments
with ade qua te pre ci si on.
AC KNOWL EDGE MENTS
FAPESP, Mas ter’s de gree - Sara dos Santos
Rocha (01/05700-3) - and Grants - (99/10276-4)
Pro fes sor, Ph.D., Marcelo G. P. Cavalcanti.
REF ER ENCES
1. Arbenz GO. Me di ci na le gal e an tro po lo gia fo ren se. Rio de
Ja ne i ro: Athe neu; 1988.
2. Ca val can ti MGP, Yang J, Ru precht A, Van ni er MW. Va li da -
ti on of spi ral com pu ted to mo graphy for den tal im plants.
Den to ma xil lo fac Ra di ol 1998;27:329-33.
3. Ca val can ti MGP, Van ni er MW. Qu an ti ta ti ve analy sis of spi -
ral com pu ted to mo graphy for cra ni o fa ci al cli ni cal ap pli ca -
ti ons. Den to ma xil lo fac Ra di ol 1998;27:344-50.
4. Ca val can ti MGP, Van ni er MW. Me a su re ment of the vo lu me
of oral tu mors by three-di men si o nal spi ral com pu ted to mo -
graphy. Den to ma xil lo fac Ra di ol 2000;29:35-40.
5. Cavalcanti MGP. Tomografia computadorizada: reconstru -
ções em 2D e em 3D. In: Freitas A, Rosa EJ, Souza IF.
Radiologia odontológica. 5
a
ed. São Paulo: Artes Médicas;
2000. p.681-726.
6. Ca val can ti MGP, Ru precht A, Van ni er MW. Eva lu a ti on of
an os sif ying fi bro ma using three-di men si o nal com pu ted
to mo graphy. Den to ma xil lo fac Ra di ol 2001;30:342-5.
7. Far kas LG. Anthro po metry of the head and face. New York:
Ra ven Press; 1994.
8. Fish man EK, Ma gid D, Ney DR, Cha ney EL, Pi zer SM, Ro -
sen man JG, et al. Three-di men si o nal ima ging. Ra di o logy
1991;181:321-37.
9. Hil de bolt CF, Van ni er MW. Three-di men si o nal me a su re -
ment ac cu racy of skull sur fa ce land marks. Am J Phys
Anthro pol 1988;76:497-503.
10. Hil de bolt CF, Van ni er MW, Knapp RH. Va li da ti on study of
skull three-di men si o nal com pu te ri zed to mo graphy me a su -
re ments. Am J Phys Anthro pol 1990;82:283-94.
11. Offutt CJ. Vo lu me tric 3-D ima ging of com pu te ri zed to mo -
graphy scans. Ra di ol Tech nol 1990;61:312-9.
12. Ono I, Ohu ra T, Na ru mi E, Ka was hi ma K, Mat su no I, Na ka -
mu ra S, et al. Three-di men si o nal analy sis of cra ni o fa ci al
bo nes using three-di men si o nal com pu ted to mo graphy. J
Cra ni o ma xil lo fac Surg 1992;20:49-60.
13. Ramos DLP. Contribuição ao estudo da aplicação de foto -
grafias de face em perícias de identificação antropo lógica
quando comparadas com telerradiografias frontais [Tese de
Doutorado]. São Paulo: Faculdade de Odontologia da USP;
1995.
14. Silva M. Compêndio de Odontologia Le gal. São Paulo:
MEDSI; 1997.
15. Van ni er MW, Con roy GC, Marsh JL, Knapp RH. Three-di -
men si o nal cra ni al sur fa ce re cons truc ti ons using high-re -
so lu ti on com pu ted to mo graphy. Am J Phys Anthro pol
1985;67:299-311.
Recebido para publicação em 26/04/02
Enviado para reformulação em 28/08/02
Aceito para publicação em 27/01/03
28
Ro cha S dos S, Ra mos DL de P, Ca val can ti M de GP. Appli ca bi lity of 3D-CT fa ci al re cons truc ti on for fo ren sic in di vi du al iden ti fi ca ti on.
Pes qui Odon tol Bras 2003;17(1):24-8.
Dostları ilə paylaş: |