ORIGINAL ARTICLE

Fetal age estimation using MSCT scans of deciduous toothgerms

Marie Minier & Delphine Maret & Fabrice Dedouit &Marion Vergnault & Fathima-Zohra Mokrane & Hervé Rousseau &

Pascal Adalian & Norbert Telmon & Daniel Rougé

Received: 20 March 2013 /Accepted: 20 June 2013# Springer-Verlag Berlin Heidelberg 2013

Abstract Evaluation of fetal age is an essential element inmany fields such as anthropology, odontology, paleopatholo-gy, and forensic sciences. This study examines the correlationbetween fetal age, femoral diaphyseal length (considered asthe gold standard), and deciduous tooth germs of fetuses aged22 to 40 weeks amenorrhea (WA) based on computed tomog-raphy (MSCT) reconstructions. Qualitative and quantitativestudies of femoral and deciduous tooth germ lengths wereperformed on 81 fetuses (39 females and 42 males). R soft-ware was used for statistical analyses. Intra-observer and inter-observer variabilities and the interclass correlation coefficient(ICC) were calculated. Correlation coefficients (R2) and linearregression equations were calculated. Intra- and inter-observervariabilities were very satisfactory (intra-observer ICC≥0.96,inter-observer ICC≥0.95). Femoral length was significantly

correlated with age (R2=0.9). The correlation coefficient be-tween age and height, width, and dental volume was R2≥0.73.Tooth germs were good indicators of fetal age. Our methodappears to be reliable and reproducible, and the results of thisstudy agreed with those of the literature. The dental formulaprovided a precise estimation of fetal age between 25 and 32WA. Tooth germs were reliable indicators of fetal age, andmultislice computed tomography was shown to be an innova-tive and reliable technology for this purpose.

Keywords Fetal age . Age estimation . Anthropology .

Deciduous teeth . Femur . MSCT

Introduction

Evaluation of fetal age is an essential element in forensicscience. Fetal viability criteria established by theWorld HealthOrganization are a weight of at least 500 g and age superior orequal to 22 weeks amenorrhea (WA). In France, the law ofJanuary 8 1993 officially defines fetal viability as 22 WA andthe child can be entered in civil registers from that age.

In obstetrics, fetal abdominal perimeter, cranial perimeter,and femoral length (FL) are used to date pregnancy. Thesemetric parameters are not applicable with partial remains.

Ultrasound is the main reference imaging tool for age deter-mination of the living fetus. However, this technique cannot betransposed to anthropological studies. To overcome this prob-lem, we used multislice computer tomography (MSCT). Mea-surement of the femoral diaphysis is currently the referencemethod for fetal age determination [1–4]. Many studies usingdirect measurements of dry bones have demonstrated a strongcorrelation between age and femoral length, and charts andregression formulas have been developed [1–4].

What can be done when the femur is missing? Bodies arenot always found complete or may be damagedwhen found. Itis then essential to have other age determination tools at our

M. Minier :D. Maret : F. Dedouit :M. Vergnault : F.<Z. Mokrane :N. Telmon :D. RougéLaboratoire d’Anthropologie Moléculaire et Imagerie de Synthèse,AMIS, UMR 5288 CNRS, Faculté de Médecine, Université deToulouse III, 37 Allées Jules Guesde, 31073 Toulouse, France

D. MaretFaculté de Chirurgie Dentaire, CHU Toulouse, 3 Chemin desMaraîchers, 31062 Toulouse, France

M. Minier (*) : F. Dedouit :M. Vergnault :N. Telmon :D. RougéService de Médecine Légale, Centre Hospitalier UniversitaireRangueil, 1 Avenue du Professeur Jean Poulhès, TSA 500032,31059 Toulouse Cedex 9, Francee-mail: minier.m@chu-toulouse.fr

F. Dedouit : F.<Z. Mokrane :H. RousseauService de Radiologie, Centre Hospitalier Universitaire Rangueil, 1Avenue du Professeur Jean Poulhès, TSA 50032, 31059 ToulouseCedex 9, France

P. AdalianAnthropologie Bioculturelle, Université de la Méditerranée, UMR6578 CNRS-EFS, Faculté de Médecine Nord, 51 Boulevard PierreDramard, 13916 Marseille, France

Int J Legal MedDOI 10.1007/s00414-013-0890-z

disposal [5–7]. We studied the fetal head and, more specifi-cally, the deciduous tooth germs. These anatomic elements areknown to be resistant to taphonomic processes. They alsoappear early in fetal development.

Research has been carried out on the use of tooth germs forestimation of prenatal age. Several atlases have been compiledillustratingdental formulas for fetuses at5and7months [8–10].However, they do not allow precise fetal age evaluation.

The aim of this work was to study correlations betweenfetal age and femoral length and tooth germs based on MSCTreconstructions. This is a very valuable technique because it isnon-invasive and does not require bone preparation. Fetal agecan be estimated when the body is not entire or is damaged.

Materials and methods

Sample

The initial sample was composed of 167 fetuses (74 femalesand 93 males), aged between 20 and 40 WA. Owing to thecreation of prenatal diagnostic centers established by theFrench law, the legal medicine laboratory of the MedicalFaculty, Marseille, France, has the possibility to study anon-ymous fetopathological records and selected those whosecause of death was identified as in utero death or spontane-ous abortion. This sample corresponded to postmortem ex-aminations performed between 1997 and 2000. They con-sidered fetuses as “normal” on the basis of several clinicalexaminations (X-ray examination, normal caryotype, andnormal aspect of the viscera for both macroscopic and his-tologic examinations). Gestational age was based on mater-nal data (last menses: Naegele’s rule) and completed withultrasound data obtained at 10 weeks gestation (this exami-nation is obligatory under French law). We excluded allcases in which these two methods gave discordant age esti-mations, as well as those that were too poorly preserved. Thiscollection of fetuses has already been the basis of a numberof studies at the legal medicine laboratory of the MedicalFaculty, Marseille, France [2, 4, 7, 11–13]. Age at death(WA), sex, and weight (grams) were known for all fetuses.

Data acquisition

The fetuses were scanned in the radiology department ofRangueil University Hospital, Toulouse, France. CT exami-nations were performed using a Sensation 16 scanner (Sie-mens, Erlangen, Germany) with a collimation of 16×0.75.The matrix was 500×500 pixels. Two filter reconstructionswere performed: soft filter (B31F) and bone filters (B60,B80). The axial reconstructions were performed with a thick-ness of 0.75 mm every 0.4 mm.

Data exporting, segmentation, and 3D reconstruction

The images were saved as Digital Imaging and Communi-cations in Medicine files then processed using Amira® 5.2.2software (Mercury Computer System, Inc., Chelmsford,MA). Semi-automatic segmentation of data sets was carriedout. After segmentation, the 3D triangle-based surface ofeach tooth was reconstructed in PLY format without smooth-ing to preserve its raw volume measurement.

Methods

For eachWA, an identical number of male and female fetuseswas randomly selected. All measurements were carried outby the same operator (MM). All qualitative and quantitativedata were tested. Six fetuses were selected at random, and theprincipal investigator (MM) carried out two series of mea-surements at an interval of 3 weeks. Inter-observer variabil-ity was tested in the same conditions (MV). The secondobserver was trained on the Amira® software. All cases weretested for error due to intra-observer bias using analysis ofvariance.

The R program was used for statistical analysis [14].Intra- and inter-observer reproducibility was assessed, andthe intra-class correlation coefficient (ICC) was calculated.Differences were considered significant at P<0.0.5.

Bone age assessment: femoral length

In anthropology, femoral diaphysis measurement is the refer-ence technique not only for fetal age determination but also forthe living. The strong correlation between fetal age and fem-oral length has been repeatedly demonstrated [1–4]. Referencetables and correlation equations have been established. Totalfemoral length (from proximal end to distal end) was mea-sured on a frontal view and expressed in millimeters.

Dental age assessment

Our purpose was to determine a correlation between fetal ageand teeth germs. Visual and morphometric studies wereconducted. For visual study, all deciduous maxillary (51 to55 and 61 to 65) and mandibular teeth germs (71 to 75 and 81to 85) were examined on a dynamic frontal anteroposteriorview of the mandible and the maxilla. The presence of toothgerm was scored 1 and absence was scored 0. The metricstudy was conducted on teeth 51 and 61 (first maxillary in-cisors). As visual study showed that germs 51 and 61 werepresent in all 81 fetuses examined, the metric study of thesetwo early developing germs appeared appropriate. Maximumheights and widths of teeth germs 51 and 61 were measured(mm) on a frontal view. These two germs were segmented tocalculate their respective volumes (mm3) (Fig. 1).

Int J Legal Med

Results

The femur and deciduous teeth germs were examined in 81fetuses aged between 22 and 40 WA (39 females and 42males) (Fig. 2). Fetuses aged 20 and 21 WA were excludedbecause the spatial resolution of MSCT images was poor inthese cases.

Intra- and inter-observer variabilities

Intra-observer variability (MM) and inter-observer (MV)variability were tested for all qualitative and quantitativemeasurements at a 3-week interval in six fetuses aged 30–40 WA. For femoral diaphysis measurement, intra-observerICC was 0.99 and inter-observer ICC was 0.97 (Table 1).The ICC was 1 for visual study (number of teeth germspresent). For all morphometric study of teeth germs (height,width, and volume), intra-observer ICC was 0.96 or greater,and inter-observer ICC was 0.95 or greater (Table 1).

Bone age assessment: femoral length

Thecorrelationcoefficientwasveryhighat0.89.Thefollowingcorrelation formulawas established (Fig. 3):

Fetal age ¼ 0:44 FL mmð Þ þ 6:45:

Dental age assessment

Qualitative study

For each fetus, all maxillary and mandibular deciduous teethgerms were studied. Visual study showed that mandibularteeth germs developed later than maxillary teeth germs.Teeth 51, 61, 71, and 81 (first maxillary and mandibularincisors) were present in all fetuses. Using the dental formu-la, the age was established at between 25 and 32 WA. At 25WA, the deciduous teeth germs 52 and 62 (second maxillaryincisors) were present in 100 % of the fetuses. We thusconsidered that if germ 52 or 62 was present, fetal age wasat least 25 WA. At 26 WA, the second mandibular incisors(72 and 82) were present. At 27 WA, maxillary canines (53and 63), first maxillary molars (54 and 64), and mandibularcanines (73 and 83) were present. At 28WA, first mandibularmolars (74 and 84) were present. At 30 WA, second maxil-lary molars (55 and 65) were present. At 32 WA, the second

Fig. 1 a Anterior teeth evaluation in frontal view on a 3D MSCTreconstruction in maximal intensity projection mode, (fetus aged 40WA). b 3D MSCT reconstruction of the first maxillary incisors (51, 61)(anterior view). Dental volumes were measured from thesesegmentations

Fig. 2 Sample distribution of femoral diaphysis and deciduous toothgerms according to age and sex (81 fetuses aged 22 to 40 WA)

Table 1 Intra-observer (MM) and inter-observer (MV) ICCs for mea-surements of the different parameters

Parameter measured Intra-observerICC

Inter-observerICC

p

Left femoraldiaphysis

0.99 0.97 <0.001

H51 0.98 0.98 <0.001

H61 0.96 0.95 <0.001

L51 0.98 0.97 <0.001

L61 0.98 0.97 <0.001

Volume 51 0.99 0.98 <0.001

Volume 61 0.99 0.99 <0.001

Int J Legal Med

mandibular molars (75 and 85) were present. Dental formulaand fetal age were therefore correlated, and the chronologicaldevelopment of teeth could be illustrated as in Schour andMassler and Ubelaker [8] (Fig. 4).

Metric study

The maximum height and width of teeth 51 and 61 (firstmaxillary incisors) were measured on a frontal view. Thesemeasurements seemed to be good indicators of fetal age. Fortooth 51, the R2 was 0.73 for width and 0.83 for height. Fortooth 61, the results were similar: R2=0.78 for width andR2=0.80 for height. Volumetric study also yielded a veryhigh R2 of 0.75 for tooth 51 and 0.77 for tooth 61. Correla-tion formulas were established (Table 2; Fig. 5).

Discussion

Our aim was to estimate the fetal age from two anatomic ele-ments: the femur and deciduous teeth germs. The originality of

this study is that it was based on MSCT reconstructions. Theliterature on the subject has essentially used direct measure-ments on dry bones or indirect measurements on radiographs[1, 2, 5, 15]. Direct measurements on dry bones are time-consumingbecause they requirebonepreparationbeforeexam-ination, and they are also invasive.We proposed an innovative,simple, and non-invasivemethod that preserved the integrity ofthe anatomical specimen. Our results highlighted importantcorrelations between fetal age and femoral and teeth germmeasurements.

Intra- and inter-observer variabilities

Intra- and inter-observer variabilities were very good for allvisual parameters and for the bone and tooth measurements.Our method thus appears to be reliable and reproducible.

Bone age assessment: femoral length

Charts and correlation equations have been established frommeasurements taken in large samples [1–4]. The followingregression formula was proposed by Adalian based on indi-rect radiographical measurement of femoral diaphyseallength [2]:

Fetal age ¼ 0:43 FL mmð Þ þ 6:94:

The correlation coefficient between age and femoral lengththat we obtained in our population was very high (R2=0.90).The resulting regression formula was the following:

Fetal age ¼ 0:44 FL mmð Þ þ 6:45

This is very close to that established by Adalian [2]. Theresults obtained by indirect measurement of the femoral di-aphysis on radiography or MSCT reconstruction were verysimilar. However, MSCT has the advantage of being indepen-dent of the skeleton’s position. The femur’s robustness in fetalage determination was demonstrated once again. We conclud-ed that measurements performed on MSCT images werereliable.

Fig. 3 Linear regression formula of left femoral length (mm) accordingto age (WA)

Fig. 4 Summary drawing of thefetal age estimation from dentalformula of 25 to 32 WA based onvisual assessment. The internalsurface of the left mandible andmaxillary was schematized as inSchour and Massler andUbelaker

Int J Legal Med

Dental age assessment

Qualitative study

Very few studies have been done on fetuses. Atlases of dentalformulas at different ages have been developed. The first,proposed by Massler et al., described two stages of develop-ment, at 5 and 7 months [8]. On the same model, Ubelakerproposed a second atlas presenting two more precise stages,at 5 and 7 months ± 2 months. AlQahtani et al. defined threestages, at 32, 36, and 38WA [10]. These three stages differednot by dental formula, but by dental development stage.

MSCTisofgreatvalue for toothgermstudy,as thegermscanbe visualized at an earlier stage thanwith ultrasound.However,thevariousdentalatlasesdidnotallowpreciseestimationoffetalage estimation. In our study, dental formulas for each week ofamenorrheaenablecorrecttheestimationoffetalagebetween25and 32 WA. Our results were similar to those obtained in theliterature. Six diagrams done in this study allowed earlier andmore precise fetal age estimation (Fig. 4).

Visual examination showed that mandibular teeth germsdeveloped later than maxillary teeth germs. This later ap-pearance of teeth germs was not described in previous stud-ies [8, 10].

Metric study

There are few dental metric studies of fetuses in the medicalliterature. Direct measurements or measurements on scanreconstructions have demonstrated a correlation with fetalage [1, 12, 15]. There was a very strong correlation betweenmaximum height and width of the first maxillary incisors (51and 61; 0.73≤R2≤0.83). Since these two teeth germs appearvery early in fetal development, they are valuable anatomiccriteria for fetal age estimation. The correlation coefficient ofheight was higher than that of width. Once again, our resultswere in agreement with the literature.

In addition to these two metric criteria, dental volume wascalculated from MSCT images using AMIRA® software.This new examination demonstrated the strength of the cor-relation between fetal age and dental volume. This correla-tion has not previously been studied. A correlation formula

was also established for each dental metric criterion, whichhas not been found in the literature.

Sexual dimorphism

We found no evidence of fetal sexual dimorphism in thefemur or teeth germs, in agreement with the literature [16].A study of sexual dimorphism of the fetal ilium carried out inthe same sample showed no evidence of a significant differ-ence [17]. These results support the absence of sexual di-morphism in fetuses.

Right–left asymmetry

We found no evidence of right–left asymmetry. The deter-mination of fetal dental formulas from 22 to 40 WA showedno evidence of maxillary or mandibular right–left asymme-try. In view of these results, fetal age could be preciselyestimated in cases where only part of the maxillary or man-dible is usable.

Table 2 Regression formulas for fetal age (WA) and correlation coefficient (R2) for length, height, and volume of first maxillary incisors (teeth 51and 61)

Tooth measured Width (mm) Height (mm) Volume (mm3)

Age formula R2 Age formula R2 Age formula R2

51 4.16 (W51)+11.65 0.73 4.32 (H51)+16.70 0.83 0.23 (V51)+25.86 0.75

61 4.41 (W61)+10.87 0.78 4.27 (H61)+16.92 0.80 0.24 (V61)+25.94 0.77

Fig. 5 Linear regression formula for tooth 51 (first maxillary incisor)height (mm) according to age

Int J Legal Med

Conclusion

Our work was the first to associate femoral length anddeciduous teeth germs by MSCT scans for fetal age estima-tion. Intra- and inter-observer reproducibility was very satis-factory. The element most strongly correlated with fetal agewas femoral diaphyseal length. This had already been ob-served in previous studies on a large sample of dry bones[1–4]. One of the most important features of the present workwas the establishment of a dental formula for each week ofamenorrhea from 22 to 40 WA. Previous atlases did notallow precise fetal age estimation. The visual scale that weestablished allowed reliable fetal age estimation between 25and 32 WA. Metric study demonstrated a very strong corre-lation between fetal age and deciduous tooth germ volume.This study was original because it was based on MSCTreconstructions, and all results were in agreement with theliterature. MSCT images appear to be a valuable basis forexamination because of their reliability and non-invasivecharacter. This morphometric study carried out on the fetalhead is of value for forensics, odontology, anthropology, andpaleopathology.

References

1. Deutsch D, Goultschin J, Anteby S (1981) Determination of humanfetal age from the length of femur, mandible and maxillary incisor.Growth 45:232–238

2. Adalian P (2001) Evaluation multiparamétrique de la croissancefœtale: application à la détermination de l’âge et du sexe. Thesis,Faculty of Medicine of Marseille, Marseille, France

3. Scheuer L, Black S (2002) Developmental juvenile osteology.Academic, California

4. Adalian P, Piercecchi-Marti MD, Bourliere-Najean B, Panuel M,Fredouille C, Dutour O, Leonetti G (2001) Postmortem assessmentof fetal diaphysis femoral length: validation of a radiographicmethodology. J Forensic Sci 46:215–219

5. Kósa F (2002) Anthropological study for the determination of theEuropid and Negroid characteristics on facial bones of humanfetuses. Act Biol Szeged 46:83–90

6. Kósa F (2000) Application and role of anthropological research inthe practice of forensic medicine. Act Biol Szeged 44:179–188

7. Adalian P, Piercecchi-Marti MD, Bourlière-Najean B, Panuel M,Leonetti G, Dutour O (2002) Nouvelle formule de détermination del’âge d’un fœtus. CR Biol 325:261–269

8. Massler L, Schour I, Poncher HG (1941) Developmental pattern ofthe child as reflected in the calcification pattern of the teeth. Am JDis Child 62:33–67

9. Smith EL (1999) A test of Ubelaker’s method of estimating sub-adult age from the dentition. Thesis, University of Texas at Austin,Texas

10. AlQahtani SJ, Hector MP, Liversidge HM (2010) Brief communi-cation: the London atlas of human tooth development and eruption.Am J Phys Anthropol 142:481–490

11. Foti B, Perez-Guevaras S, Adalian P, Piercechi-Marti MD, BartoliC, Leonetti G (2007) Estimated age of gestation by biometric studyon fetal mandible (part 2). J Maxillofac Oral Surg 6:36–43

12. Lalys L, Ruquet M, Tardivo D, Laibi S, Bartoli C, Adalian P, PanuelM, Leonetti G, Foti B (2011) Estimation of gestational age fromtooth germs: biometric study of dentascan images. J Forensic Sci56:220–223

13. Minier M, Dedouit F, Mokrane F-T, Adalian P, Leonetti G, RougéD, Rousseau H, Telmon N (2012) Estimation de l’âge foetal parétude scanographique de la pars basilaris de l’os occipital. La RevMéd Lég 3:151–156

14. Core Team R (2002) R: a language and environment for statisticalcomputing. R Foundation for Statistical Computing, Vienna

15. Deutsch D, Pe’er E, Gedalia I (1984) Changes in size, morphologyand weight of human anterior teeth during fetal period. Growth48:74–85

16. Demirjian A (1973) A new system of dental age assessment. HumBiol 45:211–227

17. Mokrane FZ, Dedouit F, Gellée S, Sans N, Rousseau H, Rougé D,Telmon N (2013) Sexual dimorphism of the fetal ilium: a 3Dgeometric morphometric approach with multislice computed to-mography. J Forensic Sci. doi:10.1111/1556-4029.12118

Int J Legal Med