| Age estimation of children is still a controversial issue in several fields. More and more challenges are upon the accuracy of the bone age(BA), since BA is the essential means to evaluate the biological age in the many fields, especially in sports. Therefore, it is urgent and important to seek for another more accurate evaluation method.Hundreds of researches have been carried out to estimate chronological age with kinds of methods. We discussed the feasibility to predict biological age with age-related morphological changes in tooth development and levels changes of methylation in gene espression, and then to expolre a favorable age assessment method with the contribution of one or more quantitative indicators of the markers. We discuss the research in the following two parts.Part 1 Age estimation based on the age-related morphological changes in toothObjectives: To carry out and compare the current dental age estimation methods in Chinese children. And to explore the age prediction model upon the quantitative indicators of the teeth.Methods: This is a cross-sectional study. 952 panoramic radiographs of children(414 boys, age ranging from 7-14 year-old) were selected from the patient record database of the Oral and Maxillofacial Radiology department of an education hospital affiliated to Shanghai Jiaotong University School of Medicine.The mean age of the children is 12.14 years old(SD 1.81). All orthopanthomographs were scored by using the criteria set by Demirjian et al.( 7-tooth, 4-tooth I and 4-tooth II) and also by Willems et al. The consistence of the developmental stage of permanent teeth of the left and right mandibular was discussed. Meanwhile, the number of teeth with root development completely closed(No), apical endsof the roots incompletely closed(S), was calculated. The distance between the inner sides of the open apex was measured. The ratios of pulp/root length, pulp/tooth length, tooth/root length and pulp/root width at A levels were also collected in lateral incisors, canines and second premolars. Cameriere method was discussed using the Europe model. And the regression models of age estimation were carried out also.Results: 1. The development of permanent teeth of the left and right mandibular is basically the same. More than 93% of concordance rate of teeth developmental stages were detected and the Kappa values was above 0.9 between the two sides. Evidently, the solo study of left mandibula is believable in Chinese children. However, the gender differences should be considered during dental age assessment. 2. The results of the dental estimation methods 2.1. The seven teeth dental estimation system The system includes Demirjian(7-tooth) and Willems methods. Significant differences were found between Demirjian dental age, boys’ Willems dental age and the chronological age(P<0.001), respectively. Overestimation of age was obvious in the Demirjian method. The Willems method controlled the overestimation, boys decreased from 1.68 years(SD 1.29) to 0.66 years(SD 1.19) and girls from 1.28 years(SD 1.68) to-0.02 years(SD 1.18). 2.2. The four teeth dental estimation system The system includes Demirjian method(4-tooth I) and Demirjian method(4-tooth II). Significant differences were found between dental ages and the chronological age in the two methods(P<0.001), respectively. Overestimation of age was common, but less than the dental age estimated by the Demirjian(7-tooth) method. The overestimation of age was 0.88years(SD 1.46) for boys and 0.21 years(SD 1.39) for girls in Demirjian method(4-tooth I). The overestimation of age was 1.24 years(SD 1.22) for boys and 0.53 years(SD 1.21) for girls in Demirjian method(4-tooth II). 2.3. Cameriere methodWe used the following equation,the Cameriere(Europe model), Age=8.387+0.282*sex-1.692*X5+ 0.835*No- 0.116*S-0.139*Sx No. The estimated age was 9.70 years(SD 2.32), with the underestimation of 2.49 years(SD 1.70) of age. And also underestimation of 2.45 years(SD 1.75) of age for boys and underestimation of 2.52 years(SD 1.66) of age for girls were detected. The MAE was more than 2.5 years for both genders. The prediction model is not suitable for Chinese children aged 7-14 years old. 3. Comparison among the dental age estimation methods Among the five dental age estimation methods, the Willems method was the most accurate method with the MAE of 0.95 years, which overestimated 0.16 years(SD 1.19) of age and 59.7% of the absolute difference between the dental age and CA is beyond one year. Overestimation of age is serious in the Demirjian(7-tooth) method with 1.45 years(SD 1.24) of age, and 67.5% overestimated by more than one year. The most inaccurate method is the Cameriere method with MAE of 2.64 years and underestimated 2.49 years(SD 1.70) of age, in which 79.6% underestimated by more than one year. 4. Stepwise linear regression analysis was performed with the quantitative indicators of the dental ages by the tooth developmental stage and measurements, and yielded the following equations, Age=9.738+0.284*W-0.019*S*No-3.737*X5-3.746*X3-0.293*RPM2(boy) Age=6.705+0.38* W-7.053*X4-0.052*S*No+2.863*RC(girl) The Willems dental age contributed the most to the two equtions. The model explained about 82.1% of total variance for boys and 73.5% for girls. MAE was 0.56 years for boys and 0.65 years for girls. We also explored another model with the variable of height in a limited sample, and yielded the following equations, Age=-5.799+0.242*W+0.096*Height-2.83*X4(boy) Age=-1.267 +0.229*W+0.080*Height+6.561*X6-0.976*X7-0.301*S(girl) The power of prediction was improved in both genders. The model explained about89.3% of total variance for boys and 88.0% for girls. MAE was 0.43 years for boys and 0.56 years for girls. 5. Relationship between the bone age and dental age A significant correlations between the developmental stage of every tooth and the bone age were found in several teeth. The second molar, the second premolar, the first premolar, and canine has the significant correlations with the age(P<0.001),and the worst correlation one is the center incistor. The significant correlations were found between the Willems dental age,the bone age and the chronological age(P<0.01). The Willems method had a higher correlation with the CA than the bone age. Overestimation of age is also found in bone age with 0.71years(SD 1.74)(P>0.05), with 1.28years(SD 1.40) in Willems dental age(P<0.05), and the most obvious overestimation of was found in Demirjian method(7-tooth)with 2.42 years(SD 1.68)(P<0.05).Conclusions: 1. The development of permanent teeth of the left and right mandibular is basically the same. Study of the left mandibula is reliable. However, the gender difference should be considered during dental age assessment. 2. It is feasibile to predict biological age with age-related morphological changes in tooth development. The Willems method is the most accurate method among the five methods, but modification is recommended before use. The most inaccurate method is the Cameriere method. The prediction model of the Cameriere method is not suitable for Chinese children. 3. We recommend the following two formulas of the age estimation for Chinese children aged 7-14 years old. The equation as following, Age=9.738+0.284*W-0.019*S*No-3.737*X5-3.746*X3-0.293*RPM2(boy) Age=6.705+0.38* W-7.053*X4-0.052*S*No+2.863*RC(girl) 4. Integrated multiple factors that can improve the power of age prediction. This study is mainly based on the dental panoramic film, and we also detected improved models with height considered in a small sample. We suggested that Integrated multiple factors willprovide future research directions.Part 2 The correlation between DNA methylation levels and chronological age in childrenObjectives: To better understand and clarify the relationship of DNA methylation levels and chronological age, based on the genome-wide DNA methylation analysis, and to explore the age prediction model upon the DNA methylation levels and bone age and other biological informations.Methods: Sixty eight health children aged 7 to 14 years old were enrolled. The finger-tip blood was drawn and finally 65 samples were qualified. We performed genome-wide methylation analysis in peripheral finger-tip blood white, using 450 K Infinium Methylation Bead Chip.Results: 1. Forty four age-related DNA methylated locus were detected, which were defined by the criteria of the absolute value of the correlation coefficient greater than or edqual to 0.5. There are 31 known genes and others still unkown. The 31 known genes are a protein-coding genes, involved in human cell growth, transcription, differentiation, metabolism, transport and many other features. The location of the known genes are in the gene body and also in the upstrean and downstream of the genes. 2. The methylation levels of the 41 genetic locus were negatively correlated with chronological age, only three methylation locus were positively correlated with the chronoligical age. 3. Stepwise linear regression analysis was performed with the quantitative indicators of the methylation levels of genetic locus and bone age and other information, and yielded the following equation, Age=13.828+0.526*sex+0.425*BA+0.0001*BW+0.276*Tanner+20.418*cg1-11.828*cg2-5.881*cg3-10.951*cg4+7.828*cg5-5.768*cg8-10.854*cg9-6.048*cg10-9.174*cg11+6.529*cg16-24.577*cg20-8.782*cg23+7.594*cg24+8.399*cg26+7.451*cg27-9.605*cg28+17.739*c g30+18.913*cg38+14.427*cg40-15.420*cg41-9.602*cg42(boy and girl combined) Bone age contributed the most to the equation, and the following was gene PRICKLE2(cg28) and gene APBA2(cg38).The model explained about 89.5% of total variance. The estimation age was 10.37 years(SD 1.59), and underestimated 0.1 years of age, with a MAE of 0.33 years. The total prediction error was lower than 0.6 years, which was better than other equations we had made.Conclusions: 1. In this study, we initially explored age-related gene methylation sites in the Chinese children with the genome-wide methylation chip, and gained considerable gene methylation sites which significantly associated with age. The methods were proved reliable and scientifical. 2. Forty four age-related DNA methylated locus were detected. The location of the known genes are in the gene body and also in the upstrean and downstream of the genes. The 31 known genes are a protein-coding genes, involved in human growth and development. The methylation levels of the 41 genetic locuses were negatively correlated with chronological age. 3. We recommend the following two formulas for the age estimation in Chinese young aged 7-14-year old. Equations as following, Age=13.828+0.526*sex+0.425*BA+0.0001*BW+0.276*Tanner+20.418*cg1-11.828*cg2-5.881*cg3-10.951*cg4+7.828*cg5-5.768*cg8-10.854*cg9-6.048*cg10-9.174*cg11 +6.529*cg16-24.577*cg20-8.782*cg23+7.594*cg24+8.399*cg26+7.451*cg27-9.605*cg28+17.739*cg30+18.913*cg38+14.427*cg40-15.420*cg41-9.602*cg42The total prediction error was lower than 0.6 years, which was better than other equations we had made. But verification was suggested before use. |
PDF Full Text Request