Prediction Of Growth Potential Of Mandibular Retromolar Region In Adolescents Based On Genetic Algorithms Optimization

Objective:To explore the related influencing factors of the growth potential of mandibular retromolar region in adolescents in Jilin Province.To identify key impact factors and build prediction equations.The prediction equations were optimized using genetic algorithms(GAS)to provide reference for the analysis of orthodontic total arch space.Method:A total of 306 cases(152 males and 154 females)were screened according to the criteria for newly diagnosed patients aged 8-18 years at the department of orthodontics in Jilin University Stomatological Hospital from 2017 to 2019.The lateral cephalometric radiographs were used to obtain the space distal to the permanent molars(Retromolar Space,RMS)and cervical vertebrae measurements.Multiple linear regression analysis was used to evaluate the correlation between RMS and age,dental age,cervical vertebral maturation,etc.Screen out the most relevant factors.Predictive equations were initially established using linear regression analysis,and nonlinear equations were optimized using GAS.Results:1.There was no statistically significant difference in RMS mea surement results between different gender groups(P> 0.05).There were no statistically significant differences between the RMS measu ements of different vertical skeletal facial types(P> 0.05).The R MS measurements of different antero-posterior skeletal facial types s howed that there was no statistically significant difference between t he skeletal class?and ? groups(P> 0.05);the skeletal class?was all statistically different with the skeletal class?and ? groups(P <0.01),and the skeletal class?group was smaller than the sk eletal class?and ? groups.2.RMS was correlated with dental age and antero-posterior skeletal facial types.The most relevant factor was dental age of the mandibular third molar.3.RMS was positively correlated with dental age of the mandibular third molar.The correlation equations between the two after GAS optimization were: skeletal class?+? group:RMS =2.36 YL(0.81)+2.686;skeletal class?Group:RMS =2.36 YL (0.81)+0.723.4.Comparison between GAS optimized nonlinear equations and linear regression equations: The correlation coefficients between predicted values and the measured values of the test samples of the nonlinear equations were higher than those of the linear regression equations.The average errors of the nonlinear equations were smaller than those of the linear regression equations.Conclusion:1.The accuracy of nonlinear prediction equations optimized by GAS was better than that of linear regression equations.It could predict the growth potential of the posterior area of mandibular molars in adolescents to a certain extent.2.During the diagnosis and design of full dentition space analysis of dolescents in orthodontic clinic,dental age and antero-posterior skeletal facial types should be fully considered.3.Compared with the cervical vertebral maturation and age,dental age of the mandibular third molar was a better predictor of the growth potential of mandibular retromolar region.