基于遗传算法安氏Ⅱ类、骨性Ⅱ类错患者下颌骨与颅面骨及颈椎骨的关系预测

Prediction of the relationship between the mandibular and craniofacial bone and cervical vertebrae based on a genetic algorithm in patients with skeletal class Ⅱ malocclusion

目的探讨东北地区安氏Ⅱ类、骨性Ⅱ类患者下颌骨各项参数与颈椎骨及颅面骨各参数之间的相关性,并建立相关方程,直观定量表达下颌骨与颈椎骨及颅面骨之间的关系,为正畸、正颌临床诊断及治疗提供依据并进行一定程度的预测。方法通过数码头颅侧位片对201例8~20岁的儿童和青少年的下颌骨、颅面部骨及颈椎骨进行测量。分为男性组75例与女性组126例,运用基于多因素耦合作用的全局敏感性分析方法选择与下颌骨参数敏感性较强的颅面骨及颈椎骨参数,并采用遗传算法建立相关方程。结果通过敏感性分析,选择出与下颌骨各测量值相关性最强的参数为第四颈椎高度H4、颅底长度SN,相关性较强的参数为颅底角SN-Ar、面高比SGo/NGn、Y轴角、下颌角Ar-Go-Gn。建立的方程为男性:Ar-Pg=28.415+1.818×H4+0.746×SN(r2=0.568,P<0.001);女性:Ar-Pg=15.168+1.706×H4+0.675×SN+0.31×SN-Ar-0.29×Y轴角(r^2=0.611,P<0.001)。建立的方程得出的预测值与实测值比较差异无统计学意义(P>0.05)。结论使用敏感性分析及遗传算法建立的下颌骨长度方程具有统计学意义并能进行一定程度的生长发育预测。

Objective To explore the correlation between the parameters of the mandible and parameters of cervical vertebrae and craniofacial bone in class Ⅱ skeletal patients in Northeast China and to establish correlation equations expressing the relationship between the mandible and cervical vertebrae and craniofacial bone directly and quantitatively for the clinical diagnosis and treatment of orthodontics and orthognathics and for prediction.Methods The mandible,cranial facial bone and cervical vertebrae of 201 children and adolescents aged 8 to 20 years were measured using digital cranial lateral tablets.All of the cases were divided into male(n=75)and female(n=126)groups using a sensitivity analysis method based on genetic algorithms to select the craniofacial bone and cervical bone with strong sensitivity to mandible parameters and to establish relevant equations.Results Through sensitivity analysis,the parameters with the strongest correlation between the measured values of the mandible were H4 and SN,those with a strong correlation were SN-Ar,the anterior and posterior high ratio SGo/NGn,the Y axis angle and mandibular angle Ar-Go-Gn.The established equation was as follows:males:Ar-Pg=28.415+1.818×H4+0.746×SN(r^2=0.0568,P<0.001);females:Ar-Pg=15.168+1.706×H4+0.675×SN+0.31×SN-Ar-0.29×Y axis angle(r2=0.611,P<0.001).No significant difference was found between the predicted values obtained by the established equations and measured values(P>0.05).Conclusion The mandibular length equation established by sensitivity analysis and genetic algorithms is statistically significant and can predict a certain degree of growth and development.