Multi-factor Relative Study And Prediction Model Of Operation Time Of Mandibular Third Molar Extraction

ObjectivesThe third mandibular molar extraction is one of the most common out-patient operations in oral and maxillofacial surgery.It is an interesting subject for scholars in recent years to find out the reasons potentially accounting for the difficulty of extraction of impacted mandibular third molars through clinical and imaging examinations,and to establish a convenient and efficient prediction model for clinical evaluation of the difficulty degree in extraction of mandibular third molar,which is suitable even for the community medical services.Accurate assessment of the difficulty of extraction of mandibular third molar is not only conducive to the design of reasonable and individualized treatment plan,but also to the communication between doctors and patients.Besides,it can also be helpful in control of postoperative reactions as pain,swelling and limitation of mouth opening.The prediction model can guide the clinical teaching in order to rationally arrange the progress and content of practice in a scientific and quantitative way,and at the same time,it can wisely guide the graduates correctly define the difficulty of extraction of impacted teeth in each case,so that they can make adequate preparations before operation.The purpose of this study was to investigate the anatomical factors in impacted lower third molar extraction and their influences on extraction time thereof,so as to find out the substantial relative actors influencing the extraction time;and to establish a more reliable prediction model of this extraction time of mandibular third molars,so as to help operators make more accurate prediction of surgery,and effectively formulate individualized treatment plans for patients,including:whether it is necessary to use medicine before and after operation,prepare appropriate surgical instruments and choose reasonable surgical methods,so as to shorten the operation time,reduce trauma and reduce postoperative complications.MethodsWith the informed consent of the patients,60 patients who had unilateral mandibular third molar extracted in oral and maxillofacial surgery clinic of Stomatological Hospital,Tianjin Medical University from 8^thOctober 2019 to 30th December 2019 were enrolled in this study.All surgical procedures were performed by the same one surgeon.The basic information of the patients,such as gender and age,was collected,and the anatomical status of each part of the mandibular third molar extracted by the preoperative panoramic sectional film was counted.According to the measurements,the time required for the operator to remove the proximal crown and the remaining part of the mandibular third molar after incision and inversion was recorded.The anatomical data of the impacted teeth were independent variables and the extraction time?T?was dependent variables.The LASSO?Least absolute shrinkage and selection operator?regression was used to screen the independent variables,and the factors that had little correlation with extraction time were eliminated.A reliable model was constructed to predict the extraction time of mandibular third molar.ResultsPreliminary analysis indicated that the interrelationship strengths between the independent variables and the dependent variable in order from high to low were as the following:root resistance?RB?>age?Age?>sex?Sex?>volume ratio of impacted teeth?V?>width of impacted teeth?W?=length of impacted teeth?L?>angle between impacted teeth and adjacent teeth?A?>ratio of impacted teeth to long axis?C/R?>depth of impacted teeth?D?>retro-molar space?RMS?,and most of the independent variables have a certain linear relationship.Using the penalty function idea of LASSO regression,each independent variable was added to the penalty constraint through the elastic network to select the most relevant independent variables.The three independent variables removed under the optimal Lambda coefficient were the ratio of impacted teeth to long axis?C/R?,the angle between impacted teeth and adjacent teeth?A?,and the retro-molar space?RMS?.The test results show that the model has residual normality,homogeneous variance and no strong influence point,so it accords with the linear relationship.ConclusionsIn this study,the resistance factors of impacted teeth were numeralized to be reflected by measured data.Ten independent variables were screened and analyzed by LASSO regression analysis.The LASSO model is suitable for the analysis of biological data of various types of variables,and has the advantages of less error and lower data requirements.Finally,seven factors were selected,including age?Age?,sex?Sex?,root resistance?RB?,volume ratio of impacted teeth?V?,width of impacted teeth?W?,length of impacted teeth?L?and depth of impacted teeth?D?.At the same time,the prediction model of extraction time was constructed.In our further study,more samples will be measured to compensate the small sample size which compromised the reliability of the study,and different types of mandibular third molar extraction cases will be substitute into this prediction model in order to verify its validity.