The Decision Tree Model In The Applied Research In The Doppler Ultrasound Diagnosis Of Breast Tumor

Objective:Establish color droppler ultrasound date of breast tumor diagnosis decision tree model,to assist the identification of benign and malignant in breast tumor and improve thediagnostic accuracy of tumor.Method:1The object of Study: Collect The first affiliated hospital of University of SouthChina willing to do surgical pathology biopsy of breast disease patients, total251cases From2013August-2014January2Cases Collection: Collect The first affiliated hospital of University of South Chinatreated251cases of breast ultrasound of breast cancer patients, including the patientcase number, sex, age, the patients with breast ultrasound data, pathological resultsand so on.. All patients were female,73years old, the smallest27years old.3The case patients with breast ultrasound data:Breast ultrasound data including breasttumor location, quadrant, long diameter and short diameter, number, shape, edgevalue, there is no envelope, interior echo, and the distribution of blood flow withinand around the masses as well as the parameters of the25indexes.4Application Of The Decision Tree Model:①The case data assignment entry②Thedate is applied to The decision tree model called the flow A, read.③Run the decisiontree model, the prediction of benign and malignant breast tumors.④Follow up thepathological results of patients, compared with the prediction results⑤Collection to78patients, with78cases breast ultrasound data as data source, build a decision tree model called the flow B is used to predict new cases of breast cancer (173cases), theresults were compared with pathological results⑥In the collected251patients，themodel of decision tree is set up With the data source of251cases breastultrasound data, called the flow C.5Comparing the breast tumor decision tree model：Data of the78cases,173cases,251cases were analyzed by The decision tree model A, Data of the78cases,173cases,251cases were analyzed by The decision tree model B, Data of the78cases,173cases,251cases were analyzed by The decision tree model C, get thediagnostic accuracy of each model6The statistical results：:The three decision tree model related data were analyzed bySPSS18.0statistical software。Results:1The Results Of The Flow A： Comparison between the predicted flow A andpathological results of78cases of breast cancer patients,69cases the same,9caseswere different, calculate the accurate rate of diagnosis decision tree model for88.4%2The Results Of The Flow B： Comparison between the predicted flow B andpathological results of173cases of breast cancer patients,159cases the same,14cases were different, calculate the accurate rate of diagnosis decision tree model for91.9%3Analysis of three kinds of data for decision tree data model：①In78,173,251patients,the diagnosis accurate rate of decision tree data model A were87.18%,87.86%,88.81%②In78,173,251patients,the diagnosis accurate rate of decision treedata model B were91.03%、91.90%、91.27%③In78,173,251patients,thediagnosis accurate rate of decision tree data model C were93.18%、92.31%、93.05% 4The statistical results：The three decision tree model related data were analyzed bySPSS18.0statistical software, p<0.01, results showed that there was significantdifference between The three decision tree model, showed statistical significanceDiscussion:1the decision tree model in diagnosis accuracy rate of color Doppler sonography inthe diagnosis of benign and malignant breast tumors, with higher aspect.2the new Breast ultrasound image decision tree model with good prediction in thediagnosis of benign and malignant breast tumors, providing a good reference forclinicians and color Doppler sonography in the diagnosis of doctors, patients at highrisk of breast tumor by operation to whether certain screening effect.