Comparison Of The Efficacy Of DBT And Color Doppler Ultrasound In The Diagnosis Of Dreast Diseases And Its Predictive Model Construction

Objective?1?The aim is to study the difference between digital breast tomosynthesis?DBT?and color ultrasound in the diagnosis of breast mass and the value of combined application.?2?The other aim is to to establish a Logistic regression prediction model through the comprehensive analysis of the signs on DBT images of breast mass and clinical features and to evaluation and application of the model,to help identify the benign and malignant breast lesions.Methods?1?Collect breast patients of the Second Clinical Medical College affiliated to Fujian Medical University from May 2013 to July 2015.A total of 333 breast mass lesions was study,aged 10 to 81 years,mean age?44.22±10.61?years.The aim of this study is to compare the diagnostic performance of three diagnostic models?DBT,ultrasound,DBT+ultrasound?based on pathological results.?2?A retrospective analysis of 237 patients who were admitted to the Second Clinical Medical College affiliated to Fujian Medical University from April 2017 to February2018,aged 10 to 69 years,mean age?40.38±980?years old.A total of 255 breast masses,including 154 benign and 101 malignant masses,were diagnosed by pathology.The DBT's features of breast masses and the age of patients were collected,the DBT's features including the gland type of c and d,mass side,mass quadrant,mass maximum diameter?2cm,irregular shape of the mass,unclear edge of the mass,equal and high density of the mass,microcalcification,skin retraction,skin thickening,nipple retraction,axillary lymphadenopathy.Univariate analysis were used to analyze all of the 13 sign.Then,the sign that had statistically significant differences in differential diagnosis between benign and malignant breast masses as independent variables,pathology as the dependent variable,using binary logistic regession to filter out the most significant risk factors,then to establish regression model and evaluate the model.Results?1?Benign and malignant of breast lesions can be distinguished by these three modes?P<0.05?;?2?The specificity,misdiagnosis rate,accuracy and positive predictive value of DBT are significantly better than ultrasound;?3?Ultrasound shows that cystic lesions are better than DBT,and its sensitivity is superior to DBT in the diagnosis of benign and malignant breast masses;?4?The specificity of DBT+ultrasound is 99.5%,the misdiagnosis rate is 0.5%,and the positive predictive value is also 99%.The DBT+ultrasound examination mode is better than DBT or ultrasound alone.?2??1?Univariate analysis?² test shows that signs including age?40 years old,gland type of c and d,mass maximum diameter?2cm,irregular shape of the mass,unclear edge of the mass,microcalcification,skin retraction,skin thickening,nipple retraction,axillary lymphadenopathy had a statistically significant difference in the differential diagnosis of benign and malignant breast masses.?2?Logistic regression analysis showed that the sign entering the regression model were age?40 years?OR=4.198?,the shape of the tumor was irregular?OR=10.669?,the edge of the mass was not clear?OR=4.352?,microcalcification?OR=17.682?and skin thickening?OR=10.832?.?3?Using logistic regression model to evaluate the 255 breast masses in the study,the correct rate,sensitivity,specificity,missed diagnosis rate,misdiagnosis rate,the predicted probability of AUC and 95%CI of the model were 91.37%,87.13%,94.16%,12.87%,5.84%,0.957,?0.931⁰.984?respectively;?4?This model can be used to calculate the probability of breast cancer.Conclusion 1?DBT combined with ultrasound can improve the diagnostic efficacy of breast lesions.2?The predictive model based on DBT and clinical feture for breast cancer is Logit?P?=-4.221+1.434X₁+2.367X₄+1.471X₅+2.873X₆+2.383X₈3?The Logistic regression model is important for predicting the malignant risk of breast mass and can be used to guide clinical practice.