Analysis of breast ultrasound image features in predicting the risk of breast cancer

Objective We aimed to establish a logistic regression model to evaluate the value of breast ultrasound image feature standardized by BI-RADS in identifying malignant breast masses.To explore the diagnostic value of high frequency color Doppler ultrasonography in differential diagnosis of the benign and malignant breast masses by by applying breast BI-RADS(image report and data system) grading standard to carry out standard processing of ultrasound image features and establishing logistic regression model.Method The ultrasonographic data of 239 breast masses from 200 patients who underwent breast biopsy or surgery in Shenzhen Luohu District Maternal and Child Health Hospital from January 2017 to August 2018 were analyzed retrospectively.We retrospectively analyzed.Through the univariate analysis of ultrasound image features and the establishment of unconditional logistic regression model,the ultrasound features of breast masses were summarized,and the main points of ultrasound differential diagnosis of benign and malignant masses were put forward.Results The result of univariate analysis showed that the differences of six ultrasound signs such as morphology,boundary,microcalcification,blood flow,axillary lymph node abnormalities,posterior echo were statistically significant in benign and malignant groups(P<0.001).Nevertheless,Only burr or crab foot shape and microcalcification were retained in the logistic final regression model.Conclusion High-frequency color Doppler ultrasound had a high value in the differential diagnosis of benignant and malignant breast masses.High-frequency color Doppler ultrasound detection has simple operation,and it is harmless to our body,so it is the first choice for early diagnosis of breast tumors in clinical practice.