Application Of BI-RADS FFDM Classification System In The Diagnosis Of Benign And Malignant Breast Disease

Part Ⅰ Application of the Classic BI-RADS FFDM Classification in the Diagnosis of Benign and Malignant Breast LesionsObjective:To investigate and evaluate the effect of classic Breast Imaging Reporting and Data System FFDM classification in the diagnosis of benign and malignant breast lesions.Materials and Methods:From August2012to June2013,586patients were examined by Full field digital mammography in the outpatient clinic room.The patients aged from21to79years, average49.57years. The classification of breast lesions was conducted by two or three experienced radiologists.Image analysis were assessed according to BI-RADS classification standard. Cross-check analysis on BI-RADS classification assessment results and pathology outcome were made. The accuracy, sensitivity, specificity, positive predictive value, negative predictive value were calculated.All analyses were performed using statistical software SPSS19.0. Correlation between BI-RADS classification results and pathology outcome were tested with ROC curve.AP value below0.05was considered statistically significant.Results:(1) According to BI-RADS classification standard,586patients including219category3lesions and241category4lesions were found.(2) A total of586patients were detected in601lesions, including340benign lesions,261malignant lesions, respectively, accounting for56.6%,43.4%of the total.(3)601lesions of BI-RADS assessment of classification was compared with pathological results; the accuracy is82.9%(498/601), the sensitivity was92.0%(240/261), the specificity was75.9%(258/340), the positive predictive value was74.5%(240/322), the negative predictive value was92.5%(258/279); The area under the ROC curve was0.871,the standard error was0.015, the P vaule was0.000, the95%confidence interval was(0.842,0.900).Conclusions:BI-RADS classification are helpful in diagnosis and differential diagnosis of benign and malignant breast lesions.It is more useful in responding characteristics of the lesions compared to the past simple benign and malignant two hierarchical diagnosis model. It also has higher value in qualitative diagnosis and making effective therapeutic plan. Part Ⅱ Preliminary Analysis of Full Field Digital Mammography BI-RADS Classification and its Related Impact FactorsObjective:To investigate and evaluate the extent of influence of age, weight, menstrual status, oral contraceptive taken history and breast density on BI-RADS classification.Materials and Methods:From August2012to June2013,526patients were examined by Full field digital mammography in the outpatient clinic room. The patients aged from24to78years, average48.26years. The classification of breast lesions was conducted by two or three experienced radiologists. Image analysis was assessed according to BI-RADS classification standard. Age, weight, menstrual status, history of oral contraceptive use, breast density data were chosen as independent variable. All analyses were performed using statistical software SPSS19.0. The X2test was used.A P value below0.05was considered to be statistically significant.Results:(1) According to BI-RADS classification standard,526patients including193category3lesions and218category4lesions were found.(2) Comparison between BI-RADS classification and sample age information was:X2and P value were31.595and0.000, the result was considered to be statistically significant.(3) Comparison between BI-RADS classification and sample weight information was:X2and P value were41.251and0.000, the result was considered to be statistically significant.(4) Comparison between BI-RADS classification and sample menstrual status information was:X2and P value were24.077and0.000, the result was considered to be statistically significant.(5) Comparison between BI-RADS classification and sample oral contraceptive taken history information was:X2and P value were4.999and0.287, the result was not considered to be statistically significant.(6) Comparison between BI-RADS classification and breast density information was:X2and P value were28.244and0.005, the result was considered to be statistically significant.Conclusions:The using of BI-RADS system is a comprehensive evaluation of results, besides simple analysis of the image; it also should be combined with age, weight, menstrual status, oral history, breast density and other inspection message to make comprehensive judgment. Part Ⅲ Application of Modified BI-RADS FFDM Classification in the Diagnosis of Benign and Malignant Breast LesionsObjective:To investigate and evaluate the effect of modified Breast Imaging Reporting and Data System FFDM classification in the diagnosis of benign and malignant breast lesions.Materials and Methods:From August2012to June2013,469patients were examined by full field digital mammography in the outpatient clinic room. The patients aged from20to79years, average50.89years. The classification of breast lesions was conducted by two or three experienced radiologists. Image analysis was assessed according to the classic and modified BI-RADS classification standard. Cross-check analysis on different BI-RADS classification assessment results and pathology outcome were made. The accuracy, sensitivity, specificity, positive predictive value, negative predictive values were calculated. All analyses were performed using statistical software SPSS19.0. The ROC curve was used.A P value below0.05was considered to be statistically significant.Results:(1)In469patients,507lesions including264benign lesions,243malignant lesions, respectively, accounting for52.1%,47.9%of the total.(2)507lesions of classic BI-RADS classification was compared with pathological results; the accuracy is83.6%(424/507), the sensitivity was96.7%(235/243), the specificity was71.6%(189/264), the positive predictive value was75.8%(235/310),the negative predictive value was95.9%(189/197); The area under the ROC curve was0.911,the standard error was0.012,the95%confidence interval was(0.887,0.935).(3)507lesions of modified BI-RADS classification was compared with pathological results; the accuracy is87.2%(442/507), the sensitivity was96.3%(234/243), the specificity was78.8%(208/264), the positive predictive value was80.7%(234/290), the negative predictive value was94.5%(205/217); The area under the ROC curve was0.947,the standard error was0.011,the95%confidence interval was(0.926,0.967).(4) The accuracy, specificity, positive predictive value of the modified BI-RADS classification were higher than those of classic group (P=0.000,<0.05), but The sensitivity and negative predictive value of two groups were not significant.Conclusions:Modified BI-RADS classification can improve the specificity and positive predictive value of breast lesions, especially for malignant lesions, it is more benefit for meeting clinical need, but the degree of reliability still need clinical study of large sample more.