| BackgroundThe Breast Imaging Reporting and Data System(BI-RADS)is a breast imaging reporting and data system promulgated by the American College of Radiology in 2003 to provide a uniform reference and diagnostic standard for the diagnosis and prognosis of breast lesions,and BI-RADS is now widely used in clinical practice.Among the breast ultrasound reports,BI-RADS category 3 nodules are common and a potential lesion risk that is easily controversial in clinical management and easily ignored by patients.It is generally accepted that images of BI-RADS category 3 have a ≤2% probability of being ultimately diagnosed as malignant.However,the determination of BI-RADS in clinical practice is influenced by numerous factors to the extent that the chance of a pathologically confirmed malignant diagnosis of a BI-RADS category 3 tumour may exceed 2%.ObjectiveThe clinical significance of BIRADS-3 for the diagnosis of breast cancer was systematically evaluated using a Meta-analysis approach to provide supporting evidence for whether the conclusion that BI-RADS category 3 nodes are ≤2% likely to be malignant should be changed.MethodsSystematic searches of Chinese and English databases such as Web of Science,Pubmed,MEDLINE,China National Knowledge Infrastructure(CNKI),China Biology Medicine disc(CBM),Wipu Chinese journal full text database and Wanfang Medical Journal Network to collect the diagnostic tests of BI-RADS for determining benign and malignant breast nodules.Published,rigorously designed and methodologically reliable studies and reports were included,and the study data and diagnostic results were computer extracted to calculate true positive(TP),false positive(FP),false negative(FN)and true negative(TN)data for each study.negative,TN)data were used to construct a 2×2four-compartmental league table.The Quality Assessment of Diagnostic Accuracy Studies(QUADAS)scale [3] was used to assess the quality of the literature for the studies included in this study.All of the author’s Meta-analyses during this study were performed using Rev Man 5.0 and STATA 14.0 software,combining the BI-RADS 3 diagnostic breast cancer sensitivity(SEN),specificity(SPE),positive likelihood ratio(PLR),negative likelihood ratio(NLR),diagnostic odd ratio(DOR)and their 95% CIs,and generate a combined summary receiver operating characteristic(SROC)curve.characteristic(SROC)curves were generated.ResultsThe results of the combined Meta-analysis showed that the combined SEN,SPE,PLR,NLR and DOR of BIRADS-3 for breast cancer diagnosis were 0.82(95% CI: 0.81-0.84),0.69(95% CI: 0.68-0.70),4.18(95% CI: 2.63-6.67),0.15(95% CI: 0.07-0.30),29.59(95% CI: 12.34-70.94),0.15(95% CI: 0.07-0.30)and 29.59(95% CI: 12.34-70.94),respectively.The scatter of SROC curves showed an arm and shoulder-like distribution,indicating the presence of a threshold effect,and the Spearman correlation coefficient was-0.244(P>0.05),suggesting that heterogeneity was not Cochran’s Q=133.20,P<0.001,suggesting significant non-threshold-induced heterogeneity among the included literature,and the Meta-combined results of the literature were more robust.ConclusionIn the ultrasound examination of breast nodules,BIRADS-3 has a good application for the classification of benign and malignant nodules.However,SEN,SPE,PLR,NLR,and DOR in BIRADS-3 need to be improved,and better indicators are needed to improve diagnostic accuracy and reduce the possibility of delayed diagnosis.At the same time,follow-up should be strengthened for existing BIRADS-3 patients. |
PDF Full Text Request