Diagnostic Value Of S-Detect Technique Combined With Adler Grading Method And Elastic Strain Rate Ratio Correcting BI-RADS Classification For Breast Masses

Objective To investigate the value of S-Detect technique combined with Adler grading method and elastic strain rate ratio correction BI-RADS classification in the diagnosis of benign and malignant breast masses.Methods A total of 298 patients with breast masses treated in the second affiliated Hospital of Anhui Medical University from January 2021 to October2022 were examined by conventional ultrasound(US),artificial intelligence(S-Detect),color Doppler flow imaging(CDFI)and strain elastography(SE).The results of conventional ultrasound BI-RADS classification,artificial intelligence S-Detect evaluation,Adler classification and elastic strain ratio(SR)were obtained.The BI-RADS classification was corrected by the results of Adler classification,SR and S-Detect,and the pathological results were taken as the "gold standard".The diagnostic efficacy of different diagnostic methods was evaluated by the receiver operating characteristic curve(ROC),and the area under the receiver working characteristic curve(AUC)of each diagnostic method was compared.The sensitivity,specificity,accuracy,positive predictive value(PPV)and negative predictive value(NPV)of different diagnostic methods were calculated.Results Among 298 cases of breast mass from January 2021 to October 2022,155 cases were malignant and 143 cases were benign.The classification AUC value of BI-RADS by S-Detect,CDFI,SR and their combined diagnostic correction was higher than that of conventional ultrasound.They were 0.831(0.781～0.881),0.782(0.727～0.837),0.828(0.778～0.878)and 0.866(0.821～0.921)respectively.The diagnostic efficacy of combined diagnosis was the best,and the AUC value was 0.866(0.821～ 0.921).Sensitivity 94.19%,specificity 79.02%,accuracy 86.91%,PPV 82.96%,NPV 92.62%.Z-test was used to compare different correction methods between groups.The difference between conventional ultrasound and different diagnostic methods was statistically significant(Z = 7.775,P < 0.0001;Z = 5.373,P < 0.0001;Z = 7.323,P < 0.0001;Z =12.789,P < 0.0001).Compared with S-Detect+BI-RADS,CDFI+BI-RADS and SR+BI-RADS,the differences were statistically significant(Z = 7.166,P < 0.0001;Z=8.242,P< 0.0001;Z = 7.239,P < 0.0001);The comparison of S-Detect+BI-RADS,CDFI+BI-RADS and SR+BI-RADS showed no statistical significance(Z = 1.700,P > 0.05;Z = 0.124,P > 0.05;Z = 1.592,P > 0.05).According to the 2013 edition of ACR BI-RADS≥4 lumps,puncture biopsy should be performed,and the puncture biopsy rate decreased from 80.20% to 59.06% after correction.Meanwhile,among the benign lumps wrongly diagnosed with class 4A or above given by conventional ultrasound in this study,74 cases of breast lumps were corrected after combined diagnosis,among which adenopathy was confirmed by pathology(21 cases).Adenopathy with fibladenoma(27 cases),fibladenoma(24cases),intraductal papilloma(2 cases);10 cases of malignant mass misdiagnosed as benign were detected,including invasive breast cancer(8 cases)and ductal carcinoma in situ(2 cases).Conclusion S-Detect,SR and CDFI have certain clinical value in differentiating benign and malignant breast masses,but they need to be further optimized to cooperate with clinical diagnosis and treatment.In the respective correction BI-RADS classification of S-Detect,SR and CDFI and the joint correction BI-RADS classification of the three,the combined correction BI-RADS classification is better than other modules in the diagnosis of breast masses,that is,artificial intelligence S-Detect combined with Adler classification and elastic strain rate correction BI-RADS classification can improve the diagnostic efficiency of breast masses without reducing the diagnostic sensitivity.Improve the specificity,accuracy,PPV and NPV of diagnosis,reduce unnecessary puncture biopsy and clinical intervention,and improve the detection rate of malignant breast masses.