To Explore The Application Value Of Artificial Intelligence And High-Frame Rate Contrast-Enhanced Ultrasound In The Differential Diagnosis Of Breast BI-RADS4 Nodules

Objective: Taking pathological diagnosis as the "gold standard",To compare Artificial intelligence(AI)with High frame rate contrast ultrasound(Hi FR-CEUS)and their combination in the diagnosis of breast BI-RADS The application value in the differential diagnosis of four types of nodules.Methods: Patients with BI-RADS 4 breast nodules who were hospitalized in the Department of Thyroid and Breast Surgery,Taizhou People’s Hospital from December2021 to June 2022 were selected as the research objects(The patients had been diagnosed with BI-RADS 4 breast nodules by routine breast ultrasound examination in Taizhou People’s Hospital or other hospitals).A total of 80 female patients(80 lesions)were enrolled.Taking pathological diagnosis as the "gold standard",the diagnostic results of AI,Hi FR-CEUS alone and in combination were observed,and the diagnostic efficacy(sensitivity,specificity,accuracy,positive and negative predictive value)was calculated.Pearson correlation analysis between AI and Hi FR-CEUS was performed.The Receiver operating characteristic curve(ROC)was drawn to compare the accuracy of AI,Hi FRCEUS alone and HIFR-CEUS combined with AI in differentiating benign and malignant breast BI-RADS 4 nodules.Kappa consistency analysis was used to compare the reliability of the three diagnostic methods with the "gold standard".Results:(1)Eighty lesions were confirmed by pathological diagnosis,including 18 benign lesions and 62 malignant lesions.The sensitivity,specificity,accuracy,positive and negative predictive values of the three diagnostic methods were: AI,75.81%,94.44%,80.00%,97.92%,53.13%;Hi FR-CEUS: 74.20%,94.44%,78.75%,97.91%,51.51%;98.39%,88.89%,96.25%,96.83%,94.12% when the two were combined.The sensitivity,accuracy and negative predictive value of the combined diagnosis group were higher than those of the AI group and Hi FR-CEUS group(P< 0.05),but the specificity was lower than that of the AI group and HIFR-CEUS group(P< 0.05).There was no significant difference in the diagnostic efficacy between the AI group and HIFR-CEUS group(P>0.05).(2)There was significant positive correlation between AI and Hi FR-CEUS(r=0.249,P< 0.05).The area under the ROC curve(AUC)of AI,Hi FR-CEUS and their combination were 0.851±0.039,0.815±0.047 and 0.936±0.039,respectively.The AUC of combined diagnosis group was higher than that of AI group and Hi FR-CEUS group.The difference was statistically significant(Z1 = 2.207,Z2 = 2.477,P< 0.05).The AUC of AI group was higher than that of Hi FR-CEUS group,but the difference was not statistically significant(Z3 = 0.554,P> 0.05).(3)The AI group and Hi FR-CEUS group were moderately consistent with the "gold standard"(Kappa = 0.551,0.530,respectively),and the combined diagnosis group was highly consistent with the "gold standard"(Kappa =0.890).Conclusions:(1)Both AI and Hi FR-CEUS are effective methods for the differentiation of benign and malignant breast nodules in BI-RADS 4.(2)Compared with the single use of the two methods,the combined diagnosis of the two methods has higher diagnostic efficiency in the differential diagnosis of benign and malignant breast BI-RADS 4 nodules,which can further improve the accuracy of benign and malignant evaluation of this kind of nodules,and provide important reference value for clinical diagnosis and treatment.