Objective To observe the value of artificial intelligence ultrasound combined with serum hyaluronic acid synthase 2(HAS2)and trefoil factor-1(TFF1)for early diagnosis of breast cancer.Methods Totally 176 patients with suspected breast cancer were retrospective enrolled and divided into malignant group(n=50)and benign group(n=126)according to pathological results.Artificial intelligence ultrasound and convolutional neural network algorithms were used to automatically label suspicious breast lesions.The lesions were manually graded based on breast imaging reports and data systems,classifying 0-3 grades as benign lesions,4-5 grades as malignant lesions.Clinical data and artificial intelligence ultrasound manifestations were compared between groups.Receiver operating characteristic curve was drawn,the area under the curve(AUC)was calculated to evaluate the efficacy of HAS2,TFF1,artificial intelligence ultrasound and their combination for diagnosing breast cancer.Results HAS2,TFF1,as well as the proportions of abnormal glandular thickness,low-echo lesions and abnormal blood flow morphology in malignant group were all higher than those in benign group(all P＜0.001).AUC of serum HAS2,TFF1 and artificial intelligence ultrasound for diagnosing breast cancer was 0.772,0.754 and 0.859,respectively.The combined diagnostic efficacy of the above three(AUC=0.925)was higher than single diagnostic efficacy(all P＜0.05).Conclusion Artificial intelligence ultrasound combined with serum HAS2 and TFF1 had good efficacy for early diagnosis of breast cancer.

Objective To observe the value of artificial intelligence ultrasound combined with serum hyaluronic acid synthase 2(HAS2)and trefoil factor-1(TFF1)for early diagnosis of breast cancer.Methods Totally 176 patients with suspected breast cancer were retrospective enrolled and divided into malignant group(n=50)and benign group(n=126)according to pathological results.Artificial intelligence ultrasound and convolutional neural network algorithms were used to automatically label suspicious breast lesions.The lesions were manually graded based on breast imaging reports and data systems,classifying 0-3 grades as benign lesions,4-5 grades as malignant lesions.Clinical data and artificial intelligence ultrasound manifestations were compared between groups.Receiver operating characteristic curve was drawn,the area under the curve(AUC)was calculated to evaluate the efficacy of HAS2,TFF1,artificial intelligence ultrasound and their combination for diagnosing breast cancer.Results HAS2,TFF1,as well as the proportions of abnormal glandular thickness,low-echo lesions and abnormal blood flow morphology in malignant group were all higher than those in benign group(all P＜0.001).AUC of serum HAS2,TFF1 and artificial intelligence ultrasound for diagnosing breast cancer was 0.772,0.754 and 0.859,respectively.The combined diagnostic efficacy of the above three(AUC=0.925)was higher than single diagnostic efficacy(all P＜0.05).Conclusion Artificial intelligence ultrasound combined with serum HAS2 and TFF1 had good efficacy for early diagnosis of breast cancer.