Correlation Between Breast Imaging Reporting and Data System of Ultrasound and 18F-lfuorodeoxyglucose PET/CT in Diagnosis of Breast Diseases / 中国医学影像学杂志

ABSTRACT

Purpose To evaluate the breast imaging reporting and data system (BI-RADS) of ultrasound and 18F-lfuorodeoxyglucose (FDG) PET/CT in diagnosis of breast tumor, and to analyze the correlation between them. Materials and Methods 18F-FDG PET/CT images and ultrasound images of 103 patients with suspected breast cancer were retrospectively analyzed to get correlation between the maximum standardized uptake values (SUVmax) and BI-RADS. Sensitivity, speciifcity, positive predictive value, negative predictive value were compared with histology or follow-up results as golden standard. Results Of the 103 lesions, 46 were benign and 57 were malignant. Pearson correlation coefifcient was 0.464 (P<0.01). The sensitivity, speciifcity, positive predictive value, and negative predictive value of PET/CT were 89.47%, 73.91%, 80.95%and 84.99%, respectively;those of BI-RADS were 94.70%, 69.60%, 79.42%and 91.38%, respectively. The sensitivity, speciifcity, positive predictive value, and negative predictive value in patients with BI-RADS 3-4 were 88.90%, 71.40%, 66.65%, and 90.91%, respectively. The sensitivity, speciifcity, positive predictive value, and negative predictive value for BI-RADS grading diagnosis were 88.90%, 46.40%, 51.60%and 86.67%, respectively. Conclusion There is no signiifcant correlation between SUVmax and BI-RADS. BI-RADS has low speciifcity for patients with BI-RADS grade 3-4, while PET/CT can make up this shortcoming. Combined diagnosis in the breast disease can be potentially recommended in clinics.