Digital breast tomosynthesis in diagnosis of dense breast lesions.
- Author:
A'qiao XU
1
;
Hongqin HE
1
;
Qiujun SHI
1
;
Zhiqing LI
2
;
Shengjian ZHANG
3
Author Information
1. Department of Radiology, Shaoxing Central Hospital, Shaoxing 312030, Zhejiang Province, China.
2. Department of Pathology, Shaoxing Central Hospital, Shaoxing 312030, Zhejiang Province, China.
3. Department of Radiology, Tumor Hospital Affiliated to Fudan University, Shanghai 200032, China.
- Publication Type:Journal Article
- MeSH:
Breast Neoplasms;
Female;
Humans;
Magnetic Resonance Imaging;
Mammography;
ROC Curve;
Retrospective Studies
- From:
Journal of Zhejiang University. Medical sciences
2019;48(2):186-192
- CountryChina
- Language:Chinese
-
Abstract:
OBJECTIVE:To evaluate the value of digital breast tomosynthesis (DBT) in diagnosis of dense breast lesions.
METHODS:Clinical and pathological data of 163 patients (58 benign lesions, 122 malignant lesions, and 180 lesions in total) with breast lesions undergoing surgical treatment in Shaoxing Central Hospital from January 2017 to December 2018 were retrospectively analyzed. The lesions were classified into non-homogeneous dense gland type and extremely dense gland type according to BI-RADS creterion. Breast MRI and DBT examinations were performed before the surgery. ROC curve was generated and the diagnostic efficacy of two examination methods for dense breast lesions was evaluated with pathological results as the gold standard. The detection rate, diagnostic accuracy of benign and malignant breast lesions were compared between two methods using chi-square test. The accuracy of lesion size preoperatively evaluated by MRI and DBT was analyzed by Pearson correlation.
RESULTS:The detection rate and diagnostic accuracy for benign breast lesions by MRI were higher than those by DBT (91.4% vs. 75.9%, =5.098, <0.05 and 89.7% vs. 67.2%, =8.617, <0.01). But there were no significant differences in detection rate and accuracy for malignant lesions by MRI and DBT (98.4% vs. 95.1%, =2.068, >0.05 and 94.3% vs. 91.8%, =0.569, >0.05). The areas under the ROC curves of MRI, DBT based on BI-RADS classification were 0.910 and 0.832, respectively (=1.860, >0.05). The sensitivities of MRI, DBT to breast lesions were 93.3% and 86.7%, and the specificities were 68.3% and 79.1%. DBT and MRI measurements were positively correlated with pathological measurements (=0.887 and 0.949, all <0.01).
CONCLUSIONS:DBT can effectively diagnose benign and malignant breast lesions under dense gland background, and it has similar diagnostic efficacy with MRI for breast malignant lesions.