Value of breast MRI omics features and clinical characteristics in Breast Imaging Reporting and Data System (BI-RADS) category 4 breast lesions: an analysis of radiomics-based diagnosis
Qian Cui,Liang Sun,Yu Zhang,Zimu Zhao,Shuo Li,Yajie Liu,Hongwei Ge,Dongxue Qin,Yiping Zhao
DOI: https://doi.org/10.21037/atm-21-5441
IF: 3.616
2021-11-01
Annals of Translational Medicine
Abstract:BACKGROUND: The Breast Imaging Reporting and Data System (BI-RADS) category 4 breast lesions is categorized into 4A, 4B, and 4C, which reflect an increasing malignancy potential from low (2-10%) moderate (10-50%) and high (50-95%). Determining the benign and malignant of BI-RADS category 4 breast lesions is very important for accurate diagnosis and follow-up treatment. This study aimed to explore the value of breast magnetic resonance imaging (MRI) omics features and clinical characteristics in the assessment of BI-RADS category 4 breast lesions.METHODS: This retrospective study analyzed 96 lesions (39 benign and 57 malignant) from 92 patients diagnosed with MRI BI-RADS category 4 lesions in the Second Affiliated Hospital of Dalian Medical University between May 2017 and December 2019. The lesions were sub-categorized as BI-RADS 4A, 4B, or 4C based on the MRI findings. An imaging omics analysis model was applied to extract the MRI features. The positive predictive value (PPV) of each subcategory was calculated, and the area under the curve (AUC) was used to describe the efficiency for different diagnoses. Moreover, we analyzed 17 clinical indicators to assess their diagnostic value for BI-RADS category 4 breast lesions.RESULTS: The PPVs of BI-RADS 4A, 4B, and 4C were 7.1% (2/28), 41.2% (7/17), and 94.1% (48/51), respectively. The AUC, sensitivity, and specificity were 0.919, 84.2%, and 92.3%, respectively. The combination of T<sub>1</sub>-weighted images (T<sub>1</sub>WI) with dynamic contrast-enhanced (DCE) MRI yielded the best diagnostic results among all dual sequences. Two clinical indicators [progesterone receptor (PR) and Ki-67 expression] achieved an AUC almost equal to 1.0. The radiomics and redundancy reduction methods reduced the clinical data features from 1,233 to 14.CONCLUSIONS: High diagnostic performance can be achieved in distinguishing malignant breast BI-RADS category 4 lesions using the combination of T<sub>1</sub>WI and DCE in MRI. Combining the PR and Ki-67 expression variables can further improve MRI accuracy for breast BI-RADS category 4 lesions.
oncology,medicine, research & experimental