AUTHOR=Sun Shi Yun , Ding Yingying , Li Zhuolin , Nie Lisha , Liao Chengde , Liu Yifan , Zhang Jia , Zhang Dongxue 

TITLE=Multiparameter MRI Model With DCE-MRI, DWI, and Synthetic MRI Improves the Diagnostic Performance of BI-RADS 4 Lesions

JOURNAL=Frontiers in Oncology

VOLUME=Volume 11 - 2021

YEAR=2021

URL=https://www.frontiersin.org/journals/oncology/articles/10.3389/fonc.2021.699127

DOI=10.3389/fonc.2021.699127

ISSN=2234-943X

ABSTRACT=Objectives: To evaluate the value of syMRI, DWI, DCE-MRI and clinical features in breast BI-RADS 4 lesions, and to develop a more efficient method to help patients avoid unnecessary biopsy as much as possible.
Methods: 75 patients with breast diseases classified as BI-RADS 4 (45 with malignant lesions, 30 with benign lesions) were prospectively enrolled in this study. T1-weighted imaging (T1WI), T2-weighted imaging (T2WI), diffusion-weighted imaging (DWI), and synthetic MRI (syMRI) were performed at 3.0T. Relaxation time (T1 and T2), apparent diffusion coefficient (ADC), conventional MRI features and clinical features were measured and recorded. "T" represents the relaxation time value of the ROI pre-contrast scanning, and the "T+" represents the value post-contrast scanning. The relative change rate in T value between pre- and post-contrast scanning was represented by Δ T%. 
Results: ΔT1%, T2, ADC, age, BMI, menopause, irregular margins and heterogeneous internal enhancement pattern were all were significantly associated with breast cancer diagnosis. Four prediction models were established by combining the above parameters. Model 1 (BI-RADS model), model 2 (relaxation time model), model 3 (multi-parameter MRI model), model 4 (combined image and clinical model). model 4 has the best diagnostic performance (AUC =0.989), followed by model 3, 2 and 1 (AUC =0.962, 0.872, 0.856, all P <0.05).
Conclusions: In conclusion, the mpMRI model with DCE-MRI, DWI and syMRI is a powerful tool for evaluating possible malignant findings in BI-RADS 4 lesions. The addition of clinical features can further improve the diagnostic performance of the model.