Related Experiment Video
Updated: Sep 9, 2025
Tracking the Mammary Architectural Features and Detecting Breast Cancer with Magnetic Resonance Diffusion Tensor Imaging
Published on: December 15, 2014
MRI Radiomics Signatures of 21-Gene Recurrence Score for Predicting Survival in ER+/HER2- Breast Cancer
Yang Chen1,2, Lizhi Xie3, Wei Tang1,2
1Department of Radiology, Fudan University Shanghai Cancer Center, Shanghai, China.
Background:
Oncotype DX 21-gene assays are recommended for evaluating distant recurrence and guiding decisions on the use of adjuvant therapy in ER+/HER2- breast cancers. However, it cannot be widely applied due to the high cost and time consumption.
Purpose:
To identify MRI radiomics signatures within tumor and peritumoral tissues associated with the 21-gene recurrence score (RS) and explore their value in predicting 5-year recurrence in young women with ER+/HER2- breast cancer.
Methods:
Two datasets were analyzed; all the enrolled patients were diagnosed with ER+/HER2- breast cancers and underwent preoperative breast MRIs. The radiogenomic development dataset, with RS data obtained from April 2017 to March 2019, was used to identify optimal RS-signatures based on tumoral, peritumoral, and dilation radiomics features as well as clinical-imaging characteristics with the support vector machine method. The prognosis dataset, in which patients aged 18 to 40 years received breast surgery followed by adjuvant chemotherapy from 2012 to 2016, was used to evaluate the prognostic implication of the proposed optimal RS-signatures by measuring 5-year disease-free survival (DFS) with the Cox proportional hazards model.
Results:
159 patients (111 in the training and 48 in the validation groups) and 111 young patients (a mean follow-up time of 49.6 months and a 5-year DFS of 83.8%) were enrolled in these datasets, respectively. Areas under the receiver operating characteristic curve (AUCs) of the three optimal RS-signatures were 0.74 (95% CI: 0.59-0.87), 0.75 (95% CI: 0.61-0.88) and 0.74 (95% CI: 0.59-0.87) respectively. In the prognosis dataset, there were significant differences in survival between the patients in the predicted high-risk and low-risk groups categorized by the above three signatures, and the predicted recurrence risks were independent factors for DFS.
Conclusion:
The radiomic signatures within the tumor and peritumoral region exhibited potential to guide decisions on the use of chemotherapy and predict survival.
