An MRI-based Radiomics Classifier for Preoperative Prediction of Ki-67 Status in Breast Cancer.

RATIONALE AND OBJECTIVES

This study aims to investigate the value of a magnetic resonance imaging-based radiomics classifier for preoperatively predicting the Ki-67 status in patients with breast cancer.

MATERIALS AND METHODS

We chronologically divided 318 patients with clinicopathologically confirmed breast cancer into a training dataset (n = 200) and a validation dataset (n = 118). Radiomics features were extracted from T2-weighted (T2W) and contrast-enhanced T1-weighted (T1+C) images of breast cancer. Radiomics feature selection and radiomics classifiers were generated using the least absolute shrinkage and selection operator regression analysis method. The correlation between the radiomics classifiers and the Ki-67 status in patients with breast cancer was explored. The predictive performances of the radiomics classifiers for the Ki-67 status were evaluated with receiver operating characteristic curves in the training dataset and validated in the validation dataset.

RESULTS

Through the radiomics feature selection, 16 and 14 features based on T2W and T1+C images, respectively, were selected to constitute the radiomics classifiers. The radiomics classifier based on T2W images was significantly correlated with the Ki-67 status in both the training and the validation datasets (both P < .0001). The radiomics classifier based on T1+C images was significantly correlated with the Ki-67 status in the training dataset (P < .0001) but not in the validation dataset (P = .083). The T2W image-based radiomics classifier exhibited good discrimination for Ki-67 status, with areas under the receiver operating characteristic curves of 0.762 (95% confidence interval: 0.685, 0.838) and 0.740 (95% confidence interval: 0.645, 0.836) in the training and validation datasets, respectively.

CONCLUSIONS

The T2W image-based radiomics classifier was a significant predictor of Ki-67 status in patients with breast cancer. Thus, it may serve as a noninvasive approach to facilitate the preoperative prediction of Ki-67 status in clinical practice.