MRI‐based radiomics signature for tumor grading of rectal carcinoma using random forest model

• Published in: Journal of cellular physiology Vol. 234; no. 11; pp. 20501 - 20509
• Main Authors: He, Bo, Ji, Tao, Zhang, Hong, Zhu, Yun, Shu, Ruo, Zhao, Wei, Wang, Kunhua
• Format: Journal Article
• Language: English
• Published: United States Wiley Subscription Services, Inc 01.11.2019
• Subjects: Cancer; Classifiers; Colorectal cancer; Evaluation; Feature extraction; Learning algorithms; Lesions; Machine learning; Magnetic resonance imaging; MRI; NMR; Nuclear magnetic resonance; Oversampling; Prediction models; Radiomics; radiomics feature; random forest; rectal carcinoma; Rectum; Training; tumor grading; Tumors; radiomics feature; rectal carcinoma; random forest; MRI; tumor grading
• ISSN: 0021-9541; 1097-4652; 1097-4652
• DOI: 10.1002/jcp.28650

The present study aimed to construct prospective models for tumor grading of rectal carcinoma by using magnetic resonance (MR)‐based radiomics features. A set of 118 patients with rectal carcinoma was analyzed. After imbalance‐adjustments of the data using Synthetic Minority Oversampling Technique (SMOTE), the final data set was randomized into the training set and validation set at the ratio of 3:1. The radiomics features were captured from manually segmented lesion of magnetic resonance imaging (MRI). The most related radiomics features were selected using the random forest model by calculating the Gini importance of initial extracted characteristics. A random forest classifier model was constructed using the top important features. The classifier model performance was evaluated via receive operator characteristic curve and area under the curve (AUC). A total of 1,131 radiomics features were extracted from segmented lesion. The top 50 most important features were selected to construct a random forest classifier model. The AUC values of grade 1, 2, 3, and 4 for training set were 0.918, 0.822, 0.775, and 1.000, respectively, and the corresponding AUC values for testing set were 0.717, 0.683, 0.690, and 0.827 separately. The developed feature selection method and machine learning‐based prediction models using radiomics features of MRI show a relatively acceptable performance in tumor grading of rectal carcinoma and could distinguish the tumor subjects from the healthy ones, which is important for the prognosis of cancer patients. We proposed an automatic tumor grading model accommodating the most important radiomics features with the method of random forest, and it holds the promise to be applied to help surgeon make an appropriate preoperative treatment plan and the evaluation of prognosis.