Preoperative prediction of regional lymph node metastasis of colorectal cancer based on 18F-FDG PET/CT and machine learning

• Published in: Annals of nuclear medicine 2021-05, Vol.35 (5), p.617-627
• Main Authors: He, Jiahong, Wang, Quanshi, Zhang, Yin, Wu, Hubing, Zhou, Yongsheng, Zhao, Shuangquan
• Format: Article
• Language: English
• Subjects: Algorithms; Cancer; Colorectal cancer; Colorectal carcinoma; Computed tomography; Feature extraction; Fluorine isotopes; Heterogeneity; Image processing; Imaging; Learning algorithms; Lymph nodes; Lymphatic system; Machine learning; Medicine; Medicine & Public Health; Metastases; Metastasis; Neural networks; Nomograms; Nuclear Medicine; Original Article; Positron emission; Positron emission tomography; Prediction models; Radiology; Radiomics; Regression analysis; Regression models; Support vector machines; Surgery; Test sets; Tomography; Training; Tumors
• ISSN: 0914-7187; 1864-6433; 1864-6433
• DOI: 10.1007/s12149-021-01605-8

Purpose To establish and validate a regional lymph node (LN) metastasis prediction model of colorectal cancer (CRC) based on 18 F-FDG PET/CT and radiomic features using machine-learning methods. Methods A total of 199 colorectal cancer patients underwent pre-therapy diagnostic 18 F-FDG PET/CT scans and CRC radical surgery. The Chang-Gung Image Texture Analysis toolbox (CGITA) was used to extract 70 PET radiomic features reflecting 18 F-FDG uptake heterogeneity of tumors. The least absolute shrinkage and selection operator (LASSO) algorithm was used to select radiomic features and develop a radiomic signature score (Rad-score). The training set was used to establish five machine-learning prediction models and the test set was used to test the efficacy of the models. The effectiveness of the models was compared by ROC analysis. Results The CRC patients were divided into a training set ( n  = 144) and a test set ( n  = 55). Two radiomic features were selected to build the Rad-score. Five machine-learning algorithms including logistic regression, support vector machine (SVM), random forest, neural network and eXtreme gradient boosting (XGBoost) were used to established models. Among the five machine-learning models, logistic regression (AUC 0.866, 95% CI 0.808–0.925) and XGBoost (AUC 0.903, 95% CI 0.855–0.951) models performed the best. In the training set, the AUC of these two models were significantly higher than that of the LN metastasis status reported by 18 F-FDG PET/CT for differentiating positive and negative regional LN metastases in CRC (all p