1. Preoperative identification of microvascular invasion in hepatocellular carcinoma by XGBoost and deep learning
- Author
-
Yinan Deng, Guoying Wang, Kaining Zeng, Na Cheng, Xi-Jing Yan, Yi-Quan Jiang, Wei-Min Tang, Jian-ning Chen, Wen-Jing Huan, Wen-Qi Shi, Gui-hua Chen, Yang Yang, Kai Ma, Yefeng Zheng, Shilei Cao, Yang Haozhen, Chun-Kui Shao, Su-E Cao, and Jin Wang
- Subjects
Male ,Cancer Research ,medicine.medical_specialty ,Carcinoma, Hepatocellular ,Clinical variables ,Hepatocellular carcinoma ,Neural network models ,Original Article – Clinical Oncology ,Micro-vascular invasion ,Disease-Free Survival ,Cohort Studies ,03 medical and health sciences ,0302 clinical medicine ,Radiomics ,Humans ,Medicine ,Extreme gradient boosting ,Retrospective Studies ,Models, Statistical ,Training set ,Neovascularization, Pathologic ,Receiver operating characteristic ,business.industry ,Microcirculation ,Deep learning ,Liver Neoplasms ,General Medicine ,Middle Aged ,medicine.disease ,Confidence interval ,Oncology ,030220 oncology & carcinogenesis ,Female ,030211 gastroenterology & hepatology ,Artificial intelligence ,Radiology ,business - Abstract
Purpose Microvascular invasion (MVI) is a valuable predictor of survival in hepatocellular carcinoma (HCC) patients. This study developed predictive models using eXtreme Gradient Boosting (XGBoost) and deep learning based on CT images to predict MVI preoperatively. Methods In total, 405 patients were included. A total of 7302 radiomic features and 17 radiological features were extracted by a radiomics feature extraction package and radiologists, respectively. We developed a XGBoost model based on radiomics features, radiological features and clinical variables and a three-dimensional convolutional neural network (3D-CNN) to predict MVI status. Next, we compared the efficacy of the two models. Results Of the 405 patients, 220 (54.3%) were MVI positive, and 185 (45.7%) were MVI negative. The areas under the receiver operating characteristic curves (AUROCs) of the Radiomics-Radiological-Clinical (RRC) Model and 3D-CNN Model in the training set were 0.952 (95% confidence interval (CI) 0.923–0.973) and 0.980 (95% CI 0.959–0.993), respectively (p = 0.14). The AUROCs of the RRC Model and 3D-CNN Model in the validation set were 0.887 (95% CI 0.797–0.947) and 0.906 (95% CI 0.821–0.960), respectively (p = 0.83). Based on the MVI status predicted by the RRC and 3D-CNN Models, the mean recurrence-free survival (RFS) was significantly better in the predicted MVI-negative group than that in the predicted MVI-positive group (RRC Model: 69.95 vs. 24.80 months, p p = 0.027). Conclusion The RRC Model and 3D-CNN models showed considerable efficacy in identifying MVI preoperatively. These machine learning models may facilitate decision-making in HCC treatment but requires further validation.
- Published
- 2020