Radiomics analysis enables recurrence prediction for hepatocellular carcinoma after liver transplantation

• Published in: European journal of radiology 2019-08, Vol.117, p.33-40
• Main Authors: Guo, Donghui, Gu, Dongsheng, Wang, Honghai, Wei, Jingwei, Wang, Zhenglu, Hao, Xiaohan, Ji, Qian, Cao, Shunqi, Song, Zhuolun, Jiang, Jiabing, Shen, Zhongyang, Tian, Jie, Zheng, Hong
• Format: Article
• Language: English
• Subjects: Artificial intelligence; Carcinoma, Hepatocellular - diagnostic imaging; Carcinoma, Hepatocellular - pathology; Hepatocellular carcinoma; Humans; Liver Neoplasms - diagnostic imaging; Liver Neoplasms - pathology; Liver Transplantation; Neoplasm Recurrence, Local - diagnostic imaging; Neoplasm Recurrence, Local - pathology; Predictive Value of Tests; Radiographic Image Interpretation, Computer-Assisted; Recurrence; Tomography, X-Ray Computed
• ISSN: 0720-048X; 1872-7727
• DOI: 10.1016/j.ejrad.2019.05.010

To assess whether radiomics signature can identify aggressive behavior and predict recurrence of hepatocellular carcinoma (HCC) after liver transplantation. Our study consisted of a training dataset (n = 93) and a validation dataset (40) with clinically confirmed HCC after liver transplantation from October 2011 to December 2016. Radiomics features were extracted by delineating regions-of-interest (ROIs) around the lesion in four phases of CT images. A radiomics signature was generated using the least absolute shrinkage and selection operator (LASSO) Cox regression model. The association between radiomics signature and recurrence-free survival (RFS) was assessed. Preoperative clinical characteristics potentially associated with RFS were evaluated to develop a clinical model. A combined model incorporating clinical risk factors and radiomics signature was built. The stable radiomics features associated with the recurrence of HCC were simply found in arterial phase and portal phase. The prediction model based on the radiomics features extracted from the arterial phase showed better prediction performance than the portal vein phase or the fusion signature combining both of arterial and portal vein phase. A radiomics nomogram based on combined model consisting of the radiomics signature and clinical risk factors showed good predictive performance for RFS with a C-index of 0.785 (95% confidence interval [CI]: 0.674-0.895) in the training dataset and 0.789 (95% CI: 0.620-0.957) in the validation dataset. The calibration curves showed agreement in both training (p = 0.121) and validation cohorts (p = 0.164). Radiomics signature extracted from CT images may be a potential imaging biomarker for liver cancer invasion and enable accurate prediction of HCC recurrence after liver transplantation.